Donate
Independent, objective, nonpartisan research

R 203PLR

Authors

R 203PLR

Tagged with:

Publication PDFs

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3742 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(12) "R_203PLR.pdf" ["wpmf_size"]=> string(6) "660779" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(242147) "California’s Housing Element Law: The Issue of Local Noncompliance ••• Paul G. Lewis 2003 PUBLIC POLICY INSTITUTE OF CALIFORNIA Library of Congress Cataloging-in-Publication Data Lewis, Paul George, 1966- California’s housing element law : the issue of local noncompliance / Paul G. Lewis. p. cm. Includes bibliographical references and index. ISBN: 1-58213-069-8 1. Discrimination in housing—California. 2. Housing policy— California. I. Public Policy Institute of California. II. Title. HD7288.76.U52C25 2003 363.5'09794—dc21 2002154992 Copyright © 2003 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. PPIC does not take or support positions on any ballot measure or state and federal legislation nor does it endorse or support any political parties or candidates for public office. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Foreword It is no secret that affordable housing is scarce in California. Even with record low mortgage rates and plenty of demand, the lack of conveniently located, relatively low-cost housing has frustrated planners and consumers alike. What is less well known is that many local governments have been out of compliance with California’s housing element law, which was designed to help local officials plan for adequate housing in their communities. In his latest look at local governance in California, Paul Lewis addresses the question of why so many cities and counties have been unable or unwilling to meet the state requirements for housing. Lewis analyzes a long list of reasons why a city would be out of compliance with the law, including shortages of available land, explicit antigrowth policies, and an aversion to affordable housing among wealthy communities. His findings do not support the view that only the richest communities were out of compliance; in fact, smaller cities with older housing and those with strict growth controls were more likely to be noncompliant. He then takes a closer look at communities in the metropolitan areas of San Diego, Los Angeles, and the San Francisco Bay Area to identify the relationship between noncompliance and housing production. His results will come as a surprise to some. During the 1990s, noncompliant communities were just as likely to expand their housing stock as communities that complied with the law. Furthermore, when other factors were held constant, noncompliance was not a significant predictor of the rate of multifamily development. After reviewing the policies of others states with comparable approaches, Lewis identifies three basic problems with California’s housing element law. First, it often goes against the grain of local politics by asking cities to plan for the needs of the wider region, not just those of current city residents. Second, it may represent a mismatch of goals and policy tools. Specifically, it attempts to tackle the problems of overall iii housing underproduction with a process-oriented approach developed to prod cities and counties into planning for their share of affordable units. Third, the statute itself is unwieldy, embraces multiple objectives, and is difficult for nonexperts to understand. Lewis concludes that the time is ripe for policymakers and affected interest groups to seek a more workable, transparent, and straightforward approach to housing. These policymakers may need to resolve whether their major goal is a sheer increase in residential construction or an equitable distribution of affordable housing. Lewis warns that using a fair-share planning approach as a tool to encourage overall housing production places an unrealistic burden on a fairly fragile policy. David W. Lyon President and CEO Public Policy Institute of California iv Summary California is generally perceived as producing less housing than would be expected or desired judging by its population and job gains. The statewide plan developed by the Department of Housing and Community Development (HCD) predicts that a continuation of current trends will lead to underproduction of needed housing by approximately 60 percent—likely leading in turn to a further upward spiral of home prices and rents as well as lower homeownership rates. This shortfall particularly hurts low-income families, which have more difficulty in paying the price or rent premium that results from undersupply. Some observers argue that local governments’ lack of enthusiasm for new housing in their communities is a large source of this problem. This report investigates California’s housing element law, the major tool the state government uses to ensure that city and county land-use regulators are planning appropriately for new housing development. Enacted in original form in 1969, the housing element law requires that all cities and counties in California engage in detailed planning for their residential needs by including housing as an element of their comprehensive plans. The housing element process is intended to focus the attention of city policymakers on identifying land sites for housing, and on policy actions that would make it easier or less expensive to provide additional housing units. The motivation for this study is the high degree of local noncompliance with the law. I begin by examining which types of city governments tend to have their housing elements deemed “out of compliance” by HCD staff. A second major goal is to determine whether such noncompliance can be linked statistically to a lower subsequent production of new housing. The report also reviews California’s implementation of housing element law and compares it to the experiences of other states. v How the Housing Element Process Works The housing element is the only part of local general plans that is subject to substantial oversight by the state. The state’s interest in local housing elements has been justified by the fact that housing is enshrined in state law as a matter of “vital statewide importance.” Nevertheless, housing elements in and of themselves rarely cause new housing to be built; in a market economy, private developers (or nonprofit builders of affordable housing) construct nearly all new housing units. The housing element requirement is often called a “fair-share” housing law, with the term generally referring to a regional process by which each local community works to accommodate a fair proportion of the region’s housing need. Regional councils of governments (which are planning councils representing the cities and counties in a given metropolitan area) work from state estimates of regional housing need and assign a housing unit goal, or allocation, to each city and unincorporated county area in their region. Cities and counties are then expected to update their housing elements to plan for quantified objectives for housing units over the next five years. Unlike in some states with fair-share approaches, in California housing production itself has received as much emphasis as the geographic distribution of housing for lower-income families. This state’s broader notion of fair share probably has resulted from its longstanding problem of housing affordability and underproduction, which limits the opportunities of not only the poor but also the middle class. The housing element statute requires that local planners address and reduce governmental constraints on the development of housing for all income levels. Such constraints may include local growth controls, strict building codes, developer fees, and permit procedures. California law requires that local governments revise their housing elements periodically. In the current round of revisions, San Diego (in 1999) was the first region where localities were required to update their plans. In 2000 and 2001, cities and counties in the Southern California region and the San Francisco Bay Area undertook these updates, with the vi requirement extending to the Bakersfield, Fresno, and Sacramento areas in mid-2002 and to the rest of the state’s regions by the end of 2003.1 Once a local update has been drafted, HCD reviews it to gauge whether the plan can enable the targeted number of units—including specific amounts of housing for households of very low, low, moderate, and “above moderate” incomes. If so, HCD certifies the housing element. If not, the jurisdiction may change its plan to incorporate HCD’s suggestions. If the element is adopted without satisfying HCD—or fails to be updated at all—the city or county is regarded as noncompliant. That judgment limits its eligibility for certain state and federal funds for affordable housing and renders it more vulnerable to lawsuits that can halt development in the community. There have been frequent conflicts between state and local policymakers over housing element compliance. Testing Competing Explanations for Local Noncompliance One of the most contentious aspects of the housing element requirement is the fact that nearly four cities in ten and a quarter of counties are out of compliance with the law.2 Large majorities of jurisdictions in the state have been noncompliant at some point. Explanations of this fact differ dramatically. Some communities have argued that state or regional projections fail to reflect powerful local realities, such as a lack of vacant land, which can make it difficult to identify a sufficient physical capacity to accommodate projected housing needs. Other observers argue that local noncompliance more frequently reflects an aversion to new housing—particularly affordable units—on the part of upper-income communities. Overt antigrowth policies or ____________ 1Here, Southern California includes cities in the counties of Imperial, Los Angeles, Orange, Riverside, San Bernardino, and Ventura; the Bay Area includes cities in Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma Counties. 2This tally of noncompliant jurisdictions includes those whose housing element is overdue, as well as those found noncompliant after HCD review. In both cases, the housing element is legally noncompliant. vii regulatory postures on the part of local governments—sometimes imposed on local governments by their voters—are also seen as a culprit. Still another possibility is that many jurisdictions fail to meet the housing element requirement because they lack the requisite planning capacity or experience. State officials periodically seek ways to increase compliance with housing element law by penalizing noncompliant local governments. Housing advocates have long argued that sanctions should be stronger and that housing law needs more emphasis on results and less on process. Senate Bill 910, which passed the California Senate in 2001 but later died in the assembly, would have required that the state controller fine noncompliant cities. By contrast, local officials seek to protect local autonomy over land use—long one of the major pillars of home rule— and complain of distant state authorities who fail to understand local conditions. This study examines the distinctions between cities that have been able to attain housing element compliance from those that have not. Measures of city characteristics are drawn primarily from the U.S. Census, and measures of local land-use policies and growth controls are drawn from a PPIC mail survey of local planners in the late 1990s. A simple comparison of cities in the regions that have recently been reviewed by HCD indicates that noncompliant communities are, on average, smaller and have older housing. In the rest of the state, where cities have had about a decade to reach compliance since their last housing element revisions were due, few cities were noncompliant. The simple comparison indicates that these few laggard communities tend to be wealthier and less ethnically diverse than compliant cities. The results of a more sophisticated statistical analysis reveal the determinants of noncompliance in a more systematic way. This analysis involved cities in the Southern California region, San Diego County, the San Francisco Bay Area, and the 18-county Central Valley. Evidence was strong that cities’ residential growth policies held a particularly important role for the HCD reviews. For example, cities whose planning directors report that the review process for new development proposals has been getting longer experience more trouble in attaining compliance. Each viii restrictive growth policy that has been adopted by the city approximately doubles the odds that it will be found noncompliant. An older housing stock is also associated with a greater likelihood of noncompliance. Cities with older housing may be more settled and have a more established community character; they are also likely to contain less vacant land. Cities with smaller populations are also more likely to be noncompliant, all else equal. Governments of larger cities may have a greater capacity to undertake the broad range of planning efforts needed to reach compliance and may also be more insulated from the political pressure of homeowners. An additional finding is that the length of time that has elapsed since the deadline for submitting an updated housing element is one of the most statistically significant predictors of compliance. Specifically, the results indicate that each month that has passed since the region’s deadline renders a city approximately 5 percent more likely to reach compliance. Even controlling for all these factors, cities in the Bay Area, Central Valley, and Inland Empire were more likely to be noncompliant than other cities. It is also interesting to note that the compliance status of individual communities tends to persist over time. Although many cities were able to attain compliance as the 1990s wore on, it remained the case that noncompliance in 1991, after the past round of revisions in the coastal metropolitan areas, was a fairly good predictor of noncompliance in 2002. Assessing the Relationship Between Compliance and Housing Production Defenders of the housing element requirement tend to argue that local governments that comply with the law, by demonstrating adequate plans, enable more housing to be built. Using data from the 1990s, I examine whether a city’s compliance status in 1991 helps to predict the percentage increase in the city’s housing stock by 2000. The analysis again controls for a variety of other city characteristics that might be expected to influence the level and type of housing growth. Because of ix data limitations, this analysis focuses only on cities in the Southern California, San Francisco Bay Area, and San Diego regions. Using Census data on the number of housing units in each city, I find no detectable relationship between housing element compliance and the percentage increase in housing across these communities during the 1990s. Thus, for all the potential merits and benefits of housing element compliance, one must look to other factors to explain why some cities experience rapid housing development and other cities experience little. The analysis suggests that a city’s demographic characteristics, its position in the urban hierarchy, and its physical capacity to accommodate new buildings are better predictors of housing growth. Although the housing element statute, as a fair-share approach, is especially concerned with increasing the production of affordable housing, we unfortunately lack any comparable information for all cities about the production of affordable units. However, data from the Construction Industry Research Board do allow an analysis of multifamily housing production. Multifamily housing includes apartments and condominiums, the types of housing generally most relevant for those at the lower to middle levels of the income distribution, particularly in the coastal metropolitan areas where housing is expensive. Multifamily units represented only about one-quarter of the new units produced in the 1990s. The results of this analysis once again show that housing element noncompliance as of 1991 is not a significant predictor of the rate of multifamily development when other relevant factors are held constant. Rather, cities that were job centers and that had fewer senior citizens as of 1990 tended to experience faster rates of multifamily housing development. Finally, the report investigates whether housing element compliance affected the mix of housing developed in the 1990s—multifamily versus single-family—despite its lack of effects on the rate of increase. I examine the relationship between compliance and the percentage share of new housing units that were multifamily units. Here there is evidence of a significant association of compliance with the outcome in question. Cities with noncompliant housing elements developed new housing that was weighted more toward single-family units. Holding constant for x other city and county characteristics, a noncompliant housing element was linked to an 8 percentage point lower proportion of multifamily housing among the newly built units. The results imply that cities with compliant housing elements are willing to substitute multifamily (more affordable) units for a share of single-family units. Nevertheless, it is striking that one can detect no measurable relationship between compliance and overall housing production. Rethinking the Housing Element Approach Discussions of problems with the current housing-element law among California policymakers and housing advocates have led to a number of reform proposals in recent years. However, most of these proposals take the housing element approach—a state review of local plans for future housing needs—more or less as a given. A fundamental question may be whether the various goals and values that California policymakers hold dear—increased housing production, an equitable distribution of housing responsibility across communities, special attention to the housing needs of low-income groups, local autonomy and home rule, environmental protection, and more—can all be accommodated within this area of law, or whether there are tensions among them. A secondary question is whether the current approach is the most effective use of resources to further these goals. Enticing communities to accommodate housing would not be such an uphill battle if they perceived that doing so would be in their financial self-interest. Thus, creating a component of the state fiscal system that rewards local governments for the addition of housing units, particularly affordable units, may result in less conflict and more cooperation. The Jobs/Housing Balance Incentive Program, passed in 2000, has elements of such a “rewards for performance” approach. Revenue sources that are distributed to localities on a population basis, such as the Vehicle License Fee subvention, also create indirect incentives for cities and counties to accommodate housing. Other states, including some with equally strong traditions of home rule as California, have also wrestled with issues of inadequate housing production, mandated fair shares for jurisdictions, and state oversight of local planning. In New Jersey, for instance, production of affordable xi housing rests mainly with the profit motives of private developers; they can propose a “builder’s remedy” by including a share of affordable units in otherwise market-rate projects that would normally exceed local zoning limitations. The quasijudicial state agency overseeing local housing plans also allows communities to transfer to other jurisdictions up to half of the affordable units they are expected to produce, in exchange for a payment to the “receiving” municipality. In Massachusetts, developers of affordable projects have access to a Comprehensive Permit Law that enables them to proceed through a streamlined local review process, avoiding many of the intermediate reviews that other proposed developments must go through. Builders whose projects are denied by this local process have the option of appealing to a statewide board, the Housing Appeals Committee, which can overturn the local decision and order the project to be permitted. In Oregon, the state’s Land Conservation and Development Commission imposes minimum zoning densities on residential land for cities and counties in the Portland metropolitan area. The localities in this region are also required to write plans so as to allow at least half of future residential units to be in multifamily projects. As in these other states, California’s housing element requirement has often gone against the grain of local policymaking because it asks cities to plan for the needs of the wider region, not just the needs of current city residents. Indeed, the philosophy behind fair-share housing policy is that the so-called police power of local governments to regulate land use should be directed toward the general welfare of the region, not just the general welfare of the specific locality. Those involved in California’s debate over housing policy often note how long and detailed the housing element statute is. Highly detailed statutes are often evidence of widespread disagreement on a given policy, as waves of “reform” occur in which opposing interests seek to have their specific concerns addressed and preserved in law. In the case of the housing element statute, the result is an unwieldy law that is often difficult for outside observers to comprehend in its entirety or details. In the 33 years since the housing element statute was enacted, the search for an adequate supply of housing in California has only become more elusive. It may be a ripe occasion for policymakers and affected xii interests to seek an approach to housing policy that is more workable, transparent, and straightforward, with measurable barometers of substantive local success or failure. In so doing, policymakers will need to resolve whether the major goal of such a law is a sheer increase in residential construction or an equitable distribution of affordable housing. Using a fair-share planning approach as a tool to encourage overall housing production may place an unrealistic burden on a relatively fragile policy. xiii Contents Foreword..................................... iii Summary..................................... v Figures ...................................... xix Boxes ....................................... xxi Tables ....................................... xxiii Acknowledgments ............................... xxv 1. INTRODUCTION ........................... Controversies over Local Government Noncompliance with the Law................................ Recent Legislative Proposals and Reform Efforts ......... Plan of the Report ............................ 2. WHAT IS THE HOUSING ELEMENT LAW AND HOW HAS IT BEEN IMPLEMENTED? ............ Introduction: A Fair-Share Housing Law ............. Origins: Housing as a New Component of Local Planning .. The 1970s: Elaboration of the Housing Element Requirement ............................ The 1980s and 1990s: The Present System Takes Shape .... Regional Allocation Process ..................... Update Process for Local Housing Elements .......... Growing Contentiousness over Housing Element Requirements and Compliance ................. Concerns over Litigation....................... Are Needs Met? ............................ Concerns over Projections and Regional Allocations ..... The Current Revision Cycle: The Controversy Continues .............................. The Defense of Housing Element Law ............... Senate Bill 910 (2001–02) ....................... Conclusion ................................. 1 2 7 10 11 11 13 14 16 16 18 21 22 24 24 27 30 32 34 xv 3. WHY DO SO MANY COMMUNITIES HAVE NONCOMPLIANT HOUSING ELEMENTS? ........ Why Cities? ................................ Four Sets of Factors Potentially Affecting Compliance ..... Community Social Status and Exclusion ............. Local Land-Use Characteristics and Vacant Land ....... The Resources of the Local Government............. Local Politics and Residential Growth Policies ......... Timetable for Housing Element Updates Influences Compliance Status......................... Comparing Compliant and Noncompliant Cities: A Preliminary Profile ........................ Probing Further: A Statistical Model of Noncompliance .... Key Findings of the Multivariate Analysis ............. Some Caveats ............................... Conclusion ................................. 4. DOES HOUSING ELEMENT COMPLIANCE MEAN THAT MORE HOUSING IS PRODUCED? .......... Housing Element Compliance, Circa 1991 ............ Measuring Housing Growth Rates Using Census Data ..... Controlling for Other Factors ..................... Results of the Analysis of Total Housing Unit Growth ..... Analysis of Multifamily Development Using Construction Industry Data ............................ The Mix of New Housing Development .............. Reconciling These Findings ...................... Conclusion ................................. 5. CAN STATES ENSURE ADEQUATE LOCAL PROVISION FOR HOUSING DEVELOPMENT? ...... Alternative Approaches within the Context of the Current Housing Element Law ...................... More Penalties and Prescriptions for Local Governments .. Self-Certification Based on Performance ............. Subregional Allocations and Joint Housing Elements .... Encouraging Transfers of Housing Allocations among Jurisdictions ............................. Another Approach: Rewards for Performance ......... 35 35 36 37 38 40 41 43 44 48 51 53 55 57 57 59 60 64 65 67 68 68 71 71 71 72 73 74 75 xvi Relevant Experiences from Other States .............. New Jersey ............................... Massachusetts ............................. The Portland Region ......................... Conclusion ................................. 6. CONCLUSION ............................. Goals in Conflict ............................. Equitable Distribution or Sheer Production? .......... A Mandate in a Vacuum? ...................... Devil in the Details............................ 76 76 79 82 84 87 87 88 89 89 Appendix A. Multivariate Analyses of Housing Element Noncompliance .. B. Multivariate Analyses of the Rate and Mix of New Housing Production in the 1990s ........................ 91 95 Bibliography .................................. 101 About the Author ............................... 107 Related PPIC Publications .......................... 109 xvii Figures 1.1. Cities’ Housing Element Compliance Status, as of September 25, 2002 ........................ 1.2. Counties’ Housing Element Compliance Status, as of September 25, 2002 ........................ 4.1. Cities’ Housing Element Compliance Status, as of December 31, 1991 ........................ 4 5 58 xix Boxes 1.1. State Law: Does Housing Production Conflict with Other Goals? ............................. 2.1. A Creative Way to Help Meet Housing Goals ........ 2.2. A Housing Element Challenge in Folsom ........... 2.3. Other Important Provisions of SB 910, As Amended ... 9 30 31 34 xxi Tables 1.1. Schedule for Third Revision of Local Housing Elements ............................... 3.1. Comparison of Compliant and Noncompliant Cities in the San Diego, Southern California, San Francisco Bay Area, Fresno, Kern, and Sacramento Regions ........ 3.2. Comparison of Compliant and Noncompliant Cities in the Remainder of the State .................... 3.3. Variables Used to Predict Housing Element Noncompliance ........................... 3.4. City Characteristics That Are Statistically Associated with Noncompliance ........................ 4.1. Percentage Growth Rates for Housing in Cities in the San Diego, Southern California, and San Francisco Bay Area Regions, 1990–2000..................... 4.2. Comparison of Housing Development in Compliant and Noncompliant Cities in the San Diego, Southern California, and San Francisco Bay Area Regions....... A.1. Logistic Models of Housing Element Noncompliance as of September 2002 ......................... A.2. Summary Statistics for Continuous Variables in the Preceding Models .......................... B.1. Regression Model of Housing Unit Growth in the 1990s, with County Fixed Effects ................ B.2. Regression Model of Multifamily Development Growth in the 1990s, with County Fixed Effects ........... B.3. Regression Model of Share of Housing Development That Is Multifamily, with County Fixed Effects ....... B.4. Summary Statistics for Continuous Variables in the Preceding Models .......................... 6 45 47 50 52 59 60 92 93 96 97 98 99 xxiii Acknowledgments I thank Professors Scott Bollens of UC Irvine and Dowell Myers of USC, as well as Tracy Gordon, Hans Johnson, Fred Silva, and Mark Baldassare of PPIC, for reading a draft of this report and offering helpful suggestions. At the Department of Housing and Community Development, Deputy Director Cathy Creswell and her staff helped me understand the housing element process and obtain information. At an early stage of the project, Alex Amoroso (Association of Bay Area Governments), Mark Stivers (Senate Committee on Housing and Community Development), Peter Detwiler (Senate Local Government Committee), and Dan Flynn (California Assembly) also provided useful perspectives on the issues involved. Some of the data and information necessary for the analyses in this report were collected or prepared by Elisa Barbour, Khanh D. Bui, Rachel Flood, Hugh Louch, and Max Neiman. Peter Richardson and Joyce Peterson improved the prose and presentation. Although all of these people made important contributions, the author is solely responsible for any errors of fact or interpretation. xxv 1. Introduction California is generally perceived as producing less housing than would be expected or desired judging by its population and job gains. The state’s most recent statewide housing plan, for example, finds that developers will need to produce an average of 220,000 housing units per year through 2020 to meet projected demand. Achieving such production levels will be a challenge, the plan notes, since even in recent boom years only 150,000 or so new units received building permits. Since 1970, California has never produced 220,000 housing units for more than two consecutive years (Myers, 2001, p. 388). The state plan, published by the Department of Housing and Community Development (HCD), predicts that a continuation of current trends will lead to underproduction of needed housing by approximately 60 percent—likely leading in turn to increasing home prices and rents as well as lower homeownership rates (HCD, 2000, p. 3; Myers and Park, 2002). This lack of housing supply particularly hurts the poor, who have difficulty in paying the price or rent premium that results from undersupply. Two of three low-income renter households in California pay more than half their income to put a roof over their heads (Little Hoover Commission, 2002, p. i). Some observers argue that local governments’ lack of enthusiasm for new housing in their communities is a large source of this problem, although evidence is far from definitive on this point (see Lewis and Neiman, 2002). This report examines California’s housing element law, the major means by which the state government tries to ensure that local land-use regulators are dealing with unmet housing needs and planning appropriately for new housing development. The state requires that each city and county write (and periodically revise) a general plan to guide its future growth. The housing element is one of the seven required elements of the general plan; its purposes are to identify current and future local housing needs of all income groups and to ensure the 1 preparation of a detailed schedule, consistent with broader community planning goals, for meeting such needs. The housing element is the only part of local general plans that is subject to substantial oversight by the state. The law provides that housing elements are to be revised every five years, although the legislature allowed localities to postpone their updates throughout the 1990s because of budgetary shortfalls. HCD, relying on Finance Department population projections, assigns a target number or goal for additional units to each region of the state. In a process called the Regional Housing Needs Assessment (RHNA), the council of governments (a planning body representing the cities and counties in a given metropolitan area) must allocate this total number of housing units among the cities and unincorporated county areas in its region. HCD then reviews whether city and county housing plans conform with statutory requirements—in other words, whether the local housing elements seem likely to enable each community to meet its goal for new units. This report focuses on several key issues regarding this process of housing element revision and review. Among the topics analyzed are: • California’s implementation of the housing element law and how it compares with other states’ approaches to local housing planning, • Which types of municipal governments fail to comply with the law, and • Whether local compliance appears to make a difference in terms of the rates of new housing development across cities. Controversies over Local Government Noncompliance with the Law The city and county housing element updates have become a battleground for state and local policymakers. Some state officials argue that local governments are not energetic enough in planning for housing and are trying to deflect their fair share onto other jurisdictions. For their part, local officials often claim that the RHNA “quotas” that they have been assigned are poorly justified, unrealistic, and unresponsive to the physical limitations of their communities. “Typically, as soon as 2 these numbers are proposed, they are challenged by local governments as far exceeding local realities,” a representative of the League of California Cities writes (Carrigg, 2002). Local officials also argue that affordable housing production is stymied by both high land costs and the lack of state and federal funds for this purpose. This intergovernmental debate is hardly new. A decade ago, the previous round of housing element updates led to similar contentiousness and calls for reform (Jackson, 1994; Senate Committee on Local Government, 1993; Senate Committee on Housing and Land Use, 1995). Once a local update has been drafted, HCD reviews it to gauge whether the plan can enable the targeted number of units—including specific amounts of low- and moderate-income housing. If so, HCD certifies the housing element. If not, the jurisdiction may change its plan to incorporate HCD’s suggestions. But if the element is adopted without satisfying HCD—or fails to be updated at all—the city or county is regarded as “out of compliance.” Noncompliant communities are ineligible for certain affordable housing programs administered by HCD, such as the federal Home Investment Partnerships Program and portions of the Community Development Block Grant program, and the state Jobs/Housing Balance Improvement Incentive Grant. Noncompliant jurisdictions are also much more vulnerable to lawsuits on development matters. In some cases, judges have ordered noncompliant local governments to refrain from issuing any new building permits until the matter is resolved. Noncompliant housing elements make attractive legal targets for parties who seek to invalidate local land-use or redevelopment decisions—even though some of these parties are pursuing ends that do not always support housing development (Richards, Watson & Gershon, 1998; Senate Committee on Local Government, 1993, pp. 73–74). Thus, although HCD reviews of local housing elements are officially “advisory,” most cities and counties go to some length to achieve the department’s certification, especially since the law instructs judges to presume a local housing element legally valid if it has been certified by HCD (Warner et al., 1997). Compliance with housing element law is certainly not universal. As of September 25, 2002, about one-third of all cities and more than one- 3 fifth of all counties have had their housing elements judged noncompliant by HCD staff, and some have been so for many years. Other localities are overdue in submitting their draft housing elements to HCD, which also renders them noncompliant with the law (see Figures 1.1 and 1.2). At some points in the past, large majorities of jurisdictions in the state have been noncompliant. The disappointing levels of compliance among counties and especially cities have led housing advocates and some state legislators to argue that some localities are sidestepping their responsibility to ameliorate the shelter needs of Californians. Since housing has long been treated as a matter of statewide concern by the state legislature and the courts, the potential for intergovernmental conflict is high. State officials Under review by HCD 10% Self-certified 1% Housing element due 5% Out of compliance 33% In compliance 51% Figure 1.1—Cities’ Housing Element Compliance Status, as of September 25, 2002 4 Housing element due Under review by 5% HCD 7% Out of compliance 22% In compliance 66% Figure 1.2—Counties’ Housing Element Compliance Status, as of September 25, 2002 periodically seek ways to beef up compliance with housing element law or penalize noncompliant local governments, whereas local officials seek to protect local autonomy over land use—long one of the major pillars of home rule—and complain of distant state authorities who fail to understand local realities. In the current period, California communities have been undertaking the third revision of the required local housing elements. To reduce the burdens on HCD in reviewing local housing elements, the law specifies a staggered schedule of housing element updates. In 1999, San Diego was the first region where localities were required to update their plans. In 2000 and 2001, the process took place in the Southern California region and the San Francisco Bay Area. The requirement extended to the Fresno, Bakersfield, and Sacramento regions in mid- 5 2002, and to the rest of the state’s regions by the end of 2003 (see Table 1.1). In discussing compliance, this study uses the September 25, 2002, report on local compliance status from HCD, except where noted. The Table 1.1 Schedule for Third Revision of Local Housing Elements Council of Governments San Diego Association of Governments Counties Affected (Includes all Cities Within Each County) Revision Date San Diego December 31, 1999 Southern California Association of Governments Imperial Los Angeles Orange Riverside San Bernardino Ventura December 31, 2000 Association of Bay Area Governments Alameda Contra Costa Marin Napa San Francisco San Mateo Santa Clara Solano Sonoma December 31, 2001 Council of Fresno County Government Fresno June 30, 2002 Kern County Council of Governments Kern June 30, 2002 Sacramento Area Council of Governments Sacramento Sutter Yolo Yuba (plus cities of Lincoln, Rocklin, and Roseville in Placer County) June 30, 2002 Association of Monterey Bay Area Governments Monterey Santa Cruz December 31, 2002 All other local governments All remaining December 31, 2003 SOURCES: HCD website; 2001 California Statutes, Chapter 85. 6 compliance status of individual jurisdictions can change as new revisions are submitted and HCD reviews them.1 At present, although the state’s major metropolitan areas have already gone through the third round of housing element updates, other parts of the state are at an earlier stage of the process, and their compliance status in many cases reflects the outcomes of the prior round of updates in the early 1990s. These timing issues will become important in the statistical analysis of local compliance status in Chapter 3. Recent Legislative Proposals and Reform Efforts “All groups are dissatisfied with current housing element requirements.” —Staff report for Senate Committee on Local Government (1993) Housing advocates have long argued that sanctions should be stronger and that housing law needs more emphasis on results and less on process. Senate Bill 910, which passed the state senate in 2001 but died in the assembly, would have required that the state controller fine noncompliant cities. In its original, more punitive form, the bill would have widened penalties for noncompliant jurisdictions to include ineligibility for certain state transportation funds and directed state courts to presume that noncompliant housing elements are invalid. As a result, that bill would have opened up such jurisdictions to more litigation from housing rights’ organizations and developers. Discussing the original version of the bill, a nonprofit housing developer argued that transportation funds were used “as a bargaining chip, because, frankly, why should a city get money for transportation projects to make travel easier if it isn’t going to provide housing? It seems to me that the two components should have always worked together” (quoted in Bishop and Materna, 2001). Proponents of the bill also argued that the tougher state requirements could lead to a “good cop, bad cop” relationship that would help local officials argue against ____________ 1Readers can view the most recent compliance status report at http://www.hcd.ca. gov/hpd/hrc/plan/he/status.pdf. 7 neighborhood opposition to affordable and other housing projects. One developer said, “I think the local politicians will be relieved of the controversy that seems to surround residential development, because they will defer to state mandate” (quoted in Bishop and Materna, 2001). Some central city interest groups also supported the bill as a way to prompt suburban jurisdictions to share more of the burden of affordable housing. Opponents—including some legislators—saw the original bill as too punitive in its approach to local governments. The League of California Cities also argued that other state and federal mandates and policies hinder the ability of localities to accommodate additional housing. These mandates and policies include agricultural land preservation efforts, endangered species rules, and open space acquisition purchases (see Box 1.1). Although SB 910 attracted the largest amount of attention, a number of other bills proposing changes in the housing element process were considered in the 2001–02 legislative session. For example, SB 2292 (Dutra), which was recently signed into law, requires that local governments retain the residential zoning densities that they refer to in their approved housing elements. Cities and counties that “downzone” (reduce the allowable density of) residential land must transfer the “lost” density elsewhere in the community to make up for the loss in potential housing units. SB 1634 (Figueroa) would have required that regional planning councils, in making RHNA allocations to local governments, seek to improve the geographic balance of employment and housing in each region. AB 2863 (Longville), supported by the League of California Cities and the California State Association of Counties, would have eased definitions of “substantial compliance” under the housing element law and broadened the definition of what constitutes a “residential unit” that can count toward fulfilling a locality’s housing goals. However, neither SB 1634 nor AB 2863 was ever given a committee hearing, and both bills died. In reaction to the flurry of proposed legislation, a group of lawmakers and stakeholders, calling itself the Housing Element Reform Working Group, began meeting in Sacramento to discuss possible revisions to the law. The working group included interested legislators 8 Box 1.1 State Law: Does Housing Production Conflict with Other Goals? The state housing element law calls on cities and counties to adopt plans and zoning that can accommodate their projected housing needs. However, state law also includes other requirements that may constrain the ability of localities to accommodate housing (Bishop and Materna, 2001; Carrigg, 2002; Senate Committee on Housing and Land Use, 1995). For example: • Habitat conservation plans for threatened and endangered species cover millions of acres in California. • Williamson Act contracts, which grant lower property tax assessments to farm owners who agree to restrict their land to agriculture or related open space activities, cover about one-third of all privately held land in the state. • To receive a subdivision map, builders of housing developments of greater than 500 units must demonstrate that there will be sufficient water supply in the area for the next 20 years. • The California Coastal Commission exercises additional land-use authority in the coastal zone. • Cities and counties are also required to plan for transportation and congestion management, air quality, and seismic safety; each of these goals may conflict with the desire to construct more housing in a given area. • Cities wishing to annex land for additional housing development must have the annexation approved by the Local Agency Formation Commission (LAFCO). By state law, LAFCOs are charged with ensuring efficient service delivery patterns and must work toward restricting discontiguous development. • Significant development projects, and government actions that may encourage development, are subject to lengthy environmental impact reviews and possible litigation under the California Environmental Quality Act (CEQA). At the state level, determination of housing needs and allocations are given a statutory exemption from CEQA. But cities and counties updating their housing element are given no such exemption. and staff along with important stakeholders such as the League of California Cities, regional councils of governments, and the California Building Industry Association. The group attempted to move toward greater consensus on various issues and disputes relating to housing element preparation and review. Many of their ideas became part of SB 910 as it was amended in the assembly (to be discussed in Chapter 2). Legislators involved in the discussions agreed to hold all relevant housing element bills in committee until the reform working group had a chance to move toward agreement on some basic reforms. But consensus proved elusive. 9 Plan of the Report With policy reforms under discussion and concerns about housing production high, the time is ripe for serious study of housing element compliance. To introduce the issues and institutions involved, Chapter 2 traces the history of the housing element law in California, discusses its challenges and controversies, and summarizes the current housing element process. Chapter 3 examines why jurisdictions fail to submit housing elements that are acceptable to HCD. For example, some communities have argued that state or regional projections fail to reflect powerful local realities, such as a lack of vacant land on which to accommodate housing. Other observers argue that local noncompliance more frequently reflects an aversion to new housing—particularly affordable units—on the part of upper-income communities. Antigrowth policies or regulations— sometimes imposed on local governments by their voters—are also seen as a culprit. Still another possibility is that many jurisdictions fail to meet the housing element requirement because they lack the requisite planning capacity or experience. Chapter 4 addresses the relationship of housing element compliance to housing production. Is it true that compliant jurisdictions add housing—particularly multifamily development—at a faster rate, after one takes into account other local characteristics that affect the level of housing production? Although the housing element law involves a requirement to plan rather than to build housing, it presumably was passed under the premise that effective housing planning would lead to more residential production. Chapter 5 discusses reforms that have frequently been suggested for housing element law and compares California’s experiences and rules to those of other states that review local housing planning and have fairshare requirements, such as Massachusetts, New Jersey, and Oregon. A brief consideration of the experiences of these other states in pursuing similar policy goals may be instructive at this important juncture in the legislative history of housing elements in California. Finally, Chapter 6 offers concluding observations and calls for a reconsideration of California’s approach to local housing production. 10 2. What Is the Housing Element Law and How Has It Been Implemented? This chapter presents a policy history of the housing element law in California and discusses how the policy is currently carried out. After offering a brief introduction to the purposes of the housing element law, I focus on its origins and development, the process by which it has been implemented, and ongoing debates regarding its effectiveness and possible reforms. Housing element policy had a somewhat inauspicious start in the state, and its processes and institutional responsibilities were not fully worked out for approximately a decade following the 1969 passage of the law. Once the law began to take its present shape, an extended policy debate ensued, often pitting local governments and advocates of local flexibility against HCD, homebuilders, and affordable housing advocates. This stalemate has continued to the current day, with the state now in the midst of the third wave of mandated updates of local housing elements. Introduction: A Fair-Share Housing Law Enacted in 1969, the original housing element law requires that all cities and counties in California engage in detailed planning to meet their housing needs. Housing elements do not, in and of themselves, cause new housing to be built, as this activity rests with private developers (or nonprofit builders of affordable housing). Rather, the housing element process is intended to focus the attention of city policymakers on policy actions that they might take to make it easier or less expensive for additional housing units to be built. For example, local plans may call for reducing or eliminating fees on affordable housing construction, rewarding developers of certain types of projects by allowing them to 11 build at higher densities than would otherwise be permitted, or requiring that all housing developments above a certain size reserve a portion of units for low- or moderate-income households. The housing element requirement is sometimes referred to as a “fairshare” housing law. A number of other states and metropolitan areas have developed fair-share housing legislation, with the term generally referring to a regional process by which each local community works to accommodate a fair proportion of the region’s housing need. Fair-share plans “determine where housing, especially low- and moderate-income units, should be built within a region according to such criteria as placing housing where it will expand housing opportunity, where it is most needed, and where it is most suitable” (Listokin, 1976, p. 1). Jurisdictional fair-share goals are typically defined using a formula that includes such factors as growth rates in households and jobs, socioeconomic status measures, and often some measure of the local carrying capacity for new housing. From 1977 to 1983, the U.S. Department of Housing and Urban Development (HUD) offered matching funds to regions that engaged in the creation of Areawide Housing Opportunity Plans (Brownlow, 1991, p. 1).1 California was an early innovator in the fair-share housing field. Its law on the topic is more than three decades old, although it did not take its current, relatively strict form until 1980. Unlike in some states with fair-share approaches, in California housing production itself has received as much emphasis as the geographic distribution of housing for lower-income families. This state’s broader notion of fair share probably has resulted from its longstanding problem of housing affordability and underproduction, which limits the opportunities not only of the poor but also of the middle class (Calavita, Grimes, and Mallach, 1997). Thus, the fair-share approach taken by California allocates among localities goals not only for ____________ 1In 1968, Congress required that federal planning subsidies for metropolitan planning be contingent on preparation of a regional housing plan, which led many urban regions to develop early fair-share plans. By the 1970s, many housing activists argued that fair-share plan adoption should be required for cities or regions to qualify for federal housing subsidies. HUD did not require this, but did award subsidy bonuses for metropolitan areas that developed fair-share plans (Listokin, 1976, pp. xvii, 6). 12 subsidized or “affordable” units but also for moderate-income and “above moderate” units—that is, the entire projected housing need. The philosophy behind such allocation strategies is to serve as a warning to localities to readjust their zoning ordinance if it does not allow for their allocated growth share. The units also serve as a growth management technique alerting communities to the level of future growth they can expect and enabling them to plan properly to handle development (Listokin, 1976, p. 169). In addition, because there is a widespread perception among local officials in California that housing development is fiscally burdensome, the state has used housing element law as a way to encourage planning for balanced land-use development—that is, attention to residential as well as commercial development. Some argue that the state’s system for financing local governments encourages the opposite, since cities’ dependence on local sales tax revenues, and lack of control over property tax allocation, may cause local policymakers to favor retail development over housing. Origins: Housing as a New Component of Local Planning Until 1969, land use and circulation (transportation) were the only elements of local general plans required by California law. Housing, along with other topics such as open space, was sometimes dealt with in optional general plan elements by ambitious cities and counties. In 1969, then-Assemblymember Pete Wilson successfully carried a bill to require that each city and county prepare a housing element. This law made local housing a mandatory element of local general planning and required that HCD develop guidelines for local governments to follow in preparing housing elements. In 1971, legislation was added requiring that localities follow HCD guidelines for the preparation of housing elements.2 ____________ 2According to the Select Committee on Housing and Urban Affairs (1974, p. 59), the change to the statute was necessary because the 1969 law “had inadvertently referred to these ‘mandatory guidelines’ as ‘regulations,’ and the attorney general interpreted this as requiring that regulations be adopted” in accordance with the state’s cumbersome rules on administrative procedures. Such rule-making would be interpreted as a state mandate 13 In 1970, additional legislation passed requiring that HCD prepare a statewide housing element; such action was necessary for the state to take advantage of federal funds for comprehensive planning assistance. The law required that HCD develop one- and five-year housing development goals “needed to house all residents of this state” (quoted in Select Committee on Housing and Urban Affairs, 1974, p. 50). Notably, however, this activity was organized separately from the more comprehensive statewide planning efforts of the Office of Planning and Research. The 1970s: Elaboration of the Housing Element Requirement In the early years of the housing element requirement, the state government devoted little in the way of resources or attention to guiding or evaluating local housing plans. Governor Ronald Reagan’s administration, for example, was seen as unenthusiastic about the very existence of the Department of Housing and Community Development, and it tried unsuccessfully to merge it with other state agencies (Select Committee on Housing and Urban Affairs, 1974). Budgetary commitments to housing element activities (and to HCD generally) were low during this period. HCD budgeted only $17,000—the equivalent of one full-time employee—to updating the statewide housing element in 1974–75, for example, and approximately two-thirds of HCD’s small budget was devoted to inspection of manufactured housing. A senate select committee complained that HCD had had little public input into the preparation of the statewide housing element. Moreover, it maintained that HCD had refused to purchase Census data that would allow detailed computation of housing needs for each locality. The committee also pushed the department to separate the issue of production goals for the state from the problem of affordability ________________________________________________________ that would require reimbursement of the local housing element process. Since the legislature had appropriated no money for local housing element preparation, HCD had concluded that it could not adopt “regulations.” 14 shortfalls for low-income residents (Select Committee on Housing and Urban Affairs, 1974, pp. 54–55). Furthermore, HUD complained that California’s statewide housing element performance was “less than exemplary,” as the state had prepared only a draft needs analysis and work program by October 1973 (Select Committee on Housing and Urban Affairs, 1974, p. A12). This document called for the production of 225,000 units per year statewide, as well as rehabilitation and replacement efforts for existing housing. At this time, the League of California Cities was calling for an augmented planning role for HCD, in coordination with housing planning that local governments were undertaking. “We believe that the future of DHCD . . . must include a strengthening, rather than a weakening, of its functions. . . . In the area of housing, its role should embody an institutionalization of a strong state role in helping to provide for the housing needs of all segments of the people of California” (Select Committee on Housing and Urban Affairs, 1974, p. A19). A 1975 law (AB 1X) called upon HCD to adopt housing element guidelines and to review city and county housing elements, resulting in a set of detailed standards in 1977. These guidelines stressed the regional fair-share concept (Warner et al., 1997; Senate Committee on Local Government, 1993, p. 18). During the administration of Governor Jerry Brown (1975–1983), statewide planning, including housing planning, became more active and central to the mission of the administration. Arnold Sternberg, Brown’s director of Housing and Community Development, had represented low-income advocacy groups as an attorney, and he took advantage of the provisions in the housing element law allowing the department to develop guidelines for the implementation and enforcement of the statute. The department also took a more aggressive posture toward the fair-share goals implied in housing element law, by which each local government is expected to help solve regional housing shortfalls, particularly for low- and moderate-income households. 15 The 1980s and 1990s: The Present System Takes Shape A new law, Chapter 1143 of the Statutes of 1980, resolved the issue of HCD oversight. HCD’s guidelines were deemed as advisory, but localities were required to consider HCD comments before adopting the local housing element (Warner et al., 1997; Senate Committee on Local Government, 1993, p. 18). Local housing elements were required to include a needs analysis, a discussion of resources and constraints in meeting those needs, statements of goals and specific policies regarding housing, as well as quantified objectives for construction, conservation, and rehabilitation to move toward the goal. “This mandate aims at ensuring that each community accepts responsibility for the housing needs of not only the resident population but also of those households who might reasonably be expected to live within the jurisdiction were a variety and choice of housing appropriate to their needs available” (HCD, 1988, p. 1). In other words, the concept of fair share was fully enshrined in state law (Senate Committee on Housing and Land Use, 1995, p. 25). This statute also introduced the local duty to review and update local housing elements every five years (Richards, Watson & Gershon, 1998). Regional Allocation Process The regional allocation of housing needs was also developed much more fully during this period. HCD began to assign housing goals to each region, using state Department of Finance projections regarding future household growth. HCD adjusts these goals to account for the amount of vacant housing in the region and the expected need for replacement units. HCD also takes into account advisory comments from the region’s council of governments (COG), an association of city and county officials that is responsible for regional planning. Each COG then prepares a regional housing needs assessment. In rural parts of the state that lack COGs, HCD itself performs the RHNA analysis. The first two mandated cycles of the RHNA were in 1981 and in 1984–1986 (HCD, 1988). The stated reasons for the regional needs 16 assessment are to allow consideration of regional issues, distribute responsibilities equitably among jurisdictions, and relieve local governments of some of their data-gathering burdens (HCD, 1988, p. 4). The RHNA process is specifically exempted from the California Environmental Quality Act. The regional housing need is pegged to at least the level of projected household growth, although “major economic events” may justify a revision to a lower number (HCD, 1988, p. 11). Under the statute, several factors must be considered in the COG’s allocation of housing growth goals to specific localities: • Market demand for housing (vacancy rates, housing prices, household structure, construction, absorption, etc.), • Employment opportunities (current and projected), • Availability of suitable sites (including residentially zoned as well as nonresidentially zoned land that could be used, and the possibility of redevelopment for housing or for increased densities of housing), • Availability of services (including current and future capacity, transportation, medical and recreational facilities, etc.), • Commuting patterns (time, length, transit availability), • Type and tenure of housing need (including a consideration of special populations, such as large households, the elderly, students, and the military), and • Farmworker housing needs. Planners are also directed to avoid aggravating existing concentrations of low-income households in certain communities. The reasoning HCD cites for this requirement presents an interesting perspective on housing development: The basic philosophy behind housing element law is that citizens of all economic levels should have the opportunity to live where they choose in decent, safe, and sanitary housing. . . . Some COGs have adopted [allocations] which plan for the same percentages of households in each income group in each locality. This approach is based on the position that all areas are equally suitable for each income group. The approach also reflects the view that the 17 current income differences between localities reflect the effects of past governmental actions, such as zoning and capital improvements planning, and that housing opportunities for underrepresented income groups should be promoted in governmental planning activities (HCD, 1988, pp. 23–24). In other words, the regional needs assessment is to be used not only to help enable the production of needed amounts of housing in a region but also as a device to redistribute the burdens of lower-income households more equitably across geographic areas. It is likely that many of the controversies over housing element development are rooted in the desire of some localities to preserve their existing character, while their COG and HCD attempt to move them toward a mix of residents more representative of the entire region. Still, HCD guidelines do point out that in larger and more complex regions, there may be good reasons to deviate from the general goal of having each jurisdiction approximate the diversity of its region. For example, HCD (1988, p. 24) notes that the age and differentiation of the larger regions often lead to geographic clusterings of socioeconomic groups, such as retirement communities and college-student neighborhoods. It is important to note, however, that the actual weighting scheme in evaluating these various factors is to be designed by the COG itself (HCD, 1988, pp. 20–21; Senate Committee on Housing and Land Use, 1995, p. 5). In short, there has been no uniform, required method for assigning housing goals to local governments. Update Process for Local Housing Elements A locality may revise its need from the number given to it by the COG, but the city or county is required to cite findings, backed by data, to justify its revised goal. “A local revision may be used in a locality’s housing element even if the COG has disapproved it. In such a case, however, DHCD may be less likely to consider the element to be in compliance with the housing element law” (HCD, 1988, p. 14). A city or county housing element must provide analysis of local housing needs, resources, and constraints relevant to the production of housing, and a five-year policy program detailing actions that the local government plans to take to address its housing needs. Although there is a great deal of local discretion in preparing housing elements, there are 18 also “elaborate statutory provisions” to comply with, which is not the case for other portions of the general plan (Curtin, 2000, p. 10). The state’s interest in local housing elements has been justified by the fact that housing is enshrined in state law as a matter of vital statewide importance. In determining housing needs, local planners must consider data regarding the existing number of households and housing units, the number of households overpaying for housing, and overcrowding, rehabilitation, and replacement needs. The local government must provide an estimate of various households considered to have special needs, such as the disabled, senior citizens, large households, farmworkers, female-headed families, and homeless persons. The city or county is also required to note any existing assisted-housing projects that may be at risk of losing their subsidized status during the time period covered by the housing element update. For many communities, identifying a sufficient physical capacity to accommodate projected housing needs is a major challenge. Localities are required to provide a detailed land inventory, summarizing the number of acres zoned for various types of residential development (single-family, multifamily, mobile homes, emergency shelter, farmworker housing, and mixed use) and for nonresidential uses. Local planners must specify the density range and dwelling unit capacity for these various zones and include information on the availability of services and infrastructure (HCD, n.d.). The local government’s five-year action plan must identify a sufficient set of potential sites for future housing development that could accommodate the community’s need “by right”—that is, without requiring a conditional-use permit or imposing vague conditions on potential homebuilders.3 HCD advises local officials that although a shortage of buildable land may make accommodating new housing units more complicated, it does not relieve a locality of its responsibility to provide adequate sites for new residential development. “If a locality does not have sufficient ____________ 3A conditional-use permit allows specified land uses (e.g., a mobile home park) in areas zoned for other types of uses (e.g., single-family homes), subject to additional local government discretion. 19 existing suitable sites to meet these needs, programs such as changes in land use, annexation, upzoning, or a second unit ordinance would be appropriate” (HCD, 1988, p. 53). Redevelopment policy and infrastructure improvements are also to be addressed as possible ways increase development potential (Rawson, 2000). From the standpoint of local officials, however, other policy considerations may conflict with such intensification of land uses, such as a lack of water or sewer capacity (often not under the control of the city or county), traffic congestion, conflicts with existing industrial activities, or seismic hazards. Each city and unincorporated county area, then, faces a housing element update in which it is expected to plan for construction needs and produce quantified objectives for housing units over the next five years. These objectives are further categorized by household income: • Very low (0 to 50 percent of regional median income), • Other lower (50 to 80 percent), • Moderate (80 to 120 percent), and • Above moderate (120 percent or higher). The number of needed units so identified, however, “are simply goals, not mandated acts” (Curtin, 2000, p. 11). In other words, a local housing element is hardly a self-enforcing policy and does not, in itself, create or mandate new housing. Nevertheless, the housing element statute does require that local planners address governmental constraints on the development of housing for all income levels. For example, zoning and land-use controls, building codes, developer fees, and permit procedures must be discussed, as well as local government plans to overcome or reduce public-sector barriers to housing production (Rawson, 2000). Local planners must also assess “nongovernmental constraints” to housing production, such as land and construction costs and the availability of financing. Finally, a set of miscellaneous topics noted in the statute must be addressed, including efforts to involve the public in the housing element update, the potential for energy conservation in new housing development, and consistency with other elements of the general plan. Going beyond a listing of objectives, the housing element must detail what programs and policies the local government foresees undertaking to 20 implement the plan, including efforts to provide sufficient sites for all types of housing (which may involve changes in zoning or subdivision requirements), participation in state and federal subsidy programs for affordable housing, incentives and regulatory concessions for residential developments, efforts to conserve the existing affordable housing stock and preserve units that are at risk of losing their subsidies, programs to promote equal housing opportunities, and efforts to address and remove governmental constraints on housing construction (HCD, n.d.). The city or county must send a draft of its housing element to HCD for review and consider the department’s findings on the draft. The department “considers the element to be in compliance only if every one of the statutory requirements is met” (HCD, 1991, p. 1). After the city council or county board of supervisors officially adopts the housing element, HCD again reviews it for its compliance with the law.4 Because of the detailed requirements of the statute, housing elements are subject to close scrutiny by the courts (Curtin, 2000, p. 27). Growing Contentiousness over Housing Element Requirements and Compliance During the late 1980s and early 1990s, when local governments were completing housing element updates, grassroots opposition to growth and development burgeoned in many parts of California, particularly in coastal regions. In this atmosphere, writing plans that focused on how to accommodate more housing became an increasingly contentious process. Compliance with the housing element law, as measured by HCD’s certification of city and county revisions, has been spotty at best. In 1991, only 19 percent of localities were certified as being in compliance by the department, although this proportion grew to 37 percent by 1993 ____________ 4Even if the department gives the opinion that the draft element complies with the law, the city or county must officially adopt the element before the jurisdiction is considered to be in compliance. Although most local governments in such a situation quickly adopt the approved draft element, some communities delay in their adoption or decide to submit a new draft to reflect new local concerns or political changes. 21 and to 52 percent by 1995 (Senate Committee on Local Government, 1993; Senate Committee on Housing and Land Use, 1995). The greatly increased rates of compliance were traceable in large part to an active program of technical assistance, combined with legal pressure, by HCD. HCD worked with the attorney general’s office to inform 47 delinquent localities that their housing element revisions were overdue and that they faced possible legal action. According to Timothy Coyle, then-director of HCD, “Many communities around the state . . . had failed to submit any evidence of a local housing element for, in some cases, as many as 15 years” (Senate Committee on Local Government, 1993, p. 32). Still, as Coyle and others affirmed, some of those noncompliant jurisdictions had actually experienced rapid rates of housing development, including affordable housing. Thus, the procedural burden of preparing a housing element may have been the major obstacle for many of these communities in their lack of compliance. Coyle told a Senate committee, “You can draw no correlation between housing element compliance and housing production” (Senate Committee on Local Government, 1993, p. 37). Nevertheless, two years later Coyle testified that the jurisdictions that were in compliance accounted for a disproportionate percentage of the building permits issued in the state—particularly multifamily permits (Senate Committee on Housing and Land Use, 1995, p. 31; see also Little Hoover Commission, 2002, p. 18). Concerns over Litigation According to some observers, a fear of litigation is a major motivator for local governments in expending time and resources on housing elements. Not only housing advocates but also growth opponents, and sometimes school districts and other overlapping jurisdictions, have used noncompliant housing elements as a litigation strategy in seeking to overturn local land-use decisions. The planning director of Long Beach testified in 1993 that the housing element “is the only element that is prepared defensively, rather than as a guide to local policy and decisionmaking. In most cities, the housing element is prepared as a joint effort by the city attorney and the planning department to make sure that the 22 document is defensible in court” (Senate Committee on Local Government, 1993, p. 62). The penalties for localities can be serious.5 Nevertheless, it is important to point out that Section 65589.3 of the Housing Element statute is an important line of defense for local governments that have received HCD certification of their housing elements. Added in 1990, this section establishes the “rebuttable presumption” that HCD’s finding of compliance means that the locality has drawn up a legally valid plan for its housing. In a legal case, the housing element law requires only “substantial compliance,” rather than perfection, which according to attorney Michael Colantuono “does a good job of protecting localities from frivolous challenges to their housing elements” (Senate Committee on Local Government, 1993, p. 37). Because courts have sometimes found localities to be in substantial compliance even when HCD has failed to certify their housing elements, some observers have suggested that HCD’s review is more stringent than necessary under the law. After the second round of housing element revisions, and the suspension of housing element funding in the 1992–93 state budget, a variety of interest groups began an effort to reform the housing element statute. Two major legislative hearings were held, with many voices criticizing the existing policy regime (Senate Committee on Local Government, 1993; Senate Committee on Housing and Land Use, 1995). Reform discussions were geared at streamlining the planning process for housing elements and making the RHNA process more attuned to local government concerns and capacity for growth. These discussions also emphasized performance in accommodating housing rather than the process of housing element planning (Warner et al., 1997). Nevertheless, aside from some relatively minor new legislation, ____________ 5“A prudent locality can manage this risk only by keeping its housing element up to date and building a voluminous record to ensure the element is legally defensible. Those communities which have failed to do so, often for budgetary reasons, have run significant risks and some have paid a significant price. . . . If a city loses such a case. . . , its land-use authority will be suspended. . . , it will have just 120 days to prepare a new element, often necessitating the retention of consultants, and it will likely pay not only its own legal fees, but those of its opponents. . . . During the preparation of a new element, nothing gets built, not even housing, without a court order . . . .” (Michael Colantuono, quoted in Senate Committee on Local Government, 1993, pp. 75–76, legal citations omitted.) 23 “six years of reform discussions led nowhere,” and no major reforms cleared the legislature (Richards, Watson & Gershon, 1998). Are Needs Met? Local governments tend to miss their housing unit goals, often by a wide margin. In particular, the low- and moderate-income segments of housing needs are typically unmet—not surprisingly given the frequent need for deep subsidies to fund such projects. In 1988, for example, the Bay Area Council, a business group, found that only one of the nine counties in the San Francisco Bay Area was on pace to attain even half its goals for production of low- and moderate-income units (even using a generous definition of affordable units). Napa County experienced a production of only 52 affordable units between 1980 and 1988, compared to a specified need of 3,906 units (Bay Area Council, 1988b). A more recent study of 40 fast-growing Bay Area jurisdictions found that only 34 percent of affordable housing goals were met in those communities with certified housing elements; only 9 percent of affordable housing goals were met in noncompliant jurisdictions (Dodge, 2002). Thus, the preparation of housing elements, whether compliant or not, is no guarantee that needed housing units will be built. Many fewer units may be approved, or even proposed, whether due to shifting demand, homebuilder preferences to build fewer and more expensive units, land prices in excess of what builders wish to pay, or government actions that reduce project size for reasons not contemplated in the housing element. Concerns over Projections and Regional Allocations The regional allocation of housing needs has also become the subject of increasing debate and conflict in recent decades. Some of the concerns relate to the projections used as the basis for regional and local housing goals. A 1995 analysis by senate committee staff echoed the concerns of many COG personnel and local officials in observing that the Department of Finance’s projections “are strictly mechanical and do not consider local planning factors such as local growth policies, habitat 24 preservation, clean air, and traffic congestion” (Senate Committee on Housing and Land Use, 1995, p. 19). In the current round of revisions, some commentators further argue that the Department of Finance’s projections of population increase and household formation—and, thus, projected housing needs—may be too high because of the tendency of new immigrants to have larger households.6 Another potential problem with the projections is the likelihood of decreasing birthrates among women in Latino and Asian groups in post-immigrant generations (Hill and Johnson, 2002). Over time and across generations, Latino families more closely approximate the trends among white and African American families toward fewer children, fewer persons per household, and more homeownership. Overall, the complex changes in California’s demography are consistent with any number of very different arguments about future housing needs (Myers, 2001). Some local officials have argued that population projections in regional transportation plans (mostly drawn up by COGs) should be the basis of housing need calculations, but existing state law requires the use of Department of Finance projections in preparing local housing elements. Some celebrate the flexibility of COGs in weighting factors differently in the RHNA process; others see this as an example of inconsistency that merits further state guidelines (Senate Committee on Housing and Land Use, 1995, p. 23). One criticism leveled at COGs involves their projections of future increases in jobs—upon which housing needs are, in part, based. Some of this criticism has come from ____________ 6Some observers claim that large household sizes are due to “doubling up” in otherwise unaffordable housing, but even when controlling for income, foreign-born Latinos live in significantly larger households (i.e., more persons in the housing unit) than the native-born (Myers, 2001, p. 389). Moreover, household crowding is explained almost entirely by demographic factors (such as nativity and poverty), whereas local housing market conditions have little relationship to crowding (Moller, Johnson, and Dardia, 2002). In brief, immigrant households, particularly among Latinos, are much more likely to be crowded, measured as more than one person per room. This is a major issue for housing needs assessment, since local governments must examine data on local housing crowding in assessing their current need for additional units. As Myers (2001, p. 392) notes, “Crowded housing highlights the dilemma for planners. If the situation is not viewed as a problem by the affected cultural group, should planners still treat it as a problem?” 25 advocates of increased housing production. In the 1980s, for example, the Association of Bay Area Governments (ABAG) projected that the nine-county Bay Area would add about 473,000 new jobs between 1988 and 1995. On that basis, the Bay Area Council argued, ABAG’s regional housing goal of 288,000 over that period was about 50,000 units too low, assuming a standard of 1.4 workers per household (Bay Area Council, 1988a). However, from the standpoint of local governments, job growth projections are sometimes seen as unrealistically high. Some local officials have called for a mediation process as part of the RHNA procedure, possibly with some third party evaluating disputes between local governments and COGs regarding the allocation of housing goals (Senate Committee on Local Government, 1993, p. 2). Some objections to RHNA outcomes are the result of disagreement with the decisionmaking of COGs, which ultimately reflects a political process. One critic argued that COG allocation processes allow objections by local governments after the initial allocation, but the communities that can do the staff work to mount a serious objection tend to be large cities with large staff resources. In this case, the objection continues, the unwanted units are likely to be reallocated to smaller and politically weaker communities, which are often older, poorer, or built-out inner suburbs (Senate Committee on Housing and Land Use, 1995, pp. 56–57). The zero-sum quality of RHNA allocations puts COGs, which are constituted as cooperative organizations of local governments, in a difficult position. They are left to carry out a state mandate while still feeling pressure from their local government members who feel that their housing allocations are excessive. HCD officials have said that the COG is the level at which intraregional disputes should be worked out, but given their weak political basis, most COGs have not been able to take a strong, autonomous approach. “Without independent authority, regional agencies have little wherewithal to overcome fundamental conflicts between local and statewide interests” (Barbour, 2002, p. 65). Or, as a senior planner for the Association of Bay Area Governments put it, “From ABAG’s perspective, the process has been a serious political liability for a voluntary membership organization” (Senate Committee on Local Government, 1993, p. 48). 26 Assembly Bill 438 (Torlakson), signed into law in 1998, does allow the RHNA process to be further decentralized within a region. If a county and all the cities within it (or a joint powers authority or subregional agency established by the COG) request to undertake the allocation of housing goals within a subregion, then the COG is obliged to turn that responsibility over to this smaller body, after assigning the subregion its proportionate share of the region’s housing goal. This option has not yet been widely utilized, perhaps because of time constraints during the current cycle of housing element revisions (see Richards, Watson & Gershon, 1999). The Current Revision Cycle: The Controversy Continues Concerns about the housing element law were never fully resolved after the previous round of housing element updates ended in the early 1990s. But the topic largely fell from the radar screen in the middle of the decade, because of the state’s difficulty in emerging from the recession and its associated budget problems. With state funding of housing element activity postponed during the period of budget exigencies, required housing element update activity was suspended for six fiscal years, from 1992–93 until being resumed in 1998–99. Stopgap legislation was passed during this period to extend the deadlines for regions and localities to undertake their updates. When the process started up again, however, the old disagreements quickly resurfaced, even as housing problems in much of the state had become more acute. According to planning expert William Fulton (1999, p. 114), “Most people involved in the housing element process agree that the law is ineffective: it is overly bureaucratic and exacting, has too many loopholes, and even a good housing element is no guarantee that affordable housing will actually be built in any given community. But attempts to reform the law have failed repeatedly in recent years.” One of the political tensions involved in housing element policy, Fulton argues, is standoff between housing interests (both forprofit developers and antipoverty nonprofits), who are prominent and influential at the state level, and local governments, who often see the 27 housing element law as unwarranted and clumsy meddling by state government in home rule, since land-use policy has long been the almost exclusive purview of local governments. Many local officials did not look forward to another cycle of revisions of local housing elements. In fact, both the League of California Cities and the California State Association of Counties opposed funding the RHNA mandate at all unless the statute was reworked; in turn, the governing board of the state’s largest COG, the Southern California Association of Governments (SCAG), voted to oppose state funding for the RHNA unless the local jurisdictions of the region no longer opposed it (Richards, Watson & Gershon, 1998). RHNA was funded despite these protests. After SCAG completed its allocation process in 2000, local governments in Riverside and San Bernardino Counties sued HCD as well as SCAG, arguing that the housing unit goals assigned to the Inland Empire were unrealistically and inequitably large. Although inland communities hold most of the vacant land in the metropolitan region, they argued that an unwillingness to accept housing among the coastal counties of Orange, Los Angeles, and Ventura led to excessive units being assigned to Riverside and San Bernardino. This case, which went before a Riverside County Superior Court judge in January 2002, represents perhaps the largest breakdown of intergovernmental relationships during the current round of housing element updates and has drawn particular scrutiny to the RHNA process. In short, the argument is that the regional allocation can become highly politicized and works to the detriment of less-powerful jurisdictions (Little Hoover Commission, 2002, p. 19). Elsewhere, officials in the San Luis Obispo region indicated publicly that they consider the RHNA goals assigned to that area completely unattainable (Lyons, 2002a, 2002b). More generally, some county government officials argue that it does not make planning sense for counties to be “treated like cities” in the housing element process, given the greater capacity of incorporated cities to accommodate new growth. Counties are limited in their ability to provide city-like services, and often seek to protect resource and agricultural lands from intrusion by urban uses (Senate Committee on Local Government, 1993, p. 89). 28 In the Bay Area, for example, Napa, Sonoma, and Solano Counties all attempt to direct growth away from farmland and open space and into cities. A Napa County supervisor notes, “Consider the fact that we’re an agricultural county, and the fact that the state always laments the loss of agricultural land, and the fact that HCD equates agricultural land and open space as land available for housing. There is a huge disconnect there” (quoted in Shigley, 2002, p. 1). In Napa County, the unincorporated area was assigned 28 percent of the housing unit goals for the entire county; unincorporated Sonoma received 30 percent of that county’s allocation. HCD officials, in turn, point out that it is up to the regional council of governments to allocate housing goals among cities and counties in the region, and the state would prefer that housing be developed where infrastructure is available and sprawl is minimized. COG planners note that housing goal allocations were based in large part on projected job growth, and some counties seem uninterested in restricting commercial development even as they bemoan a lack of capacity to take housing (Shigley, 2002, pp. 14–15).7 Another major debate has concerned the extent to which localities can count rehabilitation or subsidization of existing housing units for low-income households toward their housing element goals. Affordable housing advocates and developer groups have argued that new construction is the primary need, and the alteration of existing privatemarket units to subsidized or affordable units does little to ameliorate the overall supply problem in California. Local governments make the case that rehabilitation is overtly encouraged by the statute and is often the most efficient way to make affordable units available, particularly in fully developed communities. AB 438 of 1998 does allow a local government to count certain types of rehabilitation of substandard units toward its housing element goal, although its provisions are quite restrictive.8 ____________ 7In Napa, city and county officials have formed a working group to see whether some of the county’s housing unit goals can be shifted to cities, perhaps in exchange for allowing cities to annex some industrially zoned land near the county airport. Transfers of housing allocations between jurisdictions are permitted under changes to the law in the 1990s, but the requirements on such “trades” are relatively strict (Curtin, 2000, p. 11). 8For example, the housing units in question must be unfit for human habitation before the renovation, the units must be made affordable to low- and very-low-income 29 Local officials have also argued that they should be able to count the creation of shared living facilities, such as congregate care units for seniors, toward their housing goals. Generally, HCD has counted new dwellings as housing units only if they each contain their own kitchen. Some communities have converted older motels into affordable apartments, installing kitchenettes in the process to meet housing-unit requirements (see Box 2.1). Box 2.1 A Creative Way to Help Meet Housing Goals Several communities in California have acquired (or helped fund the acquisition) of old motels for conversion into affordable residential units. This may present a seemingly simpler method to generate new housing that counts toward housing element goals, without perceptibly increasing the density of the existing community. Potentially, the renovation of the motel may actually improve the neighborhood. Nevertheless, motel conversion too has its complications. Kitchens must be installed in each unit for it to qualify as a housing unit, given the existing interpretation of the housing element statute. But major modifications to an older building may necessitate costly construction changes to make the building more accessible, because of the requirements of the Americans with Disabilities Act. Planners may find that the converted motel violates local parking or fire code requirements for residential dwellings unless further major renovations are made. School fees may need to be paid, and seismic safety upgrades made. If the motel conversion displaces existing low-income tenants using the motel as temporary housing, the city or developer will likely have to pay for their relocation. Finally, there is the ever-present issue of local growth politics; neighbors—including nearby businesses worried about a loss of motel customers—may object to the conversion (discussion draws from “Motel Conversion,” 2002). The Defense of Housing Element Law Housing advocates and public-interest law firms often defend the necessity of the housing element law while still recommending reforms. They point out that the housing element law and other state requirements for local comprehensive planning emerged in large part ________________________________________________________ occupants, and former occupants must be given priority to move into the rehabilitated dwellings. Cities and counties are also now permitted to “buy down” market rate apartments (of at least 16 units) to make them affordable to low-income occupants for at least 30 years. Such local government efforts toward affordability covenants, rehabilitation, and preservation of at-risk units can be credited for up to 25 percent of the community’s housing goals. There are also strict time deadlines to implementing these programs (Richards, Watson & Gershon, 1999). 30 from frustrations with local exclusion of affordable housing or multifamily units (see Box 2.2 for an example). A representative of Catholic Charities testified that the housing element is often the driving force for localities to create policies to increase affordable housing opportunities (Senate Committee on Local Government, 1993, p. 14). An attorney for Legal Services of Northern California argued that local officials in at least four Sacramento area communities were able to use the state’s requirements to convince “not in my back yard” neighborhood groups that sites had to be made available for homeless shelters and transitional housing (Senate Committee on Local Government, 1993, pp. 118–119). And a senior attorney for the Legal Aid Society of Alameda County pointed to specific instances where the housing element requirement generated local policy changes: As a result of the process, I’ve seen the City of Alameda lift its prohibition on multifamily affordable housing development and commit funds to affordable Box 2.2 A Housing Element Challenge in Folsom Folsom is a rapidly developing suburban city in Sacramento County, growing from 11,003 residents in 1980 to 51,884 in 2000. The city had a net increase of 8,550 housing units during the 1990s, according to U.S. Census data, but failed to add affordable housing; home prices were the highest in the county. Housing activists argued that the city was ignoring its responsibilities to the state and the region. Sued by Legal Services of Northern California in 2001 for its noncompliant housing element and lack of progress toward affordable housing goals, the city entered a settlement in 2002. Under terms of the agreement, Folsom is to rezone 128 acres of land to create a “land bank” for the possible construction of up to 2,900 affordable units. The city also agreed to add an affordable housing set-aside requirement for new developments and create an impact fee for nonresidential growth to generate money for an affordable housing trust fund. At a May 2002 community meeting to receive input on the city’s new development goals, 400 residents attended an emotional meeting, “with some audience members clutching signs demanding a ‘total recall’ of the City Council” (Hecht, 2002). Although local officials assured the residents that no decisions had been made on where to locate the rezoned land, and stressed that the affordable units would house groups such as teachers, firefighters, and retail clerks, many attendees were not reassured. One homeowner argued, “This is America. These are our homes. These are our investments. I don’t need crime. Put it in your neighborhood” (Hecht, 2002; also draws upon Padmanabhan, 2002, and other Sacramento Bee articles). 31 housing development; I’ve seen the City of Healdsburg commit to expand its sewer capacity and increase densities; I’ve seen the City of East Palo Alto abandon demolition of affordable units until it identifies the resources to replace them; and I’ve seen the County of Madera commit to forming a joint housing authority with the City of Madera to facilitate affordable housing development (Senate Committee on Local Government, 1993, p. 108). Likewise, in a 2002 report, the Little Hoover Commission (p. 18) called for an “assertive stance” by the state on the housing element law. The commission took the position that “communities have more opportunities than they recognize or acknowledge” to enable housing development, pointing to such policy options as streamlined permitting procedures, public-private partnerships, and subsidized water and sewer fees for affordable housing projects (p. ii). The commission argued for stiffening penalties against noncompliant jurisdictions, “a longer wait” in the competition for transportation and park bond funds, and a limitation on use of redevelopment agency funds (p. iii). Senate Bill 910 (2001–02) As noted in Chapter 1, the most recent debate over housing element law in California centered over Senator Joseph Dunn’s proposal to strengthen the state’s hand in dealing with local governments with noncompliant housing elements. After passing the senate by a vote of 22 to 12, SB 910 was held in the assembly for much of 2001 and 2002 as a Housing Element Reform Working Group met to attempt to decide on a consensus approach. As the deadline for taking action on the bill approached, it was amended in a number of ways. The groups representing cities and counties still balked at certain provisions in the resulting bill, however. Ultimately, the bill died in the assembly, as that chamber refused to suspend the deadline for the bill to emerge from the policy and fiscal committees. Although it did not pass, the bill galvanized debate on the issue and may be resuscitated in new form in a future legislative session. SB 910 sought to alter the existing approach to noncompliance under the housing element statute in two major ways: 1. SB 910 required that the state controller fine cities and counties that either (a) fail to submit an adopted housing element within 32 six months of their deadline, (b) fail to revise their housing elements, if found noncompliant by HCD or a court, or (c) have their compliance certification rescinded by HCD because they fail to take the actions called for in the housing element. The fine was set at $5,000 per month, or 25 cents per month per resident of the jurisdiction, whichever is greater. Fines would accrue to a Housing Supply Account, to be appropriated by the legislature to assist multifamily developments. State courts could reduce fines in mitigating circumstances, such as if noncompliance were due to procedural rather than substantive problems. 2. If a court finds a housing element noncompliant, SB 910 would have directed it to (in addition to any remedy it orders) levy the fine specified above and also award attorney’s fees to the plaintiff, if the plaintiff is a public-interest organization. Courts trying housing element cases would be instructed to give HCD’s findings of noncompliance “great weight.” These provisions distressed representatives of local government, who considered the bill too punitive. They were countered by arguments that the primary shortcoming of the existing housing element statute is its lack of a “stick” to encourage or compel local compliance.9 Beyond the penalties for noncompliance, however, numerous elements in the bill were provisions that had been sought by local and regional officials to smooth the process of regional housing needs distribution. Some of the more important elements are summarized in Box 2.3. These provisions ____________ 9Indeed, some of the most exclusive communities in California can apparently thumb their noses at the issue of housing element compliance. Policy practitioners contacted by the author generally agreed that in largely built-out cities with little development activity occurring, no sizeable low-income population, and no active redevelopment effort, the local government appears largely immune from litigation under the housing element law. With no open land zoned for multifamily development or available at reasonable land prices, no nonprofit developers could propose a project that could become the basis of a court challenge. The exclusive communities also may not mind that their noncompliant housing element disqualifies them from certain housingassistance programs, given that they probably do not wish to have subsidized housing in the first place. 33 Box 2.3 Other Important Provisions of SB 910, As Amended • In allocating housing needs to the various regions of the state, HCD would be required to consider a region’s population projections under its regional transportation plan, along with projections by the state Department of Finance. A process would be set in place for resolving disagreements in situations where the Department of Finance projections and the regional projections differ by a substantial amount. To jibe with the three-year cycle of regional transportation planning, housing element updates would be required every six years, rather than the current five. • COG formulas for allocating housing needs to cities and counties would be required to consider new factors, such as localities’ previous performance in meeting needs, issues of jobs/housing balance, opportunities for infill development, lands protected under open space or farmland programs, federal and state laws restricting development, intracounty agreements to steer growth into cities, and opportunities to maximize use of public transportation. • Localities would be allowed to petition for a revised share of regional housing need. If the request is accepted, and the adjustments amount to less than 7 percent of regional needs, the units in question would be redistributed proportionately to all other local governments. • Public involvement requirements would be enhanced throughout the housing element planning process. However, in reviewing local housing elements, HCD would not be allowed to consider objections to local plans by groups or individuals if the point in question were not raised during the public comment period (unless it is based on new information or an amendment to the element). seemed likely to make what is already a detailed and complex housing element statute even more so. Conclusion The inability to come to consensus over SB 910, despite attempts to include all relevant stakeholders in discussions over the bill, illustrates the long-simmering standoff between state and local officials. Housing element policy and enforcement has become a prominent source of intergovernmental antagonism. Given the different perspectives regarding the reasons for widespread local noncompliance with housing element law, it is important to investigate which community characteristics might contribute to noncompliance. The next chapter addresses this question. 34 3. Why Do So Many Communities Have Noncompliant Housing Elements? One of the most contentious aspects of the housing element requirement is the fact that a large percentage of communities are not compliant with the law—nearly four cities in ten and more than a quarter of counties.1 As discussed in previous chapters, the perspectives as to why noncompliance is widespread differ dramatically between housing advocates and state officials, on the one hand, and defenders of local governments on the other. This chapter examines compliance status among the state’s cities to provide a more informed basis for evaluating why local governments are noncompliant. I use measures of cities’ social and demographic characteristics, land-use patterns, and local policy measures to identify factors that distinguish compliant and noncompliant communities. Why Cities? Although counties are also critical in meeting housing needs, the analysis concentrates on cities for a set of practical reasons. First, most population and housing growth occurs in cities. During the 1990s, 82 percent of the state’s population increase occurred in cities that existed as of 1990; population within city boundaries increased by an additional 12 percent as a result of 18 new municipalities that incorporated between the 1990 and 2000 Census. ____________ 1This tally of noncompliant jurisdictions includes those whose housing element is overdue as well as those found noncompliant after HCD review. In both cases, the housing element is legally noncompliant. 35 Second, statistical considerations make rigorous analysis of counties problematic. There are only 58 counties in the state (one of which, San Francisco, is also a city), which reduces the leverage of statistical analysis. By comparison, there are nearly 480 cities in California. More important, Census demographic data on counties pertain to the county as a whole, whereas county housing elements pertain only to unincorporated areas. One would thus expect that the characteristics of the unincorporated areas have a very major influence on the county’s housing policies and, ultimately, compliance. Some counties have numerous unincorporated areas in far-flung locations, each of which may be quite distinct from the others. In short, there is a “units of analysis” problem in studying county compliance with the housing element law. Finally, studying cities allows for use of PPIC’s extensive existing database on city characteristics. Compiled over several years for other projects, this dataset includes not only Census variables but some measures of city budgetary revenues and political and policy attributes. A 1998–99 mail survey of local planning officials on urban development topics also permits an analysis of local government policies and perspectives on growth.2 I have combined these earlier databases with HCD’s list of cities’ compliance status to examine relationships between community conditions and housing element compliance. Four Sets of Factors Potentially Affecting Compliance At least four different reasons are given for cities failing to comply with housing element law to HCD’s satisfaction. Note that these four reasons are not mutually exclusive; all may apply to some degree. ____________ 2The survey of planning directors was jointly undertaken by the author and Max Neiman of the University of California, Riverside. It specifically addressed local residential policies and growth management and achieved a response rate of 76 percent in the major regions of the state in which it was undertaken (the Central Valley, Southern California—here including San Diego and Santa Barbara as well as the Los Angeles area—and the San Francisco Bay Area). The survey was initially undertaken for purposes not involving housing elements, and the term “housing element” was not mentioned in the questionnaire. This means that respondents probably had little reason to let that controversial topic influence their responses. For more details on the survey and its results, see Lewis and Neiman (2000). 36 Community Social Status and Exclusion “Although they won’t say so publicly, some of these cities don’t care what their assigned [housing goal] numbers are. They just won’t do it because they don’t want low-income housing in their jurisdictions.” —Senator Joseph Dunn, quoted in Wisckol (2002). Senator Dunn authored SB 910 in an attempt to hold local governments more responsible for noncompliance. In the newspaper interview quoted above, he articulated forcefully what some critics of local housing policies have long implied: that cities are reluctant to make the necessary plans to accommodate sufficient housing—particularly multifamily or affordable developments—because of a desire for social exclusion. Such exclusion, highlighted in some national studies, is often thought to rest on the attempts of upper-income, homeowner-dominated communities to insulate their cities from lower-income populations or rental housing more generally. Whether rooted in a desire to cluster with neighbors of similar characteristics, or in racial, ethnic, or socioeconomic snobbery, these efforts are associated with what is often called a NIMBY (“not in my backyard”) attitude. It is also possible for low-income or diverse communities to pursue NIMBY policies, however. The exclusionary approach to land-use policy is often associated with homeowners’ fears of reduced property values if rental or affordable housing is permitted in the neighborhood. It is also often linked with the aspirations of suburbanites to associate with people of similar socioeconomic status and to separate themselves from the perceived problems of large central cities. For example, the mayor of Menlo Park, a Silicon Valley community, has been quoted as saying that his city’s residents “desire to have neighbors who live in homes similar to their own” (Nguyen, 2001). To understand the connection between community status and housing element compliance, one can examine such characteristics as the city’s median household income, the homeownership rate, or the 37 percentage of the population that is white and non-Hispanic. Homeownership rates are of particular interest, because some authors have attributed local antigrowth policies to cartel-like behavior on the part of homeowners, who wish to boost their housing values by restricting the supply of new housing (Brueckner, 1995). Fischel (2001) argues that homeowners’ intense desire to protect the value of their main asset—their home—underlies most local land-use policies. Local Land-Use Characteristics and Vacant Land “Some cities are land-locked with growing populations and little available land; other cities are facing booming job markets with high-paid workers where the competition for housing drives the price of the smallest cottage over half a million dollars.” —Letter from the League of California Cities and California State Association of Counties to Senator Joseph Dunn, opposing SB 910. Many city governments argue that they lack sufficient land resources to accommodate the number of new housing units required under their allocations from the regional COG. This argument can be quite powerful. Many older communities, often inner-ring suburbs, are essentially “built out,” meaning that they lack any substantial vacant land on which to construct new buildings. In such circumstances, cities with significant housing unit goals are left with the more difficult, conflictridden, and potentially expensive options of rezoning existing neighborhoods for higher-density housing or using redevelopment powers to change current land uses so as to include more housing units. In a meeting at HCD’s Division of Housing Policy Development, department staff indicated to the author that the requirement to provide a sufficient inventory of appropriately zoned land sites perhaps does the most to “trip up” local governments in their quest for compliance; many localities prove unable or unwilling to produce such an inventory. The staff indicated, however, that local governments often have an inflexible 38 notion of vacant residential land, as many cities have vacant sites that are zoned for industrial or commercial use and could be rezoned to accommodate housing. Highly developed communities may also be dominated by single-family residential areas that could accommodate additional units. For example, such cities could adopt a more liberal policy on secondary (“in-law”) units3 or could “upzone” single-family areas, where appropriate, to allow more apartments and condominiums. Nevertheless, it is also true that many of such older, inner communities in California have already felt the painful externalities of high density, as their populations have grown so rapidly as to create traffic and parking problems and strain public services. Such communities are often at the receiving end of immigration and other demographic changes, and communities like Santa Ana have expressed concerns about fire hazards and sanitation problems resulting from a profusion of apartments—often overcrowded—and legal or illegal secondary units in an already dense built environment (Perkes, 2002). Several variables can shed light on the perspective that local land-use characteristics affect housing element compliance. Such measures could include the population density of the city and the age of its housing stock (likely to be indicative of a lack of “fresh” land sites in many communities). Although there is no database of vacant land in communities statewide, the survey of local planning officials provided a helpful surrogate measure. Respondents were asked to rate the importance of land supply as a constraint on housing development in their city, on a 1-to-5 scale.4 ____________ 3It has long been state policy to ease local restrictions on secondary units. Most recently, AB 1866 (Wright), signed into law in 2002, requires that local governments provide for ministerial approval of applications (meaning that no discretionary reviews are allowed) for secondary units that meet the local ordinance. 4Although the measure is somewhat subjective, it is reassuring to note that it is highly associated with another measure of local developable land. In a 1998 survey of California city managers, respondents were asked to indicate whether their city was built out (meaning it had “little or no vacant land available”). This (dichotomous) measure from the city manager survey is correlated with the five-point measure from the planner survey of the constraint posed by low land supply at r = .45 (prob. < 0.000). 39 The Resources of the Local Government “My clients are typically smaller communities without the staff expertise or other resources of the state’s largest cities and counties, and complying with the Housing Element statute can be a Herculean task for these communities.” —Michael Colantuono (a city attorney), quoted in Senate Committee on Local Government (1993, p. 67). From almost any perspective, engaging in a full-scale revision of a local housing element is a long and complex process, calling for considerable staff time and capabilities. Many communities spend considerable sums to engage outside consultants to prepare housing elements. For small cities, those without many professional staff, or those that are fiscally strained, resource limitations may significantly hinder the preparation of an adequate and timely housing element. Furthermore, virtually all observers of the housing element law agree that “performance” in producing housing for low- and moderate-income households depends on the availability of subsidized housing funds. In particular, units for very-low-income families are almost impossible to build without heavy subsidies. Both the federal and state governments cut back housing assistance after a peak in the 1970s, and competition for such funds has since been very keen. Few communities can support housing subsidies from their own general-fund budgets, although some larger or wealthier jurisdictions do engage in some local funding of affordable housing. Thus, a lack of local government resources is seen as hindering genuine local efforts at compliance. HCD staff counter that the department has put considerable time and resources into technical assistance, with a particular effort at helping small cities and counties attain compliance. In some cases, HCD staff analysts have been dispatched to communities and taken the lead in helping the locality to prepare its initial draft. They indicated that small, rural communities often have an easier time reaching compliance than larger cities, because the former are often more enthusiastic about future growth. A number of city characteristics may be relevant in measuring local government resources and planning capacity. These include the city’s 40 population size, planning staff resources, and revenue base. Staff resources are implied by local planning officials’ responses to a question regarding the degree to which “lack of personnel to review projects” plays an important role in “constraining or slowing residential development in your city.” Fiscal resources are measured as the city’s own-source, general revenues per capita.5 As a secondary fiscal measure, I estimate the percentage share of locally generated property taxes devoted to the municipal government (as opposed to school districts, the county, or other local governments). It is often thought that cities with a larger share of local property tax revenues will be more accommodating to residential development, because such development will come closer to paying its way. Local Politics and Residential Growth Policies “I want housing in the community, but I want it to reflect the amount that can be served. . . . I’d have to throw out all our growth-management policies [to meet the city’s assigned housing goal]. . . . I think the voters would throw out the City Council.” —Allen Settle, Mayor of San Luis Obispo, quoted in Lyons (2002b). A city’s social status, land-use characteristics, and government resources may well affect its chances of achieving housing element compliance. But local growth policies and development decisions are ultimately political choices made in the context of a city’s political traditions. One aspect that HCD staff members review in examining local housing elements is whether the community has taken full account of the “governmental constraints” on housing development and whether the city has a plan of action to reduce or overcome those constraints. ____________ 5This is essentially a measure of total revenues per capita, from which are subtracted intergovernmental funds and “functional” or “enterprise” revenues, such as current service charges. 41 Evidence from a previous PPIC study indicates that antigrowth politics and efforts to manage or shape growth have emerged in strong form in a fairly small number of California communities, in a weaker or more sporadic form in many others, and not at all in some cities (Lewis and Neiman, 2002). Residential growth-management policies and citizen initiatives that seek to slow growth, along with lengthy city permitting or approval processes, may give HCD reviewers cause for concern. More indirectly, specific controversial development projects, or a history of disputes between city officials and citizen groups on growth issues, may contribute to community rancor that makes officials wary of advancing a plan for additional housing development. Where growth is a hot-button issue, production and affordability goals may take a back seat to political exigencies, which in turn could lead to a thumbs-down from HCD housing element reviewers. The survey of local planners provides several measures of local government policies and orientations toward growth and housing; specific items will be explained as they are introduced in the analysis. One of the most important measures is a count of the number of overtly restrictive growth-management policies pursued by the city. Such policies include: • A substantial recent reduction of residentially zoned land, • Annual limits on building permits, on housing units, or on multifamily dwellings, • A formula for allowable numbers of new housing units, • An official “population ceiling” for the community, or • A moratorium on building permits or water or sewer connections.6 One might assume that an antigrowth approach simply reflects community status, one of the other potential causes of housing element noncompliance. But the earlier PPIC study indicates that local ____________ 6The survey also measured whether the city had any of nine other forms of residential growth management. However, these other policies are less overtly restrictive of residential growth and showed no evidence of effects on housing element noncompliance in regression analyses. For a discussion of the various measures, see Lewis and Neiman (2002), Chapter 3. 42 socioeconomic status was only a weak predictor of whether a community engages in restrictive residential policies (Lewis and Neiman, 2002). Timetable for Housing Element Updates Influences Compliance Status An initial step in analyzing which characteristics are related to noncompliance involves comparing compliant and noncompliant cities across several dimensions. Before undertaking this comparison, however, it is important to point out a potentially severe complication for this exercise: Not all local governments are on the same schedule for updating their housing elements and having them reviewed by HCD. The Fresno, Bakersfield, and Sacramento regions, at this writing, were the most recent metropolitan areas to be scheduled for updates and reviews (see Table 1.1). Thus, localities in these areas, where housing element revisions were due by June 30, 2002, have had only a few months to receive a favorable opinion from HCD, and some jurisdictions have yet to submit their revision. Cities in the San Francisco Bay Area have had about nine months and those in the SCAG region nearly two years to attain compliance, compared to nearly three years in the San Diego region. By contrast, localities elsewhere in the state had not faced a housing element revision deadline for over a decade, as of HCD’s September 25, 2002, report on compliance status, on which this analysis is based. Thus, such cities have had a great deal of time to work with the department and bring their housing elements into compliance. Comparing all cities in the state that are compliant to those that are noncompliant, then, might be akin to comparing apples with oranges. In the regression analysis below, I will make special efforts to control statistically for such scheduling differences. Statewide, fewer than half of the jurisdictions that have submitted draft elements for review since 1999 are in compliance, whereas the vast majority of jurisdictions whose status reflects pre-1999 housing elements are compliant. 43 Comparing Compliant and Noncompliant Cities: A Preliminary Profile With this caveat in mind, this section compares compliant and noncompliant localities in the regions that have already been required to undertake housing element revisions in this round (San Diego, SCAG, San Francisco Bay Area, Fresno, Kern (Bakersfield), and Sacramento regions). A second comparison is then undertaken for cities in the other portions of the state. Included in the noncompliant category are those cities that have received an opinion of noncompliance from HCD as well as cities that are overdue in submitting their draft housing elements to the department. For many years, the department’s compliance status report referred to such localities as having an “obsolete” housing element, although they are now referred to more gently as being “due.” Legally, they are not in compliance. Table 3.1 presents the profile of compliant cities and noncompliant ones for the regions that have already been involved in housing element updates. The three measures of community status all indicate that noncompliant cities tend to be of slightly higher status. These comparisons are mildly supportive of the “social exclusion” perspective on housing element noncompliance. Nevertheless, the differences are not great, and statistical testing indicates that none of the socioeconomic characteristics are statistically significantly higher in noncompliant than in compliant communities. Regarding local land-use characteristics, the housing stock of noncompliant cities tends to be older than that in compliant cities. This difference is statistically significant. Otherwise, there is no support in this comparison for the argument that a lack of vacant land underlies noncompliance. Noncompliant cities in these regions actually have lower population densities, on average, than compliant cities, and their planners are no more likely to say that land supply has constrained residential growth. In this simple comparison, the “lack of resources” reason for noncompliance receives little support. Although noncompliant communities are significantly smaller in average population size, they 44 Table 3.1 Comparison of Compliant and Noncompliant Cities in the San Diego, Southern California, San Francisco Bay Area, Fresno, Kern, and Sacramento Regions Community Status % homeownership % white and non-Hispanic Median household income, $ Mean Value Compliant Noncompliant 60.2 48.0 54,780 62.7 51.5 59,913 Land-Use Characteristics Population per sq. mi.* Supply of land constrains development, per planning director (five-point scale) Median year housing built (as of 1990)*** 5,111 3.8 1969 4,322 3.6 1965 Resources and Capacity Population* No. of planning staff per 10,000 population City own-source general revenues per capita, 1993, $ City’s % share of property tax revenues* 64,160 1.9 507 11.5 46,788 2.1 593 13.1 Local Policies/Politics No. of restrictive growth-management policies (out of seven possible) City requires that residential projects include affordable component, % Citizen initiatives have slowed growth, according to planning director, % Level of citizen opposition to growth, according to planning director (five-point scale) 0.52 32.3 18.3 2.7 0.72 34.3 18.2 2.8 SOURCES: Author calculations based on data from the 2000 Census, the 1990 Census, a 1998–99 mail survey of city planning directors, and the HCD compliance list as of September 25, 2002. NOTES: *p < 0.1; ***p < 0.01. Asterisks indicate statistically significant differences in means between compliant and noncompliant cities, using a two-tailed ttest. Demographic measures are for 2000, except for median year of housing. For population comparison, cities of over one million residents are omitted. For city ownsource general revenue comparisons, cities of less than 1,000 residents are omitted to reduce the influence of extreme outliers in the comparisons. Geographic makeup of regions is listed in Table 1.1. There are 299 cities with usable housing element status data in these regions, although sample sizes are smaller for survey-based items because of survey nonresponse. 45 have roughly the same number of planners per 10,000 residents as compliant cities, and they also receive more own-source revenues than compliant cities. The percentage share of local property tax revenues is also greater in noncompliant cities, despite frequently expressed arguments that a greater share of the property tax base would make cities more accommodating of housing. Finally, the comparison finds muted differences between compliant and noncompliant cities regarding local growth policies and politics. None of the four measures from the planners’ survey shows significant differences between the two groups of municipalities, although the average noncompliant city has about 0.2 additional restrictive growthmanagement policies. Table 3.2 repeats this comparison, but this time for cities in parts of the state that had not yet been required to submit housing element updates. Such regions constitute only a small share of the state’s population and include the Central Coast area, portions of the Central Valley, and the rural north and mountainous counties of the state. In these regions, only a small share of cities were noncompliant (16 percent), which is perhaps not surprising since most had had about a decade to attain compliance. In these parts of the state, differences among the compliant and the (relatively few) noncompliant communities tend in many cases to be starker. For example, all three measures of community status are higher, at a statistically significant level, in the noncompliant communities. Likewise, most measures of local growth policy and politics show substantial differences between the two sets of communities, although this comparison can be only suggestive, as most cities in these regions were not sent the planner survey.7 The average noncompliant city has substantially more growth-management policies and a much higher level of citizen opposition to growth and use of citizen antigrowth initiatives than the average compliant city. The comparison between compliant and noncompliant cities is again murky when it comes to local land-use ____________ 7In the regions discussed in Table 3.2, only cities in the Central Valley and Santa Barbara County would have received the survey. Of the 125 cities compared in the table, 48 responded to the survey. 46 Table 3.2 Comparison of Compliant and Noncompliant Cities in the Remainder of the State Community Status % homeownership** % white and non-Hispanic** Median household income, $** Mean Value Compliant Noncompliant 57.2 60.9 35,341 62.1 73.7 44,441 Land Use Characteristics Population per sq. mi. Supply of land constrains development, per planning director (five-point scale)a Median year housing built (as of 1990) 2,686 2.5 1966 2,648 3.1 1967 Resources and Capacity Population No. of planning staff per 10,000 populationa City own-source general revenues per capita, 1993, $ City’s % share of property tax revenues 20,307 1.7 466 14.8 24,261 2.4 478 15.4 Local Policies/Politics No. of restrictive growth-management policies (out of seven possible)a*** City requires that residential projects include affordable component, %a Citizen initiatives have slowed growth, according to planning director, %a*** Level of citizen opposition to growth, according to planning director (five-point scale)a*** 0.43 19.5 2.5 2.3 2.57 14.3 42.9 3.9 SOURCES: Author calculations based on data from the 2000 Census, the 1990 Census, a 1998–99 mail survey of city planning directors, and the HCD compliance list as of September 25, 2002. NOTES: **p < 0.05; ***p < 0.01. Asterisks indicate statistically significant differences in means between compliant and noncompliant cities, using a two-tailed ttest. Demographic measures are for 2000, except for median year of housing. For city own-source general revenue comparisons, cities of less than 1,000 residents are omitted to reduce the influence of extreme outliers in the comparisons. This table relates to cities in all parts of the state not covered by Table 3.1. There are 125 cities with usable housing element status data in these regions, although sample sizes are considerably smaller for survey-based items because the survey was not directed at cities in some parts of the state and planners from some cities did not respond. aLimited number of planner survey responses in these regions make comparisons on these items less reliable. 47 characteristics and local government resources and capacity. Here, none of the differences are statistically significant. A decade after these cities were initially asked to submit their last updates to HCD, only a relative handful of laggard communities were unable to reach compliance. The simple comparison indicates that such cities tended to be relatively well-off and, so far as we can tell, experienced substantial local turmoil over residential growth. Probing Further: A Statistical Model of Noncompliance The comparison above is suggestive, but provides only a simple sketch of noncompliant communities. To examine more systematically the factors that are associated with noncompliance, I undertook a multivariate analysis. Tables presenting the detailed results of the final estimations are in Appendix A. This section describes the analytic approach and the next section summarizes the most relevant results. The logistic regression method used is able to identify the relationship of various city characteristics to noncompliance, while holding constant each of the other characteristics represented by variables in the model. As suggested above, the four types of city characteristics of interest include measures of community social status, local land-use and development characteristics, resources of the city government, and measures of local growth policy and politics. Cities whose housing elements are overdue for HCD review are again treated as noncompliant.8 I analyzed compliance as of September 25, 2002. Cities under review by HCD as of that date are dropped from the analysis. A set of preliminary analyses helped identify which specific variables measuring these characteristics would be most likely to contribute to an ____________ 8An additional (ordered logit) model was estimated to examine housing element compliance on a three-point scale: in compliance, overdue, and out of compliance. The results of this model were very similar to the results discussed for the dichotomous measure (compliant/noncompliant). 48 explanation of noncompliance. The variables ultimately used in the model are listed and described in Table 3.3.9 I used two techniques to control for differences that relate to the characteristics of the region rather than of the city itself. First, I account for the “timing issue” by including a measure of the number of months that elapsed since the deadline for sending housing element updates from the city’s region to HCD. This variable will indicate whether the simple passage of time makes it more likely that cities will be able to reach compliance. Second, indicator variables are used to denote cities in particular areas. I control for location in the San Francisco Bay Area, the 18-county Central Valley region, and the Inland Empire (Riverside and San Bernardino Counties). The Bay Area had gone through the review process quite recently. The Inland Empire was engaged in a legal dispute with SCAG and HCD over its housing allocation numbers. Both regions have low compliance rates. Cities in the Central Valley face challenges of rapid growth, relative poverty, and, often, fiscal constraint. The measures of local policy—such as the number of strict growth controls and degree of planning staff shortages—are taken from the PPIC mail survey of planning directors in the Bay Area, Central Valley, and metropolitan Southern California (including Santa Barbara County as well as the SCAG and San Diego regions). For this reason, the analysis must exclude cities outside those regions. (The excluded regions account for a small part of the state’s population.) Likewise, the 24 percent of cities in those regions whose planning directors did not respond to our survey cannot be included. ____________ 9Measuring socioeconomic status at the city level creates some difficulties in such a model because of the high collinearity among such measures as income, education, race, and homeownership. After some exploration, I decided to include two relevant variables: the homeownership rate of the city and the percentage of the population composed of non-Hispanic whites. These variables capture two important measures of status and diversity, respectively, that might affect local housing policies, but the variables are not too highly correlated to disentangle (r = .44). Where median income was substituted for homeownership, results did not differ greatly. There was also no evidence that the proportion of the population composed of Hispanics, African Americans, or Asians was related to noncompliance. 49 Table 3.3 Variables Used to Predict Housing Element Noncompliance Variable Definition/Notes % owner-occupied % of housing units owner-occupied % white % of city residents who are white and non-Hispanic % recreational housing % of housing units that are for recreational or seasonal use Population density Natural logarithm of the city’s population per sq. mi. Age of housing Age of median housing unit, in years, as of 1990 Low supply of land Planning director’s response, on a five-point scale, to statement that supply of land for residential growth is an important factor in constraining growth in the city Population Natural log of city population Own-source revenues Natural log of own-source revenues per capita (defined as total revenues minus intergovernmental revenues and minus current service charges), 1993 % of property tax % of locally generated property tax revenues that flow to the municipal government Lack of planning staff Planning director’s response, on a five-point scale, to statement that a lack of personnel to review project proposals is an important factor in constraining growth in the city Longer review process Planning director’s response regarding whether the time required to complete the review of residential projects in the city has been shortened, stayed the same, become somewhat longer, or become much longer (four-point scale) Overt growth restrictions Number of the following policies adopted in the city: annual limits on total building permits, annual limits on residential units authorized, annual limits on multifamily dwelling units built, recent substantial reduction in land zoned residential, policy linking local residential growth rate to a formula or external growth rate, formal population ceiling, moratorium on development or on water/sewer hookups Affordability set-aside = 1 if planning director indicates that city has a policy to require residential developments to include affordable housing Initiatives/neighborhoods Planning director’s response, on a four-point scale, regarding whether important city policies affecting residential growth have mostly been enacted by the council without neighborhood pressure, enacted by the council as a result of neighborhood pressure, enacted both by the council and the voter initiative process, or enacted pretty much exclusively as a result of initiatives Inland Empire location = 1 if city is in Riverside or San Bernardino Counties Bay Area location = 1 if city is in the jurisdiction of the Association of Bay Area Governments 50 Table 3.3 (continued) Variable Central Valley location Months since deadline Definition/Notes = 1 if city is in the counties of Butte, Colusa, Fresno, Glenn, Kern, Kings, Madera, Merced, Placer, Sacramento, San Joaquin, Shasta, Stanislaus, Sutter, Tehama, Tulare,Yolo, or Yuba Number of months elapsed between the date when the city’s housing element update was due and September 2002 SOURCES: Information on low supply of land, lack of planning staff, review of residential proposals, number of overt restrictions, affordability set-asides, and initiatives/ neighborhoods measures is derived from a 1998–99 PPIC survey of city planning directors. Own-source revenues and percentage of property tax are calculated from California State Controller data. Months since deadline was calculated by the author. Age of housing is from the 1990 Census. All other variables are derived from the 2000 Census. Key Findings of the Multivariate Analysis Table 3.4 summarizes the variables that showed a statistically significant relationship with city noncompliance. (For more detailed results, please consult Appendix A.) The results indicate that many city characteristics that one might expect to come into play in regard to housing element preparation are not, in fact, associated with whether cities comply with the housing element law, once one controls for other relevant features of the city. However, a number of patterns do stand out in this analysis. First, the measures of local housing policy seem of key importance. Cities whose planning directors report that the review process for new development proposals has been getting longer, and cities with more overt growth controls, experience more trouble in attaining compliance. For example, each restrictive growth policy that has been adopted approximately doubles the odds that the city will be found noncompliant. This relationship is perhaps not surprising, because HCD staff pay close attention to local government constraints on housing development. Communities that see themselves as a city of single-family homes may have difficulty in showing how they plan to accommodate the number and variety of units necessary to meet their housing goal. 51 Table 3.4 City Characteristics That Are Statistically Associated with Noncompliance Variable Age of housing City population Longer review process Overt growth restrictions City requires affordability set-aside Months since housing element due Inland Empire location San Francisco Bay Area location Central Valley location As Characteristic Increases/Occurs, Probability of Noncompliance . . . Increases Decreases Increases Increases Decreasesa Decreases Increases Increases Increases NOTES: The variables are defined in Table 3.1. Variables listed in Table 3.1 but not in this table were not related to noncompliance at a statistically significant level, defined as p < .1. a“City requires affordability set-aside” is statistically significant only in the model that controls for “slow growth pressure from initiatives/ neighborhoods.” Whether cities require new housing projects to include affordable housing is a significant predictor of compliance.10 HCD reviewers may take such “inclusionary” housing policies into account as a way to overcome barriers to the production of new units for low- or moderateincome households. All else equal, cities with larger populations are more likely than others to reach compliance. This relationship may indicate that large local governments have the resources necessary to complete a successful housing plan. On the other hand, there is no evidence that the fiscal strength of the city or the size of its planning staff influences the probability of compliance. Larger population size may also indicate that ____________ 10This result is significant only when controlling for whether the planning director feels that voter initiatives and neighborhood pressure have been a major source of slowgrowth policies. I do not include this latter variable in the main model because 14 percent of survey respondents answered “don’t know,” which means that such cities must be dropped from the analysis. 52 the city government is more insulated from the pressure of homeowners and more subject to the influence of progrowth business elites, who are likely to be in favor of additional housing. As Fischel (2001) shows, small local governments are more likely to be attuned to the interests of homeowners, who are risk-averse regarding new development because so much of their wealth is typically tied up in their home value. An older housing stock is also associated with a greater likelihood of noncompliance. Cities with older housing may be more “settled” and have a more established community character; they are also likely to contain less vacant land. Similarly, older housing indicates that there has been less new housing construction in recent decades, and thus the community may have come to expect a slow pace of change.11 Finally, the regional context within which each city plans for housing matters a great deal for compliance rates. As discussed, the timing of housing element revision deadlines differs among the state’s regions. The measure of the length of time that has elapsed since the deadline for submitting an updated housing element is one of the most statistically significant predictors of noncompliance. Specifically, the results indicate that each month that has elapsed since the region’s deadline renders a city approximately 5 percent more likely to reach compliance. Location in the Bay Area, the Inland Empire, or the Central Valley is also associated with noncompliance, even after controlling for the various city characteristics. Municipalities in each of these areas of the state show more difficulty in attaining compliance.12 Some Caveats The above analysis presents a useful clarification of the community characteristics that underlie noncompliance. Nevertheless, the analysis ____________ 11I use the age of the city’s housing stock as of 1990, rather than as of 2000, to avoid the statistical problem of endogeneity. That is, a noncompliant housing element may lead to less new construction; thus, one should not, in turn, use “old housing” to predict noncompliance. By using a measure of housing age from a decade ago to represent past community characteristics, one is on firmer footing in predicting current compliance status. 12The implicit comparison is to cities in other regions included in the analysis. Specifically, these include cities in San Diego, Los Angeles, Orange, Imperial, Ventura, and Santa Barbara Counties. 53 has its limits. First, the dependent variable—housing element compliance—is a dichotomous, or “yes/no,” measure that may capture local housing policy in only a rough way. Simple dichotomies often make for more uncertain statistical estimations than more “continuous” variables, such as the number of housing units produced or the number of policies passed. I did attempt to analyze noncompliance using a threepoint scale (in compliance, overdue, out of compliance) that provides an additional category (overdue) that might be viewed as an intermediate case between compliance and noncompliance. However, results did not differ much from those just discussed.13 In defense of the dichotomous measure, housing element compliance or noncompliance is worthy of attention in and of itself, because of its importance to cities for the legal defensibility of their general plans and their ability to compete for affordable housing funds. Still, those interested in a more encompassing explanation of local residential policies should supplement this report with other sources.14 Second, this analysis is of compliance status at a particular point in time—a “snapshot.” Although the analysis controls for region and for the amount of time that the city has had to reach compliance, it cannot take into account all factors that might complicate compliance. For example, Marin County and its cities collectively decided to prepare a countywide housing analysis and housing element updates. All the jurisdictions in the county were subsequently overdue, and by September 2002 all had submitted draft housing elements judged noncompliant, except for two localities whose housing plans were under review by HCD. Finally, although it is tempting to assert that factors such as small populations or a large number of growth controls “cause” noncompliance, I have not put forward any behavioral model of either ____________ 13In the ordered logit model of these three compliance statuses, the same variables were significant (and with the same signs), except for city population size, which was insignificant, and Central Valley location, also insignificant. The overall fit of the ordered logit model was less satisfactory than the dichotomous model discussed above. 14For example, Lewis and Neiman (2002, Chapter 5) present the results of other statistical analyses of local residential growth-management policies and local antigrowth politics. HCD (2001) presents an analysis of local development fees and charges that affect the cost of new housing. 54 local government activity or HCD reviews that pins down what leads to noncompliance. Moreover, over a long time horizon, the direction of causality may be muddled. Cities that have a noncompliant housing element may attract less rental housing or fail to qualify for affordable housing grants from the federal or state governments. These factors, in turn, may make it even more difficult for the community to attain compliance in the next revision period. Conclusion Among the reasons that observers have suggested for widespread noncompliance with the housing element law are the exclusionary sentiments of upper-status communities, the lack of vacant land available for housing in many cities, the lack of resources of the local government to address the issue, and the antigrowth politics or antihousing policies of some towns. In the regions that have been subject to HCD reviews for updated housing elements over the past three years, a simple comparison of compliant and noncompliant cities indicates that few of these arguments receive clear-cut support. Noncompliant cities have older housing and smaller populations but are actually lower in population density than compliant cities. Most other community characteristics do not clearly distinguish the two sets of cities. In the remainder of the state, where cities have had about a decade to reach compliance since their last housing element revisions were due, few cities are noncompliant. The simple comparison indicates that the few laggard communities do tend to be wealthier and less ethnically diverse than compliant cities. A more sophisticated analysis can isolate which local characteristics are most important for noncompliance. Such a model reveals that among the factors that best predicted noncompliance were an older housing stock, more restrictive growth policies and review processes, a smaller population, and a Bay Area, Central Valley, or Inland Empire location. The most important predictor of all, however, was simply the time elapsed since the most recent housing element revision was due. Time may not cure all ills, but it does allow many jurisdictions additional opportunities to prepare a housing element that is acceptable to HCD reviewers. 55 Thus, there is at least some evidence that several factors that have been posited to play a role in noncompliance—local land-use and building characteristics, the size of the city, and municipal growth politics—do have some influence. In the current round of updates, evidence was strong that cities’ residential growth controls and review processes held a particularly important role for the HCD reviews. By contrast, local characteristics that some have suspected play a major role—such as local homeownership rates, revenue bases, or population densities—do not show a clear link to noncompliance. In short, it is difficult to argue that any simple set of characteristics is determinative of how likely cities are to attain compliance. 56 4. Does Housing Element Compliance Mean That More Housing Is Produced? Do local governments that comply with the housing element law, by adequately planning for future residential needs, actually enable more housing to be built in California? Moreover, does compliance help predict the types of housing built, such as the mix between single-family and multifamily development? If the answer to these questions is no, the resources that state and local government currently devote to housing element compliance and review might be better directed to other policy efforts to improve housing production. Using data from the 1990s, this chapter offers a statistical analysis of whether a city’s compliance status in the early part of the decade helps to predict the percentage increase in the city’s housing stock by 2000. Although data are not available to specifically study the development of affordable units, one can examine the growth of multifamily housing and the share of housing developed that is multifamily. The analyses control for a variety of other city characteristics that might influence the level and type of housing growth. The quantitative estimates are reported in Appendix B; this chapter reports the results for a general audience. Housing Element Compliance, Circa 1991 To make this analysis possible, the Department of Housing and Community Development made available its housing element compliance report for cities and counties as of December 31, 1991. This period was chosen because jurisdictions in the most populous parts of the state—the Southern California, San Diego, and Bay Area regions—had completed their second round of housing element updates by that point and had received HCD’s judgment as to whether their plans were 57 compliant. Moreover, the 1991 date is proximate in time to the 1990 Census, allowing use of data from that year as a “starting point.” Accordingly, this analysis focuses only on communities in these three regions, where most of the state’s population resides and which tend to have the most controversies and difficulties regarding housing element compliance. Compliance rates for cities in 1991 were poor, as Figure 4.1 illustrates. Only one-quarter of cities statewide were in compliance, with 59 percent being declared either noncompliant by HCD or overdue or “obsolete” because of the lack of any updated housing element in time for the deadline. In the wake of these low compliance rates, HCD and the attorney general’s office mounted a sustained effort to encourage more cities and counties to meet their legal responsibilities. This campaign was fairly successful, as the percentage of communities attaining compliance grew substantially. Housing element obsolete 15% Housing element due 2% Out of compliance 42% Under review by HCD 16% In compliance 25% Figure 4.1—Cities’ Housing Element Compliance Status, as of December 31, 1991 58 There tends to be some persistence in compliance status. There is a moderate, statistically significant degree of correlation (r = .31) between noncompliance in 1991 and noncompliance a decade later, in 2001. In other words, although many cities attained compliance status as the 1990s wore on, it remained the case that noncompliance after the second round of revisions was a fairly good predictor of noncompliance after the third round. Measuring Housing Growth Rates Using Census Data The first analysis in this chapter draws upon U.S. Census data to examine the rate of growth in housing units between 1990 and 2000. Because communities of different sizes are likely to have very different amounts of housing built, I examine the percentage increase in the number of housing units in each city between 1990 and 2000. Table 4.1 provides summary statistics regarding this measure of housing increase. The percentage increase in housing units will serve as a dependent variable (that is, the phenomenon that I seek to explain) in a quantitative analysis. Because some very small cities experienced extreme rates of housing change, the analysis is limited to communities with at least 2,500 residents, to avoid skewing the results. The main issue of interest is whether a connection exists between a city’s compliance status as of 1991 and its rate of housing growth Table 4.1 Percentage Growth Rates for Housing in Cities in the San Diego, Southern California, and San Francisco Bay Area Regions, 1990–2000 Average increase in housing units Median increase in housing units Standard deviation Largest rate of increase Largest rate of decrease 11.4 4.9 18.3 196.3 (Brentwood) –9.7 (La Habra Heights) SOURCE: Author calculations are from the 2000 Census and the 1990 Census. NOTE: Data focus only on cities of 2,500 or greater population. 59 through the 1990s. At first blush, there is surprisingly little evidence of such a link, as Table 4.2 reports. A simple comparison shows that the average noncompliant city in the relevant regions (San Diego, Southern California, and San Francisco Bay Area) actually increased its housing stock about 1.5 percent more than the average compliant city. The median noncompliant city, however, saw about 1.7 percent fewer new housing units constructed than the median compliant city (which indicates that some noncompliant cities with very high rates of housing development pushed up the mean for that group). In neither calculation are the differences statistically significant, however. Although interesting, one cannot draw confident conclusions from this simple comparison of compliant and noncompliant cities. Housing growth rates could be expected to differ among communities for many other reasons—for example, the location of the city, its attractiveness to new residents and developers, and its demographic characteristics. Table 4.2 Comparison of Housing Development in Compliant and Noncompliant Cities in the San Diego, Southern California, and San Francisco Bay Area Regions Housing Element Status as of 1991 Compliant Noncompliant % Increase in Houing Units, 1990–2000 Average City Median City No. of Cities 9.0 5.7 48 10.5 4.0 173 NOTES: Data are limited to cities of 2,500 or greater population. Differences between compliant and noncompliant cities are statistically insignificant, using either a t-test for difference of means, or a Kruskal-Wallis test for difference of medians. Controlling for Other Factors To account for the wide variety of factors that might affect housing growth rates across cities, it is possible to estimate a multivariate regression model that controls for such characteristics while assessing whether housing element compliance status has its own independent relationship with housing growth. In such an analysis, it is important that all of the relevant city characteristics be included in the analysis, since bias can result from improperly omitted variables. 60 A useful way to account for some of an area’s housing market characteristics that may not be easy to measure is to include a control for the county in which each city is located. That is the approach taken in this analysis.1 Thus, such factors as the county’s population growth trend, its location in the path of urban growth, county government growth policies, or other unobserved countywide factors are taken into account. At the city level, several local characteristics are included to inform the predictions about housing growth. I focus on measures of city characteristics as of 1990, as these are “baseline” factors that could be expected to influence housing development trends in the 1990s. The variables included in the analysis are described below. 1. Housing growth in the prior decade: Communities in urban regions tend to persist in their tendencies to differentiate themselves (Farley, 1964; Neiman, 1980). One can expect that cities with rapid rates of increase in housing in the 1980s would continue to grow quickly, although perhaps not at an equal rate, through the 1990s. Cities that grew slowly or shrank would likely continue to stagnate. This variable is measured as the percentage increase in total housing units in the city between 1980 and 1990. 2. Population size: It is reasonable to expect that large cities grow at a different rate than small cities, all else equal. The natural logarithm of the city population is used because of the highly skewed distribution of city populations. In addition, the ten communities of less than 2,500 population in these regions, and the two cities of greater than a million population, are dropped because the unique characteristics of very large and very small cities tended to hinder the analysis. 3. Population density: Higher densities indicate that the city probably has less vacant land to accommodate additional growth and that new growth may be more likely to cause spillovers such as traffic or ____________ 1Specifically, county-level fixed effects are included in the model for all cities except San Francisco, which is the only city in its county. 61 parking problems. Density is measured as the logarithm of persons per square mile, again because of the skewed distribution. 4. Urbanization status: Cities in urbanized areas may find it more challenging to accommodate new housing because of a lack of vacant land or conflicts with existing residents and businesses. This variable is an indicator, set to equal 1 in cities in Census-identified urbanized areas as of 1990 and 0 otherwise. 5. Age of the housing stock: Cities with an older housing stock may be more settled communities that have experienced less recent pressure for new development. The variable is the number of years elapsed since the median housing unit in the city was built, as of 1990. 6. Household income: Cities with high-income households are likely to have high land costs, and developers may tend to propose mostly single-family luxury homes. Wealthier communities may also have more political resources to mobilize in opposition to new development. The variable is the median household income in the 1990 Census (measured for the 1989 calendar year). 7. Senior citizens: Communities with high proportions of seniors may experience less pressure for new home construction. A large share of seniors implies that there are smaller households and probably less new household formation. The variable is the percentage of persons age 65 and over in the city population in 1990. 8. Technical workers: As an engine of California’s growth, high-tech industries might be expected to play an especially important role in housing demand. However, aerospace, also a high-tech industry, experienced major declines in employment, so the effects on housing increases may not be straightforward. The variable measures the percentage of the local (working) population employed in technical occupations, as classified in the 1990 Census. 9. Job-to-worker ratio: Cities that are “job centers,” with a high ratio of local employment to resident workers, might be expected to attract more interest from homebuyers and developers. This variable, derived from the Census Transportation Planning Package, is the natural logarithm of the ratio of jobs within the city to workers living in the city, as of 1990. 62 10. Commuting time: Cities in which many workers must commute long distances are apt to be less attractive and may experience less pressure for home construction, all else equal. The variable measures the average one-way commute time, in minutes, of workers living in the city in 1990. 11. Distance to urban centers: Isolated municipalities located far from urban job centers are likely to be less attractive to potential residents and housing developers. A computerized mapping program was used to calculate the straight-line distance between the city in question and the nearest of a set of traditional urban central cities in California.2 12. Housing element noncompliance: The major variable of interest is whether the city was found to be compliant with the housing element statute, according to HCD, in December 1991. This is an indicator variable, with compliant cities set to equal 0 and those found noncompliant, those overdue, or those having “obsolete” housing elements set to equal 1.3 Numerous other city characteristics were tested in other versions of the model predicting housing development. These included such measures as the housing vacancy rate in 1990, the percentage of housing units not connected to sewers, the percentage of recreational housing units, controls for central city or rural status, the average number of persons per household, and the percentage of various ethnic groups in the community. Each was dropped from the analysis when it proved persistently unrelated to housing growth rates. The effects of compliance ____________ 2The central cities chosen to represent historic metropolitan centers were Bakersfield, Chico, Fresno, Los Angeles, Merced, Modesto, Monterey, Oakland, Redding, Sacramento, San Diego, San Francisco, San Jose, Santa Barbara, Santa Cruz, Stockton, and Visalia. The mapping program calculated distance between each city in California and the nearest one of these central cities, using mapping coordinates (where a unit of 1 is equivalent to approximately 60 miles). 3I also estimated a different model in which cities that had overdue or “obsolete” housing elements were represented by a separate indicator variable from those that were overtly found noncompliant. This estimation strategy was based on the theory that being late with a housing element may not demonstrate as serious a problem as being overtly noncompliant. However, no substantive differences in results were found from those reported here, with one exception to be discussed below. 63 status on housing development are not influenced by the inclusion or exclusion of these variables.4 Results of the Analysis of Total Housing Unit Growth Detailed results of the model, which accounts for 64 percent of the variation in housing growth across cities, are listed in Appendix Table B.1. For our main purposes in this chapter, however, it is easy to summarize the findings: There is no evidence of a detectable relationship between housing element compliance and the percentage increase in housing across these cities during the 1990s. If noncompliance were a good predictor of slow housing growth, all else equal, one would expect to find a negative, statistically significant relationship between noncompliant housing elements and the percentage housing increase. However, the analysis shows a positive, insignificant relationship. Thus, for all the potential merits and benefits of housing element compliance, one must look to other factors to explain why some cities experience rapid housing development and other cities experience little. The analysis suggests that certain demographic characteristics, measures of a community’s position in the urban hierarchy, and the physical capacity of the city to accommodate new buildings are better able to predict housing growth. For example, cities with higher population densities as of 1990 added housing at a slower rate in the decade that followed, as did those with older housing stock at the start of the period. These results indicate that growth in the 1990s moved away from older, denser communities. At the same time, cities with larger populations, those that were far away from urban centers, and those with long commute times also experienced slower rates of housing growth. Cities with higher-income populations grew more slowly, perhaps ____________ 4Care was also taken to choose measures of community demographics that were not excessively intercorrelated. The median household income of the city, included here, is a good proxy for the overall status of the community. Other socioeconomic variables, such as the unemployment rate, poverty rate, or homeownership rate, are closely associated with median income, but these variables do not show as much association with housing growth. Again, the (non)significance of housing element compliance is not affected by the exclusion of these variables. 64 because land was more costly or zoned for lower densities. Perhaps surprisingly, the percentage of technical workers among local resident workers was negatively related to the rate of housing increase, quite possibly a reflection of the severe declines that would occur in aerospace and related industries in the early 1990s. The best predictor of housing growth, however, was the city’s rate of housing increase in the previous decade. Within any given county, cities that grew quickly in the 1980s were the ones that tended to add to the housing stock quickly in the 1990s.5 After accounting for these trends, housing element compliance as of 1991 provides no additional information about cities’ rates of housing growth. Analysis of Multifamily Development Using Construction Industry Data It is useful to consult other data sources as a check for the above analysis, which uses Census data on the number of housing units in each city. I therefore gathered data from the Construction Industry Research Board (CIRB) on the number of housing units constructed in each city in the period from 1991 through 2000. This data source has the additional benefit of being broken down by whether the units built were single-family or multifamily. In most cities in the San Diego, Southern California, and Bay Area regions, single-family units clearly predominated. In fact, in an average city in these three regions, only 27 percent of the new units were multifamily.6 Not surprisingly, therefore, there is a very strong correlation (r = 0.90) between our Census measure of housing unit increase and the CIRB data on increases resulting from single-family construction.7 When I ran the same regression model as was used for the ____________ 5The county-level fixed effects were also significantly related to housing growth rates, indicating the regional character of housing markets. 6If data from all cities in these regions are aggregated, the share of units that were multifamily among all new units was somewhat higher, at 33 percent (28 percent in cities statewide). This is because cities that had the highest numerical increases in housing units tended also to have higher shares of multifamily development. 7Specifically, the variable used was the ratio of newly constructed single-family units in the city to all housing units existing in that city as of 1990. 65 Census measure in the CIRB single-family increase data, results were thus largely similar. Again, no relationship between housing element compliance and housing development was apparent. As a fair-share approach to housing, the housing element statute is specifically concerned with increasing the production of affordable housing. Unfortunately, comparable information for all cities about the production of affordable units appears to be nonexistent. However, the CIRB data allow us to examine multifamily housing production. Multifamily housing includes apartments and condominiums, the types of housing generally most relevant for those at the lower to middle levels of the income distribution, particularly in the coastal metropolitan areas where housing is very expensive. Moreover, analysts have been particularly concerned that multifamily production has plummeted in California since the 1980s. “Whereas multifamily housing accounted for between 45 and 49 percent of total housing construction during the 1960s, 1970s, and 1980s, last decade it fell to 25 percent of the already depressed total” (Myers and Park, 2002, p. 2). Thus, the second regression analysis featured in Appendix B examines the determinants of increases in multifamily housing. The dependent variable is the ratio of newly constructed multifamily units (from 1991 through 2000) to all units existing in the city in 1990.8 The results of this analysis again show that housing element noncompliance as of 1991 is not a significant predictor of multifamily development, once other relevant factors are held constant. Although at least the relationship is negative (indicating that noncompliance is linked to a smaller multifamily increase), it is near zero and not statistically significant. Rather, demographic and land-use characteristics of the community are better predictors of multifamily development. Cities ____________ 8Some readers may wonder why the measure is based on the ratio of new multifamily units to all existing units, rather than existing multifamily units. The reason is straightforward. Cities with a very tiny number of existing multifamily units (say, two) could add a negligible number of new multifamily units over the decade (say, another two). If one uses the existing number of multifamily units as the denominator, then such cities are unfairly credited with a huge proportionate increase (in this example, 100 percent). The ratio used in the analysis better characterizes the role of multifamily development in augmenting each city’s housing stock. 66 with more senior citizens as of 1990 tended to experience slower rates of multifamily housing development. Cities with more jobs per worker saw more rapid multifamily development. The Mix of New Housing Development Finally, it is possible that housing element compliance could affect the mix of housing developed—multifamily versus single-family—despite its lack of effects on the rate of increase. The third regression analysis shown in Appendix B examines this issue; the dependent variable in this case is simply the percentage share of new housing units that were multifamily units, according to the CIRB data. Here there is a significant association of compliance with the outcome in question. The results show clearly that cities with noncompliant housing elements develop new housing that is weighted more toward single-family units. In fact, holding constant for all the other city characteristics treated in the model, a noncompliant housing element is linked to an 8 percentage point lower proportion of multifamily housing among the new units. This reduction is considerable, given that the average city saw multifamily units account for only 27 percent of new housing. The only other variables that are clearly related to this proportion are the median income of the city, its job-to-worker ratio, and its rate of housing growth in the prior decade. Cities with lower median incomes and more jobs per worker saw a higher percentage of multifamily development. Cities with slow growth rates in the 1980s also experienced a share of multifamily development that was higher than in those that grew quickly (although the relationship was substantively small). Thus, cities with compliant housing elements in 1991 experienced a more balanced mix of new housing development in the 1990s. In a further test, I examined whether cities that had overdue or obsolete housing elements in 1991 performed as poorly in their housing mix as those judged noncompliant by HCD reviewers. The results indicated that it was the cities overtly deemed noncompliant, rather than the cities 67 with overdue or obsolete plans, that had the heavier proportion of singlefamily housing rather than multifamily.9 Reconciling These Findings What does it mean to say that housing element compliance is not linked to the rate of multifamily development but is linked to the proportion of housing developed that is multifamily? One interpretation, consistent with the fair-share approach of housing element policy, is that thorough housing planning diversifies the mix of new residential development in the community. Units that developers might have preferred to build as detached homes (or that neighborhood groups might have preferred to see become detached homes) are constructed as apartments or condos instead, quite possibly because of an emphasis in the housing element on finding sufficient sites for multifamily development. The results imply that cities with compliant housing elements, although they may not be willing or able to accommodate more total housing units, are willing to substitute multifamily (more affordable) units for a share of single-family (marketrate) units. The 8 percentage point estimate of that substitution effect in this analysis has a similar order of magnitude to the 10 or 15 percent affordable housing set-aside that many cities require of some new developments through local inclusionary housing policies. Conclusion Much attention and resources are given to housing elements, in part because there is an assumption that noncompliant cities are those that artificially slow the rate of increase to the housing stock. Compliance status from a decade ago, however, was not a good predictor of the rate of subsequent new housing development in cities in the San Diego, Southern California, and Bay Area regions. Rather, housing market and ____________ 9Specifically, the regression analysis included two indicator variables: one for cities found noncompliant and one for overdue or obsolete cities. Results showed that, all else equal, noncompliant cities experienced about an 8.7 percentage point lower share of multifamily development than compliant cities. Those with overdue or obsolete housing elements had a smaller negative (–2.5 percentage points) but statistically insignificant relationship with the percentage of new housing that was multifamily. 68 demographic factors outweigh compliance status in contributing to variations in growth rates. The pace of neither single-family nor multifamily development is associated with compliance status. Does this nonfinding imply that the resources devoted to the preparation of housing elements are a wasted effort? Not necessarily. It is probably a healthy exercise for local governments to plan for growth and to assess where various types of new housing may be accommodated in the community. It is possible that the very requirement that all local governments prepare housing elements means that more attention is paid to housing needs and more housing is built statewide, regardless of whether individual communities are found to be compliant or noncompliant. One cannot be sure, then, that there would not be deleterious effects on housing production if the requirement were eliminated. Moreover, this analysis does not attempt to sort out the relationship of compliance status to production of affordable units, which receive particular emphasis in housing element law. Indeed, when examining the proportion of all units constructed from 1991 through 2000 that were in multifamily developments—most likely to be affordable—we do find that multifamily construction tends to displace a significant portion of single-family construction in cities with compliant housing elements. Nevertheless, it is quite striking that one can detect no measurable relationship between compliance and overall housing production. Cities that made efforts to comply with the housing element statute may have played host to a wider mix of new housing but did not accommodate more units overall, at least in the 1990s.10 California policymakers concerned about housing production, and local constraints on housing, may wish to think about other potential policy approaches. The next chapter describes some alternatives. ____________ 10A reader of an earlier draft of this report commented that compliance status at the end of the 1990s might better predict housing production rates in that decade. The idea is that many cities that were noncompliant in 1991 eventually gained compliance, and also that municipal policies measured by HCD in the 2000s were likely in effect and influencing housing development in the 1990s. However, when I substituted March 2001 housing element compliance status in these regressions for 1991 compliance status, it showed insignificant effects on all measures of housing development in the 1990s. 69 5. Can States Ensure Adequate Local Provision for Housing Development? Given the turmoil over housing element law in California and its apparent lack of success in generating more housing, a discussion of alternative policy approaches seems warranted. This chapter does not recommend any single policy change but is intended to help illustrate the array of options available to state policymakers. It begins by discussing changes that have been suggested by groups in recent years. Sometimes, however, a broader perspective on a policy problem is useful, and therefore the second part of the chapter examines the experiences of three other states in crafting policies that attempt to shift local planning and land-use activities in a more housing-friendly direction. The successes and failures of these states hold potential lessons for California. Alternative Approaches within the Context of the Current Housing Element Law Discussions of problems with the current housing element law among California policymakers and housing advocates have led to a number of proposals in recent years to change tactics. What most of these reform proposals have in common, however, is that they take the housing element approach—state review of local plans—more or less as a given. More Penalties and Prescriptions for Local Governments One approach sees the problem as a lack of serious consequences for local governments that fail to meet the requirements of existing housing element law. Myers and Park (2002, p. 5) argue that “the state needs to place greater weight on localities’ production of total housing 71 construction, including a share of multifamily housing, and that emphasis needs to be backed with stronger teeth than at present.” Thus, for example, it has been suggested (as in Senate Bill 910) that financial penalties be levied against noncompliant local governments, or that their eligibility for certain state or federal grants be rescinded. The attorney general could bring legal action against communities that persistently fail to comply. Another approach of this type seeks to prescribe local land-use regulation, under the assumption that localities, left to their own devices, will fail to provide housing opportunity. Thus, the state could impose “inclusionary zoning” requirements on all communities, under which all housing developments of a certain size must reserve a share of units for low- or moderate-income households. Other suggestions have included requiring minimum zoning densities in certain areas (such as that near public transit), or a blanket exemption of affordable housing projects from local zoning or growth controls (Senate Committee on Local Government, 1993, p. 112). Innovative approaches to promoting affordable housing need not be state-prescribed, however. Some cities, for example, have developed an “affordable housing overlay” as part of their zoning ordinance that allows a developer to claim a density bonus if his or her proposed project meets affordability standards. Self-Certification Based on Performance Self-certification is the idea of granting authority to local governments to approve their own housing plans without state oversight if they meet certain criteria involving demonstrated housing production. This approach was popular among local government representatives in the mid-1990s (Senate Committee on Housing and Land Use, 1995). As Mark Pisano, longtime executive director of the Southern California Association of Governments, argued in 1993: Local governments should be allowed to self-certify their housing element reforms. Self-certification should be linked to a performance review at the subregional or regional level. The right of a third party to request a review should be also allowed. Any conflicts arising from the review should be resolved through a subregional mediation process, thereby ensuring a timely 72 resolution of disputes. If mediation is unsuccessful, [HCD] would review the housing element performance under question (Senate Committee on Local Government, 1993, p. 42). Similarly, Janet McBride, a senior planner for ABAG, testified, “Communities which show a record of success in providing their ‘fair share’ of housing should be exempt from outside scrutiny of their housing element” (Senate Committee on Local Government, 1993, p. 49). As a result of these discussions, the legislature passed a pilot program, AB 1714, in 1995. That law allowed communities in the San Diego region to self-certify their housing elements if they had met their fairshare housing needs in the period since the previous housing element revision. According to HCD’s housing element status reports, only a few cities in San Diego have availed themselves of this option to date. The pilot program will sunset in 2009 unless the legislature acts to continue it. Some have cautioned that basing self-certification on strict measures of prior performance may make such an approach somewhat inflexible. During an economic recession of the type the state experienced in the early 1990s, for example, housing production slumped because of the economy, and few jurisdictions would have been able to meet even 75 percent of their assigned housing goals. Alternative suggestions for measuring performance include assessing whether the jurisdiction has issued a share of permits above the regional average or has approved a certain share (say, 90 percent) of proposed housing projects that are consistent with the local general plan (Senate Committee on Housing and Land Use, 1995). Subregional Allocations and Joint Housing Elements A number of commentators have suggested that jurisdictions in a subregion should decide collectively how to divide up that area’s housing allocations. This approach has been advocated specifically for smaller counties in large metropolitan regions, and some have suggested setting up joint-powers authorities as an intergovernmental arrangement to adopt joint housing elements and suballocate fair shares. The aim would then be for jobs/housing balance within each subregion—probably a 73 more tractable approach than seeking balance within each jurisdiction (Senate Committee on Housing and Land Use, 1995, pp. 64–65, 79). Subregional allocation has already occurred in some cases; for example, in the decision by SCAG to allow Orange County governments to formulate allocations within that county. It is quite possible, however, that making subregional allocation more widespread would not solve the political problem of how to divide an unwanted responsibility among jurisdictions. Rather, it may merely create a new political arena to haggle over the issue at a smaller geographic level. A provision allowing for greater allocation to subregions was inserted into SB 910, the housing element reform bill, in 2002. Encouraging Transfers of Housing Allocations among Jurisdictions A number of local government representatives have sought the ability for localities to “trade” or transfer their allocations with each other, presumably in exchange for payments or other considerations. The argument is that housing may be accommodated more easily or economically in some jurisdictions than others, and allowing transfers would enable more flexibility for localities in meeting their housing obligations. For example, a small suburb with limited, expensive vacant land could transfer its allocation to a nearby community with more or cheaper vacant land or an active housing redevelopment policy, thereby possibly allowing a greater total number of units to be developed. Section 65584.5 of the government code, passed in 1994, enables transfers among jurisdictions under certain circumstances. This legislation followed two years of debate over the topic, during which various stakeholder groups agreed on a compromise. Nevertheless, the provisions for transfer are quite strict and constrained. The entities engaged in the transfer must both have compliant housing elements, and the “donor” entity must have met at least 15 percent of its housing goals. The transfer agreement must include plans to construct the units within three years (or else the units revert to the donor community), and the COG must review it. The percentage of a community’s housing goals to 74 be transferred may not exceed the percentage of its goal that the community has already satisfied (Curtin, 2000, p. 11). At issue is whether transfers violate the spirit of fair-share housing policy, which aims at an equitable distribution of responsibilities across jurisdictions. Opponents, including some housing advocate groups, argue that wealthy communities that eschew low-cost housing will be allowed to buy their way out of their responsibilities to their region if widespread transfers are permitted. Another Approach: Rewards for Performance Enticing communities to accommodate housing would not be such an uphill battle if they perceived that doing so would be in their financial self-interest. Thus, creating an element of the state fiscal system that rewards local governments for the addition of housing units, particularly affordable units, would be an approach likely to result in less conflict and more cooperation. The existing Jobs/Housing Balance Incentive Program, passed in 2000, has elements of such a “rewards for performance” approach. Until it was defunded when the state budget went into deficit, the program authorized devoting $100 million in grants to local governments that demonstrate increased issuance of building permits, if they have adopted housing elements that are in compliance with state law. The housing bond passed by voters in November 2002 earmarks $100 million to continue funding this program. Senate Bill 423, enacted in late 2002, provides some specific instructions on how the funds are to be allocated. To provide certainty for local governments, a continuing rather than one-time source of funds for such an approach may be necessary. For example, a regional fund could be established for such purposes as transportation improvements and open space protection, with funds awarded to cities and counties that clearly demonstrate that they are taking actions (beyond planning) to accommodate their fair share of units, particularly affordable units. The fund could be financed through a regional impact fee on commercial development in job-heavy, housingpoor areas (Lewis, 2002). Others have suggested recalibrating the local fiscal system so that localities would be rewarded directly (i.e., by formula) for increases in 75 population or housing units. For example, there have been discussions of distributing a portion of local sales tax revenues to cities and counties on the basis of local populations, rather than the “point of sale” of retail transactions. AB 680 (Steinberg), which failed to emerge from the state legislature in 2002, took such an approach. Of course, such a wideranging reform of local finance raises many concerns and conflicts unrelated to housing policy, and thus far such efforts have not succeeded. It is worth noting that state subventions of Vehicle License Fee (VLF) revenues are awarded to localities through a population-based formula. Local officials have expressed worries that the state’s current budget crisis may threaten the existence of the so-called VLF backfill, which attempts to ensure that cuts to the VLF in the late 1990s do not result in decreased allocations to localities. Loss of all or some VLF revenues would further reduce local incentives to provide new housing. Relevant Experiences from Other States Other states, including some with equally strong traditions of home rule as California, have also wrestled with issues of inadequate housing production, mandated fair shares for jurisdictions, and state oversight of local planning. But the approaches taken differ widely from California’s, and their experiences may be worth considering. This section briefly sketches the approaches taken in New Jersey, Massachusetts, and Oregon.1 New Jersey Along with California, New Jersey likely has the most sustained experience with policies that involve state oversight of local housing policy.2 At first glance, New Jersey’s policy posture toward fair-share ____________ 1The focus here is only on contemporary policies that have some analogies to the California situation. A comprehensive examination would need to include the nowdefunct New York State Urban Development Corporation (UDC), a powerful state agency with the power to override local zoning to build subsidized housing projects in the late 1960s and early 1970s. Its zoning override power was later stripped by that state’s legislature, and the UDC ultimately became insolvent in the mid-1970s. 2This discussion of the New Jersey experience is based upon Calavita et al. (1997), Haar (1996), Hughes and VanDoren (1990), Listokin (1976), and Weinstein (1993), as well as the website of the Council on Affordable Housing. 76 housing production appears similar to California’s. The state’s 566 municipalities are required to prepare housing elements, are allocated an obligation of low- and moderate-income housing needs, and a state entity—in this case a Council on Affordable Housing (COAH), created by the legislature in 1985—reviews local plans. At a more basic level, however, there are serious differences between the two states’ approaches. Most important, New Jersey’s requirement of a fair-share responsibility on localities to provide a reasonable amount of housing within their region was mandated by judicial action—the famous Mount Laurel decisions of 1975 and 1983. The state’s supreme court justices drew upon state constitutional language to assert that municipalities had an obligation to admit affordable housing, thus providing more legal ammunition than California’s statutory housing element requirement. The court’s motivation was largely concern about racial and socioeconomic exclusion in the suburbs, given New Jersey’s stark distinctions between older, largely African American central cities and more affluent suburbs. An initially reluctant state legislature followed up on the court’s prodding by passing a Fair Housing Act in 1985 that created COAH to oversee the process, thereby removing most of the direct judicial supervision of contested municipal housing policies. In terms of the law’s actual implementation, the major salient differences between New Jersey’s approach and California’s include the following: 1. The fair-share allocation process is state-directed, with no COG involvement, and pertains only to low- and moderate-income housing needs, not “above moderate” needs. These affordability brackets have stricter definitions in New Jersey, however, with the low-income category basically equivalent to California’s verylow-income bracket, and the moderate category equivalent to California’s low-income grouping. 2. Municipalities are not obligated to have COAH review their housing elements. However, localities that choose to approach COAH and have their housing plans certified as adequate are shielded from lawsuits by builders for a six-year period. 77 3. Municipalities can meet part of their affordable housing obligations by rehabilitating existing units. Congregate housing facilities for the disabled can also be counted toward the obligation. 4. The major mechanism for the construction of affordable units is through the so-called “builder’s remedy,” under which a developer offers to construct affordable units that go some way toward meeting a community’s obligations, in exchange for certain local concessions. These projects ordinarily are 80 percent market rate units and 20 percent low- and moderateincome units. The concession given to the developer is typically a density bonus allowing construction of more market-rate units than would otherwise be allowed under local zoning, and in some cases local government offsets of the cost of the affordable units. Communities that have not met their affordable housing obligations and have not had their housing elements certified by COAH are often approached by for-profit builders proposing such an arrangement, which can be ordered by a court or agreed to by the municipality in a settlement. To avoid such suits, many local governments adopt an “inclusionary housing” ordinance, providing for a routine affordability set-aside of 20 percent of units in any development of a threshold size. During a building boom in the late 1980s, an estimated 15 percent of all developments in New Jersey included such an affordability setaside (Calavita et al., 1997, p. 127). 5. As a quasijudicial agency, “COAH has adopted a massive body of regulations governing nearly every element or program . . . through which a New Jersey locality might seek to comply with the Mount Laurel mandate” (Calavita et al., 1997, p. 119). 6. Municipalities in New Jersey are allowed to “trade” up to half their obligated number of affordable units to another jurisdiction that agrees to build the units (or, more commonly, to renovate dilapidated units), under so-called Regional Contribution Agreements. The “sending” jurisdiction, generally a suburb, must pay the “receiving” jurisdiction, usually an older 78 central city, at least $20,000 per unit, although higher amounts can be negotiated. Some view this aspect of the New Jersey system as a “sell-out” of the original aim of distributing housing opportunity widely. But other analysts note that the trading system has the merit of effectively serving as a tax on those communities that choose to avoid their affordable housing obligations (Hughes and VanDoren, 1990). Thus, the entire community that “benefits” from such exclusion must pay for it; contrast this to the mechanism of the builder’s remedy, in which the costs of new affordable units may be cross-subsidized by the market-rate units in the same project, meaning that new entrants to a community effectively pay for the city’s past exclusionary behavior. The trading system also has fiscal benefits for distressed central cities. Although a large amount of subsidized housing has been produced under the Mount Laurel system in New Jersey, it has been faulted on a number of counts. Most notably, the emphasis in most builder’s remedy developments has been on purchase housing, not rentals, meaning that the lowest-income households often cannot take advantage of the setasides because they cannot afford a down payment or do not have good credit. Thus, studies have indicated that a typical occupant of a subsidized suburban Mount Laurel unit is a young white family that already lived in suburbia and simply has not yet achieved its full earnings potential. On the other hand, a major advantage of the New Jersey approach is its standardized procedure for meeting local fair-share housing goals—a process that after more than 15 years of operation is becoming routine. The builder’s remedy approach uses the profit motive of developers to accelerate implementation and to develop far more housing units than would have otherwise been allowed under local zoning. Massachusetts The first strong fair-share policy approach to housing development anywhere in the country was probably Massachusetts’s Comprehensive 79 Permit Law.3 The law was enacted in 1969 and aimed at opening up the suburbs to affordable housing. It has often been viewed as an attack upon “snob zoning.” The Massachusetts law is typically referred to as “40B” in that state because of its chapter location in the state legal code. Its main strategy is to radically simplify the review process for developers seeking to build residential projects in which at least 25 percent of the units have long-term affordability restrictions. Developers proposing such projects can apply to the municipality for “comprehensive” building permits, through a local Zoning Board of Appeals. If successful, the comprehensive permit issued by the zoning board cuts through local building restrictions and overrides land-use regulations that are inconsistent with providing affordable housing; the zoning board can effectively overturn local zoning in some cases. This enables the builder to use one relatively streamlined process to receive an entitlement to build the project. This local zoning board is charged with bringing to the table all the relevant municipal departments and agencies that have an interest in the housing development for comments and concerns (such as the planning board, fire department, etc.), and with conducting a series of public hearings to air community concerns. The hearings must begin within 30 days of receipt of the developer’s application for a comprehensive permit, and the board must decide whether to award the permit within 40 days after the public hearings conclude. Although state environmental and other statutes still apply, local authorities must use the principle of meeting local needs in any denial or conditions that they impose upon the project (such as density or height restrictions). The community is not allowed to impose conditions that render the proposed project uneconomical to the builder. Furthermore, developers whose projects are denied by this local process have the option of appealing to a statewide board, the Housing Appeals Committee (HAC), which can overturn the local decision and order that the project ____________ 3This discussion of the Massachusetts process draws upon numerous recent articles in the Boston Globe by housing reporter Anthony Flint, materials from the websites of the Massachusetts Department of Housing and Community Development and the Citizens Housing and Planning Association, as well as early perspectives on the law by Listokin (1976). 80 be permitted. A developer can also appeal a local decision if conditions imposed upon the project threaten its economic feasibility. However, the developer may only appeal if the locality in which they wish to build has failed to meet certain housing affordability thresholds.4 Ordinarily, the HAC first encourages the applicant and the municipality to engage in a mediation process to see if issues can be resolved without a HAC review. In the initial period after passage of the 40B law, progress was very slow, as there were court challenges to the law and frequent proposals in the legislature to weaken it. By 1972, there had been only 35 applications to local authorities for comprehensive permits, and only a few appeals to the state HAC. After six years of operation, the law had resulted in the construction of only about 1,100 new housing units, although more than 3,000 others were planned or pending (Listokin 1976, pp. 100–103). More recently, there has been a pronounced increase in the number of projects proposed under the law, with about 5,000 units produced between 1990 and 1997. Between 1970 and 2001, about 25,000 units in 170 communities had been approved (Massachusetts DHCD, 2001). According to Flint (2002), “Because many towns have put the brakes on all growth and buildable land is scarce, developers say going the 40-B route—generally large, dense projects that are 25 percent affordable under state guidelines—is the only way to do business these days.” However, a disproportionate number of the recent 40B applications have been projects for senior citizens only, which local officials often look upon more favorably. The law is also limited in that it does not apply to many commercial developers. To have the right to seek a comprehensive permit, the developer must be either a nonprofit or a limited-dividend company, and the project must initially receive approval from a federal or state subsidized housing program. In recent years, the Federal Home Loan Bank of Boston has provided substantial subsidies to developers in the program, which has helped boost construction. Although 40B ____________ 4The city is exempted from an appeal if at least 10 percent of its housing units are in projects subsidized for low- and moderate-income residents, or if such affordable units constitute at least 1.5 percent of the city’s land area zoned for residential, industrial, and commercial uses. Few jurisdictions in Massachusetts meet these criteria. 81 applications have become more numerous lately, the law has not succeeded in easing that state’s very serious housing shortage. The law has contributed to a much wider geographic distribution of subsidized housing, with the proportion of subsidized units outside large cities in Massachusetts rising from 31 percent in 1972 to 63 percent in 1997 (Citizens Housing and Planning Association, 2001). A state task force there recently proposed more radical state oversight of local planning, such as overturning local caps on growth (Flint, 2001). Given the shortage of housing subsidies in contemporary California (Little Hoover Commission, 2002), adapting the current Massachusetts statutory provisions to California is unlikely to bear immediate results in terms of housing production. The 40B law reflects its origins as a civilrights-oriented provision rather than a measure to increase housing production across the board. However, if a California version were to relax the requirement that the affordable units be subsidized by a state or federal source, and perhaps reduce the required share of affordable units to 15 or 20 percent of the project total, it would open comprehensive permitting to a far wider array of projects. Many developers in urban and suburban California already have built projects in which 15 percent or so of the units are affordable and cross-subsidized, in effect, by the market-rate units. Such approaches are mandated by some cities’ inclusionary housing requirements. In California, the right of appeal to a statewide board (or perhaps more appropriately, a regional board) could be allowed if the local government has a noncompliant housing element. The Portland Region Oregon is well-known for undertaking state-level review of local housing plans.5 In 1973, at the governor’s urging, the legislature passed a law requiring that local governments prepare comprehensive plans and establishing a state-level, quasijudicial Land Conservation and Development Commission (LCDC) to review them. One of the initial 14 statewide goals developed by LCDC concerned housing opportunity. The LCDC goal required that, as in California, localities inventory ____________ 5The Portland discussion draws upon Hammond (2002), Knaap and Nelson (1992, Chapter 3), Lewis (1996, pp. 179–182), and Toulan (1994). 82 residential land and develop plans to encourage the building of an adequate number and mix of housing units for households in various income brackets (Knaap and Nelson, 1992, p. 77). In LCDC’s initial reviews of local housing plans, every jurisdiction of more than 5,000 population in the Portland metropolitan area except the central city itself saw its housing elements rejected because they provided for an insufficient range of housing types and prices. The area’s regional government, the Metropolitan Service District (since renamed Metro) attempted to resolve this state-local conflict by enacting an Areawide Housing Opportunity Plan—a fair-share allocation plan that distributed federal housing subsidies in relation to local needs. The plan withered when federal housing funds dried up in the early 1980s, but LCDC adopted the broad outlines of the Metro plan as its own regulation, later codified into state law by the legislature. As adopted, the so-called Metropolitan Housing Rule applies only to the Portland region, which contains about half the state’s population. The rule prescribes local land-use mixing, requiring that local government plans within Metro’s boundaries allow for at least 50 percent of new residential units to be multifamily or attached housing units. It sets out a “10-8-6” formula, which requires that the largest municipalities in the region zone for an average of at least ten dwelling units per buildable acre, whereas almost all the remaining suburbs must zone for at least eight units per acre. Five very small cities were given a quota of six units per acre. Despite local government objections to the new rule, strong support at the state and regional level along with support from an unusual alliance of environmentalists and homebuilders led to the implementation of the plan in virtually every Portland-area community. The authority of LCDC to fiscally sanction uncooperative localities no doubt helped speed up the process. The amount of land zoned for multifamily housing increased fourfold to more than one-quarter of net buildable acres. By 1991, vacant residential land in the Portland region allowed for the development of 54 percent multifamily units, and some rapidly growing suburbs, such as Beaverton, heavily exceeded the 50 percent minimum. 83 During the boom of the 1990s, apartments proliferated in the suburbs of Portland, which helped lead to the recent finding that households below the poverty line in that metropolitan area became less concentrated in the central city and more likely to locate in suburbia by the time of the 2000 Census. Although neighborhood objections to multifamily housing continue in the Portland area, according to one homebuilder, “nobody has been able to keep [multifamily housing] totally out. The reason is because of the [metropolitan housing] rule. Without that, I am sure it would have followed the same pattern the rest of the country did” (quoted in Hammond, 2002). However, no systematic study of the rule has attempted to determine the degree to which it assists in housing affordability or socioeconomic mixing. Another aspect of Oregon’s statewide planning system, the requirement for urban growth boundaries around each metropolitan area, directs growth pressures inward and probably helps account for the sharp spikes in densities in the Portland region. After all, the increased zoning densities under the Metropolitan Housing Rule do not require that builders propose projects at such densities. In a different regulatory regime, developers might have found it more profitable to build lowerdensity housing in the existing suburbs while also rapidly expanding lowdensity housing beyond the urban fringe. Conclusion There have been numerous proposals to reform housing element law in California, involving penalties for noncompliance, more selfcertification, or transfers of allocations across communities. Most suggested reforms, however, would proceed within the current approach: state review of local housing plans. A more fundamental shift would be to directly reward effective performance by directing funds to localities on the basis of the number and mix of new units developed. The experiences of Massachusetts, New Jersey, and Oregon all offer different experiences regarding fair-share planning and state review of local housing policy. Even if none of these approaches is completely appropriate to the California context, some combination of the 84 innovations attempted elsewhere—comprehensive permitting, state or regional boards of review or appeals, minimum zoning densities, or the builder’s remedy—may be worthy of consideration in California. 85 6. Conclusion Many factors appear to complicate efforts by local governments to reach housing compliance. In some cases, local policies restricting growth or an older housing stock hinder compliance, although it also appears that communities that have remained persistently noncompliant tend to be high-status cities. When it comes to sorting out the effects of noncompliance, however, there is no solid evidence that noncompliant cities experience a slower rate of housing development than compliant cities, holding all else constant. Thus, given the substantial controversy over the existing housing element law, it is worth considering alternative policy approaches, some of which were discussed in Chapter 5. An important question sitting astride the whole issue is whether the various fundamental values that California policymakers hold dear— increased housing production, an equitable distribution of housing responsibility across communities, special attention to the housing needs of low-income families, local autonomy and home rule, environmental protection, and more—can all be accommodated within housing element law, or whether there are tensions among them. A secondary question is whether the current approach—state review of local planning for future housing needs—is the most effective use of resources to further these goals. Goals in Conflict Although other states have experimented with different approaches to state review of local housing plans, California has continued with its own approach to housing elements. As in the other states, California’s housing element requirement has often gone against the grain of local policymaking because of its regionalist orientation: It asks cities to plan for the needs of the wider region, not just the needs of current city residents. Indeed, the philosophy behind fair-share housing plans is that the local “police power” to regulate land use should be directed toward 87 the general welfare of the region and not just the general welfare of the specific locality (Listokin, 1976, p. 19). Equitable Distribution or Sheer Production? Beyond this conflict over the local role, there are more fundamental value conflicts in housing element policy in California. It took shape in the 1970s in an era in which there was increasing concern with civil rights and the ability of minorities and low-income families to have an opportunity to live in suburbia, not just in inner-city or rural enclaves. Most of the fair-share plans developed by regions or states in this time period had such equal-opportunity goals. Housing element policy was also concerned with improving the local planning process—a major policy goal for the state through much of the post-World War II era. In more recent decades, California policymakers have also become very concerned with overall residential production and the need to increase housing supplies, particularly but not exclusively at the low- and moderate-income levels. Thus, policymakers have increasingly turned to the housing element process as a tool to ensure adequate housing production throughout the state. But it is not clear that a fair-share allocation and planning process is necessarily the best tool for this job. In the contemporary period, many of the most racially and socioeconomically diverse jurisdictions in California are suburbs (Sandoval, Johnson, and Tafoya, 2002), although there remain numerous pockets of privilege in which low-income and minority populations are largely absent. In many cases, a greater number of affordable units could feasibly be produced in areas where there are already sizeable low-income populations than in the wealthier jurisdictions, where land prices are likely to be much higher. Thus, the goal of maximizing housing production can conflict with the goal of allocating a region’s needs as fair shares to each municipality. A scheme to allow widespread trading of housing allocations to other jurisdictions, as in New Jersey, could reduce this conflict, but many housing advocates and policymakers have little sympathy for allowing high-status communities to escape what is seen as their responsibilities to 88 the broader society.1 What the conflict implies, perhaps, is that different tools are necessary to enhance overall production from the tools necessary to encourage equal housing opportunity across jurisdictions. A Mandate in a Vacuum? Then, too, there are the conflicts between the housing element law’s single-minded pursuit of increased housing and other state laws that tend to raise barriers for housing production. As already noted, open-space and coastal protections, environmental impact review, congestion management, and farmland preservation are also all encouraged under California law and are not always easily reconciled with housing element provisions that seek to identify many land sites for future housing construction. The state’s new law requiring that local governments identify a 20-year supply of water for new housing developments before approving them also sends a different signal from housing element policy. These potential conflicts with environmental or growthmanagement goals were recognized even by an early observer of fair-share housing policies: Fair share basically accepts growth; it merely wants to direct it in a rational and equitable manner. The nongrowth or phased-growth philosophy questions whether growth, especially rapid growth, should be tolerated. This approach is therefore in contradiction with the regional allocation philosophy (Listokin, 1976, p. 137). Devil in the Details Those involved in the debate over housing policy often note how long and highly detailed the housing element statute is. Indeed, when the relevant legal code is downloaded from the Internet and printed, it stretches out over 27 pages. Highly detailed statutes are often evidence of widespread disagreement on a given policy, as waves of “reform” occur ____________ 1New Jersey’s Regional Contribution Agreements “shifted the rationale of the Mount Laurel doctrine away from the broad goal of ending geographic segregation surrounding inner-city minorities and toward the raw provision of low-income housing. . . . [Homeowners in the sending jurisdictions] are amenable, it turns out, to paying a form of ransom, through taxes, that preserves local control of new entrants while allowing lower-income housing to be built elsewhere” (Haar, 1996, p. 114). 89 in which opposing interests seek to have their concerns addressed and protections preserved in law. What has resulted, in the case of the housing element statute, is an unwieldy law that is often difficult for outside observers (including this researcher) to comprehend in its entirety or details. In the 33 years since the housing element statute was enacted, the search for an adequate supply of housing in California has become only more elusive. It may be a ripe occasion for policymakers and affected interests to regroup and reconsider the goals and approaches of state housing law, seeking an approach that is more workable, transparent, and straightforward, with measurable barometers of success or failure. In so doing, policymakers will need to resolve whether the major goal of such a law is a sheer increase in residential construction or an equitable distribution of affordable housing. Using a fair-share planning approach as a tool to encourage overall housing production may place an unrealistic burden on a relatively fragile, process-oriented policy. 90 Appendix A Multivariate Analyses of Housing Element Noncompliance This appendix reports the results of logistic regression analysis of city characteristics that might be expected to influence whether a city complies with housing element law. Chapter 3 discusses the general logic behind the analysis, and Table 3.3 describes the variables and data sources. The dependent variable is a dichotomous measure of noncompliance, equal to 1 if the city is noncompliant and 0 if compliant. If the city’s housing element draft is overdue, it is legally noncompliant and is treated as such. Cities under review by HCD at the time of the September 25, 2002, compliance report are dropped from the analysis. The table reports “odds ratios,” which represent the change in odds that an average city will be noncompliant, given a one-unit increase in the variable of interest. Odds ratios above 1.0 indicate an increased probability, and those below 1.0 a reduced probability, of noncompliance. Table A.1 provides results of a main model and an alternative model. The only difference is that the alternative model includes a measure of the degree to which the city planning director feels that the policies affecting residential growth rates have resulted from popular pressure by neighborhoods or voter initiatives (using a four-point scale). However, quite a few planners responding to the survey answered “don’t know” or skipped this question, which results in the loss of 41 cities from the analysis. Thus, I am more confident in the results of the main model. Table A.2 provides summary statistics for the variables in the models. 91 Table A.1 Logistic Models of Housing Element Noncompliance as of September 2002 92 Main Model Independent Variable Odds Ratio z-Statistic % of housing owner-occupied 1.023 1.35 % of residents white, non-Hispanic 0.998 –0.25 % of recreational housing units 1.023 0.35 Population density (natural log) Age of median housing unit (years) 1.033 1.079 0.10 2.69 Supply of land constrains growth (five-point scale) 0.910 –0.60 Population (natural log) 0.698 –1.81 Per capita own-source revenues (natural log) 1.087 0.25 City’s % share of local property tax revenues 1.037 1.03 Lack of planning staff constrains growth (five-point scale) 1.173 0.81 Review of residential proposals getting longer (four-point scale) 2.082 2.63 No. of overt growth restrictions (of possible seven) 2.373 4.39 City requires affordable units in new developments (dichotomous) 0.558 Slow-growth policies resulting from initiatives, neighborhoods (four-point scale) N/I –1.44 N/I City is in Inland Empire (dichotomous) 3.952 1.88 City is in San Francisco Bay Area (dichotomous) 3.056 2.33 City is in Central Valley (dichotomous) 31.121 2.68 No. of months since housing element revision due 0.946 –4.40 Alternative Model P>|z| Odds Ratio z-Statistic P>|z| 0.178 1.015 0.84 0.402 0.805 1.002 0.22 0.827 0.725 0.989 –0.15 0.877 0.917 0.911 0.007 1.093 –0.27 0.785 2.74 0.006 0.662 0.897 –0.60 0.546 0.070 0.645 –2.12 0.034 0.804 1.082 0.21 0.832 0.304 1.013 0.35 0.724 0.416 1.079 0.35 0.729 0.009 1.978 1.98 0.048 0.000 2.443 4.11 0.000 0.150 0.395 N/I 1.570 –1.84 0.066 1.63 0.103 0.060 3.792 1.67 0.095 0.020 2.679 1.78 0.075 0.007 32.510 2.47 0.014 0.000 0.946 –3.96 0.000 No. of cities Log likelihood Pseudo R-squared % correctly predicted 224 –98.41 0.36 79.0 183 –77.53 0.38 82.5 NOTES: Dependent variable is dichotomous (1 = noncompliant, 0 = compliant). N/I indicates variable not included in the model. Significance calculated using robust standard errors. Table A.2 Summary Statistics for Continuous Variables in the Preceding Models Variable Mean % of housing owner-occupied 60.3 % of residents white, non-Hispanic 49.5 % recreational housing units 1.4 Population density (natural log) 8.1 Age of median housing unit (years to 1990) 30.2 Supply of land constrains growth (five- point scale) 3.5 Population (natural log) 10.4 Per capita own-source revenues (natural log) 6.1 City’s % share of local property tax revenues 13.2 Lack of planning staff constrains growth (five-point scale) 2.2 Review of residential proposals getting longer (four-point scale) 1.9 No. of overt growth restrictions (of possible seven) 0.7 Slow-growth policies resulting from initiatives, neighborhoods (four-point scale) 1.6 No. of months since housing element revision due 35.7 Standard Minimum Maximum Deviation Value Value 13.1 23.8 97.1 23.8 1.0 91.2 4.1 0 34.9 0.9 3.7 10.1 8.7 11 59 1.4 1 5 1.2 6.1 14.0 0.6 4.7 8.4 7.4 0 47.6 1.1 1 5 0.7 1 4 1.0 0 5 0.8 1 40.3 3 4 117 NOTE: Values are only for cities included in model in Table A.1. 93 Appendix B Multivariate Analyses of the Rate and Mix of New Housing Production in the 1990s This appendix reports the results of ordinary least-squares regression analyses of city characteristics that might influence the rate of growth in housing units between 1990 and 2000 and the mix between single-family and multifamily units. The general logic behind the analysis is discussed in Chapter 4, where the section entitled “Controlling for Other Factors” describes the independent variables employed, virtually all of which are derived from Census sources. An exception is the measure of housing element noncompliance, derived from an HCD compliance status report dated December 31, 1991. The dependent variable in the first regression (Table B.1) is the percentage change in the number of housing units in the city between the 1990 and 2000 Censuses. The dependent variable in the second regression (Table B.2) is the ratio of newly constructed multifamily units (1991–2000) to total existing units (in 1990). The dependent variable in the third regression (Table B.3) is the percentage share of newly built units that are multifamily units. The source of the data on new construction of housing units is the Construction Industry Research Board. The analysis is limited to cities in the regions covered by the San Diego Association of Governments, the Southern California Association of Governments, and the Association of Bay Area Governments. These regions had completed their second round of housing element revisions by 1991. Finally, Table B.4 provides summary statistics for the variables in the models. 95 Table B.1 Regression Model of Housing Unit Growth in the 1990s, with County Fixed Effects Independent Variable Housing growth in prior decade Population size (log) Population density (log) Urbanized Age of housing stock Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Noncompliant housing element B0 % Growth in Housing Units Coef. t-Value P>|t| 0.174 5.64 0.000 –1.412 –1.73 0.085 –3.216 –1.68 0.096 –0.943 –0.28 0.779 –0.214 –2.16 0.032 –0.000 –1.88 0.062 –0.243 –1.07 0.285 –1.079 –1.85 0.065 0.561 0.42 0.678 –0.618 –2.08 0.039 –8.927 –2.97 0.003 1.473 1.03 0.303 83.958 3.24 0.001 No. of cities Prob. > F Adjusted R-squared 202 0.000 0.64 NOTES: Dependent variable is the percentage increase in total housing units (1990–2000). Analysis is of cities in the San Diego Association of Governments, Southern California Association of Governments, and Association of Bay Area Governments and is limited to cities that were incorporated and had a population of at least 2,500 in 1990. Robust standard errors are calculated. County fixed effects are used for all counties except San Francisco and are statistically significant (F = 4.421, p = .000). 96 Table B.2 Regression Model of Multifamily Development Growth in the 1990s, with County Fixed Effects Independent Variable Housing growth in prior decade Population size (log) Population density (log) Urbanized Age of housing stock Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Noncompliant housing element B0 Ratio of New Multifamily Units to Existing Units Coef. t–Value P>|t| –0.000 –0.17 0.868 –0.005 –0.89 0.374 –0.008 –0.78 0.435 0.009 0.59 0.558 –0.001 –1.26 0.210 –0.000 –1.25 0.213 –0.001 –2.09 0.038 –0.001 –0.69 0.494 0.021 2.45 0.015 –0.000 –0.11 0.911 –0.010 –0.70 0.485 –0.006 –1.44 0.151 0.211 1.62 0.106 No. of cities Prob. > F Adjusted R-squared 200 0.005 0.23 NOTES: Dependent variable is the ratio of newly constructed multifamily units (1991–2000) to the total number of housing units in 1990. Analysis is of cities in the San Diego Association of Governments, Southern California Association of Governments, and Association of Bay Area Governments and is limited to cities that were incorporated and had a population of at least 2,500 in 1990. Robust standard errors are calculated. County fixed effects are used for all counties except San Francisco and are statistically significant (F = 1.779, p = .041). 97 Table B.3 Regression Model of Share of Housing Development That Is Multifamily, with County Fixed Effects Independent Variable Housing growth in prior decade Population size (log) Population density (log) Urbanized Age of housing stock Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Noncompliant housing element B0 % of New Housing Units That Are Multifamily Coef. t–Value P>|t| –0.083 –2.30 0.023 1.650 0.70 0.484 4.804 1.26 0.210 –4.259 –0.62 0.535 –0.085 –0.30 0.766 –0.000 –2.96 0.004 –0.229 –0.72 0.474 0.288 0.17 0.868 12.379 3.09 0.002 –0.545 –1.19 0.235 0.658 0.07 0.946 –8.106 –2.18 0.031 19.049 0.44 0.660 No. of cities Prob. > F Adjusted R-squared 200 0.000 0.36 NOTES: Dependent variable is the percentage share of housing units developed (1991–2000) that are multifamily units. Analysis is of cities in the San Diego Association of Governments, Southern California Association of Governments, and Association of Bay Area Governments and is limited to cities that were incorporated and had a population of at least 2,500 in 1990. Robust standard errors are calculated. County fixed effects are used for all counties except San Francisco and are statistically significant (F = 1.759, p = .044). 98 Table B.4 Summary Statistics for Continuous Variables in the Preceding Models Dependent Variables % housing growth in 1990s Ratio of new multifamily units to total 1990 units % of new housing units that are multifamily Independent Variables % housing growth in prior decade Population size (log) Population density (log) Age of housing stock (to 1990) Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Standard Minimum Maximum Mean Deviation Value Value 9.6 13.7 –9.7 73.8 0.02 0.03 0 0.28 29.0 24.7 0 98.4 25.3 10.4 8.3 24.7 43,900 11.6 3.7 –0.1 24.8 0.4 39.7 1.1 0.8 0.7 18,489 6.2 1.3 0.5 4.1 0.4 –7.0 7.9 5.2 4 18,635 3.5 0 –1.8 14.1 0 389.8 13.6 10.1 51 130,734 47.9 7.9 1.8 40.5 3.2 NOTES: Values are only for cities included in the model in Tables B.1, B.2, and B.3. Commute time is measured in minutes, distance from urban center is measured in mapping units. 99 Bibliography Barbour, Elisa, Metropolitan Growth Planning in California, 1900–2000, Public Policy Institute of California, San Francisco, California, 2002. Barbour, Elisa, and Paul Lewis, Development Priorities in California Cities: Results from a PPIC Survey, Occasional Paper, Public Policy Institute of California, San Francisco, California, 1998. Bay Area Council, “ABAG Underestimates Regional Housing Needs,” Housing and Development Report, Vol. 1, No. 4, September 1988(a), p. 1. Bay Area Council, “Fair Share Housing Production: How Has the Bay Area Performed?” Housing and Development Report, Vol. 1, No. 4, September 1988(b), p. 3. Bishop, Amanda, and Jessica Materna, “Housing Bill Pressures Cities,” San Francisco Business Times, October 15, 2001. Brownlow, Judith, Fair Share Housing: A Partially Annotated Bibliography, Council of Planning Librarians, 1991. Brueckner, Jan, “Strategic Control of Growth in a System of Cities,” Journal of Political Economy, Vol. 57, 1995, pp. 393–416. Calavita, Nico, Kenneth Grimes, and Alan Mallach, “Inclusionary Housing in California and New Jersey: A Comparative Analysis,” Housing Policy Debate, Vol. 8, No. 1, 1997, pp. 109–142. Carrigg, Dan, “Balancing Housing and Growth Pressures with Limited Resources: It’s Time for Leadership,” Western City, Vol. 78, No. 4, April 2002, pp. 13, 15–16, 18–22. Citizens Housing and Planning Association, Recommendations to Improve Chapter 40B, the Massachusetts Comprehensive Permit Law, Boston, Massachusetts, 2001. Curtin, Daniel J., Jr., California Land Use and Planning Law, 20th ed., Solano Press Books, Point Arena, California, 2000. 101 Department of Housing and Community Development (HCD), Developing a Regional Housing Needs Plan, State of California, Business, Transportation, and Housing Agency, Sacramento, California, 1988. Department of Housing and Community Development (HCD), Status of Local Housing Elements (Period of October 1, 1990 through December 31, 1990), State of California, Business, Transportation, and Housing Agency, Sacramento, California, February 1991. Department of Housing and Community Development (HCD), Raising the Roof: California Housing Development Projections and Constraints, 1997–2020, State of California, Business, Transportation, and Housing Agency, Sacramento, California, May 2000. Department of Housing and Community Development (HCD), Pay To Play: Residential Development Fees in California Cities and Counties, State of California, Business, Transportation, and Housing Agency, Sacramento, California, 2001. Department of Housing and Community Development (HCD), “Housing Element Review Worksheet,” State of California, Business, Transportation, and Housing Agency, Sacramento, California, n. d. Dodge, Shannon, San Francisco Bay Area Housing Crisis Report Card, Greenbelt Alliance and Nonprofit Housing Association of Northern California, San Francisco, California, June 2002. Farley, Reynolds, “Suburban Persistence,” American Sociological Review, Vol. 29, February 1964, pp. 38–47. Fischel, William A., The Homevoter Hypothesis: How Home Values Influence Local Government Taxation, School Finance, and Land-Use Policies, Harvard University Press, Cambridge, Massachusetts, 2001. Flint, Anthony, “Report: Housing Inhibited by Rules,” Boston Globe, December 16, 2001, p. B1. Flint, Anthony, “Worries Develop as Town Expands: Affordable Housing Projects Piling Up,” Boston Globe, February 10, 2002, p. B1. Fulton, William, Guide to California Planning, 2nd ed., Solano Press Books, Point Arena, California, 1999. 102 Haar, Charles M., Suburbs under Siege: Race, Space, and Audacious Judges, Princeton University Press, Princeton, New Jersey, 1996. Hammond, Betsy, “Income Groups Intermingle,” The Oregonian, May 15, 2002, p. A1. Hecht, Peter, “Emotional Hearing on Folsom Housing,” Sacramento Bee, May 14, 2002. Hill, Laura E., and Hans P. Johnson, Understanding the Future of Californians’ Fertility: The Role of Immigrants, Public Policy Institute of California, San Francisco, California, 2002. Hughes, Mark Alan, and Peter M. VanDoren, “Social Policy Through Land Reform: New Jersey’s Mount Laurel Controversy,” Political Science Quarterly, Vol. 105, No. 1, 1990, pp. 97–111. Jackson, Carlos, “Realistic Housing Element Law Needed,” California County, September/October 1994, pp. 23–34. Knaap, Gerrit, and Arthur C. Nelson, The Regulated Landscape: Lessons on State Land Use Planning from Oregon, Lincoln Institute of Land Policy, Cambridge, Massachusetts, 1992. Lewis, Paul G., Shaping Suburbia: How Political Institutions Organize Urban Development, University of Pittsburgh Press, Pittsburgh, Pennsylvania, 1996. Lewis, Paul G., “Creating a ‘Home’ for Good Housing Policy,” Western City, Vol. 78, No. 8, August 2002, pp. 16–17, 22, 52. Lewis, Paul G., and Max Neiman, Residential Development and Growth Control Policies: Survey Results from Cities in Three California Regions, Occasional Paper, Public Policy Institute of California, San Francisco, California, 2000. Lewis, Paul G., and Max Neiman, Cities Under Pressure: Local Growth Controls and Residential Development Policy, Public Policy Institute of California, San Francisco, California, 2002. Listokin, David, Fair Share Housing Allocation, Center for Urban Policy Research, Rutgers University, New Brunswick, New Jersey, 1976. Little Hoover Commission, Rebuilding the Dream: Solving California’s Affordable Housing Crisis, Milton Marks “Little Hoover” Commission 103 on California State Organization and Economy, Sacramento, California, 2002. Lyons, Silas, “County Leaders Oppose State’s Housing Mandate,” San Luis Obispo Tribune, January 10, 2002a, p. A1. Lyons, Silas, “Raising the Roof in SLO County: Communities Struggle with Agency’s Housing Demands,” San Luis Obispo Tribune, January 27, 2002b, p. A1. Massachusetts Department of Housing and Community Development (Massachusetts DHCD), “Chapter 40B: Frequently Asked Questions,” February 2001, available at www.state.ma.us/dhcd/ Ch40B/FAQ.htm (downloaded March 28, 2002). Moller, Rosa Maria, Hans Johnson, and Michael Dardia, What Explains Crowding in California? California Research Bureau, Sacramento, California, February 2002. “Motel Conversion,” archives of an online discussion among municipal planners, posted by the League of California Cities, downloaded April 23, 2002, available at http://www.cacities.org/doc.asp?id=3896. Myers, Dowell, “Demographic Futures as a Guide to Planning: California’s Latinos and the Compact City,” Journal of the American Planning Association, Vol. 67, No. 4, 2001, pp. 383–397. Myers, Dowell, and Julie Park, The Great Housing Collapse in California, Fannie Mae Foundation, Washington, D.C., May 2002. Neiman, Max, “Zoning Policy, Income Clustering, and Suburban Change,” Social Science Quarterly, Vol. 61, 1980, pp. 666–675. Nguyen, Hang, “Cities Dismayed by Affordable Housing Measure,” Los Angeles Times, April 20, 2001, p. A3. Padmanabhan, Sekhar, “Judge Affirms Folsom Housing Pact,” Sacramento Bee, April 9, 2002, p. B1. Perkes, Courtney, “O.C. Homes Among Most Crowded in the Nation,” Orange County Register, June 16, 2002. Rawson, Mike, “Housing Element Law,” California Housing Law Project, Sacramento, 2000, available at http://www.housingadvocates. org/pdf/site/facts/HousingElementLaw.pdf. 104 Richards, Watson & Gershon, “The Housing Element Statute: A Status Report,” Los Angeles, California, March 1998, available at http:// www.rwglaw.com/Articles/Articles312. Richards, Watson & Gershon, “New Housing Element Legislation: The End of Mandate Suspension,” Los Angeles, California, January 1999, available at http://www.rwglaw.com/Articles/Articles414. Sandoval, Juan Onésimo, Hans P. Johnson, and Sonya M. Tafoya, “Who’s Your Neighbor? Residential Segregation and Diversity in California,” California Counts, Vol. 4, No. 1, August 2002. Select Committee on Housing and Urban Affairs, California Senate, Department of Housing and Community Development v. Housing Needs of the People of the State of California, Senate Publications, Sacramento, California, 1974. Senate Committee on Housing and Land Use, California Senate, Housing Element Law: A Summary Report from the Special Hearing, Senate Publications, Sacramento, California, March 27, 1995. Senate Committee on Local Government, California Senate, Housing Element Law: A Summary Report from the Interim Hearing, Senate Publications, Sacramento, California, December 8, 1993. Shigley, Paul, “Rural Counties Question Housing Policy,” California Planning & Development Report, Vol. 17, No. 2, February 2002, pp. 1, 14–15. Toulan, Nohad, “Housing as a State Planning Goal,” in Carl Abbott, Deborah Howe, and Sy Adler, eds., Planning the Oregon Way (Chapter 5), Oregon State University Press, Corvallis, Oregon, 1994. Warner, Karen, J. Alfred Dichoso, Maureen Markham, Michael McLaughlin, and Candace Stowell, “New Approaches to the Housing Element: Case Studies from California, Nevada, and Washington State,” Conference Proceedings, American Planning Association Conference, San Diego, 1997, available at http://www.asu.edu/caed/ proceedings97/warner.html. Weinstein, Alan C., “Exclusionary Zoning in the Suburbs: The Road from Mount Laurel to Chester,” Zoning and Planning Law Report, Vol. 16, No. 3, 1993, pp. 105–112. 105 Wisckol, Martin, “Affordable Housing Lacking: 12 O.C. Cities Don’t Have Required Plans,” Orange County Register, April 18, 2002. 106 About the Author PAUL G. LEWIS Paul G. Lewis is a research fellow and director of the Governance and Public Finance program at the Public Policy Institute of California. He has authored or co-authored numerous reports, journal articles, and other publications on local government structure, local sales taxes, urban development, transportation policy, regional governance, and local voter turnout. He has also presented his findings on these topics to a variety of governmental bodies and civic organizations. He holds an A.B. in political science from Indiana University and an M.A. and Ph.D. in politics from Princeton 107 Related PPIC Publications Who Pays for Development Fees and Exactions? Marla Dresch and Steven M. Sheffrin Cities Under Pressure: Local Growth Controls and Residential Development Policy Paul G. Lewis and Max Neiman Local Economic Development in Southern California’s Suburbs: 1990– 1997 Max Neiman, Gregory Andranovich, and Kenneth Fernandez Building California’s Future: Current Conditions in Infrastructure Planning, Budgeting, and Financing Michael Neuman and Jan Whittington Homelessness in California John M. Quigley, Steven Raphael, and Eugene Smolensky PPIC publications may be ordered by phone or from our website (800) 232-5343 [mainland U.S.] (415) 291-4400 [Canada, Hawaii, overseas] www.ppic.org 109" } ["___content":protected]=> string(102) "

R 203PLR

" ["_permalink":protected]=> string(107) "https://www.ppic.org/publication/californias-housing-element-law-the-issue-of-local-noncompliance/r_203plr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8298) ["ID"]=> int(8298) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:36:23" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(3469) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 203PLR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_203plr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_203PLR.pdf" ["wpmf_size"]=> string(6) "660779" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(242147) "California’s Housing Element Law: The Issue of Local Noncompliance ••• Paul G. Lewis 2003 PUBLIC POLICY INSTITUTE OF CALIFORNIA Library of Congress Cataloging-in-Publication Data Lewis, Paul George, 1966- California’s housing element law : the issue of local noncompliance / Paul G. Lewis. p. cm. Includes bibliographical references and index. ISBN: 1-58213-069-8 1. Discrimination in housing—California. 2. Housing policy— California. I. Public Policy Institute of California. II. Title. HD7288.76.U52C25 2003 363.5'09794—dc21 2002154992 Copyright © 2003 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. PPIC does not take or support positions on any ballot measure or state and federal legislation nor does it endorse or support any political parties or candidates for public office. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Foreword It is no secret that affordable housing is scarce in California. Even with record low mortgage rates and plenty of demand, the lack of conveniently located, relatively low-cost housing has frustrated planners and consumers alike. What is less well known is that many local governments have been out of compliance with California’s housing element law, which was designed to help local officials plan for adequate housing in their communities. In his latest look at local governance in California, Paul Lewis addresses the question of why so many cities and counties have been unable or unwilling to meet the state requirements for housing. Lewis analyzes a long list of reasons why a city would be out of compliance with the law, including shortages of available land, explicit antigrowth policies, and an aversion to affordable housing among wealthy communities. His findings do not support the view that only the richest communities were out of compliance; in fact, smaller cities with older housing and those with strict growth controls were more likely to be noncompliant. He then takes a closer look at communities in the metropolitan areas of San Diego, Los Angeles, and the San Francisco Bay Area to identify the relationship between noncompliance and housing production. His results will come as a surprise to some. During the 1990s, noncompliant communities were just as likely to expand their housing stock as communities that complied with the law. Furthermore, when other factors were held constant, noncompliance was not a significant predictor of the rate of multifamily development. After reviewing the policies of others states with comparable approaches, Lewis identifies three basic problems with California’s housing element law. First, it often goes against the grain of local politics by asking cities to plan for the needs of the wider region, not just those of current city residents. Second, it may represent a mismatch of goals and policy tools. Specifically, it attempts to tackle the problems of overall iii housing underproduction with a process-oriented approach developed to prod cities and counties into planning for their share of affordable units. Third, the statute itself is unwieldy, embraces multiple objectives, and is difficult for nonexperts to understand. Lewis concludes that the time is ripe for policymakers and affected interest groups to seek a more workable, transparent, and straightforward approach to housing. These policymakers may need to resolve whether their major goal is a sheer increase in residential construction or an equitable distribution of affordable housing. Lewis warns that using a fair-share planning approach as a tool to encourage overall housing production places an unrealistic burden on a fairly fragile policy. David W. Lyon President and CEO Public Policy Institute of California iv Summary California is generally perceived as producing less housing than would be expected or desired judging by its population and job gains. The statewide plan developed by the Department of Housing and Community Development (HCD) predicts that a continuation of current trends will lead to underproduction of needed housing by approximately 60 percent—likely leading in turn to a further upward spiral of home prices and rents as well as lower homeownership rates. This shortfall particularly hurts low-income families, which have more difficulty in paying the price or rent premium that results from undersupply. Some observers argue that local governments’ lack of enthusiasm for new housing in their communities is a large source of this problem. This report investigates California’s housing element law, the major tool the state government uses to ensure that city and county land-use regulators are planning appropriately for new housing development. Enacted in original form in 1969, the housing element law requires that all cities and counties in California engage in detailed planning for their residential needs by including housing as an element of their comprehensive plans. The housing element process is intended to focus the attention of city policymakers on identifying land sites for housing, and on policy actions that would make it easier or less expensive to provide additional housing units. The motivation for this study is the high degree of local noncompliance with the law. I begin by examining which types of city governments tend to have their housing elements deemed “out of compliance” by HCD staff. A second major goal is to determine whether such noncompliance can be linked statistically to a lower subsequent production of new housing. The report also reviews California’s implementation of housing element law and compares it to the experiences of other states. v How the Housing Element Process Works The housing element is the only part of local general plans that is subject to substantial oversight by the state. The state’s interest in local housing elements has been justified by the fact that housing is enshrined in state law as a matter of “vital statewide importance.” Nevertheless, housing elements in and of themselves rarely cause new housing to be built; in a market economy, private developers (or nonprofit builders of affordable housing) construct nearly all new housing units. The housing element requirement is often called a “fair-share” housing law, with the term generally referring to a regional process by which each local community works to accommodate a fair proportion of the region’s housing need. Regional councils of governments (which are planning councils representing the cities and counties in a given metropolitan area) work from state estimates of regional housing need and assign a housing unit goal, or allocation, to each city and unincorporated county area in their region. Cities and counties are then expected to update their housing elements to plan for quantified objectives for housing units over the next five years. Unlike in some states with fair-share approaches, in California housing production itself has received as much emphasis as the geographic distribution of housing for lower-income families. This state’s broader notion of fair share probably has resulted from its longstanding problem of housing affordability and underproduction, which limits the opportunities of not only the poor but also the middle class. The housing element statute requires that local planners address and reduce governmental constraints on the development of housing for all income levels. Such constraints may include local growth controls, strict building codes, developer fees, and permit procedures. California law requires that local governments revise their housing elements periodically. In the current round of revisions, San Diego (in 1999) was the first region where localities were required to update their plans. In 2000 and 2001, cities and counties in the Southern California region and the San Francisco Bay Area undertook these updates, with the vi requirement extending to the Bakersfield, Fresno, and Sacramento areas in mid-2002 and to the rest of the state’s regions by the end of 2003.1 Once a local update has been drafted, HCD reviews it to gauge whether the plan can enable the targeted number of units—including specific amounts of housing for households of very low, low, moderate, and “above moderate” incomes. If so, HCD certifies the housing element. If not, the jurisdiction may change its plan to incorporate HCD’s suggestions. If the element is adopted without satisfying HCD—or fails to be updated at all—the city or county is regarded as noncompliant. That judgment limits its eligibility for certain state and federal funds for affordable housing and renders it more vulnerable to lawsuits that can halt development in the community. There have been frequent conflicts between state and local policymakers over housing element compliance. Testing Competing Explanations for Local Noncompliance One of the most contentious aspects of the housing element requirement is the fact that nearly four cities in ten and a quarter of counties are out of compliance with the law.2 Large majorities of jurisdictions in the state have been noncompliant at some point. Explanations of this fact differ dramatically. Some communities have argued that state or regional projections fail to reflect powerful local realities, such as a lack of vacant land, which can make it difficult to identify a sufficient physical capacity to accommodate projected housing needs. Other observers argue that local noncompliance more frequently reflects an aversion to new housing—particularly affordable units—on the part of upper-income communities. Overt antigrowth policies or ____________ 1Here, Southern California includes cities in the counties of Imperial, Los Angeles, Orange, Riverside, San Bernardino, and Ventura; the Bay Area includes cities in Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma Counties. 2This tally of noncompliant jurisdictions includes those whose housing element is overdue, as well as those found noncompliant after HCD review. In both cases, the housing element is legally noncompliant. vii regulatory postures on the part of local governments—sometimes imposed on local governments by their voters—are also seen as a culprit. Still another possibility is that many jurisdictions fail to meet the housing element requirement because they lack the requisite planning capacity or experience. State officials periodically seek ways to increase compliance with housing element law by penalizing noncompliant local governments. Housing advocates have long argued that sanctions should be stronger and that housing law needs more emphasis on results and less on process. Senate Bill 910, which passed the California Senate in 2001 but later died in the assembly, would have required that the state controller fine noncompliant cities. By contrast, local officials seek to protect local autonomy over land use—long one of the major pillars of home rule— and complain of distant state authorities who fail to understand local conditions. This study examines the distinctions between cities that have been able to attain housing element compliance from those that have not. Measures of city characteristics are drawn primarily from the U.S. Census, and measures of local land-use policies and growth controls are drawn from a PPIC mail survey of local planners in the late 1990s. A simple comparison of cities in the regions that have recently been reviewed by HCD indicates that noncompliant communities are, on average, smaller and have older housing. In the rest of the state, where cities have had about a decade to reach compliance since their last housing element revisions were due, few cities were noncompliant. The simple comparison indicates that these few laggard communities tend to be wealthier and less ethnically diverse than compliant cities. The results of a more sophisticated statistical analysis reveal the determinants of noncompliance in a more systematic way. This analysis involved cities in the Southern California region, San Diego County, the San Francisco Bay Area, and the 18-county Central Valley. Evidence was strong that cities’ residential growth policies held a particularly important role for the HCD reviews. For example, cities whose planning directors report that the review process for new development proposals has been getting longer experience more trouble in attaining compliance. Each viii restrictive growth policy that has been adopted by the city approximately doubles the odds that it will be found noncompliant. An older housing stock is also associated with a greater likelihood of noncompliance. Cities with older housing may be more settled and have a more established community character; they are also likely to contain less vacant land. Cities with smaller populations are also more likely to be noncompliant, all else equal. Governments of larger cities may have a greater capacity to undertake the broad range of planning efforts needed to reach compliance and may also be more insulated from the political pressure of homeowners. An additional finding is that the length of time that has elapsed since the deadline for submitting an updated housing element is one of the most statistically significant predictors of compliance. Specifically, the results indicate that each month that has passed since the region’s deadline renders a city approximately 5 percent more likely to reach compliance. Even controlling for all these factors, cities in the Bay Area, Central Valley, and Inland Empire were more likely to be noncompliant than other cities. It is also interesting to note that the compliance status of individual communities tends to persist over time. Although many cities were able to attain compliance as the 1990s wore on, it remained the case that noncompliance in 1991, after the past round of revisions in the coastal metropolitan areas, was a fairly good predictor of noncompliance in 2002. Assessing the Relationship Between Compliance and Housing Production Defenders of the housing element requirement tend to argue that local governments that comply with the law, by demonstrating adequate plans, enable more housing to be built. Using data from the 1990s, I examine whether a city’s compliance status in 1991 helps to predict the percentage increase in the city’s housing stock by 2000. The analysis again controls for a variety of other city characteristics that might be expected to influence the level and type of housing growth. Because of ix data limitations, this analysis focuses only on cities in the Southern California, San Francisco Bay Area, and San Diego regions. Using Census data on the number of housing units in each city, I find no detectable relationship between housing element compliance and the percentage increase in housing across these communities during the 1990s. Thus, for all the potential merits and benefits of housing element compliance, one must look to other factors to explain why some cities experience rapid housing development and other cities experience little. The analysis suggests that a city’s demographic characteristics, its position in the urban hierarchy, and its physical capacity to accommodate new buildings are better predictors of housing growth. Although the housing element statute, as a fair-share approach, is especially concerned with increasing the production of affordable housing, we unfortunately lack any comparable information for all cities about the production of affordable units. However, data from the Construction Industry Research Board do allow an analysis of multifamily housing production. Multifamily housing includes apartments and condominiums, the types of housing generally most relevant for those at the lower to middle levels of the income distribution, particularly in the coastal metropolitan areas where housing is expensive. Multifamily units represented only about one-quarter of the new units produced in the 1990s. The results of this analysis once again show that housing element noncompliance as of 1991 is not a significant predictor of the rate of multifamily development when other relevant factors are held constant. Rather, cities that were job centers and that had fewer senior citizens as of 1990 tended to experience faster rates of multifamily housing development. Finally, the report investigates whether housing element compliance affected the mix of housing developed in the 1990s—multifamily versus single-family—despite its lack of effects on the rate of increase. I examine the relationship between compliance and the percentage share of new housing units that were multifamily units. Here there is evidence of a significant association of compliance with the outcome in question. Cities with noncompliant housing elements developed new housing that was weighted more toward single-family units. Holding constant for x other city and county characteristics, a noncompliant housing element was linked to an 8 percentage point lower proportion of multifamily housing among the newly built units. The results imply that cities with compliant housing elements are willing to substitute multifamily (more affordable) units for a share of single-family units. Nevertheless, it is striking that one can detect no measurable relationship between compliance and overall housing production. Rethinking the Housing Element Approach Discussions of problems with the current housing-element law among California policymakers and housing advocates have led to a number of reform proposals in recent years. However, most of these proposals take the housing element approach—a state review of local plans for future housing needs—more or less as a given. A fundamental question may be whether the various goals and values that California policymakers hold dear—increased housing production, an equitable distribution of housing responsibility across communities, special attention to the housing needs of low-income groups, local autonomy and home rule, environmental protection, and more—can all be accommodated within this area of law, or whether there are tensions among them. A secondary question is whether the current approach is the most effective use of resources to further these goals. Enticing communities to accommodate housing would not be such an uphill battle if they perceived that doing so would be in their financial self-interest. Thus, creating a component of the state fiscal system that rewards local governments for the addition of housing units, particularly affordable units, may result in less conflict and more cooperation. The Jobs/Housing Balance Incentive Program, passed in 2000, has elements of such a “rewards for performance” approach. Revenue sources that are distributed to localities on a population basis, such as the Vehicle License Fee subvention, also create indirect incentives for cities and counties to accommodate housing. Other states, including some with equally strong traditions of home rule as California, have also wrestled with issues of inadequate housing production, mandated fair shares for jurisdictions, and state oversight of local planning. In New Jersey, for instance, production of affordable xi housing rests mainly with the profit motives of private developers; they can propose a “builder’s remedy” by including a share of affordable units in otherwise market-rate projects that would normally exceed local zoning limitations. The quasijudicial state agency overseeing local housing plans also allows communities to transfer to other jurisdictions up to half of the affordable units they are expected to produce, in exchange for a payment to the “receiving” municipality. In Massachusetts, developers of affordable projects have access to a Comprehensive Permit Law that enables them to proceed through a streamlined local review process, avoiding many of the intermediate reviews that other proposed developments must go through. Builders whose projects are denied by this local process have the option of appealing to a statewide board, the Housing Appeals Committee, which can overturn the local decision and order the project to be permitted. In Oregon, the state’s Land Conservation and Development Commission imposes minimum zoning densities on residential land for cities and counties in the Portland metropolitan area. The localities in this region are also required to write plans so as to allow at least half of future residential units to be in multifamily projects. As in these other states, California’s housing element requirement has often gone against the grain of local policymaking because it asks cities to plan for the needs of the wider region, not just the needs of current city residents. Indeed, the philosophy behind fair-share housing policy is that the so-called police power of local governments to regulate land use should be directed toward the general welfare of the region, not just the general welfare of the specific locality. Those involved in California’s debate over housing policy often note how long and detailed the housing element statute is. Highly detailed statutes are often evidence of widespread disagreement on a given policy, as waves of “reform” occur in which opposing interests seek to have their specific concerns addressed and preserved in law. In the case of the housing element statute, the result is an unwieldy law that is often difficult for outside observers to comprehend in its entirety or details. In the 33 years since the housing element statute was enacted, the search for an adequate supply of housing in California has only become more elusive. It may be a ripe occasion for policymakers and affected xii interests to seek an approach to housing policy that is more workable, transparent, and straightforward, with measurable barometers of substantive local success or failure. In so doing, policymakers will need to resolve whether the major goal of such a law is a sheer increase in residential construction or an equitable distribution of affordable housing. Using a fair-share planning approach as a tool to encourage overall housing production may place an unrealistic burden on a relatively fragile policy. xiii Contents Foreword..................................... iii Summary..................................... v Figures ...................................... xix Boxes ....................................... xxi Tables ....................................... xxiii Acknowledgments ............................... xxv 1. INTRODUCTION ........................... Controversies over Local Government Noncompliance with the Law................................ Recent Legislative Proposals and Reform Efforts ......... Plan of the Report ............................ 2. WHAT IS THE HOUSING ELEMENT LAW AND HOW HAS IT BEEN IMPLEMENTED? ............ Introduction: A Fair-Share Housing Law ............. Origins: Housing as a New Component of Local Planning .. The 1970s: Elaboration of the Housing Element Requirement ............................ The 1980s and 1990s: The Present System Takes Shape .... Regional Allocation Process ..................... Update Process for Local Housing Elements .......... Growing Contentiousness over Housing Element Requirements and Compliance ................. Concerns over Litigation....................... Are Needs Met? ............................ Concerns over Projections and Regional Allocations ..... The Current Revision Cycle: The Controversy Continues .............................. The Defense of Housing Element Law ............... Senate Bill 910 (2001–02) ....................... Conclusion ................................. 1 2 7 10 11 11 13 14 16 16 18 21 22 24 24 27 30 32 34 xv 3. WHY DO SO MANY COMMUNITIES HAVE NONCOMPLIANT HOUSING ELEMENTS? ........ Why Cities? ................................ Four Sets of Factors Potentially Affecting Compliance ..... Community Social Status and Exclusion ............. Local Land-Use Characteristics and Vacant Land ....... The Resources of the Local Government............. Local Politics and Residential Growth Policies ......... Timetable for Housing Element Updates Influences Compliance Status......................... Comparing Compliant and Noncompliant Cities: A Preliminary Profile ........................ Probing Further: A Statistical Model of Noncompliance .... Key Findings of the Multivariate Analysis ............. Some Caveats ............................... Conclusion ................................. 4. DOES HOUSING ELEMENT COMPLIANCE MEAN THAT MORE HOUSING IS PRODUCED? .......... Housing Element Compliance, Circa 1991 ............ Measuring Housing Growth Rates Using Census Data ..... Controlling for Other Factors ..................... Results of the Analysis of Total Housing Unit Growth ..... Analysis of Multifamily Development Using Construction Industry Data ............................ The Mix of New Housing Development .............. Reconciling These Findings ...................... Conclusion ................................. 5. CAN STATES ENSURE ADEQUATE LOCAL PROVISION FOR HOUSING DEVELOPMENT? ...... Alternative Approaches within the Context of the Current Housing Element Law ...................... More Penalties and Prescriptions for Local Governments .. Self-Certification Based on Performance ............. Subregional Allocations and Joint Housing Elements .... Encouraging Transfers of Housing Allocations among Jurisdictions ............................. Another Approach: Rewards for Performance ......... 35 35 36 37 38 40 41 43 44 48 51 53 55 57 57 59 60 64 65 67 68 68 71 71 71 72 73 74 75 xvi Relevant Experiences from Other States .............. New Jersey ............................... Massachusetts ............................. The Portland Region ......................... Conclusion ................................. 6. CONCLUSION ............................. Goals in Conflict ............................. Equitable Distribution or Sheer Production? .......... A Mandate in a Vacuum? ...................... Devil in the Details............................ 76 76 79 82 84 87 87 88 89 89 Appendix A. Multivariate Analyses of Housing Element Noncompliance .. B. Multivariate Analyses of the Rate and Mix of New Housing Production in the 1990s ........................ 91 95 Bibliography .................................. 101 About the Author ............................... 107 Related PPIC Publications .......................... 109 xvii Figures 1.1. Cities’ Housing Element Compliance Status, as of September 25, 2002 ........................ 1.2. Counties’ Housing Element Compliance Status, as of September 25, 2002 ........................ 4.1. Cities’ Housing Element Compliance Status, as of December 31, 1991 ........................ 4 5 58 xix Boxes 1.1. State Law: Does Housing Production Conflict with Other Goals? ............................. 2.1. A Creative Way to Help Meet Housing Goals ........ 2.2. A Housing Element Challenge in Folsom ........... 2.3. Other Important Provisions of SB 910, As Amended ... 9 30 31 34 xxi Tables 1.1. Schedule for Third Revision of Local Housing Elements ............................... 3.1. Comparison of Compliant and Noncompliant Cities in the San Diego, Southern California, San Francisco Bay Area, Fresno, Kern, and Sacramento Regions ........ 3.2. Comparison of Compliant and Noncompliant Cities in the Remainder of the State .................... 3.3. Variables Used to Predict Housing Element Noncompliance ........................... 3.4. City Characteristics That Are Statistically Associated with Noncompliance ........................ 4.1. Percentage Growth Rates for Housing in Cities in the San Diego, Southern California, and San Francisco Bay Area Regions, 1990–2000..................... 4.2. Comparison of Housing Development in Compliant and Noncompliant Cities in the San Diego, Southern California, and San Francisco Bay Area Regions....... A.1. Logistic Models of Housing Element Noncompliance as of September 2002 ......................... A.2. Summary Statistics for Continuous Variables in the Preceding Models .......................... B.1. Regression Model of Housing Unit Growth in the 1990s, with County Fixed Effects ................ B.2. Regression Model of Multifamily Development Growth in the 1990s, with County Fixed Effects ........... B.3. Regression Model of Share of Housing Development That Is Multifamily, with County Fixed Effects ....... B.4. Summary Statistics for Continuous Variables in the Preceding Models .......................... 6 45 47 50 52 59 60 92 93 96 97 98 99 xxiii Acknowledgments I thank Professors Scott Bollens of UC Irvine and Dowell Myers of USC, as well as Tracy Gordon, Hans Johnson, Fred Silva, and Mark Baldassare of PPIC, for reading a draft of this report and offering helpful suggestions. At the Department of Housing and Community Development, Deputy Director Cathy Creswell and her staff helped me understand the housing element process and obtain information. At an early stage of the project, Alex Amoroso (Association of Bay Area Governments), Mark Stivers (Senate Committee on Housing and Community Development), Peter Detwiler (Senate Local Government Committee), and Dan Flynn (California Assembly) also provided useful perspectives on the issues involved. Some of the data and information necessary for the analyses in this report were collected or prepared by Elisa Barbour, Khanh D. Bui, Rachel Flood, Hugh Louch, and Max Neiman. Peter Richardson and Joyce Peterson improved the prose and presentation. Although all of these people made important contributions, the author is solely responsible for any errors of fact or interpretation. xxv 1. Introduction California is generally perceived as producing less housing than would be expected or desired judging by its population and job gains. The state’s most recent statewide housing plan, for example, finds that developers will need to produce an average of 220,000 housing units per year through 2020 to meet projected demand. Achieving such production levels will be a challenge, the plan notes, since even in recent boom years only 150,000 or so new units received building permits. Since 1970, California has never produced 220,000 housing units for more than two consecutive years (Myers, 2001, p. 388). The state plan, published by the Department of Housing and Community Development (HCD), predicts that a continuation of current trends will lead to underproduction of needed housing by approximately 60 percent—likely leading in turn to increasing home prices and rents as well as lower homeownership rates (HCD, 2000, p. 3; Myers and Park, 2002). This lack of housing supply particularly hurts the poor, who have difficulty in paying the price or rent premium that results from undersupply. Two of three low-income renter households in California pay more than half their income to put a roof over their heads (Little Hoover Commission, 2002, p. i). Some observers argue that local governments’ lack of enthusiasm for new housing in their communities is a large source of this problem, although evidence is far from definitive on this point (see Lewis and Neiman, 2002). This report examines California’s housing element law, the major means by which the state government tries to ensure that local land-use regulators are dealing with unmet housing needs and planning appropriately for new housing development. The state requires that each city and county write (and periodically revise) a general plan to guide its future growth. The housing element is one of the seven required elements of the general plan; its purposes are to identify current and future local housing needs of all income groups and to ensure the 1 preparation of a detailed schedule, consistent with broader community planning goals, for meeting such needs. The housing element is the only part of local general plans that is subject to substantial oversight by the state. The law provides that housing elements are to be revised every five years, although the legislature allowed localities to postpone their updates throughout the 1990s because of budgetary shortfalls. HCD, relying on Finance Department population projections, assigns a target number or goal for additional units to each region of the state. In a process called the Regional Housing Needs Assessment (RHNA), the council of governments (a planning body representing the cities and counties in a given metropolitan area) must allocate this total number of housing units among the cities and unincorporated county areas in its region. HCD then reviews whether city and county housing plans conform with statutory requirements—in other words, whether the local housing elements seem likely to enable each community to meet its goal for new units. This report focuses on several key issues regarding this process of housing element revision and review. Among the topics analyzed are: • California’s implementation of the housing element law and how it compares with other states’ approaches to local housing planning, • Which types of municipal governments fail to comply with the law, and • Whether local compliance appears to make a difference in terms of the rates of new housing development across cities. Controversies over Local Government Noncompliance with the Law The city and county housing element updates have become a battleground for state and local policymakers. Some state officials argue that local governments are not energetic enough in planning for housing and are trying to deflect their fair share onto other jurisdictions. For their part, local officials often claim that the RHNA “quotas” that they have been assigned are poorly justified, unrealistic, and unresponsive to the physical limitations of their communities. “Typically, as soon as 2 these numbers are proposed, they are challenged by local governments as far exceeding local realities,” a representative of the League of California Cities writes (Carrigg, 2002). Local officials also argue that affordable housing production is stymied by both high land costs and the lack of state and federal funds for this purpose. This intergovernmental debate is hardly new. A decade ago, the previous round of housing element updates led to similar contentiousness and calls for reform (Jackson, 1994; Senate Committee on Local Government, 1993; Senate Committee on Housing and Land Use, 1995). Once a local update has been drafted, HCD reviews it to gauge whether the plan can enable the targeted number of units—including specific amounts of low- and moderate-income housing. If so, HCD certifies the housing element. If not, the jurisdiction may change its plan to incorporate HCD’s suggestions. But if the element is adopted without satisfying HCD—or fails to be updated at all—the city or county is regarded as “out of compliance.” Noncompliant communities are ineligible for certain affordable housing programs administered by HCD, such as the federal Home Investment Partnerships Program and portions of the Community Development Block Grant program, and the state Jobs/Housing Balance Improvement Incentive Grant. Noncompliant jurisdictions are also much more vulnerable to lawsuits on development matters. In some cases, judges have ordered noncompliant local governments to refrain from issuing any new building permits until the matter is resolved. Noncompliant housing elements make attractive legal targets for parties who seek to invalidate local land-use or redevelopment decisions—even though some of these parties are pursuing ends that do not always support housing development (Richards, Watson & Gershon, 1998; Senate Committee on Local Government, 1993, pp. 73–74). Thus, although HCD reviews of local housing elements are officially “advisory,” most cities and counties go to some length to achieve the department’s certification, especially since the law instructs judges to presume a local housing element legally valid if it has been certified by HCD (Warner et al., 1997). Compliance with housing element law is certainly not universal. As of September 25, 2002, about one-third of all cities and more than one- 3 fifth of all counties have had their housing elements judged noncompliant by HCD staff, and some have been so for many years. Other localities are overdue in submitting their draft housing elements to HCD, which also renders them noncompliant with the law (see Figures 1.1 and 1.2). At some points in the past, large majorities of jurisdictions in the state have been noncompliant. The disappointing levels of compliance among counties and especially cities have led housing advocates and some state legislators to argue that some localities are sidestepping their responsibility to ameliorate the shelter needs of Californians. Since housing has long been treated as a matter of statewide concern by the state legislature and the courts, the potential for intergovernmental conflict is high. State officials Under review by HCD 10% Self-certified 1% Housing element due 5% Out of compliance 33% In compliance 51% Figure 1.1—Cities’ Housing Element Compliance Status, as of September 25, 2002 4 Housing element due Under review by 5% HCD 7% Out of compliance 22% In compliance 66% Figure 1.2—Counties’ Housing Element Compliance Status, as of September 25, 2002 periodically seek ways to beef up compliance with housing element law or penalize noncompliant local governments, whereas local officials seek to protect local autonomy over land use—long one of the major pillars of home rule—and complain of distant state authorities who fail to understand local realities. In the current period, California communities have been undertaking the third revision of the required local housing elements. To reduce the burdens on HCD in reviewing local housing elements, the law specifies a staggered schedule of housing element updates. In 1999, San Diego was the first region where localities were required to update their plans. In 2000 and 2001, the process took place in the Southern California region and the San Francisco Bay Area. The requirement extended to the Fresno, Bakersfield, and Sacramento regions in mid- 5 2002, and to the rest of the state’s regions by the end of 2003 (see Table 1.1). In discussing compliance, this study uses the September 25, 2002, report on local compliance status from HCD, except where noted. The Table 1.1 Schedule for Third Revision of Local Housing Elements Council of Governments San Diego Association of Governments Counties Affected (Includes all Cities Within Each County) Revision Date San Diego December 31, 1999 Southern California Association of Governments Imperial Los Angeles Orange Riverside San Bernardino Ventura December 31, 2000 Association of Bay Area Governments Alameda Contra Costa Marin Napa San Francisco San Mateo Santa Clara Solano Sonoma December 31, 2001 Council of Fresno County Government Fresno June 30, 2002 Kern County Council of Governments Kern June 30, 2002 Sacramento Area Council of Governments Sacramento Sutter Yolo Yuba (plus cities of Lincoln, Rocklin, and Roseville in Placer County) June 30, 2002 Association of Monterey Bay Area Governments Monterey Santa Cruz December 31, 2002 All other local governments All remaining December 31, 2003 SOURCES: HCD website; 2001 California Statutes, Chapter 85. 6 compliance status of individual jurisdictions can change as new revisions are submitted and HCD reviews them.1 At present, although the state’s major metropolitan areas have already gone through the third round of housing element updates, other parts of the state are at an earlier stage of the process, and their compliance status in many cases reflects the outcomes of the prior round of updates in the early 1990s. These timing issues will become important in the statistical analysis of local compliance status in Chapter 3. Recent Legislative Proposals and Reform Efforts “All groups are dissatisfied with current housing element requirements.” —Staff report for Senate Committee on Local Government (1993) Housing advocates have long argued that sanctions should be stronger and that housing law needs more emphasis on results and less on process. Senate Bill 910, which passed the state senate in 2001 but died in the assembly, would have required that the state controller fine noncompliant cities. In its original, more punitive form, the bill would have widened penalties for noncompliant jurisdictions to include ineligibility for certain state transportation funds and directed state courts to presume that noncompliant housing elements are invalid. As a result, that bill would have opened up such jurisdictions to more litigation from housing rights’ organizations and developers. Discussing the original version of the bill, a nonprofit housing developer argued that transportation funds were used “as a bargaining chip, because, frankly, why should a city get money for transportation projects to make travel easier if it isn’t going to provide housing? It seems to me that the two components should have always worked together” (quoted in Bishop and Materna, 2001). Proponents of the bill also argued that the tougher state requirements could lead to a “good cop, bad cop” relationship that would help local officials argue against ____________ 1Readers can view the most recent compliance status report at http://www.hcd.ca. gov/hpd/hrc/plan/he/status.pdf. 7 neighborhood opposition to affordable and other housing projects. One developer said, “I think the local politicians will be relieved of the controversy that seems to surround residential development, because they will defer to state mandate” (quoted in Bishop and Materna, 2001). Some central city interest groups also supported the bill as a way to prompt suburban jurisdictions to share more of the burden of affordable housing. Opponents—including some legislators—saw the original bill as too punitive in its approach to local governments. The League of California Cities also argued that other state and federal mandates and policies hinder the ability of localities to accommodate additional housing. These mandates and policies include agricultural land preservation efforts, endangered species rules, and open space acquisition purchases (see Box 1.1). Although SB 910 attracted the largest amount of attention, a number of other bills proposing changes in the housing element process were considered in the 2001–02 legislative session. For example, SB 2292 (Dutra), which was recently signed into law, requires that local governments retain the residential zoning densities that they refer to in their approved housing elements. Cities and counties that “downzone” (reduce the allowable density of) residential land must transfer the “lost” density elsewhere in the community to make up for the loss in potential housing units. SB 1634 (Figueroa) would have required that regional planning councils, in making RHNA allocations to local governments, seek to improve the geographic balance of employment and housing in each region. AB 2863 (Longville), supported by the League of California Cities and the California State Association of Counties, would have eased definitions of “substantial compliance” under the housing element law and broadened the definition of what constitutes a “residential unit” that can count toward fulfilling a locality’s housing goals. However, neither SB 1634 nor AB 2863 was ever given a committee hearing, and both bills died. In reaction to the flurry of proposed legislation, a group of lawmakers and stakeholders, calling itself the Housing Element Reform Working Group, began meeting in Sacramento to discuss possible revisions to the law. The working group included interested legislators 8 Box 1.1 State Law: Does Housing Production Conflict with Other Goals? The state housing element law calls on cities and counties to adopt plans and zoning that can accommodate their projected housing needs. However, state law also includes other requirements that may constrain the ability of localities to accommodate housing (Bishop and Materna, 2001; Carrigg, 2002; Senate Committee on Housing and Land Use, 1995). For example: • Habitat conservation plans for threatened and endangered species cover millions of acres in California. • Williamson Act contracts, which grant lower property tax assessments to farm owners who agree to restrict their land to agriculture or related open space activities, cover about one-third of all privately held land in the state. • To receive a subdivision map, builders of housing developments of greater than 500 units must demonstrate that there will be sufficient water supply in the area for the next 20 years. • The California Coastal Commission exercises additional land-use authority in the coastal zone. • Cities and counties are also required to plan for transportation and congestion management, air quality, and seismic safety; each of these goals may conflict with the desire to construct more housing in a given area. • Cities wishing to annex land for additional housing development must have the annexation approved by the Local Agency Formation Commission (LAFCO). By state law, LAFCOs are charged with ensuring efficient service delivery patterns and must work toward restricting discontiguous development. • Significant development projects, and government actions that may encourage development, are subject to lengthy environmental impact reviews and possible litigation under the California Environmental Quality Act (CEQA). At the state level, determination of housing needs and allocations are given a statutory exemption from CEQA. But cities and counties updating their housing element are given no such exemption. and staff along with important stakeholders such as the League of California Cities, regional councils of governments, and the California Building Industry Association. The group attempted to move toward greater consensus on various issues and disputes relating to housing element preparation and review. Many of their ideas became part of SB 910 as it was amended in the assembly (to be discussed in Chapter 2). Legislators involved in the discussions agreed to hold all relevant housing element bills in committee until the reform working group had a chance to move toward agreement on some basic reforms. But consensus proved elusive. 9 Plan of the Report With policy reforms under discussion and concerns about housing production high, the time is ripe for serious study of housing element compliance. To introduce the issues and institutions involved, Chapter 2 traces the history of the housing element law in California, discusses its challenges and controversies, and summarizes the current housing element process. Chapter 3 examines why jurisdictions fail to submit housing elements that are acceptable to HCD. For example, some communities have argued that state or regional projections fail to reflect powerful local realities, such as a lack of vacant land on which to accommodate housing. Other observers argue that local noncompliance more frequently reflects an aversion to new housing—particularly affordable units—on the part of upper-income communities. Antigrowth policies or regulations— sometimes imposed on local governments by their voters—are also seen as a culprit. Still another possibility is that many jurisdictions fail to meet the housing element requirement because they lack the requisite planning capacity or experience. Chapter 4 addresses the relationship of housing element compliance to housing production. Is it true that compliant jurisdictions add housing—particularly multifamily development—at a faster rate, after one takes into account other local characteristics that affect the level of housing production? Although the housing element law involves a requirement to plan rather than to build housing, it presumably was passed under the premise that effective housing planning would lead to more residential production. Chapter 5 discusses reforms that have frequently been suggested for housing element law and compares California’s experiences and rules to those of other states that review local housing planning and have fairshare requirements, such as Massachusetts, New Jersey, and Oregon. A brief consideration of the experiences of these other states in pursuing similar policy goals may be instructive at this important juncture in the legislative history of housing elements in California. Finally, Chapter 6 offers concluding observations and calls for a reconsideration of California’s approach to local housing production. 10 2. What Is the Housing Element Law and How Has It Been Implemented? This chapter presents a policy history of the housing element law in California and discusses how the policy is currently carried out. After offering a brief introduction to the purposes of the housing element law, I focus on its origins and development, the process by which it has been implemented, and ongoing debates regarding its effectiveness and possible reforms. Housing element policy had a somewhat inauspicious start in the state, and its processes and institutional responsibilities were not fully worked out for approximately a decade following the 1969 passage of the law. Once the law began to take its present shape, an extended policy debate ensued, often pitting local governments and advocates of local flexibility against HCD, homebuilders, and affordable housing advocates. This stalemate has continued to the current day, with the state now in the midst of the third wave of mandated updates of local housing elements. Introduction: A Fair-Share Housing Law Enacted in 1969, the original housing element law requires that all cities and counties in California engage in detailed planning to meet their housing needs. Housing elements do not, in and of themselves, cause new housing to be built, as this activity rests with private developers (or nonprofit builders of affordable housing). Rather, the housing element process is intended to focus the attention of city policymakers on policy actions that they might take to make it easier or less expensive for additional housing units to be built. For example, local plans may call for reducing or eliminating fees on affordable housing construction, rewarding developers of certain types of projects by allowing them to 11 build at higher densities than would otherwise be permitted, or requiring that all housing developments above a certain size reserve a portion of units for low- or moderate-income households. The housing element requirement is sometimes referred to as a “fairshare” housing law. A number of other states and metropolitan areas have developed fair-share housing legislation, with the term generally referring to a regional process by which each local community works to accommodate a fair proportion of the region’s housing need. Fair-share plans “determine where housing, especially low- and moderate-income units, should be built within a region according to such criteria as placing housing where it will expand housing opportunity, where it is most needed, and where it is most suitable” (Listokin, 1976, p. 1). Jurisdictional fair-share goals are typically defined using a formula that includes such factors as growth rates in households and jobs, socioeconomic status measures, and often some measure of the local carrying capacity for new housing. From 1977 to 1983, the U.S. Department of Housing and Urban Development (HUD) offered matching funds to regions that engaged in the creation of Areawide Housing Opportunity Plans (Brownlow, 1991, p. 1).1 California was an early innovator in the fair-share housing field. Its law on the topic is more than three decades old, although it did not take its current, relatively strict form until 1980. Unlike in some states with fair-share approaches, in California housing production itself has received as much emphasis as the geographic distribution of housing for lower-income families. This state’s broader notion of fair share probably has resulted from its longstanding problem of housing affordability and underproduction, which limits the opportunities not only of the poor but also of the middle class (Calavita, Grimes, and Mallach, 1997). Thus, the fair-share approach taken by California allocates among localities goals not only for ____________ 1In 1968, Congress required that federal planning subsidies for metropolitan planning be contingent on preparation of a regional housing plan, which led many urban regions to develop early fair-share plans. By the 1970s, many housing activists argued that fair-share plan adoption should be required for cities or regions to qualify for federal housing subsidies. HUD did not require this, but did award subsidy bonuses for metropolitan areas that developed fair-share plans (Listokin, 1976, pp. xvii, 6). 12 subsidized or “affordable” units but also for moderate-income and “above moderate” units—that is, the entire projected housing need. The philosophy behind such allocation strategies is to serve as a warning to localities to readjust their zoning ordinance if it does not allow for their allocated growth share. The units also serve as a growth management technique alerting communities to the level of future growth they can expect and enabling them to plan properly to handle development (Listokin, 1976, p. 169). In addition, because there is a widespread perception among local officials in California that housing development is fiscally burdensome, the state has used housing element law as a way to encourage planning for balanced land-use development—that is, attention to residential as well as commercial development. Some argue that the state’s system for financing local governments encourages the opposite, since cities’ dependence on local sales tax revenues, and lack of control over property tax allocation, may cause local policymakers to favor retail development over housing. Origins: Housing as a New Component of Local Planning Until 1969, land use and circulation (transportation) were the only elements of local general plans required by California law. Housing, along with other topics such as open space, was sometimes dealt with in optional general plan elements by ambitious cities and counties. In 1969, then-Assemblymember Pete Wilson successfully carried a bill to require that each city and county prepare a housing element. This law made local housing a mandatory element of local general planning and required that HCD develop guidelines for local governments to follow in preparing housing elements. In 1971, legislation was added requiring that localities follow HCD guidelines for the preparation of housing elements.2 ____________ 2According to the Select Committee on Housing and Urban Affairs (1974, p. 59), the change to the statute was necessary because the 1969 law “had inadvertently referred to these ‘mandatory guidelines’ as ‘regulations,’ and the attorney general interpreted this as requiring that regulations be adopted” in accordance with the state’s cumbersome rules on administrative procedures. Such rule-making would be interpreted as a state mandate 13 In 1970, additional legislation passed requiring that HCD prepare a statewide housing element; such action was necessary for the state to take advantage of federal funds for comprehensive planning assistance. The law required that HCD develop one- and five-year housing development goals “needed to house all residents of this state” (quoted in Select Committee on Housing and Urban Affairs, 1974, p. 50). Notably, however, this activity was organized separately from the more comprehensive statewide planning efforts of the Office of Planning and Research. The 1970s: Elaboration of the Housing Element Requirement In the early years of the housing element requirement, the state government devoted little in the way of resources or attention to guiding or evaluating local housing plans. Governor Ronald Reagan’s administration, for example, was seen as unenthusiastic about the very existence of the Department of Housing and Community Development, and it tried unsuccessfully to merge it with other state agencies (Select Committee on Housing and Urban Affairs, 1974). Budgetary commitments to housing element activities (and to HCD generally) were low during this period. HCD budgeted only $17,000—the equivalent of one full-time employee—to updating the statewide housing element in 1974–75, for example, and approximately two-thirds of HCD’s small budget was devoted to inspection of manufactured housing. A senate select committee complained that HCD had had little public input into the preparation of the statewide housing element. Moreover, it maintained that HCD had refused to purchase Census data that would allow detailed computation of housing needs for each locality. The committee also pushed the department to separate the issue of production goals for the state from the problem of affordability ________________________________________________________ that would require reimbursement of the local housing element process. Since the legislature had appropriated no money for local housing element preparation, HCD had concluded that it could not adopt “regulations.” 14 shortfalls for low-income residents (Select Committee on Housing and Urban Affairs, 1974, pp. 54–55). Furthermore, HUD complained that California’s statewide housing element performance was “less than exemplary,” as the state had prepared only a draft needs analysis and work program by October 1973 (Select Committee on Housing and Urban Affairs, 1974, p. A12). This document called for the production of 225,000 units per year statewide, as well as rehabilitation and replacement efforts for existing housing. At this time, the League of California Cities was calling for an augmented planning role for HCD, in coordination with housing planning that local governments were undertaking. “We believe that the future of DHCD . . . must include a strengthening, rather than a weakening, of its functions. . . . In the area of housing, its role should embody an institutionalization of a strong state role in helping to provide for the housing needs of all segments of the people of California” (Select Committee on Housing and Urban Affairs, 1974, p. A19). A 1975 law (AB 1X) called upon HCD to adopt housing element guidelines and to review city and county housing elements, resulting in a set of detailed standards in 1977. These guidelines stressed the regional fair-share concept (Warner et al., 1997; Senate Committee on Local Government, 1993, p. 18). During the administration of Governor Jerry Brown (1975–1983), statewide planning, including housing planning, became more active and central to the mission of the administration. Arnold Sternberg, Brown’s director of Housing and Community Development, had represented low-income advocacy groups as an attorney, and he took advantage of the provisions in the housing element law allowing the department to develop guidelines for the implementation and enforcement of the statute. The department also took a more aggressive posture toward the fair-share goals implied in housing element law, by which each local government is expected to help solve regional housing shortfalls, particularly for low- and moderate-income households. 15 The 1980s and 1990s: The Present System Takes Shape A new law, Chapter 1143 of the Statutes of 1980, resolved the issue of HCD oversight. HCD’s guidelines were deemed as advisory, but localities were required to consider HCD comments before adopting the local housing element (Warner et al., 1997; Senate Committee on Local Government, 1993, p. 18). Local housing elements were required to include a needs analysis, a discussion of resources and constraints in meeting those needs, statements of goals and specific policies regarding housing, as well as quantified objectives for construction, conservation, and rehabilitation to move toward the goal. “This mandate aims at ensuring that each community accepts responsibility for the housing needs of not only the resident population but also of those households who might reasonably be expected to live within the jurisdiction were a variety and choice of housing appropriate to their needs available” (HCD, 1988, p. 1). In other words, the concept of fair share was fully enshrined in state law (Senate Committee on Housing and Land Use, 1995, p. 25). This statute also introduced the local duty to review and update local housing elements every five years (Richards, Watson & Gershon, 1998). Regional Allocation Process The regional allocation of housing needs was also developed much more fully during this period. HCD began to assign housing goals to each region, using state Department of Finance projections regarding future household growth. HCD adjusts these goals to account for the amount of vacant housing in the region and the expected need for replacement units. HCD also takes into account advisory comments from the region’s council of governments (COG), an association of city and county officials that is responsible for regional planning. Each COG then prepares a regional housing needs assessment. In rural parts of the state that lack COGs, HCD itself performs the RHNA analysis. The first two mandated cycles of the RHNA were in 1981 and in 1984–1986 (HCD, 1988). The stated reasons for the regional needs 16 assessment are to allow consideration of regional issues, distribute responsibilities equitably among jurisdictions, and relieve local governments of some of their data-gathering burdens (HCD, 1988, p. 4). The RHNA process is specifically exempted from the California Environmental Quality Act. The regional housing need is pegged to at least the level of projected household growth, although “major economic events” may justify a revision to a lower number (HCD, 1988, p. 11). Under the statute, several factors must be considered in the COG’s allocation of housing growth goals to specific localities: • Market demand for housing (vacancy rates, housing prices, household structure, construction, absorption, etc.), • Employment opportunities (current and projected), • Availability of suitable sites (including residentially zoned as well as nonresidentially zoned land that could be used, and the possibility of redevelopment for housing or for increased densities of housing), • Availability of services (including current and future capacity, transportation, medical and recreational facilities, etc.), • Commuting patterns (time, length, transit availability), • Type and tenure of housing need (including a consideration of special populations, such as large households, the elderly, students, and the military), and • Farmworker housing needs. Planners are also directed to avoid aggravating existing concentrations of low-income households in certain communities. The reasoning HCD cites for this requirement presents an interesting perspective on housing development: The basic philosophy behind housing element law is that citizens of all economic levels should have the opportunity to live where they choose in decent, safe, and sanitary housing. . . . Some COGs have adopted [allocations] which plan for the same percentages of households in each income group in each locality. This approach is based on the position that all areas are equally suitable for each income group. The approach also reflects the view that the 17 current income differences between localities reflect the effects of past governmental actions, such as zoning and capital improvements planning, and that housing opportunities for underrepresented income groups should be promoted in governmental planning activities (HCD, 1988, pp. 23–24). In other words, the regional needs assessment is to be used not only to help enable the production of needed amounts of housing in a region but also as a device to redistribute the burdens of lower-income households more equitably across geographic areas. It is likely that many of the controversies over housing element development are rooted in the desire of some localities to preserve their existing character, while their COG and HCD attempt to move them toward a mix of residents more representative of the entire region. Still, HCD guidelines do point out that in larger and more complex regions, there may be good reasons to deviate from the general goal of having each jurisdiction approximate the diversity of its region. For example, HCD (1988, p. 24) notes that the age and differentiation of the larger regions often lead to geographic clusterings of socioeconomic groups, such as retirement communities and college-student neighborhoods. It is important to note, however, that the actual weighting scheme in evaluating these various factors is to be designed by the COG itself (HCD, 1988, pp. 20–21; Senate Committee on Housing and Land Use, 1995, p. 5). In short, there has been no uniform, required method for assigning housing goals to local governments. Update Process for Local Housing Elements A locality may revise its need from the number given to it by the COG, but the city or county is required to cite findings, backed by data, to justify its revised goal. “A local revision may be used in a locality’s housing element even if the COG has disapproved it. In such a case, however, DHCD may be less likely to consider the element to be in compliance with the housing element law” (HCD, 1988, p. 14). A city or county housing element must provide analysis of local housing needs, resources, and constraints relevant to the production of housing, and a five-year policy program detailing actions that the local government plans to take to address its housing needs. Although there is a great deal of local discretion in preparing housing elements, there are 18 also “elaborate statutory provisions” to comply with, which is not the case for other portions of the general plan (Curtin, 2000, p. 10). The state’s interest in local housing elements has been justified by the fact that housing is enshrined in state law as a matter of vital statewide importance. In determining housing needs, local planners must consider data regarding the existing number of households and housing units, the number of households overpaying for housing, and overcrowding, rehabilitation, and replacement needs. The local government must provide an estimate of various households considered to have special needs, such as the disabled, senior citizens, large households, farmworkers, female-headed families, and homeless persons. The city or county is also required to note any existing assisted-housing projects that may be at risk of losing their subsidized status during the time period covered by the housing element update. For many communities, identifying a sufficient physical capacity to accommodate projected housing needs is a major challenge. Localities are required to provide a detailed land inventory, summarizing the number of acres zoned for various types of residential development (single-family, multifamily, mobile homes, emergency shelter, farmworker housing, and mixed use) and for nonresidential uses. Local planners must specify the density range and dwelling unit capacity for these various zones and include information on the availability of services and infrastructure (HCD, n.d.). The local government’s five-year action plan must identify a sufficient set of potential sites for future housing development that could accommodate the community’s need “by right”—that is, without requiring a conditional-use permit or imposing vague conditions on potential homebuilders.3 HCD advises local officials that although a shortage of buildable land may make accommodating new housing units more complicated, it does not relieve a locality of its responsibility to provide adequate sites for new residential development. “If a locality does not have sufficient ____________ 3A conditional-use permit allows specified land uses (e.g., a mobile home park) in areas zoned for other types of uses (e.g., single-family homes), subject to additional local government discretion. 19 existing suitable sites to meet these needs, programs such as changes in land use, annexation, upzoning, or a second unit ordinance would be appropriate” (HCD, 1988, p. 53). Redevelopment policy and infrastructure improvements are also to be addressed as possible ways increase development potential (Rawson, 2000). From the standpoint of local officials, however, other policy considerations may conflict with such intensification of land uses, such as a lack of water or sewer capacity (often not under the control of the city or county), traffic congestion, conflicts with existing industrial activities, or seismic hazards. Each city and unincorporated county area, then, faces a housing element update in which it is expected to plan for construction needs and produce quantified objectives for housing units over the next five years. These objectives are further categorized by household income: • Very low (0 to 50 percent of regional median income), • Other lower (50 to 80 percent), • Moderate (80 to 120 percent), and • Above moderate (120 percent or higher). The number of needed units so identified, however, “are simply goals, not mandated acts” (Curtin, 2000, p. 11). In other words, a local housing element is hardly a self-enforcing policy and does not, in itself, create or mandate new housing. Nevertheless, the housing element statute does require that local planners address governmental constraints on the development of housing for all income levels. For example, zoning and land-use controls, building codes, developer fees, and permit procedures must be discussed, as well as local government plans to overcome or reduce public-sector barriers to housing production (Rawson, 2000). Local planners must also assess “nongovernmental constraints” to housing production, such as land and construction costs and the availability of financing. Finally, a set of miscellaneous topics noted in the statute must be addressed, including efforts to involve the public in the housing element update, the potential for energy conservation in new housing development, and consistency with other elements of the general plan. Going beyond a listing of objectives, the housing element must detail what programs and policies the local government foresees undertaking to 20 implement the plan, including efforts to provide sufficient sites for all types of housing (which may involve changes in zoning or subdivision requirements), participation in state and federal subsidy programs for affordable housing, incentives and regulatory concessions for residential developments, efforts to conserve the existing affordable housing stock and preserve units that are at risk of losing their subsidies, programs to promote equal housing opportunities, and efforts to address and remove governmental constraints on housing construction (HCD, n.d.). The city or county must send a draft of its housing element to HCD for review and consider the department’s findings on the draft. The department “considers the element to be in compliance only if every one of the statutory requirements is met” (HCD, 1991, p. 1). After the city council or county board of supervisors officially adopts the housing element, HCD again reviews it for its compliance with the law.4 Because of the detailed requirements of the statute, housing elements are subject to close scrutiny by the courts (Curtin, 2000, p. 27). Growing Contentiousness over Housing Element Requirements and Compliance During the late 1980s and early 1990s, when local governments were completing housing element updates, grassroots opposition to growth and development burgeoned in many parts of California, particularly in coastal regions. In this atmosphere, writing plans that focused on how to accommodate more housing became an increasingly contentious process. Compliance with the housing element law, as measured by HCD’s certification of city and county revisions, has been spotty at best. In 1991, only 19 percent of localities were certified as being in compliance by the department, although this proportion grew to 37 percent by 1993 ____________ 4Even if the department gives the opinion that the draft element complies with the law, the city or county must officially adopt the element before the jurisdiction is considered to be in compliance. Although most local governments in such a situation quickly adopt the approved draft element, some communities delay in their adoption or decide to submit a new draft to reflect new local concerns or political changes. 21 and to 52 percent by 1995 (Senate Committee on Local Government, 1993; Senate Committee on Housing and Land Use, 1995). The greatly increased rates of compliance were traceable in large part to an active program of technical assistance, combined with legal pressure, by HCD. HCD worked with the attorney general’s office to inform 47 delinquent localities that their housing element revisions were overdue and that they faced possible legal action. According to Timothy Coyle, then-director of HCD, “Many communities around the state . . . had failed to submit any evidence of a local housing element for, in some cases, as many as 15 years” (Senate Committee on Local Government, 1993, p. 32). Still, as Coyle and others affirmed, some of those noncompliant jurisdictions had actually experienced rapid rates of housing development, including affordable housing. Thus, the procedural burden of preparing a housing element may have been the major obstacle for many of these communities in their lack of compliance. Coyle told a Senate committee, “You can draw no correlation between housing element compliance and housing production” (Senate Committee on Local Government, 1993, p. 37). Nevertheless, two years later Coyle testified that the jurisdictions that were in compliance accounted for a disproportionate percentage of the building permits issued in the state—particularly multifamily permits (Senate Committee on Housing and Land Use, 1995, p. 31; see also Little Hoover Commission, 2002, p. 18). Concerns over Litigation According to some observers, a fear of litigation is a major motivator for local governments in expending time and resources on housing elements. Not only housing advocates but also growth opponents, and sometimes school districts and other overlapping jurisdictions, have used noncompliant housing elements as a litigation strategy in seeking to overturn local land-use decisions. The planning director of Long Beach testified in 1993 that the housing element “is the only element that is prepared defensively, rather than as a guide to local policy and decisionmaking. In most cities, the housing element is prepared as a joint effort by the city attorney and the planning department to make sure that the 22 document is defensible in court” (Senate Committee on Local Government, 1993, p. 62). The penalties for localities can be serious.5 Nevertheless, it is important to point out that Section 65589.3 of the Housing Element statute is an important line of defense for local governments that have received HCD certification of their housing elements. Added in 1990, this section establishes the “rebuttable presumption” that HCD’s finding of compliance means that the locality has drawn up a legally valid plan for its housing. In a legal case, the housing element law requires only “substantial compliance,” rather than perfection, which according to attorney Michael Colantuono “does a good job of protecting localities from frivolous challenges to their housing elements” (Senate Committee on Local Government, 1993, p. 37). Because courts have sometimes found localities to be in substantial compliance even when HCD has failed to certify their housing elements, some observers have suggested that HCD’s review is more stringent than necessary under the law. After the second round of housing element revisions, and the suspension of housing element funding in the 1992–93 state budget, a variety of interest groups began an effort to reform the housing element statute. Two major legislative hearings were held, with many voices criticizing the existing policy regime (Senate Committee on Local Government, 1993; Senate Committee on Housing and Land Use, 1995). Reform discussions were geared at streamlining the planning process for housing elements and making the RHNA process more attuned to local government concerns and capacity for growth. These discussions also emphasized performance in accommodating housing rather than the process of housing element planning (Warner et al., 1997). Nevertheless, aside from some relatively minor new legislation, ____________ 5“A prudent locality can manage this risk only by keeping its housing element up to date and building a voluminous record to ensure the element is legally defensible. Those communities which have failed to do so, often for budgetary reasons, have run significant risks and some have paid a significant price. . . . If a city loses such a case. . . , its land-use authority will be suspended. . . , it will have just 120 days to prepare a new element, often necessitating the retention of consultants, and it will likely pay not only its own legal fees, but those of its opponents. . . . During the preparation of a new element, nothing gets built, not even housing, without a court order . . . .” (Michael Colantuono, quoted in Senate Committee on Local Government, 1993, pp. 75–76, legal citations omitted.) 23 “six years of reform discussions led nowhere,” and no major reforms cleared the legislature (Richards, Watson & Gershon, 1998). Are Needs Met? Local governments tend to miss their housing unit goals, often by a wide margin. In particular, the low- and moderate-income segments of housing needs are typically unmet—not surprisingly given the frequent need for deep subsidies to fund such projects. In 1988, for example, the Bay Area Council, a business group, found that only one of the nine counties in the San Francisco Bay Area was on pace to attain even half its goals for production of low- and moderate-income units (even using a generous definition of affordable units). Napa County experienced a production of only 52 affordable units between 1980 and 1988, compared to a specified need of 3,906 units (Bay Area Council, 1988b). A more recent study of 40 fast-growing Bay Area jurisdictions found that only 34 percent of affordable housing goals were met in those communities with certified housing elements; only 9 percent of affordable housing goals were met in noncompliant jurisdictions (Dodge, 2002). Thus, the preparation of housing elements, whether compliant or not, is no guarantee that needed housing units will be built. Many fewer units may be approved, or even proposed, whether due to shifting demand, homebuilder preferences to build fewer and more expensive units, land prices in excess of what builders wish to pay, or government actions that reduce project size for reasons not contemplated in the housing element. Concerns over Projections and Regional Allocations The regional allocation of housing needs has also become the subject of increasing debate and conflict in recent decades. Some of the concerns relate to the projections used as the basis for regional and local housing goals. A 1995 analysis by senate committee staff echoed the concerns of many COG personnel and local officials in observing that the Department of Finance’s projections “are strictly mechanical and do not consider local planning factors such as local growth policies, habitat 24 preservation, clean air, and traffic congestion” (Senate Committee on Housing and Land Use, 1995, p. 19). In the current round of revisions, some commentators further argue that the Department of Finance’s projections of population increase and household formation—and, thus, projected housing needs—may be too high because of the tendency of new immigrants to have larger households.6 Another potential problem with the projections is the likelihood of decreasing birthrates among women in Latino and Asian groups in post-immigrant generations (Hill and Johnson, 2002). Over time and across generations, Latino families more closely approximate the trends among white and African American families toward fewer children, fewer persons per household, and more homeownership. Overall, the complex changes in California’s demography are consistent with any number of very different arguments about future housing needs (Myers, 2001). Some local officials have argued that population projections in regional transportation plans (mostly drawn up by COGs) should be the basis of housing need calculations, but existing state law requires the use of Department of Finance projections in preparing local housing elements. Some celebrate the flexibility of COGs in weighting factors differently in the RHNA process; others see this as an example of inconsistency that merits further state guidelines (Senate Committee on Housing and Land Use, 1995, p. 23). One criticism leveled at COGs involves their projections of future increases in jobs—upon which housing needs are, in part, based. Some of this criticism has come from ____________ 6Some observers claim that large household sizes are due to “doubling up” in otherwise unaffordable housing, but even when controlling for income, foreign-born Latinos live in significantly larger households (i.e., more persons in the housing unit) than the native-born (Myers, 2001, p. 389). Moreover, household crowding is explained almost entirely by demographic factors (such as nativity and poverty), whereas local housing market conditions have little relationship to crowding (Moller, Johnson, and Dardia, 2002). In brief, immigrant households, particularly among Latinos, are much more likely to be crowded, measured as more than one person per room. This is a major issue for housing needs assessment, since local governments must examine data on local housing crowding in assessing their current need for additional units. As Myers (2001, p. 392) notes, “Crowded housing highlights the dilemma for planners. If the situation is not viewed as a problem by the affected cultural group, should planners still treat it as a problem?” 25 advocates of increased housing production. In the 1980s, for example, the Association of Bay Area Governments (ABAG) projected that the nine-county Bay Area would add about 473,000 new jobs between 1988 and 1995. On that basis, the Bay Area Council argued, ABAG’s regional housing goal of 288,000 over that period was about 50,000 units too low, assuming a standard of 1.4 workers per household (Bay Area Council, 1988a). However, from the standpoint of local governments, job growth projections are sometimes seen as unrealistically high. Some local officials have called for a mediation process as part of the RHNA procedure, possibly with some third party evaluating disputes between local governments and COGs regarding the allocation of housing goals (Senate Committee on Local Government, 1993, p. 2). Some objections to RHNA outcomes are the result of disagreement with the decisionmaking of COGs, which ultimately reflects a political process. One critic argued that COG allocation processes allow objections by local governments after the initial allocation, but the communities that can do the staff work to mount a serious objection tend to be large cities with large staff resources. In this case, the objection continues, the unwanted units are likely to be reallocated to smaller and politically weaker communities, which are often older, poorer, or built-out inner suburbs (Senate Committee on Housing and Land Use, 1995, pp. 56–57). The zero-sum quality of RHNA allocations puts COGs, which are constituted as cooperative organizations of local governments, in a difficult position. They are left to carry out a state mandate while still feeling pressure from their local government members who feel that their housing allocations are excessive. HCD officials have said that the COG is the level at which intraregional disputes should be worked out, but given their weak political basis, most COGs have not been able to take a strong, autonomous approach. “Without independent authority, regional agencies have little wherewithal to overcome fundamental conflicts between local and statewide interests” (Barbour, 2002, p. 65). Or, as a senior planner for the Association of Bay Area Governments put it, “From ABAG’s perspective, the process has been a serious political liability for a voluntary membership organization” (Senate Committee on Local Government, 1993, p. 48). 26 Assembly Bill 438 (Torlakson), signed into law in 1998, does allow the RHNA process to be further decentralized within a region. If a county and all the cities within it (or a joint powers authority or subregional agency established by the COG) request to undertake the allocation of housing goals within a subregion, then the COG is obliged to turn that responsibility over to this smaller body, after assigning the subregion its proportionate share of the region’s housing goal. This option has not yet been widely utilized, perhaps because of time constraints during the current cycle of housing element revisions (see Richards, Watson & Gershon, 1999). The Current Revision Cycle: The Controversy Continues Concerns about the housing element law were never fully resolved after the previous round of housing element updates ended in the early 1990s. But the topic largely fell from the radar screen in the middle of the decade, because of the state’s difficulty in emerging from the recession and its associated budget problems. With state funding of housing element activity postponed during the period of budget exigencies, required housing element update activity was suspended for six fiscal years, from 1992–93 until being resumed in 1998–99. Stopgap legislation was passed during this period to extend the deadlines for regions and localities to undertake their updates. When the process started up again, however, the old disagreements quickly resurfaced, even as housing problems in much of the state had become more acute. According to planning expert William Fulton (1999, p. 114), “Most people involved in the housing element process agree that the law is ineffective: it is overly bureaucratic and exacting, has too many loopholes, and even a good housing element is no guarantee that affordable housing will actually be built in any given community. But attempts to reform the law have failed repeatedly in recent years.” One of the political tensions involved in housing element policy, Fulton argues, is standoff between housing interests (both forprofit developers and antipoverty nonprofits), who are prominent and influential at the state level, and local governments, who often see the 27 housing element law as unwarranted and clumsy meddling by state government in home rule, since land-use policy has long been the almost exclusive purview of local governments. Many local officials did not look forward to another cycle of revisions of local housing elements. In fact, both the League of California Cities and the California State Association of Counties opposed funding the RHNA mandate at all unless the statute was reworked; in turn, the governing board of the state’s largest COG, the Southern California Association of Governments (SCAG), voted to oppose state funding for the RHNA unless the local jurisdictions of the region no longer opposed it (Richards, Watson & Gershon, 1998). RHNA was funded despite these protests. After SCAG completed its allocation process in 2000, local governments in Riverside and San Bernardino Counties sued HCD as well as SCAG, arguing that the housing unit goals assigned to the Inland Empire were unrealistically and inequitably large. Although inland communities hold most of the vacant land in the metropolitan region, they argued that an unwillingness to accept housing among the coastal counties of Orange, Los Angeles, and Ventura led to excessive units being assigned to Riverside and San Bernardino. This case, which went before a Riverside County Superior Court judge in January 2002, represents perhaps the largest breakdown of intergovernmental relationships during the current round of housing element updates and has drawn particular scrutiny to the RHNA process. In short, the argument is that the regional allocation can become highly politicized and works to the detriment of less-powerful jurisdictions (Little Hoover Commission, 2002, p. 19). Elsewhere, officials in the San Luis Obispo region indicated publicly that they consider the RHNA goals assigned to that area completely unattainable (Lyons, 2002a, 2002b). More generally, some county government officials argue that it does not make planning sense for counties to be “treated like cities” in the housing element process, given the greater capacity of incorporated cities to accommodate new growth. Counties are limited in their ability to provide city-like services, and often seek to protect resource and agricultural lands from intrusion by urban uses (Senate Committee on Local Government, 1993, p. 89). 28 In the Bay Area, for example, Napa, Sonoma, and Solano Counties all attempt to direct growth away from farmland and open space and into cities. A Napa County supervisor notes, “Consider the fact that we’re an agricultural county, and the fact that the state always laments the loss of agricultural land, and the fact that HCD equates agricultural land and open space as land available for housing. There is a huge disconnect there” (quoted in Shigley, 2002, p. 1). In Napa County, the unincorporated area was assigned 28 percent of the housing unit goals for the entire county; unincorporated Sonoma received 30 percent of that county’s allocation. HCD officials, in turn, point out that it is up to the regional council of governments to allocate housing goals among cities and counties in the region, and the state would prefer that housing be developed where infrastructure is available and sprawl is minimized. COG planners note that housing goal allocations were based in large part on projected job growth, and some counties seem uninterested in restricting commercial development even as they bemoan a lack of capacity to take housing (Shigley, 2002, pp. 14–15).7 Another major debate has concerned the extent to which localities can count rehabilitation or subsidization of existing housing units for low-income households toward their housing element goals. Affordable housing advocates and developer groups have argued that new construction is the primary need, and the alteration of existing privatemarket units to subsidized or affordable units does little to ameliorate the overall supply problem in California. Local governments make the case that rehabilitation is overtly encouraged by the statute and is often the most efficient way to make affordable units available, particularly in fully developed communities. AB 438 of 1998 does allow a local government to count certain types of rehabilitation of substandard units toward its housing element goal, although its provisions are quite restrictive.8 ____________ 7In Napa, city and county officials have formed a working group to see whether some of the county’s housing unit goals can be shifted to cities, perhaps in exchange for allowing cities to annex some industrially zoned land near the county airport. Transfers of housing allocations between jurisdictions are permitted under changes to the law in the 1990s, but the requirements on such “trades” are relatively strict (Curtin, 2000, p. 11). 8For example, the housing units in question must be unfit for human habitation before the renovation, the units must be made affordable to low- and very-low-income 29 Local officials have also argued that they should be able to count the creation of shared living facilities, such as congregate care units for seniors, toward their housing goals. Generally, HCD has counted new dwellings as housing units only if they each contain their own kitchen. Some communities have converted older motels into affordable apartments, installing kitchenettes in the process to meet housing-unit requirements (see Box 2.1). Box 2.1 A Creative Way to Help Meet Housing Goals Several communities in California have acquired (or helped fund the acquisition) of old motels for conversion into affordable residential units. This may present a seemingly simpler method to generate new housing that counts toward housing element goals, without perceptibly increasing the density of the existing community. Potentially, the renovation of the motel may actually improve the neighborhood. Nevertheless, motel conversion too has its complications. Kitchens must be installed in each unit for it to qualify as a housing unit, given the existing interpretation of the housing element statute. But major modifications to an older building may necessitate costly construction changes to make the building more accessible, because of the requirements of the Americans with Disabilities Act. Planners may find that the converted motel violates local parking or fire code requirements for residential dwellings unless further major renovations are made. School fees may need to be paid, and seismic safety upgrades made. If the motel conversion displaces existing low-income tenants using the motel as temporary housing, the city or developer will likely have to pay for their relocation. Finally, there is the ever-present issue of local growth politics; neighbors—including nearby businesses worried about a loss of motel customers—may object to the conversion (discussion draws from “Motel Conversion,” 2002). The Defense of Housing Element Law Housing advocates and public-interest law firms often defend the necessity of the housing element law while still recommending reforms. They point out that the housing element law and other state requirements for local comprehensive planning emerged in large part ________________________________________________________ occupants, and former occupants must be given priority to move into the rehabilitated dwellings. Cities and counties are also now permitted to “buy down” market rate apartments (of at least 16 units) to make them affordable to low-income occupants for at least 30 years. Such local government efforts toward affordability covenants, rehabilitation, and preservation of at-risk units can be credited for up to 25 percent of the community’s housing goals. There are also strict time deadlines to implementing these programs (Richards, Watson & Gershon, 1999). 30 from frustrations with local exclusion of affordable housing or multifamily units (see Box 2.2 for an example). A representative of Catholic Charities testified that the housing element is often the driving force for localities to create policies to increase affordable housing opportunities (Senate Committee on Local Government, 1993, p. 14). An attorney for Legal Services of Northern California argued that local officials in at least four Sacramento area communities were able to use the state’s requirements to convince “not in my back yard” neighborhood groups that sites had to be made available for homeless shelters and transitional housing (Senate Committee on Local Government, 1993, pp. 118–119). And a senior attorney for the Legal Aid Society of Alameda County pointed to specific instances where the housing element requirement generated local policy changes: As a result of the process, I’ve seen the City of Alameda lift its prohibition on multifamily affordable housing development and commit funds to affordable Box 2.2 A Housing Element Challenge in Folsom Folsom is a rapidly developing suburban city in Sacramento County, growing from 11,003 residents in 1980 to 51,884 in 2000. The city had a net increase of 8,550 housing units during the 1990s, according to U.S. Census data, but failed to add affordable housing; home prices were the highest in the county. Housing activists argued that the city was ignoring its responsibilities to the state and the region. Sued by Legal Services of Northern California in 2001 for its noncompliant housing element and lack of progress toward affordable housing goals, the city entered a settlement in 2002. Under terms of the agreement, Folsom is to rezone 128 acres of land to create a “land bank” for the possible construction of up to 2,900 affordable units. The city also agreed to add an affordable housing set-aside requirement for new developments and create an impact fee for nonresidential growth to generate money for an affordable housing trust fund. At a May 2002 community meeting to receive input on the city’s new development goals, 400 residents attended an emotional meeting, “with some audience members clutching signs demanding a ‘total recall’ of the City Council” (Hecht, 2002). Although local officials assured the residents that no decisions had been made on where to locate the rezoned land, and stressed that the affordable units would house groups such as teachers, firefighters, and retail clerks, many attendees were not reassured. One homeowner argued, “This is America. These are our homes. These are our investments. I don’t need crime. Put it in your neighborhood” (Hecht, 2002; also draws upon Padmanabhan, 2002, and other Sacramento Bee articles). 31 housing development; I’ve seen the City of Healdsburg commit to expand its sewer capacity and increase densities; I’ve seen the City of East Palo Alto abandon demolition of affordable units until it identifies the resources to replace them; and I’ve seen the County of Madera commit to forming a joint housing authority with the City of Madera to facilitate affordable housing development (Senate Committee on Local Government, 1993, p. 108). Likewise, in a 2002 report, the Little Hoover Commission (p. 18) called for an “assertive stance” by the state on the housing element law. The commission took the position that “communities have more opportunities than they recognize or acknowledge” to enable housing development, pointing to such policy options as streamlined permitting procedures, public-private partnerships, and subsidized water and sewer fees for affordable housing projects (p. ii). The commission argued for stiffening penalties against noncompliant jurisdictions, “a longer wait” in the competition for transportation and park bond funds, and a limitation on use of redevelopment agency funds (p. iii). Senate Bill 910 (2001–02) As noted in Chapter 1, the most recent debate over housing element law in California centered over Senator Joseph Dunn’s proposal to strengthen the state’s hand in dealing with local governments with noncompliant housing elements. After passing the senate by a vote of 22 to 12, SB 910 was held in the assembly for much of 2001 and 2002 as a Housing Element Reform Working Group met to attempt to decide on a consensus approach. As the deadline for taking action on the bill approached, it was amended in a number of ways. The groups representing cities and counties still balked at certain provisions in the resulting bill, however. Ultimately, the bill died in the assembly, as that chamber refused to suspend the deadline for the bill to emerge from the policy and fiscal committees. Although it did not pass, the bill galvanized debate on the issue and may be resuscitated in new form in a future legislative session. SB 910 sought to alter the existing approach to noncompliance under the housing element statute in two major ways: 1. SB 910 required that the state controller fine cities and counties that either (a) fail to submit an adopted housing element within 32 six months of their deadline, (b) fail to revise their housing elements, if found noncompliant by HCD or a court, or (c) have their compliance certification rescinded by HCD because they fail to take the actions called for in the housing element. The fine was set at $5,000 per month, or 25 cents per month per resident of the jurisdiction, whichever is greater. Fines would accrue to a Housing Supply Account, to be appropriated by the legislature to assist multifamily developments. State courts could reduce fines in mitigating circumstances, such as if noncompliance were due to procedural rather than substantive problems. 2. If a court finds a housing element noncompliant, SB 910 would have directed it to (in addition to any remedy it orders) levy the fine specified above and also award attorney’s fees to the plaintiff, if the plaintiff is a public-interest organization. Courts trying housing element cases would be instructed to give HCD’s findings of noncompliance “great weight.” These provisions distressed representatives of local government, who considered the bill too punitive. They were countered by arguments that the primary shortcoming of the existing housing element statute is its lack of a “stick” to encourage or compel local compliance.9 Beyond the penalties for noncompliance, however, numerous elements in the bill were provisions that had been sought by local and regional officials to smooth the process of regional housing needs distribution. Some of the more important elements are summarized in Box 2.3. These provisions ____________ 9Indeed, some of the most exclusive communities in California can apparently thumb their noses at the issue of housing element compliance. Policy practitioners contacted by the author generally agreed that in largely built-out cities with little development activity occurring, no sizeable low-income population, and no active redevelopment effort, the local government appears largely immune from litigation under the housing element law. With no open land zoned for multifamily development or available at reasonable land prices, no nonprofit developers could propose a project that could become the basis of a court challenge. The exclusive communities also may not mind that their noncompliant housing element disqualifies them from certain housingassistance programs, given that they probably do not wish to have subsidized housing in the first place. 33 Box 2.3 Other Important Provisions of SB 910, As Amended • In allocating housing needs to the various regions of the state, HCD would be required to consider a region’s population projections under its regional transportation plan, along with projections by the state Department of Finance. A process would be set in place for resolving disagreements in situations where the Department of Finance projections and the regional projections differ by a substantial amount. To jibe with the three-year cycle of regional transportation planning, housing element updates would be required every six years, rather than the current five. • COG formulas for allocating housing needs to cities and counties would be required to consider new factors, such as localities’ previous performance in meeting needs, issues of jobs/housing balance, opportunities for infill development, lands protected under open space or farmland programs, federal and state laws restricting development, intracounty agreements to steer growth into cities, and opportunities to maximize use of public transportation. • Localities would be allowed to petition for a revised share of regional housing need. If the request is accepted, and the adjustments amount to less than 7 percent of regional needs, the units in question would be redistributed proportionately to all other local governments. • Public involvement requirements would be enhanced throughout the housing element planning process. However, in reviewing local housing elements, HCD would not be allowed to consider objections to local plans by groups or individuals if the point in question were not raised during the public comment period (unless it is based on new information or an amendment to the element). seemed likely to make what is already a detailed and complex housing element statute even more so. Conclusion The inability to come to consensus over SB 910, despite attempts to include all relevant stakeholders in discussions over the bill, illustrates the long-simmering standoff between state and local officials. Housing element policy and enforcement has become a prominent source of intergovernmental antagonism. Given the different perspectives regarding the reasons for widespread local noncompliance with housing element law, it is important to investigate which community characteristics might contribute to noncompliance. The next chapter addresses this question. 34 3. Why Do So Many Communities Have Noncompliant Housing Elements? One of the most contentious aspects of the housing element requirement is the fact that a large percentage of communities are not compliant with the law—nearly four cities in ten and more than a quarter of counties.1 As discussed in previous chapters, the perspectives as to why noncompliance is widespread differ dramatically between housing advocates and state officials, on the one hand, and defenders of local governments on the other. This chapter examines compliance status among the state’s cities to provide a more informed basis for evaluating why local governments are noncompliant. I use measures of cities’ social and demographic characteristics, land-use patterns, and local policy measures to identify factors that distinguish compliant and noncompliant communities. Why Cities? Although counties are also critical in meeting housing needs, the analysis concentrates on cities for a set of practical reasons. First, most population and housing growth occurs in cities. During the 1990s, 82 percent of the state’s population increase occurred in cities that existed as of 1990; population within city boundaries increased by an additional 12 percent as a result of 18 new municipalities that incorporated between the 1990 and 2000 Census. ____________ 1This tally of noncompliant jurisdictions includes those whose housing element is overdue as well as those found noncompliant after HCD review. In both cases, the housing element is legally noncompliant. 35 Second, statistical considerations make rigorous analysis of counties problematic. There are only 58 counties in the state (one of which, San Francisco, is also a city), which reduces the leverage of statistical analysis. By comparison, there are nearly 480 cities in California. More important, Census demographic data on counties pertain to the county as a whole, whereas county housing elements pertain only to unincorporated areas. One would thus expect that the characteristics of the unincorporated areas have a very major influence on the county’s housing policies and, ultimately, compliance. Some counties have numerous unincorporated areas in far-flung locations, each of which may be quite distinct from the others. In short, there is a “units of analysis” problem in studying county compliance with the housing element law. Finally, studying cities allows for use of PPIC’s extensive existing database on city characteristics. Compiled over several years for other projects, this dataset includes not only Census variables but some measures of city budgetary revenues and political and policy attributes. A 1998–99 mail survey of local planning officials on urban development topics also permits an analysis of local government policies and perspectives on growth.2 I have combined these earlier databases with HCD’s list of cities’ compliance status to examine relationships between community conditions and housing element compliance. Four Sets of Factors Potentially Affecting Compliance At least four different reasons are given for cities failing to comply with housing element law to HCD’s satisfaction. Note that these four reasons are not mutually exclusive; all may apply to some degree. ____________ 2The survey of planning directors was jointly undertaken by the author and Max Neiman of the University of California, Riverside. It specifically addressed local residential policies and growth management and achieved a response rate of 76 percent in the major regions of the state in which it was undertaken (the Central Valley, Southern California—here including San Diego and Santa Barbara as well as the Los Angeles area—and the San Francisco Bay Area). The survey was initially undertaken for purposes not involving housing elements, and the term “housing element” was not mentioned in the questionnaire. This means that respondents probably had little reason to let that controversial topic influence their responses. For more details on the survey and its results, see Lewis and Neiman (2000). 36 Community Social Status and Exclusion “Although they won’t say so publicly, some of these cities don’t care what their assigned [housing goal] numbers are. They just won’t do it because they don’t want low-income housing in their jurisdictions.” —Senator Joseph Dunn, quoted in Wisckol (2002). Senator Dunn authored SB 910 in an attempt to hold local governments more responsible for noncompliance. In the newspaper interview quoted above, he articulated forcefully what some critics of local housing policies have long implied: that cities are reluctant to make the necessary plans to accommodate sufficient housing—particularly multifamily or affordable developments—because of a desire for social exclusion. Such exclusion, highlighted in some national studies, is often thought to rest on the attempts of upper-income, homeowner-dominated communities to insulate their cities from lower-income populations or rental housing more generally. Whether rooted in a desire to cluster with neighbors of similar characteristics, or in racial, ethnic, or socioeconomic snobbery, these efforts are associated with what is often called a NIMBY (“not in my backyard”) attitude. It is also possible for low-income or diverse communities to pursue NIMBY policies, however. The exclusionary approach to land-use policy is often associated with homeowners’ fears of reduced property values if rental or affordable housing is permitted in the neighborhood. It is also often linked with the aspirations of suburbanites to associate with people of similar socioeconomic status and to separate themselves from the perceived problems of large central cities. For example, the mayor of Menlo Park, a Silicon Valley community, has been quoted as saying that his city’s residents “desire to have neighbors who live in homes similar to their own” (Nguyen, 2001). To understand the connection between community status and housing element compliance, one can examine such characteristics as the city’s median household income, the homeownership rate, or the 37 percentage of the population that is white and non-Hispanic. Homeownership rates are of particular interest, because some authors have attributed local antigrowth policies to cartel-like behavior on the part of homeowners, who wish to boost their housing values by restricting the supply of new housing (Brueckner, 1995). Fischel (2001) argues that homeowners’ intense desire to protect the value of their main asset—their home—underlies most local land-use policies. Local Land-Use Characteristics and Vacant Land “Some cities are land-locked with growing populations and little available land; other cities are facing booming job markets with high-paid workers where the competition for housing drives the price of the smallest cottage over half a million dollars.” —Letter from the League of California Cities and California State Association of Counties to Senator Joseph Dunn, opposing SB 910. Many city governments argue that they lack sufficient land resources to accommodate the number of new housing units required under their allocations from the regional COG. This argument can be quite powerful. Many older communities, often inner-ring suburbs, are essentially “built out,” meaning that they lack any substantial vacant land on which to construct new buildings. In such circumstances, cities with significant housing unit goals are left with the more difficult, conflictridden, and potentially expensive options of rezoning existing neighborhoods for higher-density housing or using redevelopment powers to change current land uses so as to include more housing units. In a meeting at HCD’s Division of Housing Policy Development, department staff indicated to the author that the requirement to provide a sufficient inventory of appropriately zoned land sites perhaps does the most to “trip up” local governments in their quest for compliance; many localities prove unable or unwilling to produce such an inventory. The staff indicated, however, that local governments often have an inflexible 38 notion of vacant residential land, as many cities have vacant sites that are zoned for industrial or commercial use and could be rezoned to accommodate housing. Highly developed communities may also be dominated by single-family residential areas that could accommodate additional units. For example, such cities could adopt a more liberal policy on secondary (“in-law”) units3 or could “upzone” single-family areas, where appropriate, to allow more apartments and condominiums. Nevertheless, it is also true that many of such older, inner communities in California have already felt the painful externalities of high density, as their populations have grown so rapidly as to create traffic and parking problems and strain public services. Such communities are often at the receiving end of immigration and other demographic changes, and communities like Santa Ana have expressed concerns about fire hazards and sanitation problems resulting from a profusion of apartments—often overcrowded—and legal or illegal secondary units in an already dense built environment (Perkes, 2002). Several variables can shed light on the perspective that local land-use characteristics affect housing element compliance. Such measures could include the population density of the city and the age of its housing stock (likely to be indicative of a lack of “fresh” land sites in many communities). Although there is no database of vacant land in communities statewide, the survey of local planning officials provided a helpful surrogate measure. Respondents were asked to rate the importance of land supply as a constraint on housing development in their city, on a 1-to-5 scale.4 ____________ 3It has long been state policy to ease local restrictions on secondary units. Most recently, AB 1866 (Wright), signed into law in 2002, requires that local governments provide for ministerial approval of applications (meaning that no discretionary reviews are allowed) for secondary units that meet the local ordinance. 4Although the measure is somewhat subjective, it is reassuring to note that it is highly associated with another measure of local developable land. In a 1998 survey of California city managers, respondents were asked to indicate whether their city was built out (meaning it had “little or no vacant land available”). This (dichotomous) measure from the city manager survey is correlated with the five-point measure from the planner survey of the constraint posed by low land supply at r = .45 (prob. < 0.000). 39 The Resources of the Local Government “My clients are typically smaller communities without the staff expertise or other resources of the state’s largest cities and counties, and complying with the Housing Element statute can be a Herculean task for these communities.” —Michael Colantuono (a city attorney), quoted in Senate Committee on Local Government (1993, p. 67). From almost any perspective, engaging in a full-scale revision of a local housing element is a long and complex process, calling for considerable staff time and capabilities. Many communities spend considerable sums to engage outside consultants to prepare housing elements. For small cities, those without many professional staff, or those that are fiscally strained, resource limitations may significantly hinder the preparation of an adequate and timely housing element. Furthermore, virtually all observers of the housing element law agree that “performance” in producing housing for low- and moderate-income households depends on the availability of subsidized housing funds. In particular, units for very-low-income families are almost impossible to build without heavy subsidies. Both the federal and state governments cut back housing assistance after a peak in the 1970s, and competition for such funds has since been very keen. Few communities can support housing subsidies from their own general-fund budgets, although some larger or wealthier jurisdictions do engage in some local funding of affordable housing. Thus, a lack of local government resources is seen as hindering genuine local efforts at compliance. HCD staff counter that the department has put considerable time and resources into technical assistance, with a particular effort at helping small cities and counties attain compliance. In some cases, HCD staff analysts have been dispatched to communities and taken the lead in helping the locality to prepare its initial draft. They indicated that small, rural communities often have an easier time reaching compliance than larger cities, because the former are often more enthusiastic about future growth. A number of city characteristics may be relevant in measuring local government resources and planning capacity. These include the city’s 40 population size, planning staff resources, and revenue base. Staff resources are implied by local planning officials’ responses to a question regarding the degree to which “lack of personnel to review projects” plays an important role in “constraining or slowing residential development in your city.” Fiscal resources are measured as the city’s own-source, general revenues per capita.5 As a secondary fiscal measure, I estimate the percentage share of locally generated property taxes devoted to the municipal government (as opposed to school districts, the county, or other local governments). It is often thought that cities with a larger share of local property tax revenues will be more accommodating to residential development, because such development will come closer to paying its way. Local Politics and Residential Growth Policies “I want housing in the community, but I want it to reflect the amount that can be served. . . . I’d have to throw out all our growth-management policies [to meet the city’s assigned housing goal]. . . . I think the voters would throw out the City Council.” —Allen Settle, Mayor of San Luis Obispo, quoted in Lyons (2002b). A city’s social status, land-use characteristics, and government resources may well affect its chances of achieving housing element compliance. But local growth policies and development decisions are ultimately political choices made in the context of a city’s political traditions. One aspect that HCD staff members review in examining local housing elements is whether the community has taken full account of the “governmental constraints” on housing development and whether the city has a plan of action to reduce or overcome those constraints. ____________ 5This is essentially a measure of total revenues per capita, from which are subtracted intergovernmental funds and “functional” or “enterprise” revenues, such as current service charges. 41 Evidence from a previous PPIC study indicates that antigrowth politics and efforts to manage or shape growth have emerged in strong form in a fairly small number of California communities, in a weaker or more sporadic form in many others, and not at all in some cities (Lewis and Neiman, 2002). Residential growth-management policies and citizen initiatives that seek to slow growth, along with lengthy city permitting or approval processes, may give HCD reviewers cause for concern. More indirectly, specific controversial development projects, or a history of disputes between city officials and citizen groups on growth issues, may contribute to community rancor that makes officials wary of advancing a plan for additional housing development. Where growth is a hot-button issue, production and affordability goals may take a back seat to political exigencies, which in turn could lead to a thumbs-down from HCD housing element reviewers. The survey of local planners provides several measures of local government policies and orientations toward growth and housing; specific items will be explained as they are introduced in the analysis. One of the most important measures is a count of the number of overtly restrictive growth-management policies pursued by the city. Such policies include: • A substantial recent reduction of residentially zoned land, • Annual limits on building permits, on housing units, or on multifamily dwellings, • A formula for allowable numbers of new housing units, • An official “population ceiling” for the community, or • A moratorium on building permits or water or sewer connections.6 One might assume that an antigrowth approach simply reflects community status, one of the other potential causes of housing element noncompliance. But the earlier PPIC study indicates that local ____________ 6The survey also measured whether the city had any of nine other forms of residential growth management. However, these other policies are less overtly restrictive of residential growth and showed no evidence of effects on housing element noncompliance in regression analyses. For a discussion of the various measures, see Lewis and Neiman (2002), Chapter 3. 42 socioeconomic status was only a weak predictor of whether a community engages in restrictive residential policies (Lewis and Neiman, 2002). Timetable for Housing Element Updates Influences Compliance Status An initial step in analyzing which characteristics are related to noncompliance involves comparing compliant and noncompliant cities across several dimensions. Before undertaking this comparison, however, it is important to point out a potentially severe complication for this exercise: Not all local governments are on the same schedule for updating their housing elements and having them reviewed by HCD. The Fresno, Bakersfield, and Sacramento regions, at this writing, were the most recent metropolitan areas to be scheduled for updates and reviews (see Table 1.1). Thus, localities in these areas, where housing element revisions were due by June 30, 2002, have had only a few months to receive a favorable opinion from HCD, and some jurisdictions have yet to submit their revision. Cities in the San Francisco Bay Area have had about nine months and those in the SCAG region nearly two years to attain compliance, compared to nearly three years in the San Diego region. By contrast, localities elsewhere in the state had not faced a housing element revision deadline for over a decade, as of HCD’s September 25, 2002, report on compliance status, on which this analysis is based. Thus, such cities have had a great deal of time to work with the department and bring their housing elements into compliance. Comparing all cities in the state that are compliant to those that are noncompliant, then, might be akin to comparing apples with oranges. In the regression analysis below, I will make special efforts to control statistically for such scheduling differences. Statewide, fewer than half of the jurisdictions that have submitted draft elements for review since 1999 are in compliance, whereas the vast majority of jurisdictions whose status reflects pre-1999 housing elements are compliant. 43 Comparing Compliant and Noncompliant Cities: A Preliminary Profile With this caveat in mind, this section compares compliant and noncompliant localities in the regions that have already been required to undertake housing element revisions in this round (San Diego, SCAG, San Francisco Bay Area, Fresno, Kern (Bakersfield), and Sacramento regions). A second comparison is then undertaken for cities in the other portions of the state. Included in the noncompliant category are those cities that have received an opinion of noncompliance from HCD as well as cities that are overdue in submitting their draft housing elements to the department. For many years, the department’s compliance status report referred to such localities as having an “obsolete” housing element, although they are now referred to more gently as being “due.” Legally, they are not in compliance. Table 3.1 presents the profile of compliant cities and noncompliant ones for the regions that have already been involved in housing element updates. The three measures of community status all indicate that noncompliant cities tend to be of slightly higher status. These comparisons are mildly supportive of the “social exclusion” perspective on housing element noncompliance. Nevertheless, the differences are not great, and statistical testing indicates that none of the socioeconomic characteristics are statistically significantly higher in noncompliant than in compliant communities. Regarding local land-use characteristics, the housing stock of noncompliant cities tends to be older than that in compliant cities. This difference is statistically significant. Otherwise, there is no support in this comparison for the argument that a lack of vacant land underlies noncompliance. Noncompliant cities in these regions actually have lower population densities, on average, than compliant cities, and their planners are no more likely to say that land supply has constrained residential growth. In this simple comparison, the “lack of resources” reason for noncompliance receives little support. Although noncompliant communities are significantly smaller in average population size, they 44 Table 3.1 Comparison of Compliant and Noncompliant Cities in the San Diego, Southern California, San Francisco Bay Area, Fresno, Kern, and Sacramento Regions Community Status % homeownership % white and non-Hispanic Median household income, $ Mean Value Compliant Noncompliant 60.2 48.0 54,780 62.7 51.5 59,913 Land-Use Characteristics Population per sq. mi.* Supply of land constrains development, per planning director (five-point scale) Median year housing built (as of 1990)*** 5,111 3.8 1969 4,322 3.6 1965 Resources and Capacity Population* No. of planning staff per 10,000 population City own-source general revenues per capita, 1993, $ City’s % share of property tax revenues* 64,160 1.9 507 11.5 46,788 2.1 593 13.1 Local Policies/Politics No. of restrictive growth-management policies (out of seven possible) City requires that residential projects include affordable component, % Citizen initiatives have slowed growth, according to planning director, % Level of citizen opposition to growth, according to planning director (five-point scale) 0.52 32.3 18.3 2.7 0.72 34.3 18.2 2.8 SOURCES: Author calculations based on data from the 2000 Census, the 1990 Census, a 1998–99 mail survey of city planning directors, and the HCD compliance list as of September 25, 2002. NOTES: *p < 0.1; ***p < 0.01. Asterisks indicate statistically significant differences in means between compliant and noncompliant cities, using a two-tailed ttest. Demographic measures are for 2000, except for median year of housing. For population comparison, cities of over one million residents are omitted. For city ownsource general revenue comparisons, cities of less than 1,000 residents are omitted to reduce the influence of extreme outliers in the comparisons. Geographic makeup of regions is listed in Table 1.1. There are 299 cities with usable housing element status data in these regions, although sample sizes are smaller for survey-based items because of survey nonresponse. 45 have roughly the same number of planners per 10,000 residents as compliant cities, and they also receive more own-source revenues than compliant cities. The percentage share of local property tax revenues is also greater in noncompliant cities, despite frequently expressed arguments that a greater share of the property tax base would make cities more accommodating of housing. Finally, the comparison finds muted differences between compliant and noncompliant cities regarding local growth policies and politics. None of the four measures from the planners’ survey shows significant differences between the two groups of municipalities, although the average noncompliant city has about 0.2 additional restrictive growthmanagement policies. Table 3.2 repeats this comparison, but this time for cities in parts of the state that had not yet been required to submit housing element updates. Such regions constitute only a small share of the state’s population and include the Central Coast area, portions of the Central Valley, and the rural north and mountainous counties of the state. In these regions, only a small share of cities were noncompliant (16 percent), which is perhaps not surprising since most had had about a decade to attain compliance. In these parts of the state, differences among the compliant and the (relatively few) noncompliant communities tend in many cases to be starker. For example, all three measures of community status are higher, at a statistically significant level, in the noncompliant communities. Likewise, most measures of local growth policy and politics show substantial differences between the two sets of communities, although this comparison can be only suggestive, as most cities in these regions were not sent the planner survey.7 The average noncompliant city has substantially more growth-management policies and a much higher level of citizen opposition to growth and use of citizen antigrowth initiatives than the average compliant city. The comparison between compliant and noncompliant cities is again murky when it comes to local land-use ____________ 7In the regions discussed in Table 3.2, only cities in the Central Valley and Santa Barbara County would have received the survey. Of the 125 cities compared in the table, 48 responded to the survey. 46 Table 3.2 Comparison of Compliant and Noncompliant Cities in the Remainder of the State Community Status % homeownership** % white and non-Hispanic** Median household income, $** Mean Value Compliant Noncompliant 57.2 60.9 35,341 62.1 73.7 44,441 Land Use Characteristics Population per sq. mi. Supply of land constrains development, per planning director (five-point scale)a Median year housing built (as of 1990) 2,686 2.5 1966 2,648 3.1 1967 Resources and Capacity Population No. of planning staff per 10,000 populationa City own-source general revenues per capita, 1993, $ City’s % share of property tax revenues 20,307 1.7 466 14.8 24,261 2.4 478 15.4 Local Policies/Politics No. of restrictive growth-management policies (out of seven possible)a*** City requires that residential projects include affordable component, %a Citizen initiatives have slowed growth, according to planning director, %a*** Level of citizen opposition to growth, according to planning director (five-point scale)a*** 0.43 19.5 2.5 2.3 2.57 14.3 42.9 3.9 SOURCES: Author calculations based on data from the 2000 Census, the 1990 Census, a 1998–99 mail survey of city planning directors, and the HCD compliance list as of September 25, 2002. NOTES: **p < 0.05; ***p < 0.01. Asterisks indicate statistically significant differences in means between compliant and noncompliant cities, using a two-tailed ttest. Demographic measures are for 2000, except for median year of housing. For city own-source general revenue comparisons, cities of less than 1,000 residents are omitted to reduce the influence of extreme outliers in the comparisons. This table relates to cities in all parts of the state not covered by Table 3.1. There are 125 cities with usable housing element status data in these regions, although sample sizes are considerably smaller for survey-based items because the survey was not directed at cities in some parts of the state and planners from some cities did not respond. aLimited number of planner survey responses in these regions make comparisons on these items less reliable. 47 characteristics and local government resources and capacity. Here, none of the differences are statistically significant. A decade after these cities were initially asked to submit their last updates to HCD, only a relative handful of laggard communities were unable to reach compliance. The simple comparison indicates that such cities tended to be relatively well-off and, so far as we can tell, experienced substantial local turmoil over residential growth. Probing Further: A Statistical Model of Noncompliance The comparison above is suggestive, but provides only a simple sketch of noncompliant communities. To examine more systematically the factors that are associated with noncompliance, I undertook a multivariate analysis. Tables presenting the detailed results of the final estimations are in Appendix A. This section describes the analytic approach and the next section summarizes the most relevant results. The logistic regression method used is able to identify the relationship of various city characteristics to noncompliance, while holding constant each of the other characteristics represented by variables in the model. As suggested above, the four types of city characteristics of interest include measures of community social status, local land-use and development characteristics, resources of the city government, and measures of local growth policy and politics. Cities whose housing elements are overdue for HCD review are again treated as noncompliant.8 I analyzed compliance as of September 25, 2002. Cities under review by HCD as of that date are dropped from the analysis. A set of preliminary analyses helped identify which specific variables measuring these characteristics would be most likely to contribute to an ____________ 8An additional (ordered logit) model was estimated to examine housing element compliance on a three-point scale: in compliance, overdue, and out of compliance. The results of this model were very similar to the results discussed for the dichotomous measure (compliant/noncompliant). 48 explanation of noncompliance. The variables ultimately used in the model are listed and described in Table 3.3.9 I used two techniques to control for differences that relate to the characteristics of the region rather than of the city itself. First, I account for the “timing issue” by including a measure of the number of months that elapsed since the deadline for sending housing element updates from the city’s region to HCD. This variable will indicate whether the simple passage of time makes it more likely that cities will be able to reach compliance. Second, indicator variables are used to denote cities in particular areas. I control for location in the San Francisco Bay Area, the 18-county Central Valley region, and the Inland Empire (Riverside and San Bernardino Counties). The Bay Area had gone through the review process quite recently. The Inland Empire was engaged in a legal dispute with SCAG and HCD over its housing allocation numbers. Both regions have low compliance rates. Cities in the Central Valley face challenges of rapid growth, relative poverty, and, often, fiscal constraint. The measures of local policy—such as the number of strict growth controls and degree of planning staff shortages—are taken from the PPIC mail survey of planning directors in the Bay Area, Central Valley, and metropolitan Southern California (including Santa Barbara County as well as the SCAG and San Diego regions). For this reason, the analysis must exclude cities outside those regions. (The excluded regions account for a small part of the state’s population.) Likewise, the 24 percent of cities in those regions whose planning directors did not respond to our survey cannot be included. ____________ 9Measuring socioeconomic status at the city level creates some difficulties in such a model because of the high collinearity among such measures as income, education, race, and homeownership. After some exploration, I decided to include two relevant variables: the homeownership rate of the city and the percentage of the population composed of non-Hispanic whites. These variables capture two important measures of status and diversity, respectively, that might affect local housing policies, but the variables are not too highly correlated to disentangle (r = .44). Where median income was substituted for homeownership, results did not differ greatly. There was also no evidence that the proportion of the population composed of Hispanics, African Americans, or Asians was related to noncompliance. 49 Table 3.3 Variables Used to Predict Housing Element Noncompliance Variable Definition/Notes % owner-occupied % of housing units owner-occupied % white % of city residents who are white and non-Hispanic % recreational housing % of housing units that are for recreational or seasonal use Population density Natural logarithm of the city’s population per sq. mi. Age of housing Age of median housing unit, in years, as of 1990 Low supply of land Planning director’s response, on a five-point scale, to statement that supply of land for residential growth is an important factor in constraining growth in the city Population Natural log of city population Own-source revenues Natural log of own-source revenues per capita (defined as total revenues minus intergovernmental revenues and minus current service charges), 1993 % of property tax % of locally generated property tax revenues that flow to the municipal government Lack of planning staff Planning director’s response, on a five-point scale, to statement that a lack of personnel to review project proposals is an important factor in constraining growth in the city Longer review process Planning director’s response regarding whether the time required to complete the review of residential projects in the city has been shortened, stayed the same, become somewhat longer, or become much longer (four-point scale) Overt growth restrictions Number of the following policies adopted in the city: annual limits on total building permits, annual limits on residential units authorized, annual limits on multifamily dwelling units built, recent substantial reduction in land zoned residential, policy linking local residential growth rate to a formula or external growth rate, formal population ceiling, moratorium on development or on water/sewer hookups Affordability set-aside = 1 if planning director indicates that city has a policy to require residential developments to include affordable housing Initiatives/neighborhoods Planning director’s response, on a four-point scale, regarding whether important city policies affecting residential growth have mostly been enacted by the council without neighborhood pressure, enacted by the council as a result of neighborhood pressure, enacted both by the council and the voter initiative process, or enacted pretty much exclusively as a result of initiatives Inland Empire location = 1 if city is in Riverside or San Bernardino Counties Bay Area location = 1 if city is in the jurisdiction of the Association of Bay Area Governments 50 Table 3.3 (continued) Variable Central Valley location Months since deadline Definition/Notes = 1 if city is in the counties of Butte, Colusa, Fresno, Glenn, Kern, Kings, Madera, Merced, Placer, Sacramento, San Joaquin, Shasta, Stanislaus, Sutter, Tehama, Tulare,Yolo, or Yuba Number of months elapsed between the date when the city’s housing element update was due and September 2002 SOURCES: Information on low supply of land, lack of planning staff, review of residential proposals, number of overt restrictions, affordability set-asides, and initiatives/ neighborhoods measures is derived from a 1998–99 PPIC survey of city planning directors. Own-source revenues and percentage of property tax are calculated from California State Controller data. Months since deadline was calculated by the author. Age of housing is from the 1990 Census. All other variables are derived from the 2000 Census. Key Findings of the Multivariate Analysis Table 3.4 summarizes the variables that showed a statistically significant relationship with city noncompliance. (For more detailed results, please consult Appendix A.) The results indicate that many city characteristics that one might expect to come into play in regard to housing element preparation are not, in fact, associated with whether cities comply with the housing element law, once one controls for other relevant features of the city. However, a number of patterns do stand out in this analysis. First, the measures of local housing policy seem of key importance. Cities whose planning directors report that the review process for new development proposals has been getting longer, and cities with more overt growth controls, experience more trouble in attaining compliance. For example, each restrictive growth policy that has been adopted approximately doubles the odds that the city will be found noncompliant. This relationship is perhaps not surprising, because HCD staff pay close attention to local government constraints on housing development. Communities that see themselves as a city of single-family homes may have difficulty in showing how they plan to accommodate the number and variety of units necessary to meet their housing goal. 51 Table 3.4 City Characteristics That Are Statistically Associated with Noncompliance Variable Age of housing City population Longer review process Overt growth restrictions City requires affordability set-aside Months since housing element due Inland Empire location San Francisco Bay Area location Central Valley location As Characteristic Increases/Occurs, Probability of Noncompliance . . . Increases Decreases Increases Increases Decreasesa Decreases Increases Increases Increases NOTES: The variables are defined in Table 3.1. Variables listed in Table 3.1 but not in this table were not related to noncompliance at a statistically significant level, defined as p < .1. a“City requires affordability set-aside” is statistically significant only in the model that controls for “slow growth pressure from initiatives/ neighborhoods.” Whether cities require new housing projects to include affordable housing is a significant predictor of compliance.10 HCD reviewers may take such “inclusionary” housing policies into account as a way to overcome barriers to the production of new units for low- or moderateincome households. All else equal, cities with larger populations are more likely than others to reach compliance. This relationship may indicate that large local governments have the resources necessary to complete a successful housing plan. On the other hand, there is no evidence that the fiscal strength of the city or the size of its planning staff influences the probability of compliance. Larger population size may also indicate that ____________ 10This result is significant only when controlling for whether the planning director feels that voter initiatives and neighborhood pressure have been a major source of slowgrowth policies. I do not include this latter variable in the main model because 14 percent of survey respondents answered “don’t know,” which means that such cities must be dropped from the analysis. 52 the city government is more insulated from the pressure of homeowners and more subject to the influence of progrowth business elites, who are likely to be in favor of additional housing. As Fischel (2001) shows, small local governments are more likely to be attuned to the interests of homeowners, who are risk-averse regarding new development because so much of their wealth is typically tied up in their home value. An older housing stock is also associated with a greater likelihood of noncompliance. Cities with older housing may be more “settled” and have a more established community character; they are also likely to contain less vacant land. Similarly, older housing indicates that there has been less new housing construction in recent decades, and thus the community may have come to expect a slow pace of change.11 Finally, the regional context within which each city plans for housing matters a great deal for compliance rates. As discussed, the timing of housing element revision deadlines differs among the state’s regions. The measure of the length of time that has elapsed since the deadline for submitting an updated housing element is one of the most statistically significant predictors of noncompliance. Specifically, the results indicate that each month that has elapsed since the region’s deadline renders a city approximately 5 percent more likely to reach compliance. Location in the Bay Area, the Inland Empire, or the Central Valley is also associated with noncompliance, even after controlling for the various city characteristics. Municipalities in each of these areas of the state show more difficulty in attaining compliance.12 Some Caveats The above analysis presents a useful clarification of the community characteristics that underlie noncompliance. Nevertheless, the analysis ____________ 11I use the age of the city’s housing stock as of 1990, rather than as of 2000, to avoid the statistical problem of endogeneity. That is, a noncompliant housing element may lead to less new construction; thus, one should not, in turn, use “old housing” to predict noncompliance. By using a measure of housing age from a decade ago to represent past community characteristics, one is on firmer footing in predicting current compliance status. 12The implicit comparison is to cities in other regions included in the analysis. Specifically, these include cities in San Diego, Los Angeles, Orange, Imperial, Ventura, and Santa Barbara Counties. 53 has its limits. First, the dependent variable—housing element compliance—is a dichotomous, or “yes/no,” measure that may capture local housing policy in only a rough way. Simple dichotomies often make for more uncertain statistical estimations than more “continuous” variables, such as the number of housing units produced or the number of policies passed. I did attempt to analyze noncompliance using a threepoint scale (in compliance, overdue, out of compliance) that provides an additional category (overdue) that might be viewed as an intermediate case between compliance and noncompliance. However, results did not differ much from those just discussed.13 In defense of the dichotomous measure, housing element compliance or noncompliance is worthy of attention in and of itself, because of its importance to cities for the legal defensibility of their general plans and their ability to compete for affordable housing funds. Still, those interested in a more encompassing explanation of local residential policies should supplement this report with other sources.14 Second, this analysis is of compliance status at a particular point in time—a “snapshot.” Although the analysis controls for region and for the amount of time that the city has had to reach compliance, it cannot take into account all factors that might complicate compliance. For example, Marin County and its cities collectively decided to prepare a countywide housing analysis and housing element updates. All the jurisdictions in the county were subsequently overdue, and by September 2002 all had submitted draft housing elements judged noncompliant, except for two localities whose housing plans were under review by HCD. Finally, although it is tempting to assert that factors such as small populations or a large number of growth controls “cause” noncompliance, I have not put forward any behavioral model of either ____________ 13In the ordered logit model of these three compliance statuses, the same variables were significant (and with the same signs), except for city population size, which was insignificant, and Central Valley location, also insignificant. The overall fit of the ordered logit model was less satisfactory than the dichotomous model discussed above. 14For example, Lewis and Neiman (2002, Chapter 5) present the results of other statistical analyses of local residential growth-management policies and local antigrowth politics. HCD (2001) presents an analysis of local development fees and charges that affect the cost of new housing. 54 local government activity or HCD reviews that pins down what leads to noncompliance. Moreover, over a long time horizon, the direction of causality may be muddled. Cities that have a noncompliant housing element may attract less rental housing or fail to qualify for affordable housing grants from the federal or state governments. These factors, in turn, may make it even more difficult for the community to attain compliance in the next revision period. Conclusion Among the reasons that observers have suggested for widespread noncompliance with the housing element law are the exclusionary sentiments of upper-status communities, the lack of vacant land available for housing in many cities, the lack of resources of the local government to address the issue, and the antigrowth politics or antihousing policies of some towns. In the regions that have been subject to HCD reviews for updated housing elements over the past three years, a simple comparison of compliant and noncompliant cities indicates that few of these arguments receive clear-cut support. Noncompliant cities have older housing and smaller populations but are actually lower in population density than compliant cities. Most other community characteristics do not clearly distinguish the two sets of cities. In the remainder of the state, where cities have had about a decade to reach compliance since their last housing element revisions were due, few cities are noncompliant. The simple comparison indicates that the few laggard communities do tend to be wealthier and less ethnically diverse than compliant cities. A more sophisticated analysis can isolate which local characteristics are most important for noncompliance. Such a model reveals that among the factors that best predicted noncompliance were an older housing stock, more restrictive growth policies and review processes, a smaller population, and a Bay Area, Central Valley, or Inland Empire location. The most important predictor of all, however, was simply the time elapsed since the most recent housing element revision was due. Time may not cure all ills, but it does allow many jurisdictions additional opportunities to prepare a housing element that is acceptable to HCD reviewers. 55 Thus, there is at least some evidence that several factors that have been posited to play a role in noncompliance—local land-use and building characteristics, the size of the city, and municipal growth politics—do have some influence. In the current round of updates, evidence was strong that cities’ residential growth controls and review processes held a particularly important role for the HCD reviews. By contrast, local characteristics that some have suspected play a major role—such as local homeownership rates, revenue bases, or population densities—do not show a clear link to noncompliance. In short, it is difficult to argue that any simple set of characteristics is determinative of how likely cities are to attain compliance. 56 4. Does Housing Element Compliance Mean That More Housing Is Produced? Do local governments that comply with the housing element law, by adequately planning for future residential needs, actually enable more housing to be built in California? Moreover, does compliance help predict the types of housing built, such as the mix between single-family and multifamily development? If the answer to these questions is no, the resources that state and local government currently devote to housing element compliance and review might be better directed to other policy efforts to improve housing production. Using data from the 1990s, this chapter offers a statistical analysis of whether a city’s compliance status in the early part of the decade helps to predict the percentage increase in the city’s housing stock by 2000. Although data are not available to specifically study the development of affordable units, one can examine the growth of multifamily housing and the share of housing developed that is multifamily. The analyses control for a variety of other city characteristics that might influence the level and type of housing growth. The quantitative estimates are reported in Appendix B; this chapter reports the results for a general audience. Housing Element Compliance, Circa 1991 To make this analysis possible, the Department of Housing and Community Development made available its housing element compliance report for cities and counties as of December 31, 1991. This period was chosen because jurisdictions in the most populous parts of the state—the Southern California, San Diego, and Bay Area regions—had completed their second round of housing element updates by that point and had received HCD’s judgment as to whether their plans were 57 compliant. Moreover, the 1991 date is proximate in time to the 1990 Census, allowing use of data from that year as a “starting point.” Accordingly, this analysis focuses only on communities in these three regions, where most of the state’s population resides and which tend to have the most controversies and difficulties regarding housing element compliance. Compliance rates for cities in 1991 were poor, as Figure 4.1 illustrates. Only one-quarter of cities statewide were in compliance, with 59 percent being declared either noncompliant by HCD or overdue or “obsolete” because of the lack of any updated housing element in time for the deadline. In the wake of these low compliance rates, HCD and the attorney general’s office mounted a sustained effort to encourage more cities and counties to meet their legal responsibilities. This campaign was fairly successful, as the percentage of communities attaining compliance grew substantially. Housing element obsolete 15% Housing element due 2% Out of compliance 42% Under review by HCD 16% In compliance 25% Figure 4.1—Cities’ Housing Element Compliance Status, as of December 31, 1991 58 There tends to be some persistence in compliance status. There is a moderate, statistically significant degree of correlation (r = .31) between noncompliance in 1991 and noncompliance a decade later, in 2001. In other words, although many cities attained compliance status as the 1990s wore on, it remained the case that noncompliance after the second round of revisions was a fairly good predictor of noncompliance after the third round. Measuring Housing Growth Rates Using Census Data The first analysis in this chapter draws upon U.S. Census data to examine the rate of growth in housing units between 1990 and 2000. Because communities of different sizes are likely to have very different amounts of housing built, I examine the percentage increase in the number of housing units in each city between 1990 and 2000. Table 4.1 provides summary statistics regarding this measure of housing increase. The percentage increase in housing units will serve as a dependent variable (that is, the phenomenon that I seek to explain) in a quantitative analysis. Because some very small cities experienced extreme rates of housing change, the analysis is limited to communities with at least 2,500 residents, to avoid skewing the results. The main issue of interest is whether a connection exists between a city’s compliance status as of 1991 and its rate of housing growth Table 4.1 Percentage Growth Rates for Housing in Cities in the San Diego, Southern California, and San Francisco Bay Area Regions, 1990–2000 Average increase in housing units Median increase in housing units Standard deviation Largest rate of increase Largest rate of decrease 11.4 4.9 18.3 196.3 (Brentwood) –9.7 (La Habra Heights) SOURCE: Author calculations are from the 2000 Census and the 1990 Census. NOTE: Data focus only on cities of 2,500 or greater population. 59 through the 1990s. At first blush, there is surprisingly little evidence of such a link, as Table 4.2 reports. A simple comparison shows that the average noncompliant city in the relevant regions (San Diego, Southern California, and San Francisco Bay Area) actually increased its housing stock about 1.5 percent more than the average compliant city. The median noncompliant city, however, saw about 1.7 percent fewer new housing units constructed than the median compliant city (which indicates that some noncompliant cities with very high rates of housing development pushed up the mean for that group). In neither calculation are the differences statistically significant, however. Although interesting, one cannot draw confident conclusions from this simple comparison of compliant and noncompliant cities. Housing growth rates could be expected to differ among communities for many other reasons—for example, the location of the city, its attractiveness to new residents and developers, and its demographic characteristics. Table 4.2 Comparison of Housing Development in Compliant and Noncompliant Cities in the San Diego, Southern California, and San Francisco Bay Area Regions Housing Element Status as of 1991 Compliant Noncompliant % Increase in Houing Units, 1990–2000 Average City Median City No. of Cities 9.0 5.7 48 10.5 4.0 173 NOTES: Data are limited to cities of 2,500 or greater population. Differences between compliant and noncompliant cities are statistically insignificant, using either a t-test for difference of means, or a Kruskal-Wallis test for difference of medians. Controlling for Other Factors To account for the wide variety of factors that might affect housing growth rates across cities, it is possible to estimate a multivariate regression model that controls for such characteristics while assessing whether housing element compliance status has its own independent relationship with housing growth. In such an analysis, it is important that all of the relevant city characteristics be included in the analysis, since bias can result from improperly omitted variables. 60 A useful way to account for some of an area’s housing market characteristics that may not be easy to measure is to include a control for the county in which each city is located. That is the approach taken in this analysis.1 Thus, such factors as the county’s population growth trend, its location in the path of urban growth, county government growth policies, or other unobserved countywide factors are taken into account. At the city level, several local characteristics are included to inform the predictions about housing growth. I focus on measures of city characteristics as of 1990, as these are “baseline” factors that could be expected to influence housing development trends in the 1990s. The variables included in the analysis are described below. 1. Housing growth in the prior decade: Communities in urban regions tend to persist in their tendencies to differentiate themselves (Farley, 1964; Neiman, 1980). One can expect that cities with rapid rates of increase in housing in the 1980s would continue to grow quickly, although perhaps not at an equal rate, through the 1990s. Cities that grew slowly or shrank would likely continue to stagnate. This variable is measured as the percentage increase in total housing units in the city between 1980 and 1990. 2. Population size: It is reasonable to expect that large cities grow at a different rate than small cities, all else equal. The natural logarithm of the city population is used because of the highly skewed distribution of city populations. In addition, the ten communities of less than 2,500 population in these regions, and the two cities of greater than a million population, are dropped because the unique characteristics of very large and very small cities tended to hinder the analysis. 3. Population density: Higher densities indicate that the city probably has less vacant land to accommodate additional growth and that new growth may be more likely to cause spillovers such as traffic or ____________ 1Specifically, county-level fixed effects are included in the model for all cities except San Francisco, which is the only city in its county. 61 parking problems. Density is measured as the logarithm of persons per square mile, again because of the skewed distribution. 4. Urbanization status: Cities in urbanized areas may find it more challenging to accommodate new housing because of a lack of vacant land or conflicts with existing residents and businesses. This variable is an indicator, set to equal 1 in cities in Census-identified urbanized areas as of 1990 and 0 otherwise. 5. Age of the housing stock: Cities with an older housing stock may be more settled communities that have experienced less recent pressure for new development. The variable is the number of years elapsed since the median housing unit in the city was built, as of 1990. 6. Household income: Cities with high-income households are likely to have high land costs, and developers may tend to propose mostly single-family luxury homes. Wealthier communities may also have more political resources to mobilize in opposition to new development. The variable is the median household income in the 1990 Census (measured for the 1989 calendar year). 7. Senior citizens: Communities with high proportions of seniors may experience less pressure for new home construction. A large share of seniors implies that there are smaller households and probably less new household formation. The variable is the percentage of persons age 65 and over in the city population in 1990. 8. Technical workers: As an engine of California’s growth, high-tech industries might be expected to play an especially important role in housing demand. However, aerospace, also a high-tech industry, experienced major declines in employment, so the effects on housing increases may not be straightforward. The variable measures the percentage of the local (working) population employed in technical occupations, as classified in the 1990 Census. 9. Job-to-worker ratio: Cities that are “job centers,” with a high ratio of local employment to resident workers, might be expected to attract more interest from homebuyers and developers. This variable, derived from the Census Transportation Planning Package, is the natural logarithm of the ratio of jobs within the city to workers living in the city, as of 1990. 62 10. Commuting time: Cities in which many workers must commute long distances are apt to be less attractive and may experience less pressure for home construction, all else equal. The variable measures the average one-way commute time, in minutes, of workers living in the city in 1990. 11. Distance to urban centers: Isolated municipalities located far from urban job centers are likely to be less attractive to potential residents and housing developers. A computerized mapping program was used to calculate the straight-line distance between the city in question and the nearest of a set of traditional urban central cities in California.2 12. Housing element noncompliance: The major variable of interest is whether the city was found to be compliant with the housing element statute, according to HCD, in December 1991. This is an indicator variable, with compliant cities set to equal 0 and those found noncompliant, those overdue, or those having “obsolete” housing elements set to equal 1.3 Numerous other city characteristics were tested in other versions of the model predicting housing development. These included such measures as the housing vacancy rate in 1990, the percentage of housing units not connected to sewers, the percentage of recreational housing units, controls for central city or rural status, the average number of persons per household, and the percentage of various ethnic groups in the community. Each was dropped from the analysis when it proved persistently unrelated to housing growth rates. The effects of compliance ____________ 2The central cities chosen to represent historic metropolitan centers were Bakersfield, Chico, Fresno, Los Angeles, Merced, Modesto, Monterey, Oakland, Redding, Sacramento, San Diego, San Francisco, San Jose, Santa Barbara, Santa Cruz, Stockton, and Visalia. The mapping program calculated distance between each city in California and the nearest one of these central cities, using mapping coordinates (where a unit of 1 is equivalent to approximately 60 miles). 3I also estimated a different model in which cities that had overdue or “obsolete” housing elements were represented by a separate indicator variable from those that were overtly found noncompliant. This estimation strategy was based on the theory that being late with a housing element may not demonstrate as serious a problem as being overtly noncompliant. However, no substantive differences in results were found from those reported here, with one exception to be discussed below. 63 status on housing development are not influenced by the inclusion or exclusion of these variables.4 Results of the Analysis of Total Housing Unit Growth Detailed results of the model, which accounts for 64 percent of the variation in housing growth across cities, are listed in Appendix Table B.1. For our main purposes in this chapter, however, it is easy to summarize the findings: There is no evidence of a detectable relationship between housing element compliance and the percentage increase in housing across these cities during the 1990s. If noncompliance were a good predictor of slow housing growth, all else equal, one would expect to find a negative, statistically significant relationship between noncompliant housing elements and the percentage housing increase. However, the analysis shows a positive, insignificant relationship. Thus, for all the potential merits and benefits of housing element compliance, one must look to other factors to explain why some cities experience rapid housing development and other cities experience little. The analysis suggests that certain demographic characteristics, measures of a community’s position in the urban hierarchy, and the physical capacity of the city to accommodate new buildings are better able to predict housing growth. For example, cities with higher population densities as of 1990 added housing at a slower rate in the decade that followed, as did those with older housing stock at the start of the period. These results indicate that growth in the 1990s moved away from older, denser communities. At the same time, cities with larger populations, those that were far away from urban centers, and those with long commute times also experienced slower rates of housing growth. Cities with higher-income populations grew more slowly, perhaps ____________ 4Care was also taken to choose measures of community demographics that were not excessively intercorrelated. The median household income of the city, included here, is a good proxy for the overall status of the community. Other socioeconomic variables, such as the unemployment rate, poverty rate, or homeownership rate, are closely associated with median income, but these variables do not show as much association with housing growth. Again, the (non)significance of housing element compliance is not affected by the exclusion of these variables. 64 because land was more costly or zoned for lower densities. Perhaps surprisingly, the percentage of technical workers among local resident workers was negatively related to the rate of housing increase, quite possibly a reflection of the severe declines that would occur in aerospace and related industries in the early 1990s. The best predictor of housing growth, however, was the city’s rate of housing increase in the previous decade. Within any given county, cities that grew quickly in the 1980s were the ones that tended to add to the housing stock quickly in the 1990s.5 After accounting for these trends, housing element compliance as of 1991 provides no additional information about cities’ rates of housing growth. Analysis of Multifamily Development Using Construction Industry Data It is useful to consult other data sources as a check for the above analysis, which uses Census data on the number of housing units in each city. I therefore gathered data from the Construction Industry Research Board (CIRB) on the number of housing units constructed in each city in the period from 1991 through 2000. This data source has the additional benefit of being broken down by whether the units built were single-family or multifamily. In most cities in the San Diego, Southern California, and Bay Area regions, single-family units clearly predominated. In fact, in an average city in these three regions, only 27 percent of the new units were multifamily.6 Not surprisingly, therefore, there is a very strong correlation (r = 0.90) between our Census measure of housing unit increase and the CIRB data on increases resulting from single-family construction.7 When I ran the same regression model as was used for the ____________ 5The county-level fixed effects were also significantly related to housing growth rates, indicating the regional character of housing markets. 6If data from all cities in these regions are aggregated, the share of units that were multifamily among all new units was somewhat higher, at 33 percent (28 percent in cities statewide). This is because cities that had the highest numerical increases in housing units tended also to have higher shares of multifamily development. 7Specifically, the variable used was the ratio of newly constructed single-family units in the city to all housing units existing in that city as of 1990. 65 Census measure in the CIRB single-family increase data, results were thus largely similar. Again, no relationship between housing element compliance and housing development was apparent. As a fair-share approach to housing, the housing element statute is specifically concerned with increasing the production of affordable housing. Unfortunately, comparable information for all cities about the production of affordable units appears to be nonexistent. However, the CIRB data allow us to examine multifamily housing production. Multifamily housing includes apartments and condominiums, the types of housing generally most relevant for those at the lower to middle levels of the income distribution, particularly in the coastal metropolitan areas where housing is very expensive. Moreover, analysts have been particularly concerned that multifamily production has plummeted in California since the 1980s. “Whereas multifamily housing accounted for between 45 and 49 percent of total housing construction during the 1960s, 1970s, and 1980s, last decade it fell to 25 percent of the already depressed total” (Myers and Park, 2002, p. 2). Thus, the second regression analysis featured in Appendix B examines the determinants of increases in multifamily housing. The dependent variable is the ratio of newly constructed multifamily units (from 1991 through 2000) to all units existing in the city in 1990.8 The results of this analysis again show that housing element noncompliance as of 1991 is not a significant predictor of multifamily development, once other relevant factors are held constant. Although at least the relationship is negative (indicating that noncompliance is linked to a smaller multifamily increase), it is near zero and not statistically significant. Rather, demographic and land-use characteristics of the community are better predictors of multifamily development. Cities ____________ 8Some readers may wonder why the measure is based on the ratio of new multifamily units to all existing units, rather than existing multifamily units. The reason is straightforward. Cities with a very tiny number of existing multifamily units (say, two) could add a negligible number of new multifamily units over the decade (say, another two). If one uses the existing number of multifamily units as the denominator, then such cities are unfairly credited with a huge proportionate increase (in this example, 100 percent). The ratio used in the analysis better characterizes the role of multifamily development in augmenting each city’s housing stock. 66 with more senior citizens as of 1990 tended to experience slower rates of multifamily housing development. Cities with more jobs per worker saw more rapid multifamily development. The Mix of New Housing Development Finally, it is possible that housing element compliance could affect the mix of housing developed—multifamily versus single-family—despite its lack of effects on the rate of increase. The third regression analysis shown in Appendix B examines this issue; the dependent variable in this case is simply the percentage share of new housing units that were multifamily units, according to the CIRB data. Here there is a significant association of compliance with the outcome in question. The results show clearly that cities with noncompliant housing elements develop new housing that is weighted more toward single-family units. In fact, holding constant for all the other city characteristics treated in the model, a noncompliant housing element is linked to an 8 percentage point lower proportion of multifamily housing among the new units. This reduction is considerable, given that the average city saw multifamily units account for only 27 percent of new housing. The only other variables that are clearly related to this proportion are the median income of the city, its job-to-worker ratio, and its rate of housing growth in the prior decade. Cities with lower median incomes and more jobs per worker saw a higher percentage of multifamily development. Cities with slow growth rates in the 1980s also experienced a share of multifamily development that was higher than in those that grew quickly (although the relationship was substantively small). Thus, cities with compliant housing elements in 1991 experienced a more balanced mix of new housing development in the 1990s. In a further test, I examined whether cities that had overdue or obsolete housing elements in 1991 performed as poorly in their housing mix as those judged noncompliant by HCD reviewers. The results indicated that it was the cities overtly deemed noncompliant, rather than the cities 67 with overdue or obsolete plans, that had the heavier proportion of singlefamily housing rather than multifamily.9 Reconciling These Findings What does it mean to say that housing element compliance is not linked to the rate of multifamily development but is linked to the proportion of housing developed that is multifamily? One interpretation, consistent with the fair-share approach of housing element policy, is that thorough housing planning diversifies the mix of new residential development in the community. Units that developers might have preferred to build as detached homes (or that neighborhood groups might have preferred to see become detached homes) are constructed as apartments or condos instead, quite possibly because of an emphasis in the housing element on finding sufficient sites for multifamily development. The results imply that cities with compliant housing elements, although they may not be willing or able to accommodate more total housing units, are willing to substitute multifamily (more affordable) units for a share of single-family (marketrate) units. The 8 percentage point estimate of that substitution effect in this analysis has a similar order of magnitude to the 10 or 15 percent affordable housing set-aside that many cities require of some new developments through local inclusionary housing policies. Conclusion Much attention and resources are given to housing elements, in part because there is an assumption that noncompliant cities are those that artificially slow the rate of increase to the housing stock. Compliance status from a decade ago, however, was not a good predictor of the rate of subsequent new housing development in cities in the San Diego, Southern California, and Bay Area regions. Rather, housing market and ____________ 9Specifically, the regression analysis included two indicator variables: one for cities found noncompliant and one for overdue or obsolete cities. Results showed that, all else equal, noncompliant cities experienced about an 8.7 percentage point lower share of multifamily development than compliant cities. Those with overdue or obsolete housing elements had a smaller negative (–2.5 percentage points) but statistically insignificant relationship with the percentage of new housing that was multifamily. 68 demographic factors outweigh compliance status in contributing to variations in growth rates. The pace of neither single-family nor multifamily development is associated with compliance status. Does this nonfinding imply that the resources devoted to the preparation of housing elements are a wasted effort? Not necessarily. It is probably a healthy exercise for local governments to plan for growth and to assess where various types of new housing may be accommodated in the community. It is possible that the very requirement that all local governments prepare housing elements means that more attention is paid to housing needs and more housing is built statewide, regardless of whether individual communities are found to be compliant or noncompliant. One cannot be sure, then, that there would not be deleterious effects on housing production if the requirement were eliminated. Moreover, this analysis does not attempt to sort out the relationship of compliance status to production of affordable units, which receive particular emphasis in housing element law. Indeed, when examining the proportion of all units constructed from 1991 through 2000 that were in multifamily developments—most likely to be affordable—we do find that multifamily construction tends to displace a significant portion of single-family construction in cities with compliant housing elements. Nevertheless, it is quite striking that one can detect no measurable relationship between compliance and overall housing production. Cities that made efforts to comply with the housing element statute may have played host to a wider mix of new housing but did not accommodate more units overall, at least in the 1990s.10 California policymakers concerned about housing production, and local constraints on housing, may wish to think about other potential policy approaches. The next chapter describes some alternatives. ____________ 10A reader of an earlier draft of this report commented that compliance status at the end of the 1990s might better predict housing production rates in that decade. The idea is that many cities that were noncompliant in 1991 eventually gained compliance, and also that municipal policies measured by HCD in the 2000s were likely in effect and influencing housing development in the 1990s. However, when I substituted March 2001 housing element compliance status in these regressions for 1991 compliance status, it showed insignificant effects on all measures of housing development in the 1990s. 69 5. Can States Ensure Adequate Local Provision for Housing Development? Given the turmoil over housing element law in California and its apparent lack of success in generating more housing, a discussion of alternative policy approaches seems warranted. This chapter does not recommend any single policy change but is intended to help illustrate the array of options available to state policymakers. It begins by discussing changes that have been suggested by groups in recent years. Sometimes, however, a broader perspective on a policy problem is useful, and therefore the second part of the chapter examines the experiences of three other states in crafting policies that attempt to shift local planning and land-use activities in a more housing-friendly direction. The successes and failures of these states hold potential lessons for California. Alternative Approaches within the Context of the Current Housing Element Law Discussions of problems with the current housing element law among California policymakers and housing advocates have led to a number of proposals in recent years to change tactics. What most of these reform proposals have in common, however, is that they take the housing element approach—state review of local plans—more or less as a given. More Penalties and Prescriptions for Local Governments One approach sees the problem as a lack of serious consequences for local governments that fail to meet the requirements of existing housing element law. Myers and Park (2002, p. 5) argue that “the state needs to place greater weight on localities’ production of total housing 71 construction, including a share of multifamily housing, and that emphasis needs to be backed with stronger teeth than at present.” Thus, for example, it has been suggested (as in Senate Bill 910) that financial penalties be levied against noncompliant local governments, or that their eligibility for certain state or federal grants be rescinded. The attorney general could bring legal action against communities that persistently fail to comply. Another approach of this type seeks to prescribe local land-use regulation, under the assumption that localities, left to their own devices, will fail to provide housing opportunity. Thus, the state could impose “inclusionary zoning” requirements on all communities, under which all housing developments of a certain size must reserve a share of units for low- or moderate-income households. Other suggestions have included requiring minimum zoning densities in certain areas (such as that near public transit), or a blanket exemption of affordable housing projects from local zoning or growth controls (Senate Committee on Local Government, 1993, p. 112). Innovative approaches to promoting affordable housing need not be state-prescribed, however. Some cities, for example, have developed an “affordable housing overlay” as part of their zoning ordinance that allows a developer to claim a density bonus if his or her proposed project meets affordability standards. Self-Certification Based on Performance Self-certification is the idea of granting authority to local governments to approve their own housing plans without state oversight if they meet certain criteria involving demonstrated housing production. This approach was popular among local government representatives in the mid-1990s (Senate Committee on Housing and Land Use, 1995). As Mark Pisano, longtime executive director of the Southern California Association of Governments, argued in 1993: Local governments should be allowed to self-certify their housing element reforms. Self-certification should be linked to a performance review at the subregional or regional level. The right of a third party to request a review should be also allowed. Any conflicts arising from the review should be resolved through a subregional mediation process, thereby ensuring a timely 72 resolution of disputes. If mediation is unsuccessful, [HCD] would review the housing element performance under question (Senate Committee on Local Government, 1993, p. 42). Similarly, Janet McBride, a senior planner for ABAG, testified, “Communities which show a record of success in providing their ‘fair share’ of housing should be exempt from outside scrutiny of their housing element” (Senate Committee on Local Government, 1993, p. 49). As a result of these discussions, the legislature passed a pilot program, AB 1714, in 1995. That law allowed communities in the San Diego region to self-certify their housing elements if they had met their fairshare housing needs in the period since the previous housing element revision. According to HCD’s housing element status reports, only a few cities in San Diego have availed themselves of this option to date. The pilot program will sunset in 2009 unless the legislature acts to continue it. Some have cautioned that basing self-certification on strict measures of prior performance may make such an approach somewhat inflexible. During an economic recession of the type the state experienced in the early 1990s, for example, housing production slumped because of the economy, and few jurisdictions would have been able to meet even 75 percent of their assigned housing goals. Alternative suggestions for measuring performance include assessing whether the jurisdiction has issued a share of permits above the regional average or has approved a certain share (say, 90 percent) of proposed housing projects that are consistent with the local general plan (Senate Committee on Housing and Land Use, 1995). Subregional Allocations and Joint Housing Elements A number of commentators have suggested that jurisdictions in a subregion should decide collectively how to divide up that area’s housing allocations. This approach has been advocated specifically for smaller counties in large metropolitan regions, and some have suggested setting up joint-powers authorities as an intergovernmental arrangement to adopt joint housing elements and suballocate fair shares. The aim would then be for jobs/housing balance within each subregion—probably a 73 more tractable approach than seeking balance within each jurisdiction (Senate Committee on Housing and Land Use, 1995, pp. 64–65, 79). Subregional allocation has already occurred in some cases; for example, in the decision by SCAG to allow Orange County governments to formulate allocations within that county. It is quite possible, however, that making subregional allocation more widespread would not solve the political problem of how to divide an unwanted responsibility among jurisdictions. Rather, it may merely create a new political arena to haggle over the issue at a smaller geographic level. A provision allowing for greater allocation to subregions was inserted into SB 910, the housing element reform bill, in 2002. Encouraging Transfers of Housing Allocations among Jurisdictions A number of local government representatives have sought the ability for localities to “trade” or transfer their allocations with each other, presumably in exchange for payments or other considerations. The argument is that housing may be accommodated more easily or economically in some jurisdictions than others, and allowing transfers would enable more flexibility for localities in meeting their housing obligations. For example, a small suburb with limited, expensive vacant land could transfer its allocation to a nearby community with more or cheaper vacant land or an active housing redevelopment policy, thereby possibly allowing a greater total number of units to be developed. Section 65584.5 of the government code, passed in 1994, enables transfers among jurisdictions under certain circumstances. This legislation followed two years of debate over the topic, during which various stakeholder groups agreed on a compromise. Nevertheless, the provisions for transfer are quite strict and constrained. The entities engaged in the transfer must both have compliant housing elements, and the “donor” entity must have met at least 15 percent of its housing goals. The transfer agreement must include plans to construct the units within three years (or else the units revert to the donor community), and the COG must review it. The percentage of a community’s housing goals to 74 be transferred may not exceed the percentage of its goal that the community has already satisfied (Curtin, 2000, p. 11). At issue is whether transfers violate the spirit of fair-share housing policy, which aims at an equitable distribution of responsibilities across jurisdictions. Opponents, including some housing advocate groups, argue that wealthy communities that eschew low-cost housing will be allowed to buy their way out of their responsibilities to their region if widespread transfers are permitted. Another Approach: Rewards for Performance Enticing communities to accommodate housing would not be such an uphill battle if they perceived that doing so would be in their financial self-interest. Thus, creating an element of the state fiscal system that rewards local governments for the addition of housing units, particularly affordable units, would be an approach likely to result in less conflict and more cooperation. The existing Jobs/Housing Balance Incentive Program, passed in 2000, has elements of such a “rewards for performance” approach. Until it was defunded when the state budget went into deficit, the program authorized devoting $100 million in grants to local governments that demonstrate increased issuance of building permits, if they have adopted housing elements that are in compliance with state law. The housing bond passed by voters in November 2002 earmarks $100 million to continue funding this program. Senate Bill 423, enacted in late 2002, provides some specific instructions on how the funds are to be allocated. To provide certainty for local governments, a continuing rather than one-time source of funds for such an approach may be necessary. For example, a regional fund could be established for such purposes as transportation improvements and open space protection, with funds awarded to cities and counties that clearly demonstrate that they are taking actions (beyond planning) to accommodate their fair share of units, particularly affordable units. The fund could be financed through a regional impact fee on commercial development in job-heavy, housingpoor areas (Lewis, 2002). Others have suggested recalibrating the local fiscal system so that localities would be rewarded directly (i.e., by formula) for increases in 75 population or housing units. For example, there have been discussions of distributing a portion of local sales tax revenues to cities and counties on the basis of local populations, rather than the “point of sale” of retail transactions. AB 680 (Steinberg), which failed to emerge from the state legislature in 2002, took such an approach. Of course, such a wideranging reform of local finance raises many concerns and conflicts unrelated to housing policy, and thus far such efforts have not succeeded. It is worth noting that state subventions of Vehicle License Fee (VLF) revenues are awarded to localities through a population-based formula. Local officials have expressed worries that the state’s current budget crisis may threaten the existence of the so-called VLF backfill, which attempts to ensure that cuts to the VLF in the late 1990s do not result in decreased allocations to localities. Loss of all or some VLF revenues would further reduce local incentives to provide new housing. Relevant Experiences from Other States Other states, including some with equally strong traditions of home rule as California, have also wrestled with issues of inadequate housing production, mandated fair shares for jurisdictions, and state oversight of local planning. But the approaches taken differ widely from California’s, and their experiences may be worth considering. This section briefly sketches the approaches taken in New Jersey, Massachusetts, and Oregon.1 New Jersey Along with California, New Jersey likely has the most sustained experience with policies that involve state oversight of local housing policy.2 At first glance, New Jersey’s policy posture toward fair-share ____________ 1The focus here is only on contemporary policies that have some analogies to the California situation. A comprehensive examination would need to include the nowdefunct New York State Urban Development Corporation (UDC), a powerful state agency with the power to override local zoning to build subsidized housing projects in the late 1960s and early 1970s. Its zoning override power was later stripped by that state’s legislature, and the UDC ultimately became insolvent in the mid-1970s. 2This discussion of the New Jersey experience is based upon Calavita et al. (1997), Haar (1996), Hughes and VanDoren (1990), Listokin (1976), and Weinstein (1993), as well as the website of the Council on Affordable Housing. 76 housing production appears similar to California’s. The state’s 566 municipalities are required to prepare housing elements, are allocated an obligation of low- and moderate-income housing needs, and a state entity—in this case a Council on Affordable Housing (COAH), created by the legislature in 1985—reviews local plans. At a more basic level, however, there are serious differences between the two states’ approaches. Most important, New Jersey’s requirement of a fair-share responsibility on localities to provide a reasonable amount of housing within their region was mandated by judicial action—the famous Mount Laurel decisions of 1975 and 1983. The state’s supreme court justices drew upon state constitutional language to assert that municipalities had an obligation to admit affordable housing, thus providing more legal ammunition than California’s statutory housing element requirement. The court’s motivation was largely concern about racial and socioeconomic exclusion in the suburbs, given New Jersey’s stark distinctions between older, largely African American central cities and more affluent suburbs. An initially reluctant state legislature followed up on the court’s prodding by passing a Fair Housing Act in 1985 that created COAH to oversee the process, thereby removing most of the direct judicial supervision of contested municipal housing policies. In terms of the law’s actual implementation, the major salient differences between New Jersey’s approach and California’s include the following: 1. The fair-share allocation process is state-directed, with no COG involvement, and pertains only to low- and moderate-income housing needs, not “above moderate” needs. These affordability brackets have stricter definitions in New Jersey, however, with the low-income category basically equivalent to California’s verylow-income bracket, and the moderate category equivalent to California’s low-income grouping. 2. Municipalities are not obligated to have COAH review their housing elements. However, localities that choose to approach COAH and have their housing plans certified as adequate are shielded from lawsuits by builders for a six-year period. 77 3. Municipalities can meet part of their affordable housing obligations by rehabilitating existing units. Congregate housing facilities for the disabled can also be counted toward the obligation. 4. The major mechanism for the construction of affordable units is through the so-called “builder’s remedy,” under which a developer offers to construct affordable units that go some way toward meeting a community’s obligations, in exchange for certain local concessions. These projects ordinarily are 80 percent market rate units and 20 percent low- and moderateincome units. The concession given to the developer is typically a density bonus allowing construction of more market-rate units than would otherwise be allowed under local zoning, and in some cases local government offsets of the cost of the affordable units. Communities that have not met their affordable housing obligations and have not had their housing elements certified by COAH are often approached by for-profit builders proposing such an arrangement, which can be ordered by a court or agreed to by the municipality in a settlement. To avoid such suits, many local governments adopt an “inclusionary housing” ordinance, providing for a routine affordability set-aside of 20 percent of units in any development of a threshold size. During a building boom in the late 1980s, an estimated 15 percent of all developments in New Jersey included such an affordability setaside (Calavita et al., 1997, p. 127). 5. As a quasijudicial agency, “COAH has adopted a massive body of regulations governing nearly every element or program . . . through which a New Jersey locality might seek to comply with the Mount Laurel mandate” (Calavita et al., 1997, p. 119). 6. Municipalities in New Jersey are allowed to “trade” up to half their obligated number of affordable units to another jurisdiction that agrees to build the units (or, more commonly, to renovate dilapidated units), under so-called Regional Contribution Agreements. The “sending” jurisdiction, generally a suburb, must pay the “receiving” jurisdiction, usually an older 78 central city, at least $20,000 per unit, although higher amounts can be negotiated. Some view this aspect of the New Jersey system as a “sell-out” of the original aim of distributing housing opportunity widely. But other analysts note that the trading system has the merit of effectively serving as a tax on those communities that choose to avoid their affordable housing obligations (Hughes and VanDoren, 1990). Thus, the entire community that “benefits” from such exclusion must pay for it; contrast this to the mechanism of the builder’s remedy, in which the costs of new affordable units may be cross-subsidized by the market-rate units in the same project, meaning that new entrants to a community effectively pay for the city’s past exclusionary behavior. The trading system also has fiscal benefits for distressed central cities. Although a large amount of subsidized housing has been produced under the Mount Laurel system in New Jersey, it has been faulted on a number of counts. Most notably, the emphasis in most builder’s remedy developments has been on purchase housing, not rentals, meaning that the lowest-income households often cannot take advantage of the setasides because they cannot afford a down payment or do not have good credit. Thus, studies have indicated that a typical occupant of a subsidized suburban Mount Laurel unit is a young white family that already lived in suburbia and simply has not yet achieved its full earnings potential. On the other hand, a major advantage of the New Jersey approach is its standardized procedure for meeting local fair-share housing goals—a process that after more than 15 years of operation is becoming routine. The builder’s remedy approach uses the profit motive of developers to accelerate implementation and to develop far more housing units than would have otherwise been allowed under local zoning. Massachusetts The first strong fair-share policy approach to housing development anywhere in the country was probably Massachusetts’s Comprehensive 79 Permit Law.3 The law was enacted in 1969 and aimed at opening up the suburbs to affordable housing. It has often been viewed as an attack upon “snob zoning.” The Massachusetts law is typically referred to as “40B” in that state because of its chapter location in the state legal code. Its main strategy is to radically simplify the review process for developers seeking to build residential projects in which at least 25 percent of the units have long-term affordability restrictions. Developers proposing such projects can apply to the municipality for “comprehensive” building permits, through a local Zoning Board of Appeals. If successful, the comprehensive permit issued by the zoning board cuts through local building restrictions and overrides land-use regulations that are inconsistent with providing affordable housing; the zoning board can effectively overturn local zoning in some cases. This enables the builder to use one relatively streamlined process to receive an entitlement to build the project. This local zoning board is charged with bringing to the table all the relevant municipal departments and agencies that have an interest in the housing development for comments and concerns (such as the planning board, fire department, etc.), and with conducting a series of public hearings to air community concerns. The hearings must begin within 30 days of receipt of the developer’s application for a comprehensive permit, and the board must decide whether to award the permit within 40 days after the public hearings conclude. Although state environmental and other statutes still apply, local authorities must use the principle of meeting local needs in any denial or conditions that they impose upon the project (such as density or height restrictions). The community is not allowed to impose conditions that render the proposed project uneconomical to the builder. Furthermore, developers whose projects are denied by this local process have the option of appealing to a statewide board, the Housing Appeals Committee (HAC), which can overturn the local decision and order that the project ____________ 3This discussion of the Massachusetts process draws upon numerous recent articles in the Boston Globe by housing reporter Anthony Flint, materials from the websites of the Massachusetts Department of Housing and Community Development and the Citizens Housing and Planning Association, as well as early perspectives on the law by Listokin (1976). 80 be permitted. A developer can also appeal a local decision if conditions imposed upon the project threaten its economic feasibility. However, the developer may only appeal if the locality in which they wish to build has failed to meet certain housing affordability thresholds.4 Ordinarily, the HAC first encourages the applicant and the municipality to engage in a mediation process to see if issues can be resolved without a HAC review. In the initial period after passage of the 40B law, progress was very slow, as there were court challenges to the law and frequent proposals in the legislature to weaken it. By 1972, there had been only 35 applications to local authorities for comprehensive permits, and only a few appeals to the state HAC. After six years of operation, the law had resulted in the construction of only about 1,100 new housing units, although more than 3,000 others were planned or pending (Listokin 1976, pp. 100–103). More recently, there has been a pronounced increase in the number of projects proposed under the law, with about 5,000 units produced between 1990 and 1997. Between 1970 and 2001, about 25,000 units in 170 communities had been approved (Massachusetts DHCD, 2001). According to Flint (2002), “Because many towns have put the brakes on all growth and buildable land is scarce, developers say going the 40-B route—generally large, dense projects that are 25 percent affordable under state guidelines—is the only way to do business these days.” However, a disproportionate number of the recent 40B applications have been projects for senior citizens only, which local officials often look upon more favorably. The law is also limited in that it does not apply to many commercial developers. To have the right to seek a comprehensive permit, the developer must be either a nonprofit or a limited-dividend company, and the project must initially receive approval from a federal or state subsidized housing program. In recent years, the Federal Home Loan Bank of Boston has provided substantial subsidies to developers in the program, which has helped boost construction. Although 40B ____________ 4The city is exempted from an appeal if at least 10 percent of its housing units are in projects subsidized for low- and moderate-income residents, or if such affordable units constitute at least 1.5 percent of the city’s land area zoned for residential, industrial, and commercial uses. Few jurisdictions in Massachusetts meet these criteria. 81 applications have become more numerous lately, the law has not succeeded in easing that state’s very serious housing shortage. The law has contributed to a much wider geographic distribution of subsidized housing, with the proportion of subsidized units outside large cities in Massachusetts rising from 31 percent in 1972 to 63 percent in 1997 (Citizens Housing and Planning Association, 2001). A state task force there recently proposed more radical state oversight of local planning, such as overturning local caps on growth (Flint, 2001). Given the shortage of housing subsidies in contemporary California (Little Hoover Commission, 2002), adapting the current Massachusetts statutory provisions to California is unlikely to bear immediate results in terms of housing production. The 40B law reflects its origins as a civilrights-oriented provision rather than a measure to increase housing production across the board. However, if a California version were to relax the requirement that the affordable units be subsidized by a state or federal source, and perhaps reduce the required share of affordable units to 15 or 20 percent of the project total, it would open comprehensive permitting to a far wider array of projects. Many developers in urban and suburban California already have built projects in which 15 percent or so of the units are affordable and cross-subsidized, in effect, by the market-rate units. Such approaches are mandated by some cities’ inclusionary housing requirements. In California, the right of appeal to a statewide board (or perhaps more appropriately, a regional board) could be allowed if the local government has a noncompliant housing element. The Portland Region Oregon is well-known for undertaking state-level review of local housing plans.5 In 1973, at the governor’s urging, the legislature passed a law requiring that local governments prepare comprehensive plans and establishing a state-level, quasijudicial Land Conservation and Development Commission (LCDC) to review them. One of the initial 14 statewide goals developed by LCDC concerned housing opportunity. The LCDC goal required that, as in California, localities inventory ____________ 5The Portland discussion draws upon Hammond (2002), Knaap and Nelson (1992, Chapter 3), Lewis (1996, pp. 179–182), and Toulan (1994). 82 residential land and develop plans to encourage the building of an adequate number and mix of housing units for households in various income brackets (Knaap and Nelson, 1992, p. 77). In LCDC’s initial reviews of local housing plans, every jurisdiction of more than 5,000 population in the Portland metropolitan area except the central city itself saw its housing elements rejected because they provided for an insufficient range of housing types and prices. The area’s regional government, the Metropolitan Service District (since renamed Metro) attempted to resolve this state-local conflict by enacting an Areawide Housing Opportunity Plan—a fair-share allocation plan that distributed federal housing subsidies in relation to local needs. The plan withered when federal housing funds dried up in the early 1980s, but LCDC adopted the broad outlines of the Metro plan as its own regulation, later codified into state law by the legislature. As adopted, the so-called Metropolitan Housing Rule applies only to the Portland region, which contains about half the state’s population. The rule prescribes local land-use mixing, requiring that local government plans within Metro’s boundaries allow for at least 50 percent of new residential units to be multifamily or attached housing units. It sets out a “10-8-6” formula, which requires that the largest municipalities in the region zone for an average of at least ten dwelling units per buildable acre, whereas almost all the remaining suburbs must zone for at least eight units per acre. Five very small cities were given a quota of six units per acre. Despite local government objections to the new rule, strong support at the state and regional level along with support from an unusual alliance of environmentalists and homebuilders led to the implementation of the plan in virtually every Portland-area community. The authority of LCDC to fiscally sanction uncooperative localities no doubt helped speed up the process. The amount of land zoned for multifamily housing increased fourfold to more than one-quarter of net buildable acres. By 1991, vacant residential land in the Portland region allowed for the development of 54 percent multifamily units, and some rapidly growing suburbs, such as Beaverton, heavily exceeded the 50 percent minimum. 83 During the boom of the 1990s, apartments proliferated in the suburbs of Portland, which helped lead to the recent finding that households below the poverty line in that metropolitan area became less concentrated in the central city and more likely to locate in suburbia by the time of the 2000 Census. Although neighborhood objections to multifamily housing continue in the Portland area, according to one homebuilder, “nobody has been able to keep [multifamily housing] totally out. The reason is because of the [metropolitan housing] rule. Without that, I am sure it would have followed the same pattern the rest of the country did” (quoted in Hammond, 2002). However, no systematic study of the rule has attempted to determine the degree to which it assists in housing affordability or socioeconomic mixing. Another aspect of Oregon’s statewide planning system, the requirement for urban growth boundaries around each metropolitan area, directs growth pressures inward and probably helps account for the sharp spikes in densities in the Portland region. After all, the increased zoning densities under the Metropolitan Housing Rule do not require that builders propose projects at such densities. In a different regulatory regime, developers might have found it more profitable to build lowerdensity housing in the existing suburbs while also rapidly expanding lowdensity housing beyond the urban fringe. Conclusion There have been numerous proposals to reform housing element law in California, involving penalties for noncompliance, more selfcertification, or transfers of allocations across communities. Most suggested reforms, however, would proceed within the current approach: state review of local housing plans. A more fundamental shift would be to directly reward effective performance by directing funds to localities on the basis of the number and mix of new units developed. The experiences of Massachusetts, New Jersey, and Oregon all offer different experiences regarding fair-share planning and state review of local housing policy. Even if none of these approaches is completely appropriate to the California context, some combination of the 84 innovations attempted elsewhere—comprehensive permitting, state or regional boards of review or appeals, minimum zoning densities, or the builder’s remedy—may be worthy of consideration in California. 85 6. Conclusion Many factors appear to complicate efforts by local governments to reach housing compliance. In some cases, local policies restricting growth or an older housing stock hinder compliance, although it also appears that communities that have remained persistently noncompliant tend to be high-status cities. When it comes to sorting out the effects of noncompliance, however, there is no solid evidence that noncompliant cities experience a slower rate of housing development than compliant cities, holding all else constant. Thus, given the substantial controversy over the existing housing element law, it is worth considering alternative policy approaches, some of which were discussed in Chapter 5. An important question sitting astride the whole issue is whether the various fundamental values that California policymakers hold dear— increased housing production, an equitable distribution of housing responsibility across communities, special attention to the housing needs of low-income families, local autonomy and home rule, environmental protection, and more—can all be accommodated within housing element law, or whether there are tensions among them. A secondary question is whether the current approach—state review of local planning for future housing needs—is the most effective use of resources to further these goals. Goals in Conflict Although other states have experimented with different approaches to state review of local housing plans, California has continued with its own approach to housing elements. As in the other states, California’s housing element requirement has often gone against the grain of local policymaking because of its regionalist orientation: It asks cities to plan for the needs of the wider region, not just the needs of current city residents. Indeed, the philosophy behind fair-share housing plans is that the local “police power” to regulate land use should be directed toward 87 the general welfare of the region and not just the general welfare of the specific locality (Listokin, 1976, p. 19). Equitable Distribution or Sheer Production? Beyond this conflict over the local role, there are more fundamental value conflicts in housing element policy in California. It took shape in the 1970s in an era in which there was increasing concern with civil rights and the ability of minorities and low-income families to have an opportunity to live in suburbia, not just in inner-city or rural enclaves. Most of the fair-share plans developed by regions or states in this time period had such equal-opportunity goals. Housing element policy was also concerned with improving the local planning process—a major policy goal for the state through much of the post-World War II era. In more recent decades, California policymakers have also become very concerned with overall residential production and the need to increase housing supplies, particularly but not exclusively at the low- and moderate-income levels. Thus, policymakers have increasingly turned to the housing element process as a tool to ensure adequate housing production throughout the state. But it is not clear that a fair-share allocation and planning process is necessarily the best tool for this job. In the contemporary period, many of the most racially and socioeconomically diverse jurisdictions in California are suburbs (Sandoval, Johnson, and Tafoya, 2002), although there remain numerous pockets of privilege in which low-income and minority populations are largely absent. In many cases, a greater number of affordable units could feasibly be produced in areas where there are already sizeable low-income populations than in the wealthier jurisdictions, where land prices are likely to be much higher. Thus, the goal of maximizing housing production can conflict with the goal of allocating a region’s needs as fair shares to each municipality. A scheme to allow widespread trading of housing allocations to other jurisdictions, as in New Jersey, could reduce this conflict, but many housing advocates and policymakers have little sympathy for allowing high-status communities to escape what is seen as their responsibilities to 88 the broader society.1 What the conflict implies, perhaps, is that different tools are necessary to enhance overall production from the tools necessary to encourage equal housing opportunity across jurisdictions. A Mandate in a Vacuum? Then, too, there are the conflicts between the housing element law’s single-minded pursuit of increased housing and other state laws that tend to raise barriers for housing production. As already noted, open-space and coastal protections, environmental impact review, congestion management, and farmland preservation are also all encouraged under California law and are not always easily reconciled with housing element provisions that seek to identify many land sites for future housing construction. The state’s new law requiring that local governments identify a 20-year supply of water for new housing developments before approving them also sends a different signal from housing element policy. These potential conflicts with environmental or growthmanagement goals were recognized even by an early observer of fair-share housing policies: Fair share basically accepts growth; it merely wants to direct it in a rational and equitable manner. The nongrowth or phased-growth philosophy questions whether growth, especially rapid growth, should be tolerated. This approach is therefore in contradiction with the regional allocation philosophy (Listokin, 1976, p. 137). Devil in the Details Those involved in the debate over housing policy often note how long and highly detailed the housing element statute is. Indeed, when the relevant legal code is downloaded from the Internet and printed, it stretches out over 27 pages. Highly detailed statutes are often evidence of widespread disagreement on a given policy, as waves of “reform” occur ____________ 1New Jersey’s Regional Contribution Agreements “shifted the rationale of the Mount Laurel doctrine away from the broad goal of ending geographic segregation surrounding inner-city minorities and toward the raw provision of low-income housing. . . . [Homeowners in the sending jurisdictions] are amenable, it turns out, to paying a form of ransom, through taxes, that preserves local control of new entrants while allowing lower-income housing to be built elsewhere” (Haar, 1996, p. 114). 89 in which opposing interests seek to have their concerns addressed and protections preserved in law. What has resulted, in the case of the housing element statute, is an unwieldy law that is often difficult for outside observers (including this researcher) to comprehend in its entirety or details. In the 33 years since the housing element statute was enacted, the search for an adequate supply of housing in California has become only more elusive. It may be a ripe occasion for policymakers and affected interests to regroup and reconsider the goals and approaches of state housing law, seeking an approach that is more workable, transparent, and straightforward, with measurable barometers of success or failure. In so doing, policymakers will need to resolve whether the major goal of such a law is a sheer increase in residential construction or an equitable distribution of affordable housing. Using a fair-share planning approach as a tool to encourage overall housing production may place an unrealistic burden on a relatively fragile, process-oriented policy. 90 Appendix A Multivariate Analyses of Housing Element Noncompliance This appendix reports the results of logistic regression analysis of city characteristics that might be expected to influence whether a city complies with housing element law. Chapter 3 discusses the general logic behind the analysis, and Table 3.3 describes the variables and data sources. The dependent variable is a dichotomous measure of noncompliance, equal to 1 if the city is noncompliant and 0 if compliant. If the city’s housing element draft is overdue, it is legally noncompliant and is treated as such. Cities under review by HCD at the time of the September 25, 2002, compliance report are dropped from the analysis. The table reports “odds ratios,” which represent the change in odds that an average city will be noncompliant, given a one-unit increase in the variable of interest. Odds ratios above 1.0 indicate an increased probability, and those below 1.0 a reduced probability, of noncompliance. Table A.1 provides results of a main model and an alternative model. The only difference is that the alternative model includes a measure of the degree to which the city planning director feels that the policies affecting residential growth rates have resulted from popular pressure by neighborhoods or voter initiatives (using a four-point scale). However, quite a few planners responding to the survey answered “don’t know” or skipped this question, which results in the loss of 41 cities from the analysis. Thus, I am more confident in the results of the main model. Table A.2 provides summary statistics for the variables in the models. 91 Table A.1 Logistic Models of Housing Element Noncompliance as of September 2002 92 Main Model Independent Variable Odds Ratio z-Statistic % of housing owner-occupied 1.023 1.35 % of residents white, non-Hispanic 0.998 –0.25 % of recreational housing units 1.023 0.35 Population density (natural log) Age of median housing unit (years) 1.033 1.079 0.10 2.69 Supply of land constrains growth (five-point scale) 0.910 –0.60 Population (natural log) 0.698 –1.81 Per capita own-source revenues (natural log) 1.087 0.25 City’s % share of local property tax revenues 1.037 1.03 Lack of planning staff constrains growth (five-point scale) 1.173 0.81 Review of residential proposals getting longer (four-point scale) 2.082 2.63 No. of overt growth restrictions (of possible seven) 2.373 4.39 City requires affordable units in new developments (dichotomous) 0.558 Slow-growth policies resulting from initiatives, neighborhoods (four-point scale) N/I –1.44 N/I City is in Inland Empire (dichotomous) 3.952 1.88 City is in San Francisco Bay Area (dichotomous) 3.056 2.33 City is in Central Valley (dichotomous) 31.121 2.68 No. of months since housing element revision due 0.946 –4.40 Alternative Model P>|z| Odds Ratio z-Statistic P>|z| 0.178 1.015 0.84 0.402 0.805 1.002 0.22 0.827 0.725 0.989 –0.15 0.877 0.917 0.911 0.007 1.093 –0.27 0.785 2.74 0.006 0.662 0.897 –0.60 0.546 0.070 0.645 –2.12 0.034 0.804 1.082 0.21 0.832 0.304 1.013 0.35 0.724 0.416 1.079 0.35 0.729 0.009 1.978 1.98 0.048 0.000 2.443 4.11 0.000 0.150 0.395 N/I 1.570 –1.84 0.066 1.63 0.103 0.060 3.792 1.67 0.095 0.020 2.679 1.78 0.075 0.007 32.510 2.47 0.014 0.000 0.946 –3.96 0.000 No. of cities Log likelihood Pseudo R-squared % correctly predicted 224 –98.41 0.36 79.0 183 –77.53 0.38 82.5 NOTES: Dependent variable is dichotomous (1 = noncompliant, 0 = compliant). N/I indicates variable not included in the model. Significance calculated using robust standard errors. Table A.2 Summary Statistics for Continuous Variables in the Preceding Models Variable Mean % of housing owner-occupied 60.3 % of residents white, non-Hispanic 49.5 % recreational housing units 1.4 Population density (natural log) 8.1 Age of median housing unit (years to 1990) 30.2 Supply of land constrains growth (five- point scale) 3.5 Population (natural log) 10.4 Per capita own-source revenues (natural log) 6.1 City’s % share of local property tax revenues 13.2 Lack of planning staff constrains growth (five-point scale) 2.2 Review of residential proposals getting longer (four-point scale) 1.9 No. of overt growth restrictions (of possible seven) 0.7 Slow-growth policies resulting from initiatives, neighborhoods (four-point scale) 1.6 No. of months since housing element revision due 35.7 Standard Minimum Maximum Deviation Value Value 13.1 23.8 97.1 23.8 1.0 91.2 4.1 0 34.9 0.9 3.7 10.1 8.7 11 59 1.4 1 5 1.2 6.1 14.0 0.6 4.7 8.4 7.4 0 47.6 1.1 1 5 0.7 1 4 1.0 0 5 0.8 1 40.3 3 4 117 NOTE: Values are only for cities included in model in Table A.1. 93 Appendix B Multivariate Analyses of the Rate and Mix of New Housing Production in the 1990s This appendix reports the results of ordinary least-squares regression analyses of city characteristics that might influence the rate of growth in housing units between 1990 and 2000 and the mix between single-family and multifamily units. The general logic behind the analysis is discussed in Chapter 4, where the section entitled “Controlling for Other Factors” describes the independent variables employed, virtually all of which are derived from Census sources. An exception is the measure of housing element noncompliance, derived from an HCD compliance status report dated December 31, 1991. The dependent variable in the first regression (Table B.1) is the percentage change in the number of housing units in the city between the 1990 and 2000 Censuses. The dependent variable in the second regression (Table B.2) is the ratio of newly constructed multifamily units (1991–2000) to total existing units (in 1990). The dependent variable in the third regression (Table B.3) is the percentage share of newly built units that are multifamily units. The source of the data on new construction of housing units is the Construction Industry Research Board. The analysis is limited to cities in the regions covered by the San Diego Association of Governments, the Southern California Association of Governments, and the Association of Bay Area Governments. These regions had completed their second round of housing element revisions by 1991. Finally, Table B.4 provides summary statistics for the variables in the models. 95 Table B.1 Regression Model of Housing Unit Growth in the 1990s, with County Fixed Effects Independent Variable Housing growth in prior decade Population size (log) Population density (log) Urbanized Age of housing stock Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Noncompliant housing element B0 % Growth in Housing Units Coef. t-Value P>|t| 0.174 5.64 0.000 –1.412 –1.73 0.085 –3.216 –1.68 0.096 –0.943 –0.28 0.779 –0.214 –2.16 0.032 –0.000 –1.88 0.062 –0.243 –1.07 0.285 –1.079 –1.85 0.065 0.561 0.42 0.678 –0.618 –2.08 0.039 –8.927 –2.97 0.003 1.473 1.03 0.303 83.958 3.24 0.001 No. of cities Prob. > F Adjusted R-squared 202 0.000 0.64 NOTES: Dependent variable is the percentage increase in total housing units (1990–2000). Analysis is of cities in the San Diego Association of Governments, Southern California Association of Governments, and Association of Bay Area Governments and is limited to cities that were incorporated and had a population of at least 2,500 in 1990. Robust standard errors are calculated. County fixed effects are used for all counties except San Francisco and are statistically significant (F = 4.421, p = .000). 96 Table B.2 Regression Model of Multifamily Development Growth in the 1990s, with County Fixed Effects Independent Variable Housing growth in prior decade Population size (log) Population density (log) Urbanized Age of housing stock Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Noncompliant housing element B0 Ratio of New Multifamily Units to Existing Units Coef. t–Value P>|t| –0.000 –0.17 0.868 –0.005 –0.89 0.374 –0.008 –0.78 0.435 0.009 0.59 0.558 –0.001 –1.26 0.210 –0.000 –1.25 0.213 –0.001 –2.09 0.038 –0.001 –0.69 0.494 0.021 2.45 0.015 –0.000 –0.11 0.911 –0.010 –0.70 0.485 –0.006 –1.44 0.151 0.211 1.62 0.106 No. of cities Prob. > F Adjusted R-squared 200 0.005 0.23 NOTES: Dependent variable is the ratio of newly constructed multifamily units (1991–2000) to the total number of housing units in 1990. Analysis is of cities in the San Diego Association of Governments, Southern California Association of Governments, and Association of Bay Area Governments and is limited to cities that were incorporated and had a population of at least 2,500 in 1990. Robust standard errors are calculated. County fixed effects are used for all counties except San Francisco and are statistically significant (F = 1.779, p = .041). 97 Table B.3 Regression Model of Share of Housing Development That Is Multifamily, with County Fixed Effects Independent Variable Housing growth in prior decade Population size (log) Population density (log) Urbanized Age of housing stock Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Noncompliant housing element B0 % of New Housing Units That Are Multifamily Coef. t–Value P>|t| –0.083 –2.30 0.023 1.650 0.70 0.484 4.804 1.26 0.210 –4.259 –0.62 0.535 –0.085 –0.30 0.766 –0.000 –2.96 0.004 –0.229 –0.72 0.474 0.288 0.17 0.868 12.379 3.09 0.002 –0.545 –1.19 0.235 0.658 0.07 0.946 –8.106 –2.18 0.031 19.049 0.44 0.660 No. of cities Prob. > F Adjusted R-squared 200 0.000 0.36 NOTES: Dependent variable is the percentage share of housing units developed (1991–2000) that are multifamily units. Analysis is of cities in the San Diego Association of Governments, Southern California Association of Governments, and Association of Bay Area Governments and is limited to cities that were incorporated and had a population of at least 2,500 in 1990. Robust standard errors are calculated. County fixed effects are used for all counties except San Francisco and are statistically significant (F = 1.759, p = .044). 98 Table B.4 Summary Statistics for Continuous Variables in the Preceding Models Dependent Variables % housing growth in 1990s Ratio of new multifamily units to total 1990 units % of new housing units that are multifamily Independent Variables % housing growth in prior decade Population size (log) Population density (log) Age of housing stock (to 1990) Median household income % of residents age 65+ % of workers in technical occupations Job/worker ratio (log) Average commute time Distance from urban center Standard Minimum Maximum Mean Deviation Value Value 9.6 13.7 –9.7 73.8 0.02 0.03 0 0.28 29.0 24.7 0 98.4 25.3 10.4 8.3 24.7 43,900 11.6 3.7 –0.1 24.8 0.4 39.7 1.1 0.8 0.7 18,489 6.2 1.3 0.5 4.1 0.4 –7.0 7.9 5.2 4 18,635 3.5 0 –1.8 14.1 0 389.8 13.6 10.1 51 130,734 47.9 7.9 1.8 40.5 3.2 NOTES: Values are only for cities included in the model in Tables B.1, B.2, and B.3. Commute time is measured in minutes, distance from urban center is measured in mapping units. 99 Bibliography Barbour, Elisa, Metropolitan Growth Planning in California, 1900–2000, Public Policy Institute of California, San Francisco, California, 2002. Barbour, Elisa, and Paul Lewis, Development Priorities in California Cities: Results from a PPIC Survey, Occasional Paper, Public Policy Institute of California, San Francisco, California, 1998. Bay Area Council, “ABAG Underestimates Regional Housing Needs,” Housing and Development Report, Vol. 1, No. 4, September 1988(a), p. 1. Bay Area Council, “Fair Share Housing Production: How Has the Bay Area Performed?” Housing and Development Report, Vol. 1, No. 4, September 1988(b), p. 3. Bishop, Amanda, and Jessica Materna, “Housing Bill Pressures Cities,” San Francisco Business Times, October 15, 2001. Brownlow, Judith, Fair Share Housing: A Partially Annotated Bibliography, Council of Planning Librarians, 1991. Brueckner, Jan, “Strategic Control of Growth in a System of Cities,” Journal of Political Economy, Vol. 57, 1995, pp. 393–416. Calavita, Nico, Kenneth Grimes, and Alan Mallach, “Inclusionary Housing in California and New Jersey: A Comparative Analysis,” Housing Policy Debate, Vol. 8, No. 1, 1997, pp. 109–142. Carrigg, Dan, “Balancing Housing and Growth Pressures with Limited Resources: It’s Time for Leadership,” Western City, Vol. 78, No. 4, April 2002, pp. 13, 15–16, 18–22. Citizens Housing and Planning Association, Recommendations to Improve Chapter 40B, the Massachusetts Comprehensive Permit Law, Boston, Massachusetts, 2001. Curtin, Daniel J., Jr., California Land Use and Planning Law, 20th ed., Solano Press Books, Point Arena, California, 2000. 101 Department of Housing and Community Development (HCD), Developing a Regional Housing Needs Plan, State of California, Business, Transportation, and Housing Agency, Sacramento, California, 1988. Department of Housing and Community Development (HCD), Status of Local Housing Elements (Period of October 1, 1990 through December 31, 1990), State of California, Business, Transportation, and Housing Agency, Sacramento, California, February 1991. Department of Housing and Community Development (HCD), Raising the Roof: California Housing Development Projections and Constraints, 1997–2020, State of California, Business, Transportation, and Housing Agency, Sacramento, California, May 2000. Department of Housing and Community Development (HCD), Pay To Play: Residential Development Fees in California Cities and Counties, State of California, Business, Transportation, and Housing Agency, Sacramento, California, 2001. Department of Housing and Community Development (HCD), “Housing Element Review Worksheet,” State of California, Business, Transportation, and Housing Agency, Sacramento, California, n. d. Dodge, Shannon, San Francisco Bay Area Housing Crisis Report Card, Greenbelt Alliance and Nonprofit Housing Association of Northern California, San Francisco, California, June 2002. Farley, Reynolds, “Suburban Persistence,” American Sociological Review, Vol. 29, February 1964, pp. 38–47. Fischel, William A., The Homevoter Hypothesis: How Home Values Influence Local Government Taxation, School Finance, and Land-Use Policies, Harvard University Press, Cambridge, Massachusetts, 2001. Flint, Anthony, “Report: Housing Inhibited by Rules,” Boston Globe, December 16, 2001, p. B1. Flint, Anthony, “Worries Develop as Town Expands: Affordable Housing Projects Piling Up,” Boston Globe, February 10, 2002, p. B1. Fulton, William, Guide to California Planning, 2nd ed., Solano Press Books, Point Arena, California, 1999. 102 Haar, Charles M., Suburbs under Siege: Race, Space, and Audacious Judges, Princeton University Press, Princeton, New Jersey, 1996. Hammond, Betsy, “Income Groups Intermingle,” The Oregonian, May 15, 2002, p. A1. Hecht, Peter, “Emotional Hearing on Folsom Housing,” Sacramento Bee, May 14, 2002. Hill, Laura E., and Hans P. Johnson, Understanding the Future of Californians’ Fertility: The Role of Immigrants, Public Policy Institute of California, San Francisco, California, 2002. Hughes, Mark Alan, and Peter M. VanDoren, “Social Policy Through Land Reform: New Jersey’s Mount Laurel Controversy,” Political Science Quarterly, Vol. 105, No. 1, 1990, pp. 97–111. Jackson, Carlos, “Realistic Housing Element Law Needed,” California County, September/October 1994, pp. 23–34. Knaap, Gerrit, and Arthur C. Nelson, The Regulated Landscape: Lessons on State Land Use Planning from Oregon, Lincoln Institute of Land Policy, Cambridge, Massachusetts, 1992. Lewis, Paul G., Shaping Suburbia: How Political Institutions Organize Urban Development, University of Pittsburgh Press, Pittsburgh, Pennsylvania, 1996. Lewis, Paul G., “Creating a ‘Home’ for Good Housing Policy,” Western City, Vol. 78, No. 8, August 2002, pp. 16–17, 22, 52. Lewis, Paul G., and Max Neiman, Residential Development and Growth Control Policies: Survey Results from Cities in Three California Regions, Occasional Paper, Public Policy Institute of California, San Francisco, California, 2000. Lewis, Paul G., and Max Neiman, Cities Under Pressure: Local Growth Controls and Residential Development Policy, Public Policy Institute of California, San Francisco, California, 2002. Listokin, David, Fair Share Housing Allocation, Center for Urban Policy Research, Rutgers University, New Brunswick, New Jersey, 1976. Little Hoover Commission, Rebuilding the Dream: Solving California’s Affordable Housing Crisis, Milton Marks “Little Hoover” Commission 103 on California State Organization and Economy, Sacramento, California, 2002. Lyons, Silas, “County Leaders Oppose State’s Housing Mandate,” San Luis Obispo Tribune, January 10, 2002a, p. A1. Lyons, Silas, “Raising the Roof in SLO County: Communities Struggle with Agency’s Housing Demands,” San Luis Obispo Tribune, January 27, 2002b, p. A1. Massachusetts Department of Housing and Community Development (Massachusetts DHCD), “Chapter 40B: Frequently Asked Questions,” February 2001, available at www.state.ma.us/dhcd/ Ch40B/FAQ.htm (downloaded March 28, 2002). Moller, Rosa Maria, Hans Johnson, and Michael Dardia, What Explains Crowding in California? California Research Bureau, Sacramento, California, February 2002. “Motel Conversion,” archives of an online discussion among municipal planners, posted by the League of California Cities, downloaded April 23, 2002, available at http://www.cacities.org/doc.asp?id=3896. Myers, Dowell, “Demographic Futures as a Guide to Planning: California’s Latinos and the Compact City,” Journal of the American Planning Association, Vol. 67, No. 4, 2001, pp. 383–397. Myers, Dowell, and Julie Park, The Great Housing Collapse in California, Fannie Mae Foundation, Washington, D.C., May 2002. Neiman, Max, “Zoning Policy, Income Clustering, and Suburban Change,” Social Science Quarterly, Vol. 61, 1980, pp. 666–675. Nguyen, Hang, “Cities Dismayed by Affordable Housing Measure,” Los Angeles Times, April 20, 2001, p. A3. Padmanabhan, Sekhar, “Judge Affirms Folsom Housing Pact,” Sacramento Bee, April 9, 2002, p. B1. Perkes, Courtney, “O.C. Homes Among Most Crowded in the Nation,” Orange County Register, June 16, 2002. Rawson, Mike, “Housing Element Law,” California Housing Law Project, Sacramento, 2000, available at http://www.housingadvocates. org/pdf/site/facts/HousingElementLaw.pdf. 104 Richards, Watson & Gershon, “The Housing Element Statute: A Status Report,” Los Angeles, California, March 1998, available at http:// www.rwglaw.com/Articles/Articles312. Richards, Watson & Gershon, “New Housing Element Legislation: The End of Mandate Suspension,” Los Angeles, California, January 1999, available at http://www.rwglaw.com/Articles/Articles414. Sandoval, Juan Onésimo, Hans P. Johnson, and Sonya M. Tafoya, “Who’s Your Neighbor? Residential Segregation and Diversity in California,” California Counts, Vol. 4, No. 1, August 2002. Select Committee on Housing and Urban Affairs, California Senate, Department of Housing and Community Development v. Housing Needs of the People of the State of California, Senate Publications, Sacramento, California, 1974. Senate Committee on Housing and Land Use, California Senate, Housing Element Law: A Summary Report from the Special Hearing, Senate Publications, Sacramento, California, March 27, 1995. Senate Committee on Local Government, California Senate, Housing Element Law: A Summary Report from the Interim Hearing, Senate Publications, Sacramento, California, December 8, 1993. Shigley, Paul, “Rural Counties Question Housing Policy,” California Planning & Development Report, Vol. 17, No. 2, February 2002, pp. 1, 14–15. Toulan, Nohad, “Housing as a State Planning Goal,” in Carl Abbott, Deborah Howe, and Sy Adler, eds., Planning the Oregon Way (Chapter 5), Oregon State University Press, Corvallis, Oregon, 1994. Warner, Karen, J. Alfred Dichoso, Maureen Markham, Michael McLaughlin, and Candace Stowell, “New Approaches to the Housing Element: Case Studies from California, Nevada, and Washington State,” Conference Proceedings, American Planning Association Conference, San Diego, 1997, available at http://www.asu.edu/caed/ proceedings97/warner.html. Weinstein, Alan C., “Exclusionary Zoning in the Suburbs: The Road from Mount Laurel to Chester,” Zoning and Planning Law Report, Vol. 16, No. 3, 1993, pp. 105–112. 105 Wisckol, Martin, “Affordable Housing Lacking: 12 O.C. Cities Don’t Have Required Plans,” Orange County Register, April 18, 2002. 106 About the Author PAUL G. LEWIS Paul G. Lewis is a research fellow and director of the Governance and Public Finance program at the Public Policy Institute of California. He has authored or co-authored numerous reports, journal articles, and other publications on local government structure, local sales taxes, urban development, transportation policy, regional governance, and local voter turnout. He has also presented his findings on these topics to a variety of governmental bodies and civic organizations. He holds an A.B. in political science from Indiana University and an M.A. and Ph.D. in politics from Princeton 107 Related PPIC Publications Who Pays for Development Fees and Exactions? Marla Dresch and Steven M. Sheffrin Cities Under Pressure: Local Growth Controls and Residential Development Policy Paul G. Lewis and Max Neiman Local Economic Development in Southern California’s Suburbs: 1990– 1997 Max Neiman, Gregory Andranovich, and Kenneth Fernandez Building California’s Future: Current Conditions in Infrastructure Planning, Budgeting, and Financing Michael Neuman and Jan Whittington Homelessness in California John M. Quigley, Steven Raphael, and Eugene Smolensky PPIC publications may be ordered by phone or from our website (800) 232-5343 [mainland U.S.] (415) 291-4400 [Canada, Hawaii, overseas] www.ppic.org 109" ["post_date_gmt"]=> string(19) "2017-05-20 09:36:23" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_203plr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:36:23" ["post_modified_gmt"]=> string(19) "2017-05-20 09:36:23" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_203PLR.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }