Donate
Independent, objective, nonpartisan research

R 304TGR

Authors

R 304TGR

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_304TGR.pdf" ["wpmf_size"]=> string(7) "1110683" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(135667) "Planned Developments in California: Private Communities and Public Life ••• Tracy M. Gordon 2004 Library of Congress Cataloging-in-Publication Data Gordon, Tracy M. Planned developments in California : private communities and public life / Tracy M. Gordon. p. cm. ISBN: 1-58213-081-7 1. Planned communities—California. 2. Housing, Cooperative— California. 3. Condominium associations—California. 4. Homeowners’ associations—California. 5. California—Ethnic relations. 6. Social classes—California. 7. Political participation— California. 8. Voting—California. I. Title. HT167.5.C2G57 2004 307.76'8'09794—dc22 2004001904 Copyright © 2004 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 on any local, state, or federal legislation, nor does it endorse, support, or oppose 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 Over the last three decades, there has been a notable increase in common interest developments (CIDs) in California. Variously characterized as “Fortress America,” “privatopias,” “secession of the successful,” or, more neutrally, “private government,” CIDs in the broadest sense (condominiums, cooperatives, and planned developments) now make up 60 percent of new housing starts in California. However, the image of CIDs commonly projected in critical scholarship and journalism is the gated suburban community closed to public access. These communities provide an array of traditional public services through private, commonly held organizations governed by their residents. The most extreme characterizations suggest that CIDs are nothing less than income and racial segregation reinvented in a more modern, legal form. Less extreme characterizations suggest that such communities might well signal an effective withdrawal of residents from traditional civic society—contributing to declining voting rates and the general movement toward fiscal conservatism at the state and local levels of government. PPIC research fellow Tracy Gordon has taken a careful look at CIDs as an increasingly popular form of privatization. In Planned Developments in California: Private Communities and Public Life, Gordon focuses primarily on planned developments, the prevalent form of CIDs in California. Drawing on a unique dataset of real estate activity in California, she concludes that some “sorting” is apparent when residents of planned developments are compared to other residents in central cities and suburbs. Residents in planned developments have a higher percentage of non-Hispanic whites, higher incomes, and more education and they are generally older than the comparison group. Her findings hold up even when she looks at planned developments for central cities and suburbia separately. The concern over social and economic segregation is thus given some support by the report. She is iii careful to note, however, that the contribution of planned developments to overall residential segregation is minor, and she found little evidence that residents of planned developments are withdrawing from civic life— at least as measured by voting behavior. It is striking that condominiums have also grown in importance to the new homebuyer—representing 37 percent of all CIDs in California compared to 61 percent for planned developments. Although condominiums are often seen as the first rung in the ladder of eventual home ownership, they are also fraught with challenges when disagreements arise between developers and owners and later on when owners cannot agree on key common interest issues. In a review of the governance processes used by CIDs, Gordon notes that the definition of community varies from example to example, but managing a community interest may be just as difficult as reaching a decision in local planning commission meetings. In sum, CIDs are here to stay, and Gordon has made a valuable contribution to our understanding of them in California. Although CIDs are still a relatively small part of the overall housing stock in California, their steep growth path and tendency toward residential selectivity suggest that they are well worth monitoring in the years and decades ahead. De Tocqueville’s observation that America has a penchant for voluntary associations may now be evolving into yet another form—voluntary local government imposed by the decision to purchase a home. This is certainly not a new idea, but this latest example serves as a reminder of the challenges facing California’s governing processes at various levels. David W. Lyon President and CEO Public Policy Institute of California iv Summary Nearly three million California homes, one-quarter of the state housing stock, are located within common interest developments (CIDs) such as planned developments, condominiums, and cooperatives. Moreover, this proportion is ever-increasing, with CIDs accounting for 60 percent of residential construction starts during the 1990s. One type of CID, the planned development, alone represented more than 40 percent of new single-family home sales over this period (see Figure S.1). More than a pattern of residential development, CIDs are an emerging form of privatization. Homeowner associations in CIDs provide many goods and services traditionally supplied by local government, including garbage collection, street cleaning, street lighting, and security patrol. They also levy assessments, adjudicate disputes, and regulate land use as well as other aspects of life within their boundaries. As a result, CIDs are often called “quasi-” or “private governments.” 2,500 Housing units (1000s) 2,000 1,500 1,000 500 0 1960 1970 1975 1980 1985 1990 1995 2000 Planned developments Cooperatives Condominiums SOURCE: HOA-Info (2002). NOTE: Figure does not include CIDs of unknown type. Figure S.1—Growth of Common Interest Developments in California v The growth of so-called private governments has sparked a popular debate. Proponents claim that CIDs aid cash-strapped localities in accommodating the service and infrastructure costs of growth. They further advocate extending this governance structure to existing neighborhoods as a means of improving local amenities and neighborhood quality. Critics charge that CIDs generate adverse consequences for nonresidents. Primary among these feared outcomes is heightened racial and economic segregation and diminished civic engagement. Such criticisms have prompted recurrent calls for enhanced regulation of these communities. Although the bulk of proposed legislation has addressed the internal governance of CIDs, a frequent complaint is that there are no current and reliable studies of common interest developments in California. In 1999, for example, the state legislature authorized the California Law Revision Commission to undertake a multiyear investigation to address, among other questions, “to what extent [common interest housing developments] should be subject to regulation” (California Law Revision Commission, 2001, p. 3). This report seeks to fill the gap in knowledge about CIDs in California. It focuses on planned developments, in particular, because these are the most prevalent and fastest-growing type of CID in both California and the nation as a whole. These developments also bear the strongest resemblance to traditional local government and hence attract the most scrutiny for their potential policy consequences. A Large and Diverse Group There are over 36,000 common interest developments in California. These communities range in size from three to 30,000 units and can accommodate several different housing types and legal ownership structures. The majority of CID units (61 percent) are located in planned developments, in which each homeowner owns his or her individual unit and lot and a homeowner association owns and maintains all common property. The other major type of CID (37 percent of all units) is the condominium, in which each owner holds title to his or her individual unit and a percentage interest in the common areas. vi Monthly assessments in CIDs vary widely, from merely nominal amounts to thousands of dollars per unit per month. In 2002, median monthly assessments were $112 per unit in planned developments and $186 per unit in condominiums and cooperatives. Total homeowner association revenues in California were estimated at $6.3 billion in 2003 (HOA-Info, 2003). Successful But Not Seceding Social commentators have referred to the spread of CIDs as the “secession of the successful” (Reich, 1991). An examination of the demographic and socioeconomic composition of planned developments lends support to this view. Nevertheless, the data also suggest that the picture is more complicated. Planned developments are less diverse with respect to race and ethnicity than other neighborhoods. In both central city and suburban areas, planned developments include significantly more residents who are non-Hispanic white (60 and 65 percent) and fewer who are black (4 percent in central cities and 3 percent in suburban areas) and Hispanic (20 and 19 percent, respectively). Interestingly, suburban planned developments tend to have larger Asian populations than other neighborhoods (Table S.1). Residents of planned developments are on average older than other central city or suburban dwellers. Planned developments include more individuals age 65 and over than other neighborhoods (differences of 2 to 3 percentage points, depending on location). They also include more residents ages 40 to 64 (differences of 3 and 4 percentage points for central cities and suburbs, respectively). Planned developments have correspondingly smaller population shares in the younger age categories and slightly fewer married couples with children (Table S.2). Planned developments are more diverse with respect to income than their image might suggest. To be sure, planned developments house more of the highest-income Californians than other neighborhoods. In central cities and suburbs, 22 and 26 percent of planned development households earned more than $100,000 in 1999, compared to 15 to 17 percent of households in other neighborhoods. However, planned developments also include nearly as many residents in the middle- to vii Table S.1 Selected Demographic Characteristics of California Planned Developments (in percent) Characteristic Black Hispanic Asian White Age 0 to 19 Age 20 to 39 Age 40 to 64 Age 65 and up Central City Planned Non-PD Developments Neighborhoods 4.2 9.8** 20.1 33.5** 11.7 12.3 60.2 40.8** 25.1 28.3** 30.2 33.2** 30.7 27.6** 14.0 10.9** Suburb Planned Non-PD Developments Neighborhoods 2.7 5.1** 18.8 32.6** 10.3 9.2** 64.6 49.5** 27.0 30.3** 26.1 28.7** 33.6 29.6** 13.3 11.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. **Denotes statistically significant difference at the 1 percent level. higher-income categories ($35,000 to $49,999, $50,000 to $74,999, and $75,000 to $99,999) as do other neighborhoods (Table S.2). Despite their relative racial and ethnic homogeneity, the contribution of planned developments to residential segregation is minor. Differences between planned developments and other neighborhoods account for only 2 percent of statewide metropolitan area segregation. This finding is primarily due to the relatively small proportion of the population living within planned developments. Although they represent roughly 40 percent of new single-family homes, planned developments constitute only 12 percent of the existing housing stock. Planned developments do not exhibit markedly different patterns of voting behavior once other relevant characteristics are taken into account. At first glance, planned developments have higher rates of voter registration and turnout than other neighborhoods. They also appear more likely to affiliate with the Republican Party and to vote for statewide ballot propositions favoring the privatization of government services (Propositions 35 and 38) and higher vote requirements for fees and bonds (Propositions 37 and 39). viii Table S.2 Selected Socioeconomic Characteristics of California Planned Developments Central City Suburb Planned Non-PD Planned Non-PD Characteristic Developments Neighborhoods Developments Neighborhoods Median household income in 1999, $ 58,687.8 46,688.3** 67,045.9 53,297.9** Household income, % Less than $35,000 32.9 42.9** 26.8 35.6** $35,000–$49,999 14.2 15.0** 13.7 15.5** $50,000–$74,999 19.1 17.5** 19.8 20.1 $75,000–$99,999 12.4 10.0** 13.5 12.0** $100,000–$199,999 16.6 11.6** 19.5 13.5** $200,000 and up 4.9 3.0** 6.6 3.3** % Married couples with children 30.9 31.9** 34.4 36.4* % with no college 28.0 40.4** 26.5 38.8** % in managerial and professional occupations 42.6 33.8** 42.3 32.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. However, planned developments also differ from other neighborhoods along a number of dimensions that are related to voting behavior (e.g., age, income, education, and residential mobility). Regression analyses holding these differences constant demonstrate that planned developments themselves do not strongly affect either political participation or patterns of voting behavior in statewide general elections. Remaining differences are small, at 1 to 2 percentage points. These conclusions are somewhat unexpected. Thus far, it would appear that concerns about planned developments siphoning off the wealthiest Californians and their participation in public life are unwarranted. Planned developments still represent a relatively small proportion of the population overall, however, and as their numbers increase, their effects on neighborhood composition and civic engagement may become more pronounced. ix Contents Foreword ..................................... Summary..................................... Figures ...................................... Tables ....................................... Acknowledgments ............................... iii v xiii xv xix 1. INTRODUCTION ........................... 2. WHAT ARE COMMON INTEREST DEVELOPMENTS? .......................... A Brief History .............................. Definition and Taxonomy ....................... Scope of Activities ............................ Policy Issues ................................ 3. WHAT DO CALIFORNIA CIDs LOOK LIKE? ........ Note on Data Sources .......................... A Large and Diverse Group ...................... 4. FOCUS ON PLANNED DEVELOPMENTS .......... Geographic Distribution ........................ Demographic and Socioeconomic Characteristics ........ Diversity and Segregation ....................... 5. VOTING BEHAVIOR ........................ Voter Turnout, Registration, and Party Affiliation ........ Voting on Statewide Ballot Propositions .............. A Closer Look at Voting Behavior .................. 6. CONCLUSIONS ............................ Appendix A. Data and Methods ............................ B. Measuring Diversity and Segregation ................ C. Technical Considerations and Regression Results ........ 1 9 9 11 13 14 19 19 20 27 27 28 35 39 41 41 42 49 51 55 61 xi References .................................... About the Authors ............................... Related PPIC Publications .......................... 73 79 81 xii Figures S.1. Growth of Common Interest Developments in California ............................... 1.1. Growth of Common Interest Develoments in the United States .................................. 1.2. Growth of Common Interest Developments and Suburban Housing Units in California and the United States, 1940–1999 ......................... 2.1. Growth of Common Interest Developments in the United States, by Type....................... 3.1. Growth of Common Interest Developments in California, by Type ......................... 3.2. Size Distribution of Common Interest Developments in California, by Type ......................... 3.3. Revenues of Common Interest Developments in California ............................... 5.1. Estimated Effects of Planned Developments on Voter Turnout, Registration, and Party Affiliation ......... 5.2. Estimated Effects of Planned Developments on Votes for Statewide Ballot Propositions .................. A.1. Example of Geocoding Procedure ................ v 2 4 12 22 23 25 45 47 52 xiii Tables S.1. Selected Demographic Characteristics of California Planned Developments ...................... viii S.2. Selected Socioeconomic Characteristics of California Planned Developments ...................... ix 2.1. An Example of Nonterritorial and Territorial CIDs .... 12 2.2. Services Provided by Common Interest Developments in the United States and California ................ 13 3.1. Geographic Distribution of Common Interest Developments ............................ 21 3.2. Types of Common Interest Developments in California ............................... 22 3.3. Average Size of Common Interest Developments, by Year of Construction ........................ 23 3.4. Distribution of Monthly Fees of Common Interest Developments ............................ 24 4.1. Planned Developments, by Region ............... 28 4.2. Planned Developments, by Central City/Suburban Location and Incorporation Status ............... 29 4.3. Planned Developments, by City Size and City Incorporation Date ......................... 29 4.4. Selected Demographic Characteristics of California Planned Developments ...................... 30 4.5. Selected Socioeconomic Characteristics of California Planned Developments ...................... 31 4.6. Selected Housing Characteristics of California Planned Developments ............................ 32 4.7. Percentage Black, by Planned Development Status and Housing Characteristics ...................... 33 4.8. Percentage Hispanic, by Planned Development Status and Housing Characteristics ................... 34 xv 4.9. Average Diversity Scores of California Neighborhoods, by Metropolitan Area, Planned Development, and Central City/Suburban Status .................. 4.10. Decomposition of Statewide Metropolitan Racial Segregation .............................. 5.1. Voter Turnout and Party Affiliation, by Planned Development Status ........................ 5.2. Approval Rates for Statewide Ballot Propositions, by Planned Development Status ................... 5.3. Summary of Demographic and Socioeconomic Differences Between Planned Developments and Other Neighborhoods and Predicted Effects on Voter Turnout ................................ 5.4. Voter Turnout, by Planned Development Status and Racial and Ethnic Diversity.................... 5.5. Voter Registration, by Planned Development Status and Racial and Ethnic Diversity.................... A.1. Selected Characteristics of Planned Developments in California ............................... A.2. Regional Distribution of Planned Developments Units, by Sample Definition ....................... B.1. Decomposition of Segregation, by Planned Development and Central City/Suburban Status ............... C.1. Estimated Effects of Planned Developments on Voter Turnout, 2000............................ C.2. Estimated Effects of Planned Developments on Percentage Registered, 2000 ................... C.3. Estimated Effects of Planned Developments on Percentage Registered Democrat, 2000 ............ C.4. Estimated Effects of Planned Developments on Percentage Registered Republican, 2000 ........... C.5. Estimated Effects of Planned Developments on Percentage Registered Independent, 2000........... C.6. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 32, 2000 ................. C.7. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 35, 2000 ................. 36 37 42 43 44 44 45 53 54 56 62 63 64 65 66 67 68 xvi C.8. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 37, 2000 ................. C.9. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 38, 2000 ................. C.10. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 39, 2000 ................. 69 70 71 xvii Acknowledgments This report benefited from the contributions of several individuals. David Levy of Levy & Company, CPAs, generously provided access to the HOA-Info dataset and shared his extensive knowledge of the CID industry. Julia Lave-Johnston and Dean Misczynski of the California Research Bureau, Peter Detwiler of the Senate Local Government Committee, J. Stacey Sullivan of the Assembly Local Government Committee, and Nathaniel Sterling of the California Law Revision Commision shed light on legal and policy issues relating to CIDs. Professor Robert Jay Dilger of West Virginia University, David Levy, and Paul Lewis, Christopher Weare, and Junfu Zhang of PPIC gave thoughful reviews of the initial manuscript. Joyce Peterson, Peter Richardson, and Patricia Bedrosian improved the clarity of exposition. David Haskel and Rachel Flood provided excellent research assistance. Any remaining errors, as well as opinions or interpretations expressed in this report, are those of the author alone and do not reflect the views of the Public Policy Institute of California. xix 1. Introduction Rancho Bernardo, a master-planned community located about 25 miles north of downtown San Diego, houses approximately 45,000 residents in a mix of single-family detached homes, townhouses, condominiums, and apartments. The community also features service and retail businesses, light industry, shopping areas, parks, and golf courses. Council members elected from districts oversee working committees on public safety, government relations, utilities, traffic, transportation, finance, services, and elections. A planning board enforces land use codes, and a recreation council maintains the community park. Although it bears a striking resemblance to a city, Rancho Bernardo is in fact a private corporation. It is a large, but by no means unusual, example of what has been called “the most significant privatization of local government responsibilities in recent times” (Advisory Commission on Intergovernmental Relations, 1989, p. 18). It also represents a shift in the usual concept of privatization. In contrast to previous models, where a government contracts with a private firm for the provision of a single service or function, individuals today are increasingly contracting directly with so-called “private governments” for many, if not all, traditional municipal services. According to one study, this trend represents nothing less than “a quiet revolution,” with profound implications for community organization, local government, land use, neighbor relations, and other aspects of public life (Barton and Silverman, 1994, p. xi). Rancho Bernardo is an example of a common interest development, or CID.1 Common interest developments encompass several housing ____________ 1Common interest developments are also known as community associations (CAs), common interest communities (CICs), common interest realty associations (CIRAs), and residential community associations (RCAs) (Treese, 1999, p. 3). A related phenomenon is the business or downtown improvement district (BID or DID) (Pack, 1992). 1 types and ownership structures, including planned developments, condominiums, and cooperatives. The distinguishing feature of these communities is common property ownership. Beyond an individual house or unit, property owners in these communities also hold an interest in common areas—such as recreation facilities, streets, lawns, or parking lots—either as individuals or through a mandatory homeowner association. Deed restrictions known as conditions, covenants, and restrictions (CCRs) authorize the association to collect lien-based assessments in exchange for providing services and regulating land use as well as other aspects of life within CID boundaries. Over the past 30 years, common interest developments have proliferated throughout the United States. In 1962, there were fewer than 500 common interest developments nationally; today there are nearly 250,000 CIDs housing an estimated 50 million Americans and constituting 15 percent of the U.S. housing stock (Community Associations Institute, 2003) (Figure 1.1).2 250 16 CIDs as a % of 200 total housing units Total CIDs 150 100 50 0 1962 1970 1975 1980 1985 1990 1995 14 12 10 8 6 4 2 0 1999 SOURCE: Treese (1999). Figure 1.1—Growth of Common Interest Developments in the United States ____________ 2Gated communities constitute a subset of this total. The Community Associations Institute (CAI) estimates that 19 percent of its members, or 8.4 million people, live within gated communities (Blakely and Snyder, 1997, p. 3). Other sources put this figure at four million people, or roughly 10 percent of all CID residents (Egan, 1995, pp. A1, A22). The CAI estimate is likely an upper bound because its members tend to be larger associations. Number of CIDs (1000s) CIDs as a % of total units 2 In California, these trends are pronounced: There are three million homes within common interest developments, encompassing nearly onequarter of the state housing stock. Moreover, CIDs constituted 60 percent of new construction starts during the 1990s (HOA-Info, 2002; U.S. Census Bureau, 1994, 1995, 1996, 1998, 1999, 2002b). Planned developments alone represent over 40 percent of new single-family home sales in the state (Construction Industry Research Board, 2000). To put this trend in perspective, compare the growth of CIDs to the trend of suburbanization more generally. Since 1970, common interest developments have grown faster than the suburbs as a share of all housing units in both California and the United States (Figure 1.2). This growth rate exceeds the pace of suburbanization during the peak years of 1940 to 1960 by a factor of five.3 Although a vast literature has explored the social, economic, and political implications of suburbanization, the consequences of this most recent transformation are largely unknown. The growth of common interest developments has nonetheless sparked a popular debate. Proponents claim that CIDs aid cash-strapped localities in meeting demand for public services.4 They further advocate extending this governance structure to existing neighborhoods as a way to improve local amenities and neighborhood quality (e.g., Nelson, 1999; Ellickson, 1998). Critics charge that common interest developments have adverse consequences for nonresidents. Primary among these feared outcomes is heightened racial and economic segregation. Robert Reich (1991), for example, has referred to the growth of CIDs as the “secession of the successful.” Similarly, Evan McKenzie (1994), in his comprehensive study of the rise of CIDs, links the CCRs governing these communities to the history of exclusionary zoning practices in the United States. Although the Supreme Court ruled in 1948 that racially restrictive covenants were illegal (Shelley v. Kraemer), McKenzie argues “the principle is still the same: certain ____________ 3The average annual growth rates for each type of housing share in each period were 10 and 2 percent, respectively. 4In theory, common interest developments may also improve efficiency in the allocation of public goods by allowing individuals to “vote with their feet” among an expanded array of public and private options for funding and receiving collective services (e.g., Tiebout, 1956). 3 CIDs as a % of total units Suburban housing as a % of total units United States 50 40 Suburb units CID units 30 20 10 0 1940 1960 1975 1990 16 14 12 10 8 6 4 2 0 CIDs as a % of total units Suburban housing as a % of total units 60 50 40 30 20 10 0 1940 1950 California 1960 1970 1980 1990 25 20 15 10 5 0 SOURCE: Treese (1999); HOA-Info (2002); Devaney (1994); U.S. Census Bureau (1970, 1980, and 1990). Figure 1.2—Growth of Common Interest Developments and Suburban Housing Units in California and the United States, 1940–1999 groups of people are considered a threat to property values and are excluded. The result is still increased homogeneity and . . . continuing racial segregation” (p. 78). Other social commentators have expressed concern about the effects of CIDs on political participation and civic engagement. Robert Putnam (2000, p. 210), for instance, links the growth of CIDs to declines in social capital: Not only are canvassing politicians and Girl Scouts selling cookies excluded from exclusive communities, but the affluent residents themselves also appear to have a surprisingly low rate of civic engagement and neighborliness even within their boundaries. One reason to expect lower rates of political participation in CIDs is that the homogeneity of these communities reduces the incentives, or 4 controversies, that draw citizens into civic affairs (Oliver, 2001).5 Alternatively, if planned developments are substitutes for traditional local government, then residents of these developments may perceive less benefit to taking part in elections and other aspects of public life. For instance, in their survey of gated communities in particular, Blakely and Snyder assert that residents of these developments “have less need to participate [in the wider community] . . . because they live in gated enclaves, with private recreation, roads, parks, and security” (1997, p. 72). A related concern is that members of common interest developments may vote differently from nonmembers. In particular, CID residents may favor public spending that complements private expenditures (e.g., police protection) but oppose that which is duplicative (e.g., infrastructure) or redistributive (e.g., welfare). Simply put (“Government by the Nice . . . ,” 1992, p. 25): If affluent Americans choose to live in private communities which raise their own taxes but do not redistribute them outside their walls, they are likely to vote to cut spending on public services that they do not use, ignoring the needs of people who cannot afford to go private. Such criticisms have promoted repeated calls for enhanced regulation of these communities. Over the years, the state legislature has convened several committees and working groups to review policies toward common interest developments. Although the bulk of proposed legislation has focused on the internal governance of CIDs, a frequent complaint is that there are no current and reliable studies of common interest developments in California (e.g., Senate Housing and Community Development Committee, 2002; California Law Revision Commission, 2002a). A California Research Bureau report commissioned by the Senate Housing and Land Use Committee recommended a new legislative subcommittee “to provide more up-todate empirical research on [homeowner associations] and related issues” (Roland, 1998, p. 44). ____________ 5See Alesina and La Ferrara (2000) for an opposing view, i.e., that heterogeneity depresses participation in voluntary associations. 5 The primary reason for the lack of empirical research on CIDs is limited data. The federal government only recently began collecting information on gated communities and residential associations in its 2001 American Housing Survey. As a result, most research on common interest developments has relied on small-scale surveys and case studies (e.g., McKenzie, 1994; Foldvary, 1994; Blakely and Snyder, 1997), with the exception of Barton and Silverman (1994) whose results, based on a 1987 survey, are largely out of date. Given the absence of research in this area, there is a great deal of uncertainty regarding common interest developments among state and local lawmakers, policy analysts, and CID managers and other professionals, boards of directors, and residents themselves. An improved understanding of these communities is an important prerequisite to the formation of sound policies toward private government. This report seeks to fill the gap in knowledge about common interest developments in California. It focuses on planned developments, in particular, because these are the most prevalent and fastest-growing type of CID in both California and the nation as a whole. These developments also represent the closest analogue to traditional local government and hence the greatest source of controversy regarding common interest developments. Among the questions addressed in this report are: • What are common interest developments? What types of housing structures do they include? What kinds of services do they provide? How are they similar to and different from traditional local governments? • What do common interest developments in California look like? How many CIDs are there? How are they distributed regionally? How do they vary in type, size, age, and financial resources? • What about planned developments in particular? Where are they located? How do they compare to other neighborhoods in terms of their demographic and socioeconomic characteristics (e.g., racial and ethnic composition, age structure, income distribution)? Do they contribute to overall residential segregation? 6 • Do planned developments represent the “secession of the successful”? Do these communities exhibit lower levels of civic engagement, as measured by rates of voter turnout and registration? Do they have a different political ideology, as indicated through party affiliation and votes on statewide ballot propositions? The plan for the remainder of the report is as follows. Chapter 2 compares the origins, institutional features, and powers and responsibilities of common interest developments to traditional local government. Chapter 3 presents a portrait of common interest developments in California, highlighting the diversity among these communities. Chapter 4 analyzes the demographic and socioeconomic composition of planned developments, in particular, and their contribution to overall residential segregation. Chapter 5 compares voter turnout and patterns of voting behavior in planned developments to similar neighborhoods, using regression analysis to control for relevant demographic and socioeconomic factors. Chapter 6 concludes and offers directions for state and local public policy. 7 2. What Are Common Interest Developments? This chapter presents an overview of common interest developments in the United States. It traces the evolution of these communities from Ebenezer Howard’s “garden cities of tomorrow” to the modern CID. The chapter also develops a definition and taxonomy of common interest developments. It concludes with a discussion of the scope of activities of these so-called “private governments” and issues for policymakers. A Brief History Alexis de Tocqueville observed that the principle of voluntary association had nowhere in the world “been more successfully used, or unsparingly applied to a multitude of different objects, than in America.” He further described these associations as “a separate nation in the midst of the nation, a government within the government” (1969 [1835], pp. 513–517). Yet the idea of residential associations based on collective property ownership originated in England. During the mid-1700s, Lord Leicester established a park in Leicester Square and charged adjacent property owners an assessment for their exclusive enjoyment of it. In 1837, private developers built Victoria Park, a community with deed restrictions tied to the sale of each lot to protect local amenities (Foldvary, 2002, p. 274). Perhaps the most important antecedent to the modern common interest development was Ebenezer Howard’s utopian ideal of the garden city. Howard sought to combine the best aspects of city and country life in his “new towns” of Welwyn and Letchworth, England. Both cities were based on a leasehold concept, whereby a group of trustees would own the municipality and collect rents from its residents to pay off initial construction loans, build infrastructure, and provide collective services. Howard characterized the political organization as a technocracy, 9 “‘modelled upon that of a large and well-appointed business.’” In his view, this governance structure would be more powerful than other local bodies and “‘solve to a large extent the problem of local selfgovernment’” (McKenzie, 1994, pp. 5–6). Precursors to the CID in the United States—Gramercy Park in New York City (established in 1831), Louisburg Square in Boston (1844), and South Park in San Francisco (1852)—were formed primarily to maintain private amenities. Builder Jesse Clyde Nichols established the first fullservice common interest development, Mission Hills, Missouri, in 1914 with the express intention of supplanting local government. According to one study, “he was afraid if the subdivision were part of a larger village or town organization the political unit would not be sufficiently responsive to the needs of his residents” (Worley, 1990, pp. 166–167). During the 1920s, the reform-minded Regional Planning Association of America revived the idea of a “garden city” with the new town of Radburn, New Jersey. In addition to its innovative physical plan, the community included a homeowner association with broad land use regulation and service responsibilities (Barton and Silverman, 1994, p. 9). The popularity of common interest developments burgeoned during the post–World War II housing boom and rise of “community builders,” who favored large-scale, dense forms of development. In 1961 and 1963, the Federal Housing Administration encouraged this form of housing by publishing guidelines for condominiums and planned developments to obtain mortgage insurance. The Federal National Mortgage Association and Federal Home Loan Mortgage Corporation later approved condominium and planned development purchases for secondary mortgage loan markets. State and local governments also codified real estate practices and modified building and zoning laws in the 1960s to accommodate CID housing.1 This period of growth culminated in the emergence of large-scale master-planned communities such as Irvine, California, and Reston, Virginia. ____________ 1For instance, many localities implemented planned unit development (PUD) zoning, allowing more dense or clustered housing than would otherwise be permitted in exchange for developer-provided amenities, such as parks and recreation. It is important to note, however, that not all planned developments are zoned PUD and vice versa. 10 Following a few highly visible insolvencies, the late 1960s and early 1970s brought a series of reforms to common interest developments. The Urban Land Institute and National Association of Home Builders jointly established the Community Association Institute to provide technical assistance to homeowner associations on financial management and governance. At the same time, several states adopted model condominium and planned development codes establishing financial reporting requirements for these communities. Faced with the fiscal strains of the 1970s and 1980s, local governments further promoted CID housing as a means of offsetting the service and infrastructure costs of new development. In some instances, jurisdictions even required that developers establish homeowner associations as a condition of new project approval (Barton and Silverman, 1994, p. 11). Definition and Taxonomy There are three main kinds of common interest developments: planned developments, condominiums, and cooperatives. Each type involves a different legal assignment of property rights. In a planned development, the homeowner owns his or her individual unit and lot, and the homeowner association, usually incorporated as a nonprofit mutual benefit corporation, owns and maintains all common property. In a condominium, the association manages but does not own the common elements; each unit owner holds an undivided interest in these areas as a “tenant in common.” In a cooperative, a corporation owns and maintains the entire property. Individuals do not own units but hold proprietary lease agreements or membership documents granting exclusive use of a particular unit for a specified period of time. Figure 2.1 shows that nearly 64 percent of common interest development units in the United States are within planned developments, whereas 31 percent are in condominiums and 5 percent are in cooperatives. Planned developments are also the fastest-growing type of CID, with the number of planned development units increasing fourfold since 1985. 11 Number of units (1000s) 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 1962 1970 1975 1980 1985 1990 1995 1999 Planned developments Cooperatives Condominiums SOURCE: Treese (1999). Figure 2.1—Growth of Common Interest Developments in the United States, by Type These legal definitions can mask substantial variation. Groups of detached single-family homes can be legally organized as condominiums and, more rarely, apartment complexes can be planned developments (Barton and Silverman, 1994, p. 5). Table 2.1 illustrates a more useful distinction between “territorial” CIDs encompassing multiple buildings on large tracts of land and “nonterritorial” CIDs Table 2.1 An Example of Nonterritorial and Territorial CIDs Number of buildings Number of units Amenities Nonterritorial 1 6 None Services Common property Governance Management City garbage dumpster, gardener Halls, outside, 1 acre lot Each unit has one member on homeowner asosciation board; majority rule Board SOURCE: Roland (1998, p. 9). Territorial 105 100 Pool, tennis, private parks, health club Garbage, security, street maintenance All except home lots (75 acres) Eight board members elected from 100 wth full authority unless law requires vote Off-site management company 12 within single high-rise buildings (Advisory Commission on Intergovernment Relations, 1989). The former category constitutes 60 to 80 percent of all common interest developments (Barton and Silverman, 1994; Treese, 1999). Scope of Activities Territorial common interest developments resemble traditional local governments in that they are responsible for providing services in exchange for regular assessments from residents. Many of these services are similar to those that would be provided by a municipality (see Table 2.2). Homeowner associations are also responsible for enforcing the “constitutions,” or CCRs of the community. CCRs typically regulate property uses such as exterior renovations, paint colors, fences, and antennas. They may also restrict individual conduct or “lifestyle,” Table 2.2 Services Provided by Common Interest Developments in the United States and California California Service Open space or lawns Lawn care or gardening Parks or playgrounds Meeting places Garbage disposal Parking lots or structures Swimming facililites Roads Water or sewer lines Other recreational facilities Entry guards or security patrols United States % of CIDs Service % of CIDs 92 Snow removal 77 87 Pools 76 84 Garbage disposal 64 72 Clubhouse 56 65 Street cleaning 54 62 Street lighting 53 58 Tennis 45 43 Playgrounds 37 42 Lake or pond 30 21 Park area 29 15 Security patrols 12 SOURCES: Barton and Silverman (1994, p. 5); Treese (1999, pp. 7, 13). NOTES: Table reports services commonly provided by respondents in two separate survyes. National survey responses are selected from developments of 151 to 350 units. 13 including pet ownership and noise levels.2 Crafted by the developer, who retains control of the association until 50 to 75 percent of units are sold, these restrictions “run with the land,” meaning that they are binding on all subsequent property owners. Amendments to CCRs usually require approval from a supermajority of homeowners. Common interest developments differ from public entities most notably in that they lack the police power of local governments. They must rely on the courts to enforce CCRs or collect delinquent fees.3 Typically, associations are also not subject to the same constitutional and statutory requirements as public entities regarding elections, public meetings, and public access.4 Moreover, CID boards of directors usually do not share public officials’ immunity from personal liability. In California, however, amendments to the 1985 Davis-Stirling Act institute open meeting requirements for CIDs parallel to those of the Brown Act for state and local governments. They also ensure that board members cannot be held personally liable if they meet certain standards of care and maintain minimal liability insurance (Roland, 1998, pp. 22–24). Policy Issues The dual public and private nature of common interest developments raises a host of policy questions. At the state level, there ____________ 2Legal scholars have noted the seemingly arbitrary and capricious nature of these rules. For example: • In Boca Raton, Florida, a homeowners association resident was required to attend a court-supervised weigh-in of his 29-1/2-pound dog because the association’s CCRs specified a 30-pound limit for pets. • In Santa Ana, California, a 51-year-old grandmother received a warning citation from her condominium association for kissing a friend good night in her driveway. • Bylaw VI 7(k) of the Bailey, a condominium in Washington, D.C., provides that “No Barry Manilow records, tapes or CDs may be owned or played on the premises.” See Kennedy (1995) and Kress (1995). 3CCRs are “equitable servitudes,” or land use restrictions enforceable in a court of equity. See Black’s Law Dictionary (1979, p. 484). 4For instance, voting rights in CIDs are limited to homeowners only and may be apportioned on the basis of unit size or value rather than equal representation (see Barzel and Sass, 1990). 14 have been recurrent calls for the enhanced regulation of these communities. The bulk of proposed legislation concerns the internal governance of CIDs. Among the major bills passed during the 2003– 2004 legislative session were: • AB 104 (Lowenthal)—requires that homeowner associations make accounting books and records as well as minutes of proceedings available to all members for inspection and copying; • AB 1423 (Dutra)—amends and clarifies a bill passed in the prior session (AB 555) instituting continuing education and examination requirements for property managers to be called “certified common interest development managers”; and • AB 1525 (Longville)—prohibits homeowner associations from banning the posting or display of noncommercial signs or banners except for the protection of public health or safety or if the posting violates a local, state, or federal law. Over the years, the legislature has also appointed several committees and working groups to review state policies toward common interest developments. In 1983, the assembly convened a select committee to streamline previous codes. Their efforts culminated in the 1985 DavisStirling Common Interest Development Act (California Civil Code §§ 1350 et seq.), the main body of law governing CIDs in California. Since its adoption, the act has been amended more than 40 times. As a result, it is a cumbersome piece of legislation that has been criticized as complicated and hard to understand, contradictory, uneven in its coverage, difficult to enforce and, offering weak protection for individual rights (French, 2000). In 1999, the legislature authorized the California Law Revision Commission to undertake a multiyear study of common interest law, including the Davis-Stirling Act and other key statutes. The objectives of the study are: • to set a clear, consistent, and unified policy with regard to the formation and management of CIDs and the transaction of real property interests located within them; 15 • to clarify the law, eliminate unnecessary or obsolete provisions, consolidate existing statutes in one place in the codes; and, • to determine to what extent common interest housing developments should be subject to regulation (California Law Revision Commission, 2002b, p. 3). Thus far, the commission has recommended instituting requirements for procedural fairness in homeowner association rulemaking and decisionmaking, as well as a nonsubstantive reorganization of the DavisStirling Act. Assemblywoman Patricia Bates introduced implementing legislation for these changes (AB 512) in the 2003–2004 session. The commission has also recommended improving existing alternative dispute resolution requirements and requiring that common interest development associations provide internal dispute resolution mechanisms. Several emerging policy issues also have yet to be addressed by the state legislature. A major issue for CID residents is “double taxation,” or their responsibility for local property taxes even if they are consuming private alternatives to public services. Some state and local governments have agreed to property tax rebates for CID residents subject to certain public access requirements. For example, New Jersey’s Municipal Services Act requires that local governments reimburse CID homeowner associations for street sweeping, garbage collection, snow plowing, and street lights provided that the CID accepts public dedication of its roads or streets (New Jersey Statutes, § 40:67-23-3). Similar arrangements exist in Houston, Texas; Kansas City, Missouri; and Montgomery County, Maryland (Dilger, 1992).5 A related concern is that local real property taxes are deductible under the federal Internal Revenue Code but homeowner assessments are not. According to Executive Council of Home Owners (ECHO) lobbyist Robyn Boyer Stewart, “It is only a matter of time before the tax-and-equity bomb blows. . . . The politician ____________ 5Where a CID is located exclusively on private property, local governments may not provide routine services such as police patrols, trash collection, and animal control because of litigation concerns unless they have negotiated prior agreements with the CID (Advisory Commission on Intergovernmental Relations, 1989). 16 who manages to capture this constituency, speak to its needs, and offer it a voice will be amply rewarded” (Kennedy, 1995, p. 775). Nonresidents, on the other hand, are concerned about the social consequences of common interest developments. Critics of CIDs argue that these communities “compete with cities for the affluent, siphoning off their tax dollars, their expertise and participation, and their sense of identification with the community” (McKenzie, 1994, p. 23). The chapters to follow provide an empirical basis for this debate by analyzing the demographic, socioeconomic, and political features of common interest developments in California. 17 3. What Do California CIDs Look Like? This chapter surveys common interest developments in California. It highlights the diversity among CIDs, presenting breakdowns of the main attributes of these developments by type and over time. It concludes with an examination of the financial resources of common interest developments and their relative position among local governments in California. Note on Data Sources The chapters above noted the lack of previous empirical studies on CIDs because of data limitations. For example, the federal government only recently began collecting information on community associations in the 2001 American Housing Survey. Various states, however, do collect this information as part of their regulatory oversight.1 In California, developers of subdivided land consisting of five or more lots or units must file a public report with the Department of Real Estate (DRE) before offering units for sale (California’s Subdivided Lands Act, California Business and Professions Code §§ 11000-1200). The report must include the location and size of the development as well as the governing documents and proposed budget of the homeowner association (Roland, 1998, p. 17). Each addition or material change to the development requires an amendment to the initial DRE report. Although not required by law, most homeowner associations also incorporate as nonprofit mutual benefit organizations.2 The Davis____________ 1Florida, Virginia, Maryland, and Nevada have reporting requirements similar to those in California. 2Unincorporated homeowner associations do not receive the tax benefits or liability protections of incorporated CIDs, although they are required to follow most of the same laws (Lave-Johnston and Johnston-Dodds, 2002). 19 Stirling Act requires that incorporated associations file articles of incorporation with the Secretary of State (SOS), along with a statement identifying the corporation as a CID, the location of the association's business or corporate office, and the name and address of the association's managing agent. A separate statement of “principal business activity,” listing three officers, their addresses, and an agent's name and address for service of process, must be filed within 90 days of incorporation and every two years thereafter. Information from these two sources is available through HOA-Info, a proprietary database compiled by Levy and Company, CPAs of Oakland, California. These data were supplemented with other sources, such as homeowner association websites and directories as well as property tax assessment, deed transfer, and mortgage records. The result is the most comprehensive existing database on common interest developments in California.3 A Large and Diverse Group There are currently over 36,000 common interest developments— including planned developments, condominiums, and cooperatives—in California. These communities include over three million CID housing units, equivalent to approximately one-quarter of the state’s housing stock (Table 3.1). The prevalence of common interest developments varies substantially by region. The South Coast has the most developments and CID units in the state, whereas the Sacramento metropolitan area and San Diego region exhibit the highest concentration of common interest developments as a share of all housing units (42 percent). Roughly one-third of all CIDs in California are PDs and two-thirds are condominiums and cooperatives (Table 3.2). These proportions are nearly reversed when one considers units rather than developments. About 61 percent of all California CID units are in planned developments, and 39 percent are in condominiums or cooperatives. ____________ 3There are remaining gaps in the data. Specifically, the number of units is missing for 22 percent of all records, acreage is missing for 31 percent, and fee and revenues are missing for 61 percent. 20 Table 3.1 Geographic Distribution of Common Interest Developments Region Sacramento Metro San Diego Inland Empire Sierras Far North South Coast Bay Area Central Coast San Joaquin Valley Total No. of CIDs 1,325 4,772 2,250 312 663 16,311 7,769 1,459 1,321 36,182 CID Units 301,329 450,256 450,584 31,308 106,504 956,613 491,523 77,731 144,822 3,010,670 % of Total Housing Units 42 42 38 32 22 21 19 16 13 25 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: The counties included in each region are as follows: Far North: Butte, Colusa, Del Norte, Glenn, Humboldt, Lake, Lassen, Mendocino, Modoc, Nevada, Plumas, Shasta, Sierra, Siskiyou, Sutter, Tehama, Trinity, and Yuba. Sacramento Metro: El Dorado, Placer, Sacramento, and Yolo. Sierras: Alpine, Amador, Calaveras, Inyo, Mariposa, Mono, and Tuolumne. Bay Area: Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma. San Joaquin Valley: Fresno, Kern, Kings, Madera, Merced, San Joaquin, Stanislaus, and Tulare. Central Coast: Monterey, San Benito, San Luis Obispo, Santa Barbara, and Santa Cruz. Inland Empire: Riverside and San Bernardino. South Coast: Los Angeles, Orange, and Ventura. San Diego: Imperial and San Diego. Common interest developments have grown exponentially in California since the early 1960s. As in the rest of the nation, planned developments are the fastest-growing type of CID, increasing by 84 percent since 1985 (Figure 3.1). Although planned developments are on average larger than condominiums and cooperatives, there is considerable overlap in the size distributions of these two types of CID (Figure 3.2). Planned 21 Table 3.2 Types of Common Interest Developments in California Type Planned developments % of total Condominiums and cooperatives % of total Total Developments 9,313 33 19,310 67 28,623 Housing Units 1,495,913 61 938,857 39 2,434,770 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Figures do not include developments of unknown or “other” types of associations in business parks, mobile homes, lofts, timeshares, roads, airports, and docks. 2,500 Number of units (1000s) 2,000 1,500 1,000 500 0 1962 1970 1975 1980 1985 1990 1995 1999 Planned developments Cooperatives Condominiums SOURCE: HOA-Info (2002). NOTE: Figure does not include CIDs of unknown type. Figure 3.1—Growth of Common Interest Developments in California, by Type developments can include as few as three units and condominiums can encompass as many as 27,000 units at build-out. The average planned development includes 168 units and the average condominium or cooperative has 52 units. There is a trend toward smaller common interest developments over time, as evidenced in Table 3.3. However, recent years have witnessed a slight up-tick in the average size of CIDs, as measured by both the number of units and total acreage. 22 Number of developments 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0 1–5 6–10 11– 16– 21– 26– 51– 76– 101– 151– 201– 301– 501– 1,001+ 15 20 25 50 75 100 150 200 300 500 1,000 Number of units Planned developments Condominiums and cooperatives Source: HOA-Info (2002). Figure 3.2—Size Distribution of Common Interest Developments in California, by Type Table 3.3 Average Size of Common Interest Developments, by Year of Construction Year Built Pre-1940 1940–1949 1950–1959 1960–1969 1970–1979 1980–1989 1990–1995 1996–2002 Planned Developments Units Acres 350 135 190 14 155 40 111 28 76 14 47 11 39 10 60 15 Condominiums and Cooperatives Units Acres 11 1 11 1 24 1 36 2 24 1 15 1 14 1 18 1 SOURCE: HOA-Info (2002). NOTE: Cells include the median number of units or acres for each type of CID. 23 The dataset used in this study does not include information on the services provided by each common interest development. One proxy measure of service activity is revenue collected by the homeowner association through monthly assessments per unit. (This measure does not include funds raised through special assessments for nonroutine expenses.)4 As with other CID characteristics, monthly assessments exhibit wide variation, ranging from merely nominal amounts to thousands of dollars per unit per month. Three-quarters of all reporting planned developments charged at least $68 per unit per month in 2002, with a median monthly assessment of $112. Similarly, three-quarters of all condominiums and cooperatives charged monthly fees of at least $138, for a median of $186 per unit per month (Table 3.4).5 Monthly fees translate into varying annual revenues for homeowner associations (Figure 3.3). Total association revenues were estimated at $6.3 billion in 2003 (HOA-Info, 2003). Thus, homeowner associations are a major presence in California. Subsequent chapters will focus on planned developments—the CIDs that most resemble local governments—and their consequences for neighborhood diversity and segregation as well as civic engagement. Table 3.4 Distribution of Monthly Fees of Common Interest Developments Planned developments Condominiums or cooperatives 5 $25 $81 SOURCE: HOA-Info (2002). 25 $68 $139 Percentile 50 $112 $186 75 $168 $249 95 $313 $443 ____________ 4The HOA-Info database includes assessments reported by the developer to the California Department of Real Estate before offering units for sale. These figures were supplemented with current financial data, where available, or inflated to current dollars assuming a 4 percent annual inflation rate. 5Anecdotal evidence suggests that condominium fees are higher because they generally offer more common areas and amenities. 24 3,000 2,500 Number of developments 2,000 1,500 1,000 500 0 0–25 26–50 51–75 76–100 101– 201– 200 300 Thousands of dollars 301– 400 401– 1,000+ 500 Planned developments Condominiums and cooperatives SOURCE: HOA-Info (2002). Figure 3.3—Revenues of Common Interest Developments in California 25 4. Focus on Planned Developments The remainder of this report focuses on planned developments. This is the most prevalent and fastest-growing type of common interest development housing in both California and the nation. These CIDs also bear the strongest resemblance to traditional local government and attract the most scrutiny for their potential policy consequences. This chapter presents a portrait of planned developments in California.1 It first surveys the regional distribution of these communities and the types of places in which they are located. The chapter then compares the demographic and socioeconomic characteristics of planned developments to characteristics of other neighborhoods. It concludes with a detailed examination of the racial and ethnic composition of PDs and their contribution to neighborhood diversity and metropolitan area segregation. Geographic Distribution Planned developments are primarily concentrated in urbanized regions of the state, including the South Coast, Bay Area, and San Diego region (Table 4.1). They represent particularly large shares of the housing stock in fast-growing regions such as the Inland Empire and Sacramento metropolitan area as well as in certain less urbanized areas, such as the Sierras and Far North. ____________ 1The analysis presented in this chapter and the remainder of this report relies on the results of a geographic matching procedure described in Appendix A. Comparisons between planned developments and other neighborhoods are based on census block groups with and without PDs. Census block groups are the smallest level of geography at which socioeconomic variables from the U.S. Census are publicly available. There were 22,133 block groups in California in 2000, of which about 16 percent had at least one planned development in 2002. 27 Table 4.1 Planned Developments, by Region Region Bay Area Central Coast Far North Inland Empire Sacramento Metro San Diego San Joaquin Valley Sierras South Coast Total No. of Planned Developments 2,475 722 316 907 646 960 648 96 2,530 9,300 % Single- % Population Population % of All Family Growth, per Square Housing Housing 1990–2000 Mile, 2000 10 17 13 980 9 13 12 121 15 22 11 26 17 24 26 119 15 21 21 353 17 29 13 353 10 14 20 121 26 34 16 9 10 18 10 1,959 12 19 14 217 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). Approximately two-thirds of planned developments are in suburbs, or in metropolitan areas but beyond the central city (Table 4.2). Contrary to the notion that homeowner associations in PDs are substitutes for traditional local government, over three-quarters of PDs are within incorporated cities. Most of these cities were incorporated before 1950 and have populations of less than 100,000 (Table 4.3). Demographic and Socioeconomic Characteristics Planned developments differ from other neighborhoods along a number of dimensions.2 Table 4.4 indicates that, on average, planned developments have significantly higher percentages of the population who are white, in both central city and suburban areas (60 and 65 percent, respectively).3 Correspondingly, they have lower proportions who are black (4 percent in central cities and 3 percent in suburban ____________ 2Comparisons between planned developments and other neighborhoods refer to census block groups with and without PDs. See Appendix A for details on the procedure used to match PD locations to census geography. 3All reported demographic and socioeconomic information is from the 2000 Census unless otherwise noted. Racial and ethnic groups include only individuals who identify themselves as belonging to one race. The terms “white,” “black,” and “Asian” in the text refer to non-Hispanic members of these groups. 28 Table 4.2 Planned Developments, by Central City/Suburban Location and City Incorporation Status Incorporated city Unincorporated area Total Central City 1,648 — 1,648 Suburb 2,550 1,088 3,638 Nonmetropolitan Area 65 161 226 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). Total 4,263 1,249 5,512 Table 4.3 Planned Develoments, by City Size and City Incorporation Date City Population Less than 1,500 1,500–1,999 2,000–2,499 2,500–4,999 5,000–9,999 10,000–19,999 20,000–24,999 25,000–49,999 50,000–99,999 100,000–249,999 250,000–499,999 500,000–999,999 1,000,000–2,499,999 2,500,000–4,999,999 Total Before 1950 2 4 6 69 110 45 355 823 704 257 170 152 131 2,828 City Incorporation Date 1950– 1960– 1970– 1980– 1990– 1959 1969 1979 1989 2000 Total 2 13 4 4 7 11 2 26 8 19 2 18 2 118 23 28 34 16 18 229 6 20 10 31 4 116 173 63 119 124 21 855 169 65 12 83 43 1,195 151 58 55 20 988 257 170 152 131 538 267 232 292 90 4,247 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). areas) and Hispanic (20 and 19 percent, respectively). Differences in Asian population shares are not significant between planned developments and other neighborhoods in central cities. They are significant, but in the opposite direction for the suburbs. That is, planned developments tend to have larger Asian populations in these areas. 29 Table 4.4 Selected Demographic Characteristics of California Planned Developments (in percent) Characteristic Black Hispanic Asian White Ages 0 to 19 Ages 20 to 39 Ages 40 to 64 Age 65 and up Central City Planned Non-PD Developments Neighborhoods 4.2 9.8** 20.1 33.5** 11.7 12.3 60.2 40.8** 25.1 28.3** 30.2 33.2** 30.7 27.6** 14.0 10.9** Suburb Planned Non-PD Developments neighborhoods 2.7 5.1** 18.8 32.6** 10.3 9.2** 64.6 49.5** 27.0 30.3** 26.1 28.7** 33.6 29.6** 13.3 11.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. **Denotes statistically significant difference at the 1 percent level. Planned developments also exhibit significant differences in age structure compared to other neighborhoods. Although a few well-known developments cater primarily to seniors (e.g., Leisure World, incorporated as Laguna Woods in 1999), planned developments as a group are more heterogeneous with respect to age. On average, planned developments include more individuals age 65 and older than other neighborhoods (differences of 2 to 3 percentage points, depending on location). However, they also include more residents ages 40 to 64 (differences of 3 and 4 percentage points for central cities and suburbs, respectively). They have correspondingly smaller population shares of younger persons and also slightly fewer married couples with children than other neighborhoods (Table 4.5). There are major socioeconomic differences between planned developments and other neighborhoods as well (Table 4.5). Planned developments had significantly higher median annual household incomes in 1999, by about $12,000 in central cities and $14,000 in suburban areas. They also tend to have more households in the highest income categories and fewer households in the lowest categories. For example, 30 Table 4.5 Selected Socioeconomic Characteristics of California Planned Developments Central City Suburb Planned Non-PD Planned Non-PD Characteristic Developments Neighborhoods Developments Neighborhoods Median household income in 1999, $ 58,687.8 46,688.3** 67,045.9 53,297.9** Household income, % Less than $35,000 32.9 42.9** 26.8 35.6** $35,000–$49,999 14.2 15.0** 13.7 15.5** $50,000–$74,999 19.1 17.5** 19.8 20.1 $75,000–$99,999 12.4 10.0** 13.5 12.0** $100,000–$199,999 16.6 11.6** 19.5 13.5** $200,000 and up 4.9 3.0** 6.6 3.3** % Married couples with children 30.9 31.9** 34.4 36.4* % with no college 28.0 40.4** 26.5 38.8** % in managerial and professional occupations 42.6 33.8** 42.3 32.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. nearly 22 percent of PD households in central cities earned more than $100,000 in 1999, and 26 percent in suburban areas earned at least this amount (compared to 15 and 17 percent for non-PD neighborhoods). Individuals residing in planned developments are also more likely to have attended college (differences of roughly 12 percentage points in both central city and suburban areas) and to work in managerial and professional occupations (differences of 9 to 10 percentage points in each area). In addition, there are several distinguishing features of planned development housing itself. Homes in these communities are generally newer and of a higher median value ($280,000 in central cities and $324,000 in suburban areas). They also are more likely to be singlefamily structures and owner-occupied (Table 4.6). 31 Table 4.6 Selected Housing Characteristics of California Planned Developments Characteristic Housing density (units per square mile) % single-family units % owner-occupied units Median year built Median home value Central City Suburb Planned Non-PD Planned Non-PD Developments Neighborhoods Developments Neighborhoods 2,853.9 64.8 59.8 1,968.3 280,189.8 4,806.2** 60.2** 49.5** 1,947.8** 233,449.4** 1,839.3 75.1 69.7 1,971.9 323,518.1 2,482.0** 73.7* 63.3** 1,960.2** 231,384.0** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. Many observed demographic differences between planned developments and other neighborhoods persist even after controlling for housing characteristics. Notably, planned developments have fewer blacks and Hispanics, holding constant the median home value and the decade when the majority of units were built. Differences between planned developments and other neighborhoods narrow at higher housing market values, however (Tables 4.7 and 4.8). In sum, planned developments conform in many ways to the popular image of “the secession of the successful.” Residents of these communities are on average white, older, and better educated, and earn higher incomes. Nevertheless, PDs are also more heterogeneous than their image as exclusive luxury enclaves would suggest. Although they house more of the highest-income Californians, planned developments also include many residents in the middle- to higher-income categories (those with annual earnings of $35,000 to $49,999, $50,000 to $74,999, and $75,000 to $99,999). In addition to the age 65 and older category, there are sizable population shares in the younger age groups. Differences in the racial composition of planned developments also diminish somewhat after controlling for housing characteristics such as year of construction and median value. 32 Table 4.7 Percentage Black, by Planned Development Status and Housing Characteristics 33 Year of Construction Before 1960 1960–1969 1970–1979 1980–1989 1990–1999 Low Non-PD Planned Neighborhoods Developments 11.1 6.2** 7.5 3.4** 6.1 4.2** 8.1 4.9** 7.5 5.8 Median Home Value Medium Non-PD Neighborhoods Planned Developments 8.8 3.7** 5.8 3.5 4.7 3.0** 5.5 4.2* 5.4 4.0* High Non-PD Planned Neighborhoods Developments 4.1 2.3** 2.2 1.6 2.6 1.7** 2.7 2.0 3.2 2.3* SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTES: “Low” home values are the first through 32nd percentile for the state ($149,100) in the 2000 Census; “medium” home values are the 33rd through 65th percentile for the state ($149,200 to $244,700); and “high” home values are the 66th through 100th percentile for the state ($244,800 and above). *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. Table 4.8 Percentage Hispanic, by Planned Development Status and Housing Characteristics 34 Year of Construction Before 1960 1960–1969 1970–1979 1980–1989 1990–1999 Low Non-PD Planned Neighborhoods Developments 49.5 34.5** 38.2 26.3** 31.4 23.1** 36.9 29.7** 37.7 29.0** Median Home Value Medium Non-PD Neighborhoods Planned Developments 44.1 33.0** 32.0 24.0** 26.0 20.7** 23.9 16.7** 22.4 16.1** High Non-PD Planned Neighborhoods Developments 15.8 13.6** 14.5 11.7* 14.3 11.3** 11.9 10.5 15.2 12.0** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTES: “Low” home values are the first through 32nd percentile for the state ($149,100) in the 2000 Census; “medium” home values are the 33rd through 65th percentile for the state ($149,200 to $244,700); and “high” home values are the 66th through 100th percentile for the state ($244,800 and above). *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. Diversity and Segregation This section traces the consequences of the different racial and ethnic composition of planned developments for neighborhood diversity and overall metropolitan area segregation. It relies on the entropy index of diversity, which measures how evenly the population of any given area is distributed across the major racial and ethnic groups.4 The index reaches a maximum of 100 when all groups are equally represented and a minimum of zero when only one group is present (see Appendix B for a fuller discussion of the index). Planned developments are often less racially and ethnically diverse than other neighborhoods, even after controlling for whether the neighborhood is in a suburb or a central city (Table 4.9). At the most extreme, in the Sacramento area, suburban planned developments have an average diversity score of 43, compared to 55 for other neighborhoods. On the other hand, in some metropolitan areas, planned developments are more diverse than or about the same as other neighborhoods (e.g., Los Angeles–Riverside–Orange County). A general definition of residential segregation is the extent to which individuals belonging to different groups live separately from one another (Massey and Denton, 1988). The entropy index of segregation relates the diversity of an area as a whole to the average diversity of its constituent parts (see Appendix B). It describes, for example, how different the racial and ethnic composition of a city or metropolitan area is from its neighborhoods or, equivalently, how neighborhoods within a given city differ from each other. The index ranges from a minimum of 0, when all neighborhoods have the same racial and ethnic composition and hence the same diversity, to a maximum of 100, under complete segregation or where the diversity of each neighborhood is 0. An attractive property of the entropy index is that it can be decomposed into “between” and “within” components (see Appendix B). The “within” component is the portion of segregation attributable to ____________ 4This report considers the four largest racial and ethnic groups in California— Hispanic and non-Hispanic white, black, Asian and Pacific Islander—as well as an “other” category encompassing individuals who do not fit any of the above categories, those of more than one race, and American Indians. 35 Table 4.9 Average Diversity Scores of California Neighborhoods, by Metropolitan Area, Planned Development, and Central City/Suburban Status Central City Suburb Planned Non-PD Planned Non-PD Metropolitan Area Developments Neighborhoods Developments Neighborhoods Bakersfield 61 61 43 47 Chico–Paradise 36 39 32 43 Fresno 63 65 45 46 Los Angeles–Riverside– Orange County 57 56 55 54 Merced 63 68 56 57 Modesto 58 60 51 53 Redding 33 35 25 30 Sacramento–Yolo 70 74 43 55 Salinas 50 53 53 47 San Diego 56 59 50 57 San Francisco–Oakland– San Jose 62 65 57 60 San Luis Obispo– Atascadero–Paso Robles 41 46 39 41 Santa Barbara–Santa Maria–Lompoc 51 50 41 51 Stockton-Lodi 68 71 59 59 Visalia–Tulare–Porterville 52 56 53 41 Yuba City 67 59 61 59 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Each table entry is the population-weighted average diversity of census block groups in that category. racial and ethnic differences among neighborhoods of a given type, for example, central cities or suburbs. The “between” component is the remaining portion of segregation, which cannot be reduced without individuals relocating. By extension, central city and suburban segregation can be decomposed into differences among planned developments, as well as between planned developments and other neighborhoods. The latter component is the portion of overall segregation that would be eliminated if individuals moved out of planned developments and into other suburban or central city neighborhoods. Results from this decomposition reveal that the contribution of planned developments to overall metropolitan area segregation in 36 California is actually quite small (Table 4.10). In suburbs, differences in the racial and ethnic composition between planned developments and other neighborhoods explain only 2 percent of overall segregation. In central cities, the share attributable to planned developments is even smaller at 1 percent. (Appendix B includes decompositions of segregation indices by metropolitan area.) This finding is perhaps unsurprising because of the small proportion of the population currently living in planned developments. Recall that decompositions of the entropy measure of segregation reflect the relative population shares in each geographic component. Although planned developments represent more than 40 percent of new single-family homes, they include less than 20 percent of the existing stock of these homes and 12 percent of total units (HOA-Info, 2002; U.S. Census Bureau, 2002a). In light of racial and ethnic differences between planned developments and other neighborhoods presented in this chapter, it is conceivable that as these communities proliferate, they could exert a greater effect on residential segregation over time. Table 4.10 Decomposition of Statewide Metropolitan Racial Segregation Component Between central cities and suburbs Central cities Between planned developments and other neighborhoods Within planned developments Within other neighborhoods Suburbs Between planned developments and other neighborhoods Within planned developments Within other neighborhoods % Share 2 43 1 4 37 56 2 9 44 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Numbers may not sum to 100 because of rounding. 37 5. Voting Behavior This chapter investigates the voting behavior of Californians living in planned developments. It begins with a look at linkages between socalled “private governments” and political participation, as measured by voter turnout, registration, and party affiliation. It then considers differences between planned developments and other neighborhoods in “yes” votes on statewide ballot propositions. Finally, it examines how robust these differences are even after controlling for other factors that might influence voting behavior. Many social commentators have expressed concerns about residents of common interest developments opting out of the public sphere. Perhaps most famously, Robert Reich has referred to the growth of CIDs as the “secession of the successful.” Similarly, Charles Murray (1991, p. 26) has expressed concern about the democratizing potential of CIDs to allow “10 or 20 percent of the population… to bypass social institutions it doesn’t like in ways that only the top fraction of 1 percent used to be able to do.” One argument is that the homogeneity of these communities reduces the incentives, or controversies, that draw citizens into civic affairs (e.g., Putnam, 2000; Oliver, 2001).1 Alternatively, if planned developments are substitutes for traditional local government, then residents of these developments may perceive less benefit to taking part in elections and other aspects of public life (e.g., Blakely and Snyder, 1997). These concerns have taken on added meaning in light of recent evidence for declining “social capital” in the United States, as documented in Robert Putnam’s Bowling Alone. Analogous to physical or human capital, social capital refers to the stock of social norms, trust, and civic engagement that enhances both individual productivity and ____________ 1See Alesina and La Ferrara (2000) for an opposing view, i.e., that heterogeneity depresses participation in voluntary associations. 39 aggregate welfare. Indicators include voter turnout and other forms of political participation, membership in voluntary associations, and surveys regarding trust in government. Turnout in U.S. presidential elections has fallen from roughly 63 percent of the voting age population in 1960 to 51 percent in 2000 (U.S. Federal Elections Commission, 2000). In California, voter turnout has fallen from roughly 50 percent in nonpresidential statewide elections during the 1970s to 36 percent in November 2002 (California Secretary of State, 2002). Moreover, as in the rest of the nation, few Californians are involved in political parties, elections, or campaigns (e.g., Baldassare, 2000, p. 34; Verba, Scholzman, and Brady, 1995). These trends could have lasting consequences to the extent that government policies tend to reflect the preferences of those who participate in public life. There are several reasons, however, to believe that residents of planned developments will be more rather than less politically active. As seen in the previous chapter, residents of these communities tend to be older, better educated, and more affluent—factors that are all associated with higher rates of voter turnout (e.g., Rosenstone and Hansen, 1993). Beyond these factors, homeowners in general are more likely to take part in elections and other forms of civic participation because of an interest in preserving neighborhood quality and property values (DiPasquale and Glaeser, 1999).2 According to one gated community resident interviewed by Blakely and Snyder (1997, p. 72), for instance, The reason for the activism in Rancho Mirage and Indian Wells is the residents are rich and used to having their voices heard, and they are retired and have time. Also, the gated community has a stronger investment in expressing their needs because they identify with that community. Neighborhoods would have drawn that kind of allegiance in the past. There has been a sense of belonging and of shared destiny between an individual and where they live throughout history which has been lost. The gated community replaces that. I’m wondering if they aren’t a market response to that desire for belonging. . . . The gate and the homogeneity reinforces a feeling of shared destiny and [shared interest in] property values . ____________ 2Homeowners may also self-select into these communities because of an interest in preserving property values, for example, through more stringent land use and other regulation (Helsley and Strange, 2000). See Appendix C for a discussion of the selection issue. 40 This motivation further suggests that members of planned developments may vote differently from other Californians. In particular, in keeping with the “double taxation” issue discussed at the end of Chapter 2, they may favor public spending that complements private expenditures (e.g., police protection) while opposing that which is duplicative or redistributive (e.g., infrastructure and welfare). This chapter tests the above hypotheses using voting and registration information from the 2000 statewide general election.3 Results indicate that members of planned developments do not exhibit significantly different voting behavior once relevant demographic characteristics are taken into account. These findings undermine the popular view that socalled “private governments” crowd out political participation or influence the direction of voting behavior on individual ballot propositions, at least in statewide general elections. Voter Turnout, Registration, and Party Affiliation Planned developments exhibit higher rates of voter turnout and registration than other California neighborhoods (Table 5.1). Differences in turnout between planned developments and other neighborhoods are 6 percentage points, on average, whereas those in voter registration are even higher at 8 percentage points. Planned developments are more politically conservative than other neighborhoods, with higher Republican Party registration (by about 9 percentage points) and lower Democratic Party affiliation (by roughly 8 percentage points). (These differences are statistically significant at the 1 percent level). Voting on Statewide Ballot Propositions Differences in political ideology are apparent not only in the party affiliation of planned development residents but also in their patterns of voting on statewide ballot propositions (Table 5.2). In the 2000 statewide general election, residents of planned developments were more likely to support measures related to privatization, such as school ____________ 3These data were matched to census levels of geography by the Statewide Database at UC Berkeley’s Institute of Governmental Studies. See Statewide Database (2002). 41 Table 5.1 Voter Turnout and Party Affiliation, by Planned Development Status (in percent) Voter turnout Registered Democrat Republican Independent Other Planned Developments 74.1 70.7 40.7 39.8 14.4 5.1 Other Neighborhoods 68.4 62.8 49.1 31.2 14.2 5.5 Difference 5.7** 7.8** –8.5** 8.6** 0.2* –0.4** SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Rates of voter registration are expressed as a percentage of the voting age population. All other percentages are of registered voters. Independent voters are those who decline to state a party affiliation. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. vouchers (Proposition 38) and legislation enabling state and local governments to contract with private firms for public works projects (Proposition 35). They also voted in favor of fiscally conservative measures such as a two-thirds majority vote requirement for regulatory fees (Proposition 37) and against a measure lowering the voter approval requirement for local school bonds from two-thirds to 55 percent (Proposition 39). A Closer Look at Voting Behavior A key question is whether discrepancies in voter turnout and behavior persist after controlling for the demographic and socioeconomic characteristics of planned developments. As discussed in Chapter 4, these communities tend to differ from other neighborhoods along a number of dimensions. Table 5.3 summarizes these differences and their predicted effects on voter turnout based on previous studies. Taking account of just one factor—the proportion of the population that is non-white—narrows the gap in voter turnout considerably, 42 Table 5.2 Approval Rates for Statewide Ballot Propositions, by Planned Development Status (in percent) Prop. Planned Other No. Purpose Developments Neighborhoods Difference 32 $500,000,000 bond issue for farm and home aid to California veterans 66 69 –3** 33 Allows California legislature to participate in the Public Employees' Retirement System 37 41 –3** 34 Limits individual campaign contributions; establishes voluntary spending limits; expands public disclosure requirements 60 61 –1** 35 Provides that state and local governments may contract with private entities for engineering and architectural services 58 53 5** 36 Requires probation and treatment for conviction of possession, use, transportation, or being under influence of controlled substances 60 62 –3** 37 Requires two-thirds vote of state legislature, or either majority or two- thirds of local electorate, to impose any activity fees to monitor, study, or mitigate effects of that activity 49 47 2** 38 Authorizes state-funded school vouchers of $4,000 per pupil per year for private and religious schools 31 28 3** 39 Authorizes bonds for repair, construction, or replacement of school facilities, classrooms, if approved by a 55 percent local vote 52 56 –5** SOURCES: HOA-Info (2002); Statewide Database (2000). NOTE: Differences may not be precise because of rounding. **Denotes statistically significant difference at the 1 percent level. although differences remain statistically significant at the 1 percent level (Table 5.4). More strikingly, the gap in voter registration between planned developments and other neighborhoods is all but eliminated after controlling for the proportion of the population that is non-white (Table 5.5). 43 Table 5.3 Summary of Demographic and Socioeconomic Differences Between Planned Developments and Other Neighborhoods and Predicted Effects on Voter Turnout (in percent) Characteristic White Income above $75,000 With college degree Ages 40 to 64 Age 65 and up Homeowner Moved in last 5 years Planned Other Predicted Effect Developments Neighborhoods Difference on Turnout 64 47 17** Increases 37 26 11** Increases 21 15 6** Increases 33 29 4** Increases 14 11 2** Increases 66 57 9** Increases 49 48 1** Decreases SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Differences may not be precise because of rounding. **Denotes statistically significant difference at the 1 percent level. Table 5.4 Voter Turnout, by Planned Development Status and Racial and Ethnic Diversity (in percent) Percentage Non-White 0–25 25–50 50–75 75–100 Planned Developments 79 74 69 63 Other Neighborhoods Difference 77 2** 72 2** 66 2** 60 3** SOURCES: HOA-Info (2002); Statewide Database (2000). NOTE: Differences may not be precise because of rounding. **Denotes statistically significant difference at the 1 percent level. Regression analysis enables us to control for many demographic and socioeconomic characteristics simultaneously, thereby parsing out the effect of planned developments alone, all else being equal. Figure 5.1 summarizes the estimated effect of planned developments on voting behavior, holding constant other neighborhood features such as age structure, levels of educational attainment, income mix, racial and ethnic composition, homeownership, and urban location.4 ____________ 4Appendix C includes detailed regression results and a discussion of modeling issues. 44 Table 5.5 Voter Registration, by Planned Development Status and Racial and Ethnic Diversity (in percent) Percentage Non-White 0–25 25–50 50–75 75–100 Planned Developments 81 72 59 46 Other Neighborhoods Difference 80 1 71 1 59 0 47 –1* SOURCES: HOA-Info (2002); Statewide Database (2000). NOTE: Differences may not be precise because of rounding. *Denotes statistically significant difference at the 5 percent level. 10 8 6 4 2 0 –2 –4 –6 –8 –10 Voter turnout Percent registered Percent Republican Percent Democrat Percent Independent Raw difference in means Controlled difference in means Percentage points SOURCES: HOA-Info (2002); Statewide Database (2000). Figure 5.1—Estimated Effects of Planned Developments on Voter Turnout, Registration, and Party Affiliation 45 Differences in voter turnout, registration, and party affiliation between planned development and other neighborhoods are reduced substantially after controlling for these influences. Results from the regression model indicate that planned developments have rates of voter turnout that are higher by two-thirds of 1 percentage point; this is in contrast to the 6 percentage points detected in the uncontrolled difference. More dramatically, differences in voter registration rates switch from a positive 8 percentage points to a negative 0.77 percentage points. In other words, planned developments exhibit slightly lower rates of voter registration than we would expect given residents’ socioeconomic status. Differences in party affiliation, although muted, are still apparent. Residents of planned developments are more likely to register with the Republican Party, for an average difference of approximately 2 percentage points, with a nearly offsetting reduction in Democratic Party affiliation. Regressions seeking to explain voting on statewide ballot propositions hold constant Republican Party registration in addition to the demographic and socioeconomic neighborhood attributes described above. Again, the gap in voting behavior between planned developments and other areas narrows considerably (Figure 5.2). Differences in percentage voting in favor of “conservative” measures—such as school vouchers (Proposition 38), state and local government contracting out for public works projects (Proposition 35), and a two-thirds vote requirement for fees (Proposition 37)—drop to 1 percentage point or less. In contrast to the uncontrolled differences presented above, planned developments are associated with slightly higher approval rates for measures to lower the vote requirement for local school bonds (Proposition 39) and to issue a bond for veterans (Proposition 32). In general, these differences, although statistically significant, are quite small. Taken together, they suggest that planned developments do not strongly affect either political participation or patterns of voting behavior in statewide general elections. Residents of planned developments vote similarly to other Californians of comparable age, income, education, race, and ethnicity. 46 Percentage points 6 4 2 0 –2 –4 –6 Proposition 32 Proposition 35 Proposition 37 Proposition 38 Proposition 39 Raw difference in means Controlled difference in means SOURCES: HOA-Info (2002); Statewide Database (2000). Figure 5.2—Estimated Effects of Planned Developments on Votes for Statewide Ballot Propositions 47 6. Conclusions Common interest developments are a dominant source of new housing in California. This report is one of the first empirical analyses of these communities. It presents a comprehensive overview of common interest developments in general, including their origins, institutional features, and powers and responsibilities and compares them to traditional local government. It also surveys the features of common interest development housing in California, documenting trends in the location, size, and financial resources of these communities. The report focuses on one particular type of common interest development, the planned development. This is the fastest-growing and most prevalent CID as well as the most direct form of potential competition with traditional local government. The study relies on a variety of data sources to describe these communities and discern their potential consequences for residential segregation and civic engagement. Results indicate that concerns regarding negative social consequences of planned developments are, for the most part, unwarranted. Planned developments are less diverse than other neighborhoods, yet their contribution to area segregation is minor. Similarly, planned developments do not exhibit differences in political participation or voting behavior once relevant demographic and socioeconomic factors are taken into consideration. There are several important caveats to bear in mind, however. As recent trends in the growth of the population and planned developments continue, it is conceivable that the effects of these communities on residential segregation may become more pronounced. Although planned developments currently represent only 12 percent of all housing units, they constitute roughly 30 percent of new housing starts and over 40 percent of new single-family home sales (HOA-Info, 2002; U.S. Census Bureau, 1994, 1995, 1996, 1998, 1999, 2002b; Construction Industry Research Board, 2000). 49 Local authorities may wish to consider such potential costs before encouraging this form of development to offset the costs of additional public services and infrastructure. Nevertheless, eliminating planned developments altogether would not necessarily reduce segregation even by the modest amounts suggested in Chapter 4. These developments are but one way that individuals can sort themselves based on desires for either enhanced services or racial and economic homogeneity. In addition, voting behavior in statewide general elections is only one form of political participation, albeit one that is more tractable for empirical analysis. Other important indicators of civic engagement include joining in voluntary associations, contacting elected officials, and attending public hearings or meetings. An important area for future research is the level of involvement within common interest developments themselves through boards of directors and committees. Only this information will illuminate the total effect of planned developments on participation in public life. Finally, a consistent and important finding throughout the report is the considerable heterogeneity among common interest developments in general and planned developments specifically. This finding in particular militates against any “one-size-fits-all” policy solution. 50 Appendix A Data and Methods This appendix describes the procedure used to match locations of planned developments to U.S. Census Bureau geography. It also explores the potential consequences of the limitations of this procedure for results presented in the report. In some cases, it presents results separately based on alternative sample definitions. Planned development addresses in the HOA-Info database were geocoded—matched to latitude and longitude coordinates using geographic information systems (GIS). Matching was successful for about 8,100, or 94 percent, of the approximately 8,600 planned developments for which address information was reported.1 These coordinates were then used to link each property to census levels of geography (e.g., blocks, block groups, and tracts) and political jurisdictions (e.g., cities, counties) (Figure A.1). Geographic matches were based on the reported street address of the homeowner association president. Properties reporting only a manager address were excluded from the sample, as were any properties reporting anomalous addresses that could not be reliably distinguished from off-site management companies.2 This procedure yielded roughly 6,200, or 66 percent, of all planned developments that existed in California as of December 2002. ____________ 1Of the 9,318 planned developments in the dataset, address information was missing for 667 observations. Searches of homeowner association websites and directories as well as public tax assessment, deed transfer, and mortgage records yielded locations for several missing PDs. These developments were statistically indistinguishable from planned developments that did report address information on all observable dimensions except size (i.e., number of units and acreage), most likely reflecting the procedure used to locate missing addresses rather than systematic reasons for reporting or not reporting to the California Department of Real Estate or Secretary of State. 2These include multiple records with identical addresses except for homeowner associations that are affiliated with one another (e.g., separate building phases of the same development). Unique observations in areas that were zoned predominantly for commercial or industrial use according to county assessor data were also purged from the analysis. 51 Figure A.1—Example of Geocoding Procedure This sample differs somewhat from the population of planned developments as a whole. Notably, they are significantly smaller and older than other planned developments (Table A.1). These differences may reflect the tendency for larger and newer developments to report offsite management companies as their primary address; these would be excluded from the sample. Efforts to obtain address information from large property management companies were unsuccessful because of confidentiality concerns. Table A.2 shows that excluding these developments from the study is unlikely to bias results toward zero because they constitute a small proportion of total housing. In other words, they do not constitute a large enough share of all units to contaminate the “control group.” 52 Table A.1 Selected Characteristics of Planned Developments in California Characteristic 5th 25th Minimum Percentile Percentile Median Study Sample Units Acres Year founded Monthly fees ($) Annual revenues 3 0 1924 0 62 6 1 1968 21 6,908 16 42 3 10 1978 1984 65 113 19,022 48,947 Remaining PDs Units Acres Year founded Monthly fees ($) Annual revenues 2 0 1929 0 86 10 1 1971 30 13,617 39 7 1980 71 45,714 84 17 1987 110 105,356 SOURCE: HOA-Info (2002). 75th 95th Percentile Percentile Maximum 108 33 1990 168 132,011 460 30,000 250 11,000 1999 2002 310 4,006 591,784 35,235,192 170 662 11,335 50 329 5,716 1993 2000 2002 167 317 1,654 240,995 1,072,115 13,645,714 Nevertheless, the exclusion of these PDs could limit the external validity or generalizability of this study. A second issue is matching planned developments to census levels of geography. Ideally, we would like to assume that census block groups and planned developments are co-extensive and to use block groups to compare planned developments with other neighborhoods. However, planned development and block group boundaries do not perfectly coincide. The average block group contains roughly 550 housing units, whereas the average planned development includes 150 units. In addition, planned developments may span multiple block groups. These concerns are somewhat mitigated by the Census Bureau requirements that census block groups be as homogeneous as possible and that they reflect existing neighborhood boundaries (U.S. Census Bureau, 2000). Nevertheless, we replicated the analysis in Chapter 4 using a narrower definition of what constitutes a planned development 53 Table A.2 Regional Distribution of Planned Development Units, by Sample Definition Region Bay Area Central Coast North Coast and Mountains Inland Empire Sacramento Metro San Diego San Joaquin Valley Sierras South Coast Total Study Sample 159,501 6 40,444 8 57,946 12 111,695 9 73,926 10 88,866 8 79,259 7 21,587 22 216,719 5 849,943 7 Remaining PDs 106,307 4 5,388 1 16,971 3 93,894 8 35,183 5 98,134 9 31,167 3 3,493 4 245,324 5 635,861 5 Total Housing Units 2,552,402 492,256 488,006 1,186,043 714,981 1,084,040 1,107,260 97,456 4,492,105 12,214,549 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: The numbers in the second row for each region are percentages. block group.3 Findings were not substantially different in sign or in magnitude. Results are available from the author upon request. ____________ 3Specifically, to avoid mistakenly characterizing block groups as planned developments, we restricted the sample to block groups where PDs constituted a majority (60 percent) of the housing stock. Many of these block groups included more planned development units than total housing units, suggesting a PD that spans multiple block groups. To address this second type of error, we coded all block groups within a census tract as planned developments if the “spillover” from the initial PD constituted more than 70 percent of all housing units in the tract. 54 Appendix B Measuring Diversity and Segregation This appendix reviews the entropy measures of diversity and segregation. It also presents results for the decomposition of segregation measures by metropolitan area (Table B.1). The entropy index of diversity for a city or metropolitan area is E = n ∑ Qr r =1 log ⎛ ⎝⎜ 1 Qr ⎞ ⎠⎟ (B.1) where Qr is the proportion of the total population in group r, and n is the number of groups. Analogously, for a neighborhood the index is defined as Ei = n ∑ qri r =1 ⎛ log⎝⎜ 1 q ri ⎞ ⎠⎟ (B.2) where qri is the proportion of individuals belonging to racial or ethnic group r in neighborhood i. Conceptually, the index expresses the level of uncertainty as to the racial or ethnic identity of any individual selected at random from the area as a whole. It reaches its maximum of log(n) when all groups are equally represented in the population and its minimum of 0 when there is only one group present. Index values can be normalized to range between 0 and 100 by using the number of groups, n, as the base of the logarithm and multiplying by 100. The entropy index of segregation relates the diversity of a city or metropolitan area as a whole to the population-weighted average diversity of its constituent neighborhoods: k ∑ [ti /T (E − Ei )] H = i =1 E (B.3) 55 Table B.1 Decomposition of Segregation, by Planned Development and Central City/Suburban Status 56 Metropolitan Area Bakersfield Chico–Paradise Fresno Los Angeles–Riverside– Orange County Merced Modesto Redding Sacramento–Yolo Salinas San Diego Between Central Cities and Suburbs (1) 0.84 3 0.15 1 1.67 7 Total Central Cities (2) 6.14 25 4.59 35 10.39 43 Within Central Cities Between PDs and Within Within Non-PDs PDs Non-PDs (3) (4) (5) 0.58 0.77 4.79 23 19 0.06 0.79 3.74 06 28 0.72 1.08 8.59 35 36 0.43 11.24 1 37 0.97 5.11 6 33 0.86 7.23 6 54 0.98 2.58 14 37 2.9 5.45 17 31 1.25 9.93 4 35 0.93 13.68 4 59 0.42 0.6 12 0.34 0.11 21 0.32 0.68 25 0.06 0.65 19 0.18 0.73 14 0.06 3.8 0 13 0.55 1.46 26 10.22 34 4.66 30 6.23 47 1.87 27 4.54 26 6.07 21 11.67 51 Total Suburbs (6) 17.66 72 8.46 64 11.86 50 18.59 61 9.47 61 5.2 39 3.48 50 9.14 52 17.18 61 8.42 37 Within Suburbs Between PDs and Within Within Non-PDs PDs Non-PDs (7) (8) (9) 1.54 1.32 14.8 65 60 0.39 0.50 7.57 34 57 0.8 0.8 10.26 33 43 0.45 2.57 18 0.14 0.25 12 0.24 0.19 21 0.12 0.14 22 0.59 1.57 39 0.92 4.62 3 16 0.59 1.6 37 15.57 51 9.08 58 4.77 36 3.22 46 6.98 40 11.64 41 6.23 27 Total (10) 24.4 100 13.19 100 23.94 100 30.25 100 15.54 100 13.3 100 7.03 100 17.5 100 28.36 100 23.02 100 Table B.1 (continued) 57 Metropolitan Area San Francisco– Oakland–San Jose San Luis Obispo– Atascadero–Paso Robles Santa Barbara–Santa Maria–Lompoc Stockton–Lodi Visalia–Tulare– Porterville Yuba City Total Between Central Cities and Suburbs Total Central Cities 0.94 12.22 4 51 0.08 4.2 1 29 2.78 11.68 13 54 2.45 10.98 13 60 2.26 8.70 11 42 0.18 3.35 2 28 0.45 12.14 2 43 Within Central Cities Between PDs and Within Within Non-PDs PDs Non-PDs 0.16 1.91 18 10.15 42 0.15 0.83 16 3.22 22 0.37 2.98 2 14 0.89 1.1 56 8.33 39 8.99 50 0.64 1.64 38 0.13 0.35 13 0.42 1.23 14 6.42 31 2.87 24 10.49 37 Total Suburbs 10.81 45 10.37 71 7.15 33 4.72 26 9.68 47 8.31 70 15.92 56 Within Suburbs Between PDs and Within Within Non-PDs PDs Non-PDs Total 0.23 3.05 1 13 7.53 23.97 31 100 0.42 3.9 3 27 6.05 14.65 41 100 0.36 2.22 2 10 0.07 0.49 03 4.57 21.6 21 100 4.16 18.15 23 100 0.31 0.55 23 0.2 0.36 23 0.62 2.65 29 8.82 43 7.75 66 12.65 44 20.64 100 11.83 100 28.51 100 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTES: The first entry in each column represents the “between” or “within” component of overall metropolitan segregation. The second entry is the percentage share of metropolitan segregation attributable to this component. Columns 1, 2, and 6 sum to column 10. Columns 3, 4, and 5 sum to column 2, and Columns 7, 8, and 9 sum to column 6. Index values reported in this table are multiplied by 100. where T and ti are the respective area and subunit populations and k is the number of subunits. Since the entropy of an area as a whole (i.e., E ) can never be less than the average entropy of its parts (i.e., ∑ik=1(ti /T )Ei )), the segregation index necessarily ranges between 0 and 1. At 0, the racial composition of each subunit is identical to that of the area as a whole, whereas at 1 each subunit contains only one group. An attractive feature of the entropy index of segregation is that it may be decomposed into segregation between and within groups. The general formula for any decomposition is H = ∑ p∈P ⎛ ⎝⎜ Tp T ⎞ ⎠⎟ ⎛ ⎝⎜ Ep E ⎞ ⎠⎟ H p (B.4) where p represents racial or geographic components of the decomposition (Reardon et al., 2000, p. 354). For instance, the segregation of a metropolitan area comprising central city and suburban block groups can be written as H = H c ,s + ⎛⎝⎜TTc ⎞ ⎠⎟ ⎛ ⎝⎜ Ec E ⎞ ⎠⎟ H c + ⎛⎝⎜TTs ⎞ ⎠⎟ ⎛ ⎝⎜ Es E ⎞ ⎠⎟ H s (B.5) where Tc and Ts, Ec and Es, and Hc and Hs are, respectively, the total populations, entropy measures of diversity, and entropy measures of segregation for city and suburban block groups (Reardon et al., 2000, p. 355). The first term, Hc,s, indicates the portion of overall segregation attributable to differences between the central city and the suburbs. That is H c ,s = 1 − ⎛⎝⎜Tc E c +Ts TE E s ⎞ ⎠⎟ . (B.6) The second and third terms represent portions of segregation attributable to differences among central city and suburban subunits. These latter terms can be further decomposed into “between” and “within” components with respect to planned development status. Specifically, each “within” term in Equation (B.5) above includes 58 H PD,non−PD + ⎛⎝⎜TTPD ⎞⎛ ⎠⎟ ⎝⎜ E PD E ⎞ ⎠⎟ H PD + ⎛⎝⎜T noTn − PD ⎞⎛ ⎠⎟ ⎝⎜ E non−PD E ⎞ ⎠⎟ H non − PD (B.7) where the first term is segregation between planned developments and other block groups defined analogously to Equation (B.6), the second term represents the share of segregation among planned developments, and the third term represents the share of segregation among non-PD block groups. 59 Appendix C Technical Considerations and Regression Results A methodological complication in any study of neighborhood effects on political participation is the potential endogeneity of residential choice. That is, individuals may choose to live in a planned development because of observable or unobservable factors that are also related to voting behavior. As a result, apparent relationships between planned developments and voter turnout may be spurious. This selection bias can work in either direction. For example, individuals may sort into a planned development based on a taste for involvement in both public and private government. Alternatively, they may move into a PD to opt out of the public sphere without sacrificing collective goods and services. In either case, estimates of the effect of living in a planned development will be both biased and inconsistent. Gordon (2003) explores this possibility using several specifications to control for potential selection bias. Methods include least squares regression with block-group fixed effects, a Heckman selection correction model, and a propensity-score approach. Results from these models are not substantially different from the least squares regression with demographic covariates presented in this report.1 The remainder of this appendix includes detailed results from the multivariate analysis underlying the charts in Chapter 5. ____________ 1Gordon (2003) also includes results of regressions based on a more restricted sample definition as in Appendix A. 61 Table C.1 Estimated Effects of Planned Developments on Voter Turnout, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 5.700 (0.166)** 68.374 (0.079)** 22,055 0.04 Ordinary Least Squares 0.661 (0.107)** 0.093 (0.012)** 0.173 (0.007)** –0.118 (0.004)** –0.053 (0.004)** –0.123 (0.003)** –0.160 (0.014)** –0.049 (0.016)** 0.023 (0.010)* 0.086 (0.011)** 0.068 (0.004)** –0.043 (0.005)** 0.110 (0.009)** 0.127 (0.010)** 0.108 (0.009)** 0.059 (0.009)** –1.696 (0.226)** 63.193 (0.855)** 21,802 0.63 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 62 Table C.2 Estimated Effects of Planned Developments on Percentage Registered, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 7.827 (0.328)** 62.833 (0.148)** 21,039 0.02 Ordinary Least Squares –0.765 (0.206)** 0.083 (0.022)** 0.162 (0.014)** 0.077 (0.007)** –0.266 (0.008)** –0.345 (0.006)** –0.168 (0.024)** –0.108 (0.028)** 0.176 (0.019)** 0.379 (0.021)** 0.150 (0.007)** –0.186 (0.009)** 0.010 (0.017) –0.006 (0.022) –0.009 (0.015) 0.050 (0.018)** 4.507 (0.440)** 60.758 (1.572)** 20,874 0.66 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 63 Table C.3 Estimated Effects of Planned Developments on Percentage Registered Democrat, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means –8.460 (0.220)** 49.116 (0.107)** 22,060 0.05 Ordinary Least Squares –1.807 (0.175)** 0.136 (0.014)** 0.095 (0.010)** 0.547 (0.004)** 0.243 (0.006)** 0.055 (0.006)** –0.023 (0.016) –0.061 (0.018)** –0.272 (0.014)** 0.077 (0.015)** –0.065 (0.004)** –0.144 (0.006)** –0.015 (0.011) –0.049 (0.013)** –0.100 (0.012)** –0.143 (0.017)** 5.243 (0.265)** 46.224 (1.116)** 21,807 0.60 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 64 Table C.4 Estimated Effects of Planned Developments on Percentage Registered Republican, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 8.604 (0.252)** 31.223 (0.111)** 22,060 0.04 Ordinary Least Squares 2.031 (0.202)** –0.126 (0.016)** 0.044 (0.010)** –0.476 (0.005)** –0.185 (0.006)** –0.160 (0.006)** 0.026 (0.017) 0.061 (0.018)** 0.284 (0.015)** –0.153 (0.018)** 0.152 (0.005)** 0.134 (0.007)** 0.007 (0.013) 0.053 (0.014)** 0.116 (0.013)** 0.192 (0.019)** –4.899 (0.333)** 26.724 (1.219)** 21,807 0.54 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 65 Table C.5 Estimated Effects of Planned Developments on Percentage Registered Independent, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 0.211 (0.079)** 14.165 (0.037)** 22,060 0.00 Ordinary Least Squares –0.148 (0.061)* –0.013 (0.006)* –0.070 (0.004)** –0.070 (0.002)** –0.023 (0.002)** 0.134 (0.003)** –0.017 (0.008)* –0.002 (0.008) –0.002 (0.006) 0.061 (0.007)** –0.065 (0.002)** 0.011 (0.003)** 0.027 (0.004)** 0.018 (0.005)** 0.019 (0.005)** 0.007 (0.006) –0.077 (0.109) 16.855 (0.479)** 21,807 0.45 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 66 Table C.6 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 32, 2000 Difference Ordinary in Means Least Squares Planned development –2.752 0.314 = 1 if > 1 PD (0.096)** (0.066)** = 0 otherwise Percentage Ages 40–64 0.004 (0.007) Age 65 and older 0.096 (0.005)** Non–Hispanic black 0.140 (0.003)** Hispanic 0.019 (0.002)** Non–Hispanic Asian –0.050 (0.003)** With less than high school 0.012 (0.010) High school diploma –0.007 (0.011) With some college –0.011 (0.007) With college degree –0.021 (0.007)** Homeowner –0.016 (0.002)** Moved in last 5 years –0.006 (0.003) With income $50,000–$74,999 –0.016 (0.005)** With income $75,000–$99,999 –0.027 (0.006)** With income $100,000–$199,999 –0.048 (0.006)** With income $200,000 and up 0.011 (0.007) Urban (= 1 if in an urbanized area) 0.678 (0.128)** Percent registered Republican –0.208 (0.003)** Constant 68.993 75.398 (0.048)** (0.567)** Observations R2 22,036 0.03 21,801 0.64 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 32 is a $500,000,000 bond issue for farm and home aid to California veterans. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 67 Table C.7 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 35, 2000 Difference Ordinary in Means Least Squares Planned development 5.272 1.068 = 1 if > 1 PD (0.124)** (0.077)** = 0 otherwise Percentage Ages 40–64 0.000 (0.007) Age 65 and older –0.039 (0.005)** Non–Hispanic black –0.143 (0.003)** Hispanic –0.001 (0.003) Non–Hispanic Asian –0.025 (0.003)** With less than high school –0.041 (0.008)** High school diploma –0.027 (0.009)** With some college 0.013 (0.006)* With college degree 0.052 (0.007)** Homeowner –0.046 (0.002)** Moved in last 5 years 0.055 (0.003)** With income $50,000–$74,999 0.017 (0.005)** With income $75,000–$99,999 0.051 (0.007)** With income $100,000–$199,999 0.100 (0.005)** With income $200,000 and up 0.161 (0.008)** Urban (= 1 if in an urbanized area) –0.546 (0.140)** Percent registered Republican 0.287 (0.003)** Constant 52.509 42.972 (0.058)** (0.508)** No. of observations R2 22,032 0.06 21,799 0.70 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 35 allows state and local governments to contract with private entities for engineering and architectural services. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 68 Table C.8 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 37, 2000 Difference Ordinary in Means Least Squares Planned development 2.331 0.167 = 1 if > 1 PD (0.141)** (0.078)* = 0 otherwise Percentage Ages 40–64 0.013 (0.008) Age 65 and up –0.021 (0.005)** Non–Hispanic black –0.042 (0.004)** Hispanic 0.097 (0.003)** Non–Hispanic Asian 0.128 (0.003)** With less than high school 0.108 (0.010)** High school diploma 0.135 (0.010)** With some college 0.089 (0.007)** With college degree –0.075 (0.009)** Homeowner 0.009 (0.003)** Moved in last 5 years 0.034 (0.003)** With income $50,000–$74,999 –0.002 (0.007) With income $75,000–$99,999 0.003 (0.007) With income $100,000–$199,999 –0.022 (0.005)** With income $200,000 and up –0.066 (0.007)** Urban (= 1 if in an urbanized area) –0.553 (0.138)** Percent registered Republican 0.503 (0.004)** Constant 46.991 21.736 (0.066)** (0.598)** No. of observations R2 22,029 0.01 21,800 0.72 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 37 requires a vote of the state legislature or local electorate to impose fees to monitor, study, or mitigate an activity. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 69 Table C.9 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 38, 2000 Difference in Means Ordinary Least Squares Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and up Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (=1 if in an urbanized area) Percent registered Republican Constant No. of observations R2 2.662 (0.113)** 28.282 (0.052)** 22,036 0.020 0.112 (0.071) 0.020 (0.007)** –0.033 (0.005)** –0.027 (0.003)** 0.024 (0.003)** 0.069 (0.003)** 0.062 (0.008)** 0.087 (0.011)** 0.064 (0.007)** –0.043 (0.006)** –0.018 (0.002)** 0.029 (0.003)** 0.003 (0.005) 0.004 (0.005) –0.022 (0.007)** 0.045 (0.007)** –1.84 (0.132)** 0.379 (0.003)** 13.645 (0.532)** 21,801 0.670 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 38 authorizes state-funded school vouchers of $4,000 per pupil per year for private and religious schools. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 70 Table C.10 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 39, 2000 Difference Ordinary in Means Least Squares Planned development –4.536 0.931 = 1 if > 1 PD (0.173)** (0.099)** = 0 otherwise Percentage Ages 40–64 –0.041 (0.011)** Age 65 and up 0.028 (0.006)** Non–Hispanic black 0.156 (0.004)** Hispanic 0.065 (0.004)** Non–Hispanic Asian –0.011 (0.003)** With less than high school –0.072 (0.011)** High school diploma –0.175 (0.018)** With some college –0.109 (0.010)** With college degree 0.079 (0.009)** Homeowner –0.058 (0.003)** Moved in last 5 years 0.036 (0.004)** With income $50,000–$74,999 –0.031 (0.008)** With income $75,000–$99,999 –0.039 (0.008)** With income $100,000–$199,999 –0.015 (0.009) With income $200,000 and up 0.115 (0.009)** Urban (=1 if in an urbanized area) 2.770 (0.179)** Percent registered Republican –0.473 (0.004)** Constant 56.307 73.201 (0.089)** (0.787)** No. of observations R2 22,035 0.02 21,801 0.78 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 39 authorizes bonds for school facilities, classrooms if approved by a 55 percent local vote. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 71 References Advisory Commission on Intergovernmental Relations, Residential Community Associations: Private Governments in the Intergovernmental System? Washington, D.C., May 1989. Alesina, Alberto, and Eliana La Ferrara, “Participation in Heterogeneous Communities,” Quarterly Journal of Economics, Vol. 115, No. 3, August 2000, pp. 847–904. Baldassare, Mark, California in the New Millennium, University of California Press, Berkeley, California, and the Public Policy Institute of California, San Francisco, California, 2000. Barton, Stephen E., and Carol J. Silverman, “History and Structure of the Common Interest Community,” in Stephen E. Barton and Carol J. Silverman, eds., Common Interest Communities: Private Governments and the Public Interest, Institute of Governmental Studies Press, Berkeley, California, 1994. Barzel, Yoram, and Tim R. Sass, “The Allocation of Resources by Voting,” Quarterly Journal of Economics, Vol. 16, August, 1990, pp. 745–771. Black’s Law Dictionary, 5th ed., West Publishing Company, St. Paul, Minnesota, 1979. Blakely, Edward J., and Mary Gail Snyder, Fortress America: Gated Communities in the United States, Brookings Institution Press, Washington, D.C., 1997. California Law Revision Commission, “Common Interest Development Law: Scope of Project (Background Study),” Staff Memorandum 2001-19, Palo Alto, California, January 22, 2001. 73 California Law Revision Commission, “Nonjudicial Dispute Resolution Under CID Law: Alternative Dispute Resolution (Additional Information),” Second Supplement to Memorandum 2002-10, Palo Alto, California, 2002a. California Law Revision Commission, “Alternative Dispute Resolution in Common Interest Developments,” Tentative Recommendation, Palo Alto, California, 2002b. California Secretary of State, Statement of Vote, 2002, available at http://ss.ca.gov/elections/sov/2002_general/contents.htm. Community Associations Institute, Data on U.S. Community Associations, 2003, available at http://www.caionline.org/about/facts.cfm. Construction Industry Research Board, Characteristics of New Homes Sold: California and Selected Counties, 1982–1999, Burbank, California, 2000. Devaney, John F., Tracking the American Dream: 50 Years of Housing History from the Census Bureau, 1940–1990, U.S. Department of Commerce, Washington, D.C., 1994. Dilger, Robert J., Neighborhood Politics: Residential Community Associations in American Governance, New York University Press, New York, 1992. DiPasquale, Denise, and Edward L. Glaeser, “Incentives and Social Capital: Are Homeowners Better Citizens?” Journal for Urban Economics, Vol. 45, March 1999, pp. 354–384. Egan, Timothy, “Many Seek Security in Private Communities,” New York Times, September 3, 1995, pp. A1, A22. Ellickson, Robert C., “New Institutions for Old Neighborhoods,” Duke Law Journal, Vol. 48, No. 1, 1998, pp. 75–110. Foldvary, Fred, Public Goods and Private Communities: The Market Provision of Social Services, Edward Elgar, Brookfield, Vermont, 1994. 74 Foldvary, Fred E., “Private Communities and Community Associations,” in David T. Beito, Peter Gordon, and Alex Tabarrok, eds., The Voluntary City: Restoring Urban Life in Crisis Times, Michigan University Press and The Independent Institute, Ann Arbor, Michigan, 2002. French, Susan F., Scope of Laws Affecting Common Interest Developments, background study prepared for the California Law Revision Commission, November 2000. Gordon, Tracy M., “Crowd Out or Crowd In? The Effects of Common Interest Developments on Political Participation in California,” The Annals of Regional Science, Vol. 37, No. 2, May 2003, pp. 203–233. “Government by the Nice for the Nice,” The Economist, July 25, 1992, pp. A25–A27. Helsley, Robert W., and William C. Strange, “Social Interactions and the Institutions of Local Government,” American Economic Review, Vol. 90, No. 5, 2000, pp. 1477–1490. HOA-Info Information Services Division, HOA-Info, Oakland, California, 2002. HOA-Info Information Services Division, California Community Association Statistics, Oakland, California, 2003. Kennedy, David J., “Note: Residential Associations as State Actors: Regulating the Impact of Gated Communities on Nonmembers,” Yale Law Journal, Vol. 105, December 1995, pp. 761–793. Kress, Carl B., “Comment: Beyond Nahrstedt: Reviewing Restrictions Governing Life in a Property Owner Association,” UCLA Law Review, Vol. 42, 1995, pp. 837–883. Lave-Johnston, Julia, and Kimberly Johnston-Dodds, Common Interest Developments: Housing at Risk? California Research Bureau, Sacramento, California, 2002. Massey, Douglas S., and Nancy Denton, “The Dimensions of Racial Segregation,” Social Forces, Vol. 67, 1988, pp. 281–315. 75 McKenzie, Evan, Privatopia, Yale University Press, New Haven, Connecticut, 1994. Murray, C., “The Shape of Things to Come,” National Review, Vol. 43, 1991, pp. 26–31. Nelson, Robert H., “Privatizing the Neighborhood: A Proposal to Replace Zoning with Private Collective Property Rights to Existing Neighborhoods,” George Mason Law Review, Vol. 7, No. 4, Summer 1999, pp. 827–880. Oliver, Eric J., Democracy in Suburbia, Princeton University Press, Princeton, New Jersey, 2001. Pack, Janet R., “BIDs, DIDs, SIDs, and SADs: Private Governments in Urban America,” Brookings Review, Fall 1992, pp. 18–21. Putnam, Robert D., Bowling Alone: The Collapse and Revival of American Community, Simon and Schuster, New York, 2000. Reardon, Sean F., John T. Yun, and Tamela M. Eitle, “The Changing Structure of School Segregation: Measurement and Evidence of Multiracial Metropolitan-Area School Segregation, 1989–1995,” Demography, Vol. 37, No. 3, August 2000, pp. 351–364. Reich, Robert B., “Secession of the Successful,” New York Times, January 20, 1991, Section 6, p. 16. Roland, Helen E., Residential Common Interest Developments: An Overview, California Research Bureau, Sacramento, California, March 1998. Rosenstone, S. J., and J. M. Hansen, Mobilization, Participation, and Democracy in America, Macmillan, New York, 1993. Senate Housing and Community Development Committee, Hearing, “Common Interest Developments (CID’s): Incorporation and Registration,” AB 643 Assembly Bill-Bill Analysis, July 1, 2002. Statewide Database, Statewide Database, 2000, available at http:// swdb.berkeley.edu/info/PDRDocument.txt. 76 Statewide Database, Documentation, 2002, available at http:// swdb.berkeley.edu/data/data.html. Theil, Henri, Statistical Decomposition Analysis, Amsterdam: NorthHolland Publishing Co., 1972. Tiebout, Charles M., “A Pure Theory of Local Public Expenditures,” Journal of Political Economy, Vol. 64, October 1956, pp. 416–424. Tocqueville, Alexis de, Democracy in America, in J. P. Mayer, ed., Anchor, Garden City, New Jersey, 1969 (1835), pp. 513–517. Treese, Clifford J., Community Associations Factbook, Community Associations Institute Research Foundation, Alexandria, Virginia, 1999. U.S. Census Bureau, Census of Housing, General Housing Characteristics California, U.S. Department of Commerce, Washington, D.C., 1970, 1980, and 1990. U.S. Census Bureau, Geographic Areas Reference Manual, U.S. Department of Commerce, Washington, D.C., 2000. U.S. Census Bureau, Census 2000 Summary File 3—California, Washington, D.C., 2002a. U.S. Census Bureau, Statistical Abstract, U.S. Department of Commerce, Washington, D.C., 1994, 1995, 1996, 1998, 1999, 2002b. U.S. Federal Elections Commission, National Voter Turnout in Federal Elections: 1960–1996, 2000, available at http://www.fec.gov/pages/ htmlto5.htm. Verba, S., K. L. Scholzman, and H. E. Brady, Voice and Equality: Civic Voluntarism in American Politics, Harvard University Press, Cambridge, Massachusetts, 1995. Worley, William S., J. C. Nichols and the Shaping of a Kansas City: Innovation in Planned Residential Communities, University of Missouri Press, Columbia, Missouri, 1990. 77 About the Author TRACY GORDON Tracy Gordon is a research fellow at the Public Policy Institute of California, where her research interests include state and local public finance and political economy. She has authored several journal articles and conference papers on topics relating to local government, including common interest developments and their effects on residential segregation and voting behavior, the local initiative process, and local property taxes. She holds a B.A. in human biology from Stanford University and a Ph.D. in public policy with a concurrent M.A. in economics from the University of California, Berkeley. 79 Related PPIC Publications Who Pays for Development Fees and Exactions? Marla Dresch and Steven M. Sheffrin The Ties That Bind: Changing Demographics and Civic Engagement in California S. Karthick Ramakrishnan and Mark Baldassare “Who’s Your Neighbor? Residential Segregation and Diversity in California” California Counts: Population Trends and Profiles Volume 4, Number 1, August 2002 Juan Onésimo Sandoval, Hans P. Johnson, and Sonya M.Tafoya “California’s Newest Neighborhoods” California Counts: Population Trends and Profiles Volume 5, Number 1, August 2003 Hans P. Johnson and Joseph M. Hayes 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 81" } ["___content":protected]=> string(102) "

R 304TGR

" ["_permalink":protected]=> string(113) "https://www.ppic.org/publication/planned-developments-in-california-private-communities-and-public-life/r_304tgr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8297) ["ID"]=> int(8297) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:36:22" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(3467) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 304TGR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_304tgr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_304TGR.pdf" ["wpmf_size"]=> string(7) "1110683" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(135667) "Planned Developments in California: Private Communities and Public Life ••• Tracy M. Gordon 2004 Library of Congress Cataloging-in-Publication Data Gordon, Tracy M. Planned developments in California : private communities and public life / Tracy M. Gordon. p. cm. ISBN: 1-58213-081-7 1. Planned communities—California. 2. Housing, Cooperative— California. 3. Condominium associations—California. 4. Homeowners’ associations—California. 5. California—Ethnic relations. 6. Social classes—California. 7. Political participation— California. 8. Voting—California. I. Title. HT167.5.C2G57 2004 307.76'8'09794—dc22 2004001904 Copyright © 2004 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 on any local, state, or federal legislation, nor does it endorse, support, or oppose 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 Over the last three decades, there has been a notable increase in common interest developments (CIDs) in California. Variously characterized as “Fortress America,” “privatopias,” “secession of the successful,” or, more neutrally, “private government,” CIDs in the broadest sense (condominiums, cooperatives, and planned developments) now make up 60 percent of new housing starts in California. However, the image of CIDs commonly projected in critical scholarship and journalism is the gated suburban community closed to public access. These communities provide an array of traditional public services through private, commonly held organizations governed by their residents. The most extreme characterizations suggest that CIDs are nothing less than income and racial segregation reinvented in a more modern, legal form. Less extreme characterizations suggest that such communities might well signal an effective withdrawal of residents from traditional civic society—contributing to declining voting rates and the general movement toward fiscal conservatism at the state and local levels of government. PPIC research fellow Tracy Gordon has taken a careful look at CIDs as an increasingly popular form of privatization. In Planned Developments in California: Private Communities and Public Life, Gordon focuses primarily on planned developments, the prevalent form of CIDs in California. Drawing on a unique dataset of real estate activity in California, she concludes that some “sorting” is apparent when residents of planned developments are compared to other residents in central cities and suburbs. Residents in planned developments have a higher percentage of non-Hispanic whites, higher incomes, and more education and they are generally older than the comparison group. Her findings hold up even when she looks at planned developments for central cities and suburbia separately. The concern over social and economic segregation is thus given some support by the report. She is iii careful to note, however, that the contribution of planned developments to overall residential segregation is minor, and she found little evidence that residents of planned developments are withdrawing from civic life— at least as measured by voting behavior. It is striking that condominiums have also grown in importance to the new homebuyer—representing 37 percent of all CIDs in California compared to 61 percent for planned developments. Although condominiums are often seen as the first rung in the ladder of eventual home ownership, they are also fraught with challenges when disagreements arise between developers and owners and later on when owners cannot agree on key common interest issues. In a review of the governance processes used by CIDs, Gordon notes that the definition of community varies from example to example, but managing a community interest may be just as difficult as reaching a decision in local planning commission meetings. In sum, CIDs are here to stay, and Gordon has made a valuable contribution to our understanding of them in California. Although CIDs are still a relatively small part of the overall housing stock in California, their steep growth path and tendency toward residential selectivity suggest that they are well worth monitoring in the years and decades ahead. De Tocqueville’s observation that America has a penchant for voluntary associations may now be evolving into yet another form—voluntary local government imposed by the decision to purchase a home. This is certainly not a new idea, but this latest example serves as a reminder of the challenges facing California’s governing processes at various levels. David W. Lyon President and CEO Public Policy Institute of California iv Summary Nearly three million California homes, one-quarter of the state housing stock, are located within common interest developments (CIDs) such as planned developments, condominiums, and cooperatives. Moreover, this proportion is ever-increasing, with CIDs accounting for 60 percent of residential construction starts during the 1990s. One type of CID, the planned development, alone represented more than 40 percent of new single-family home sales over this period (see Figure S.1). More than a pattern of residential development, CIDs are an emerging form of privatization. Homeowner associations in CIDs provide many goods and services traditionally supplied by local government, including garbage collection, street cleaning, street lighting, and security patrol. They also levy assessments, adjudicate disputes, and regulate land use as well as other aspects of life within their boundaries. As a result, CIDs are often called “quasi-” or “private governments.” 2,500 Housing units (1000s) 2,000 1,500 1,000 500 0 1960 1970 1975 1980 1985 1990 1995 2000 Planned developments Cooperatives Condominiums SOURCE: HOA-Info (2002). NOTE: Figure does not include CIDs of unknown type. Figure S.1—Growth of Common Interest Developments in California v The growth of so-called private governments has sparked a popular debate. Proponents claim that CIDs aid cash-strapped localities in accommodating the service and infrastructure costs of growth. They further advocate extending this governance structure to existing neighborhoods as a means of improving local amenities and neighborhood quality. Critics charge that CIDs generate adverse consequences for nonresidents. Primary among these feared outcomes is heightened racial and economic segregation and diminished civic engagement. Such criticisms have prompted recurrent calls for enhanced regulation of these communities. Although the bulk of proposed legislation has addressed the internal governance of CIDs, a frequent complaint is that there are no current and reliable studies of common interest developments in California. In 1999, for example, the state legislature authorized the California Law Revision Commission to undertake a multiyear investigation to address, among other questions, “to what extent [common interest housing developments] should be subject to regulation” (California Law Revision Commission, 2001, p. 3). This report seeks to fill the gap in knowledge about CIDs in California. It focuses on planned developments, in particular, because these are the most prevalent and fastest-growing type of CID in both California and the nation as a whole. These developments also bear the strongest resemblance to traditional local government and hence attract the most scrutiny for their potential policy consequences. A Large and Diverse Group There are over 36,000 common interest developments in California. These communities range in size from three to 30,000 units and can accommodate several different housing types and legal ownership structures. The majority of CID units (61 percent) are located in planned developments, in which each homeowner owns his or her individual unit and lot and a homeowner association owns and maintains all common property. The other major type of CID (37 percent of all units) is the condominium, in which each owner holds title to his or her individual unit and a percentage interest in the common areas. vi Monthly assessments in CIDs vary widely, from merely nominal amounts to thousands of dollars per unit per month. In 2002, median monthly assessments were $112 per unit in planned developments and $186 per unit in condominiums and cooperatives. Total homeowner association revenues in California were estimated at $6.3 billion in 2003 (HOA-Info, 2003). Successful But Not Seceding Social commentators have referred to the spread of CIDs as the “secession of the successful” (Reich, 1991). An examination of the demographic and socioeconomic composition of planned developments lends support to this view. Nevertheless, the data also suggest that the picture is more complicated. Planned developments are less diverse with respect to race and ethnicity than other neighborhoods. In both central city and suburban areas, planned developments include significantly more residents who are non-Hispanic white (60 and 65 percent) and fewer who are black (4 percent in central cities and 3 percent in suburban areas) and Hispanic (20 and 19 percent, respectively). Interestingly, suburban planned developments tend to have larger Asian populations than other neighborhoods (Table S.1). Residents of planned developments are on average older than other central city or suburban dwellers. Planned developments include more individuals age 65 and over than other neighborhoods (differences of 2 to 3 percentage points, depending on location). They also include more residents ages 40 to 64 (differences of 3 and 4 percentage points for central cities and suburbs, respectively). Planned developments have correspondingly smaller population shares in the younger age categories and slightly fewer married couples with children (Table S.2). Planned developments are more diverse with respect to income than their image might suggest. To be sure, planned developments house more of the highest-income Californians than other neighborhoods. In central cities and suburbs, 22 and 26 percent of planned development households earned more than $100,000 in 1999, compared to 15 to 17 percent of households in other neighborhoods. However, planned developments also include nearly as many residents in the middle- to vii Table S.1 Selected Demographic Characteristics of California Planned Developments (in percent) Characteristic Black Hispanic Asian White Age 0 to 19 Age 20 to 39 Age 40 to 64 Age 65 and up Central City Planned Non-PD Developments Neighborhoods 4.2 9.8** 20.1 33.5** 11.7 12.3 60.2 40.8** 25.1 28.3** 30.2 33.2** 30.7 27.6** 14.0 10.9** Suburb Planned Non-PD Developments Neighborhoods 2.7 5.1** 18.8 32.6** 10.3 9.2** 64.6 49.5** 27.0 30.3** 26.1 28.7** 33.6 29.6** 13.3 11.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. **Denotes statistically significant difference at the 1 percent level. higher-income categories ($35,000 to $49,999, $50,000 to $74,999, and $75,000 to $99,999) as do other neighborhoods (Table S.2). Despite their relative racial and ethnic homogeneity, the contribution of planned developments to residential segregation is minor. Differences between planned developments and other neighborhoods account for only 2 percent of statewide metropolitan area segregation. This finding is primarily due to the relatively small proportion of the population living within planned developments. Although they represent roughly 40 percent of new single-family homes, planned developments constitute only 12 percent of the existing housing stock. Planned developments do not exhibit markedly different patterns of voting behavior once other relevant characteristics are taken into account. At first glance, planned developments have higher rates of voter registration and turnout than other neighborhoods. They also appear more likely to affiliate with the Republican Party and to vote for statewide ballot propositions favoring the privatization of government services (Propositions 35 and 38) and higher vote requirements for fees and bonds (Propositions 37 and 39). viii Table S.2 Selected Socioeconomic Characteristics of California Planned Developments Central City Suburb Planned Non-PD Planned Non-PD Characteristic Developments Neighborhoods Developments Neighborhoods Median household income in 1999, $ 58,687.8 46,688.3** 67,045.9 53,297.9** Household income, % Less than $35,000 32.9 42.9** 26.8 35.6** $35,000–$49,999 14.2 15.0** 13.7 15.5** $50,000–$74,999 19.1 17.5** 19.8 20.1 $75,000–$99,999 12.4 10.0** 13.5 12.0** $100,000–$199,999 16.6 11.6** 19.5 13.5** $200,000 and up 4.9 3.0** 6.6 3.3** % Married couples with children 30.9 31.9** 34.4 36.4* % with no college 28.0 40.4** 26.5 38.8** % in managerial and professional occupations 42.6 33.8** 42.3 32.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. However, planned developments also differ from other neighborhoods along a number of dimensions that are related to voting behavior (e.g., age, income, education, and residential mobility). Regression analyses holding these differences constant demonstrate that planned developments themselves do not strongly affect either political participation or patterns of voting behavior in statewide general elections. Remaining differences are small, at 1 to 2 percentage points. These conclusions are somewhat unexpected. Thus far, it would appear that concerns about planned developments siphoning off the wealthiest Californians and their participation in public life are unwarranted. Planned developments still represent a relatively small proportion of the population overall, however, and as their numbers increase, their effects on neighborhood composition and civic engagement may become more pronounced. ix Contents Foreword ..................................... Summary..................................... Figures ...................................... Tables ....................................... Acknowledgments ............................... iii v xiii xv xix 1. INTRODUCTION ........................... 2. WHAT ARE COMMON INTEREST DEVELOPMENTS? .......................... A Brief History .............................. Definition and Taxonomy ....................... Scope of Activities ............................ Policy Issues ................................ 3. WHAT DO CALIFORNIA CIDs LOOK LIKE? ........ Note on Data Sources .......................... A Large and Diverse Group ...................... 4. FOCUS ON PLANNED DEVELOPMENTS .......... Geographic Distribution ........................ Demographic and Socioeconomic Characteristics ........ Diversity and Segregation ....................... 5. VOTING BEHAVIOR ........................ Voter Turnout, Registration, and Party Affiliation ........ Voting on Statewide Ballot Propositions .............. A Closer Look at Voting Behavior .................. 6. CONCLUSIONS ............................ Appendix A. Data and Methods ............................ B. Measuring Diversity and Segregation ................ C. Technical Considerations and Regression Results ........ 1 9 9 11 13 14 19 19 20 27 27 28 35 39 41 41 42 49 51 55 61 xi References .................................... About the Authors ............................... Related PPIC Publications .......................... 73 79 81 xii Figures S.1. Growth of Common Interest Developments in California ............................... 1.1. Growth of Common Interest Develoments in the United States .................................. 1.2. Growth of Common Interest Developments and Suburban Housing Units in California and the United States, 1940–1999 ......................... 2.1. Growth of Common Interest Developments in the United States, by Type....................... 3.1. Growth of Common Interest Developments in California, by Type ......................... 3.2. Size Distribution of Common Interest Developments in California, by Type ......................... 3.3. Revenues of Common Interest Developments in California ............................... 5.1. Estimated Effects of Planned Developments on Voter Turnout, Registration, and Party Affiliation ......... 5.2. Estimated Effects of Planned Developments on Votes for Statewide Ballot Propositions .................. A.1. Example of Geocoding Procedure ................ v 2 4 12 22 23 25 45 47 52 xiii Tables S.1. Selected Demographic Characteristics of California Planned Developments ...................... viii S.2. Selected Socioeconomic Characteristics of California Planned Developments ...................... ix 2.1. An Example of Nonterritorial and Territorial CIDs .... 12 2.2. Services Provided by Common Interest Developments in the United States and California ................ 13 3.1. Geographic Distribution of Common Interest Developments ............................ 21 3.2. Types of Common Interest Developments in California ............................... 22 3.3. Average Size of Common Interest Developments, by Year of Construction ........................ 23 3.4. Distribution of Monthly Fees of Common Interest Developments ............................ 24 4.1. Planned Developments, by Region ............... 28 4.2. Planned Developments, by Central City/Suburban Location and Incorporation Status ............... 29 4.3. Planned Developments, by City Size and City Incorporation Date ......................... 29 4.4. Selected Demographic Characteristics of California Planned Developments ...................... 30 4.5. Selected Socioeconomic Characteristics of California Planned Developments ...................... 31 4.6. Selected Housing Characteristics of California Planned Developments ............................ 32 4.7. Percentage Black, by Planned Development Status and Housing Characteristics ...................... 33 4.8. Percentage Hispanic, by Planned Development Status and Housing Characteristics ................... 34 xv 4.9. Average Diversity Scores of California Neighborhoods, by Metropolitan Area, Planned Development, and Central City/Suburban Status .................. 4.10. Decomposition of Statewide Metropolitan Racial Segregation .............................. 5.1. Voter Turnout and Party Affiliation, by Planned Development Status ........................ 5.2. Approval Rates for Statewide Ballot Propositions, by Planned Development Status ................... 5.3. Summary of Demographic and Socioeconomic Differences Between Planned Developments and Other Neighborhoods and Predicted Effects on Voter Turnout ................................ 5.4. Voter Turnout, by Planned Development Status and Racial and Ethnic Diversity.................... 5.5. Voter Registration, by Planned Development Status and Racial and Ethnic Diversity.................... A.1. Selected Characteristics of Planned Developments in California ............................... A.2. Regional Distribution of Planned Developments Units, by Sample Definition ....................... B.1. Decomposition of Segregation, by Planned Development and Central City/Suburban Status ............... C.1. Estimated Effects of Planned Developments on Voter Turnout, 2000............................ C.2. Estimated Effects of Planned Developments on Percentage Registered, 2000 ................... C.3. Estimated Effects of Planned Developments on Percentage Registered Democrat, 2000 ............ C.4. Estimated Effects of Planned Developments on Percentage Registered Republican, 2000 ........... C.5. Estimated Effects of Planned Developments on Percentage Registered Independent, 2000........... C.6. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 32, 2000 ................. C.7. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 35, 2000 ................. 36 37 42 43 44 44 45 53 54 56 62 63 64 65 66 67 68 xvi C.8. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 37, 2000 ................. C.9. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 38, 2000 ................. C.10. Estimated Effects of Planned Developments on “Yes” Votes for Proposition 39, 2000 ................. 69 70 71 xvii Acknowledgments This report benefited from the contributions of several individuals. David Levy of Levy & Company, CPAs, generously provided access to the HOA-Info dataset and shared his extensive knowledge of the CID industry. Julia Lave-Johnston and Dean Misczynski of the California Research Bureau, Peter Detwiler of the Senate Local Government Committee, J. Stacey Sullivan of the Assembly Local Government Committee, and Nathaniel Sterling of the California Law Revision Commision shed light on legal and policy issues relating to CIDs. Professor Robert Jay Dilger of West Virginia University, David Levy, and Paul Lewis, Christopher Weare, and Junfu Zhang of PPIC gave thoughful reviews of the initial manuscript. Joyce Peterson, Peter Richardson, and Patricia Bedrosian improved the clarity of exposition. David Haskel and Rachel Flood provided excellent research assistance. Any remaining errors, as well as opinions or interpretations expressed in this report, are those of the author alone and do not reflect the views of the Public Policy Institute of California. xix 1. Introduction Rancho Bernardo, a master-planned community located about 25 miles north of downtown San Diego, houses approximately 45,000 residents in a mix of single-family detached homes, townhouses, condominiums, and apartments. The community also features service and retail businesses, light industry, shopping areas, parks, and golf courses. Council members elected from districts oversee working committees on public safety, government relations, utilities, traffic, transportation, finance, services, and elections. A planning board enforces land use codes, and a recreation council maintains the community park. Although it bears a striking resemblance to a city, Rancho Bernardo is in fact a private corporation. It is a large, but by no means unusual, example of what has been called “the most significant privatization of local government responsibilities in recent times” (Advisory Commission on Intergovernmental Relations, 1989, p. 18). It also represents a shift in the usual concept of privatization. In contrast to previous models, where a government contracts with a private firm for the provision of a single service or function, individuals today are increasingly contracting directly with so-called “private governments” for many, if not all, traditional municipal services. According to one study, this trend represents nothing less than “a quiet revolution,” with profound implications for community organization, local government, land use, neighbor relations, and other aspects of public life (Barton and Silverman, 1994, p. xi). Rancho Bernardo is an example of a common interest development, or CID.1 Common interest developments encompass several housing ____________ 1Common interest developments are also known as community associations (CAs), common interest communities (CICs), common interest realty associations (CIRAs), and residential community associations (RCAs) (Treese, 1999, p. 3). A related phenomenon is the business or downtown improvement district (BID or DID) (Pack, 1992). 1 types and ownership structures, including planned developments, condominiums, and cooperatives. The distinguishing feature of these communities is common property ownership. Beyond an individual house or unit, property owners in these communities also hold an interest in common areas—such as recreation facilities, streets, lawns, or parking lots—either as individuals or through a mandatory homeowner association. Deed restrictions known as conditions, covenants, and restrictions (CCRs) authorize the association to collect lien-based assessments in exchange for providing services and regulating land use as well as other aspects of life within CID boundaries. Over the past 30 years, common interest developments have proliferated throughout the United States. In 1962, there were fewer than 500 common interest developments nationally; today there are nearly 250,000 CIDs housing an estimated 50 million Americans and constituting 15 percent of the U.S. housing stock (Community Associations Institute, 2003) (Figure 1.1).2 250 16 CIDs as a % of 200 total housing units Total CIDs 150 100 50 0 1962 1970 1975 1980 1985 1990 1995 14 12 10 8 6 4 2 0 1999 SOURCE: Treese (1999). Figure 1.1—Growth of Common Interest Developments in the United States ____________ 2Gated communities constitute a subset of this total. The Community Associations Institute (CAI) estimates that 19 percent of its members, or 8.4 million people, live within gated communities (Blakely and Snyder, 1997, p. 3). Other sources put this figure at four million people, or roughly 10 percent of all CID residents (Egan, 1995, pp. A1, A22). The CAI estimate is likely an upper bound because its members tend to be larger associations. Number of CIDs (1000s) CIDs as a % of total units 2 In California, these trends are pronounced: There are three million homes within common interest developments, encompassing nearly onequarter of the state housing stock. Moreover, CIDs constituted 60 percent of new construction starts during the 1990s (HOA-Info, 2002; U.S. Census Bureau, 1994, 1995, 1996, 1998, 1999, 2002b). Planned developments alone represent over 40 percent of new single-family home sales in the state (Construction Industry Research Board, 2000). To put this trend in perspective, compare the growth of CIDs to the trend of suburbanization more generally. Since 1970, common interest developments have grown faster than the suburbs as a share of all housing units in both California and the United States (Figure 1.2). This growth rate exceeds the pace of suburbanization during the peak years of 1940 to 1960 by a factor of five.3 Although a vast literature has explored the social, economic, and political implications of suburbanization, the consequences of this most recent transformation are largely unknown. The growth of common interest developments has nonetheless sparked a popular debate. Proponents claim that CIDs aid cash-strapped localities in meeting demand for public services.4 They further advocate extending this governance structure to existing neighborhoods as a way to improve local amenities and neighborhood quality (e.g., Nelson, 1999; Ellickson, 1998). Critics charge that common interest developments have adverse consequences for nonresidents. Primary among these feared outcomes is heightened racial and economic segregation. Robert Reich (1991), for example, has referred to the growth of CIDs as the “secession of the successful.” Similarly, Evan McKenzie (1994), in his comprehensive study of the rise of CIDs, links the CCRs governing these communities to the history of exclusionary zoning practices in the United States. Although the Supreme Court ruled in 1948 that racially restrictive covenants were illegal (Shelley v. Kraemer), McKenzie argues “the principle is still the same: certain ____________ 3The average annual growth rates for each type of housing share in each period were 10 and 2 percent, respectively. 4In theory, common interest developments may also improve efficiency in the allocation of public goods by allowing individuals to “vote with their feet” among an expanded array of public and private options for funding and receiving collective services (e.g., Tiebout, 1956). 3 CIDs as a % of total units Suburban housing as a % of total units United States 50 40 Suburb units CID units 30 20 10 0 1940 1960 1975 1990 16 14 12 10 8 6 4 2 0 CIDs as a % of total units Suburban housing as a % of total units 60 50 40 30 20 10 0 1940 1950 California 1960 1970 1980 1990 25 20 15 10 5 0 SOURCE: Treese (1999); HOA-Info (2002); Devaney (1994); U.S. Census Bureau (1970, 1980, and 1990). Figure 1.2—Growth of Common Interest Developments and Suburban Housing Units in California and the United States, 1940–1999 groups of people are considered a threat to property values and are excluded. The result is still increased homogeneity and . . . continuing racial segregation” (p. 78). Other social commentators have expressed concern about the effects of CIDs on political participation and civic engagement. Robert Putnam (2000, p. 210), for instance, links the growth of CIDs to declines in social capital: Not only are canvassing politicians and Girl Scouts selling cookies excluded from exclusive communities, but the affluent residents themselves also appear to have a surprisingly low rate of civic engagement and neighborliness even within their boundaries. One reason to expect lower rates of political participation in CIDs is that the homogeneity of these communities reduces the incentives, or 4 controversies, that draw citizens into civic affairs (Oliver, 2001).5 Alternatively, if planned developments are substitutes for traditional local government, then residents of these developments may perceive less benefit to taking part in elections and other aspects of public life. For instance, in their survey of gated communities in particular, Blakely and Snyder assert that residents of these developments “have less need to participate [in the wider community] . . . because they live in gated enclaves, with private recreation, roads, parks, and security” (1997, p. 72). A related concern is that members of common interest developments may vote differently from nonmembers. In particular, CID residents may favor public spending that complements private expenditures (e.g., police protection) but oppose that which is duplicative (e.g., infrastructure) or redistributive (e.g., welfare). Simply put (“Government by the Nice . . . ,” 1992, p. 25): If affluent Americans choose to live in private communities which raise their own taxes but do not redistribute them outside their walls, they are likely to vote to cut spending on public services that they do not use, ignoring the needs of people who cannot afford to go private. Such criticisms have promoted repeated calls for enhanced regulation of these communities. Over the years, the state legislature has convened several committees and working groups to review policies toward common interest developments. Although the bulk of proposed legislation has focused on the internal governance of CIDs, a frequent complaint is that there are no current and reliable studies of common interest developments in California (e.g., Senate Housing and Community Development Committee, 2002; California Law Revision Commission, 2002a). A California Research Bureau report commissioned by the Senate Housing and Land Use Committee recommended a new legislative subcommittee “to provide more up-todate empirical research on [homeowner associations] and related issues” (Roland, 1998, p. 44). ____________ 5See Alesina and La Ferrara (2000) for an opposing view, i.e., that heterogeneity depresses participation in voluntary associations. 5 The primary reason for the lack of empirical research on CIDs is limited data. The federal government only recently began collecting information on gated communities and residential associations in its 2001 American Housing Survey. As a result, most research on common interest developments has relied on small-scale surveys and case studies (e.g., McKenzie, 1994; Foldvary, 1994; Blakely and Snyder, 1997), with the exception of Barton and Silverman (1994) whose results, based on a 1987 survey, are largely out of date. Given the absence of research in this area, there is a great deal of uncertainty regarding common interest developments among state and local lawmakers, policy analysts, and CID managers and other professionals, boards of directors, and residents themselves. An improved understanding of these communities is an important prerequisite to the formation of sound policies toward private government. This report seeks to fill the gap in knowledge about common interest developments in California. It focuses on planned developments, in particular, because these are the most prevalent and fastest-growing type of CID in both California and the nation as a whole. These developments also represent the closest analogue to traditional local government and hence the greatest source of controversy regarding common interest developments. Among the questions addressed in this report are: • What are common interest developments? What types of housing structures do they include? What kinds of services do they provide? How are they similar to and different from traditional local governments? • What do common interest developments in California look like? How many CIDs are there? How are they distributed regionally? How do they vary in type, size, age, and financial resources? • What about planned developments in particular? Where are they located? How do they compare to other neighborhoods in terms of their demographic and socioeconomic characteristics (e.g., racial and ethnic composition, age structure, income distribution)? Do they contribute to overall residential segregation? 6 • Do planned developments represent the “secession of the successful”? Do these communities exhibit lower levels of civic engagement, as measured by rates of voter turnout and registration? Do they have a different political ideology, as indicated through party affiliation and votes on statewide ballot propositions? The plan for the remainder of the report is as follows. Chapter 2 compares the origins, institutional features, and powers and responsibilities of common interest developments to traditional local government. Chapter 3 presents a portrait of common interest developments in California, highlighting the diversity among these communities. Chapter 4 analyzes the demographic and socioeconomic composition of planned developments, in particular, and their contribution to overall residential segregation. Chapter 5 compares voter turnout and patterns of voting behavior in planned developments to similar neighborhoods, using regression analysis to control for relevant demographic and socioeconomic factors. Chapter 6 concludes and offers directions for state and local public policy. 7 2. What Are Common Interest Developments? This chapter presents an overview of common interest developments in the United States. It traces the evolution of these communities from Ebenezer Howard’s “garden cities of tomorrow” to the modern CID. The chapter also develops a definition and taxonomy of common interest developments. It concludes with a discussion of the scope of activities of these so-called “private governments” and issues for policymakers. A Brief History Alexis de Tocqueville observed that the principle of voluntary association had nowhere in the world “been more successfully used, or unsparingly applied to a multitude of different objects, than in America.” He further described these associations as “a separate nation in the midst of the nation, a government within the government” (1969 [1835], pp. 513–517). Yet the idea of residential associations based on collective property ownership originated in England. During the mid-1700s, Lord Leicester established a park in Leicester Square and charged adjacent property owners an assessment for their exclusive enjoyment of it. In 1837, private developers built Victoria Park, a community with deed restrictions tied to the sale of each lot to protect local amenities (Foldvary, 2002, p. 274). Perhaps the most important antecedent to the modern common interest development was Ebenezer Howard’s utopian ideal of the garden city. Howard sought to combine the best aspects of city and country life in his “new towns” of Welwyn and Letchworth, England. Both cities were based on a leasehold concept, whereby a group of trustees would own the municipality and collect rents from its residents to pay off initial construction loans, build infrastructure, and provide collective services. Howard characterized the political organization as a technocracy, 9 “‘modelled upon that of a large and well-appointed business.’” In his view, this governance structure would be more powerful than other local bodies and “‘solve to a large extent the problem of local selfgovernment’” (McKenzie, 1994, pp. 5–6). Precursors to the CID in the United States—Gramercy Park in New York City (established in 1831), Louisburg Square in Boston (1844), and South Park in San Francisco (1852)—were formed primarily to maintain private amenities. Builder Jesse Clyde Nichols established the first fullservice common interest development, Mission Hills, Missouri, in 1914 with the express intention of supplanting local government. According to one study, “he was afraid if the subdivision were part of a larger village or town organization the political unit would not be sufficiently responsive to the needs of his residents” (Worley, 1990, pp. 166–167). During the 1920s, the reform-minded Regional Planning Association of America revived the idea of a “garden city” with the new town of Radburn, New Jersey. In addition to its innovative physical plan, the community included a homeowner association with broad land use regulation and service responsibilities (Barton and Silverman, 1994, p. 9). The popularity of common interest developments burgeoned during the post–World War II housing boom and rise of “community builders,” who favored large-scale, dense forms of development. In 1961 and 1963, the Federal Housing Administration encouraged this form of housing by publishing guidelines for condominiums and planned developments to obtain mortgage insurance. The Federal National Mortgage Association and Federal Home Loan Mortgage Corporation later approved condominium and planned development purchases for secondary mortgage loan markets. State and local governments also codified real estate practices and modified building and zoning laws in the 1960s to accommodate CID housing.1 This period of growth culminated in the emergence of large-scale master-planned communities such as Irvine, California, and Reston, Virginia. ____________ 1For instance, many localities implemented planned unit development (PUD) zoning, allowing more dense or clustered housing than would otherwise be permitted in exchange for developer-provided amenities, such as parks and recreation. It is important to note, however, that not all planned developments are zoned PUD and vice versa. 10 Following a few highly visible insolvencies, the late 1960s and early 1970s brought a series of reforms to common interest developments. The Urban Land Institute and National Association of Home Builders jointly established the Community Association Institute to provide technical assistance to homeowner associations on financial management and governance. At the same time, several states adopted model condominium and planned development codes establishing financial reporting requirements for these communities. Faced with the fiscal strains of the 1970s and 1980s, local governments further promoted CID housing as a means of offsetting the service and infrastructure costs of new development. In some instances, jurisdictions even required that developers establish homeowner associations as a condition of new project approval (Barton and Silverman, 1994, p. 11). Definition and Taxonomy There are three main kinds of common interest developments: planned developments, condominiums, and cooperatives. Each type involves a different legal assignment of property rights. In a planned development, the homeowner owns his or her individual unit and lot, and the homeowner association, usually incorporated as a nonprofit mutual benefit corporation, owns and maintains all common property. In a condominium, the association manages but does not own the common elements; each unit owner holds an undivided interest in these areas as a “tenant in common.” In a cooperative, a corporation owns and maintains the entire property. Individuals do not own units but hold proprietary lease agreements or membership documents granting exclusive use of a particular unit for a specified period of time. Figure 2.1 shows that nearly 64 percent of common interest development units in the United States are within planned developments, whereas 31 percent are in condominiums and 5 percent are in cooperatives. Planned developments are also the fastest-growing type of CID, with the number of planned development units increasing fourfold since 1985. 11 Number of units (1000s) 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 1962 1970 1975 1980 1985 1990 1995 1999 Planned developments Cooperatives Condominiums SOURCE: Treese (1999). Figure 2.1—Growth of Common Interest Developments in the United States, by Type These legal definitions can mask substantial variation. Groups of detached single-family homes can be legally organized as condominiums and, more rarely, apartment complexes can be planned developments (Barton and Silverman, 1994, p. 5). Table 2.1 illustrates a more useful distinction between “territorial” CIDs encompassing multiple buildings on large tracts of land and “nonterritorial” CIDs Table 2.1 An Example of Nonterritorial and Territorial CIDs Number of buildings Number of units Amenities Nonterritorial 1 6 None Services Common property Governance Management City garbage dumpster, gardener Halls, outside, 1 acre lot Each unit has one member on homeowner asosciation board; majority rule Board SOURCE: Roland (1998, p. 9). Territorial 105 100 Pool, tennis, private parks, health club Garbage, security, street maintenance All except home lots (75 acres) Eight board members elected from 100 wth full authority unless law requires vote Off-site management company 12 within single high-rise buildings (Advisory Commission on Intergovernment Relations, 1989). The former category constitutes 60 to 80 percent of all common interest developments (Barton and Silverman, 1994; Treese, 1999). Scope of Activities Territorial common interest developments resemble traditional local governments in that they are responsible for providing services in exchange for regular assessments from residents. Many of these services are similar to those that would be provided by a municipality (see Table 2.2). Homeowner associations are also responsible for enforcing the “constitutions,” or CCRs of the community. CCRs typically regulate property uses such as exterior renovations, paint colors, fences, and antennas. They may also restrict individual conduct or “lifestyle,” Table 2.2 Services Provided by Common Interest Developments in the United States and California California Service Open space or lawns Lawn care or gardening Parks or playgrounds Meeting places Garbage disposal Parking lots or structures Swimming facililites Roads Water or sewer lines Other recreational facilities Entry guards or security patrols United States % of CIDs Service % of CIDs 92 Snow removal 77 87 Pools 76 84 Garbage disposal 64 72 Clubhouse 56 65 Street cleaning 54 62 Street lighting 53 58 Tennis 45 43 Playgrounds 37 42 Lake or pond 30 21 Park area 29 15 Security patrols 12 SOURCES: Barton and Silverman (1994, p. 5); Treese (1999, pp. 7, 13). NOTES: Table reports services commonly provided by respondents in two separate survyes. National survey responses are selected from developments of 151 to 350 units. 13 including pet ownership and noise levels.2 Crafted by the developer, who retains control of the association until 50 to 75 percent of units are sold, these restrictions “run with the land,” meaning that they are binding on all subsequent property owners. Amendments to CCRs usually require approval from a supermajority of homeowners. Common interest developments differ from public entities most notably in that they lack the police power of local governments. They must rely on the courts to enforce CCRs or collect delinquent fees.3 Typically, associations are also not subject to the same constitutional and statutory requirements as public entities regarding elections, public meetings, and public access.4 Moreover, CID boards of directors usually do not share public officials’ immunity from personal liability. In California, however, amendments to the 1985 Davis-Stirling Act institute open meeting requirements for CIDs parallel to those of the Brown Act for state and local governments. They also ensure that board members cannot be held personally liable if they meet certain standards of care and maintain minimal liability insurance (Roland, 1998, pp. 22–24). Policy Issues The dual public and private nature of common interest developments raises a host of policy questions. At the state level, there ____________ 2Legal scholars have noted the seemingly arbitrary and capricious nature of these rules. For example: • In Boca Raton, Florida, a homeowners association resident was required to attend a court-supervised weigh-in of his 29-1/2-pound dog because the association’s CCRs specified a 30-pound limit for pets. • In Santa Ana, California, a 51-year-old grandmother received a warning citation from her condominium association for kissing a friend good night in her driveway. • Bylaw VI 7(k) of the Bailey, a condominium in Washington, D.C., provides that “No Barry Manilow records, tapes or CDs may be owned or played on the premises.” See Kennedy (1995) and Kress (1995). 3CCRs are “equitable servitudes,” or land use restrictions enforceable in a court of equity. See Black’s Law Dictionary (1979, p. 484). 4For instance, voting rights in CIDs are limited to homeowners only and may be apportioned on the basis of unit size or value rather than equal representation (see Barzel and Sass, 1990). 14 have been recurrent calls for the enhanced regulation of these communities. The bulk of proposed legislation concerns the internal governance of CIDs. Among the major bills passed during the 2003– 2004 legislative session were: • AB 104 (Lowenthal)—requires that homeowner associations make accounting books and records as well as minutes of proceedings available to all members for inspection and copying; • AB 1423 (Dutra)—amends and clarifies a bill passed in the prior session (AB 555) instituting continuing education and examination requirements for property managers to be called “certified common interest development managers”; and • AB 1525 (Longville)—prohibits homeowner associations from banning the posting or display of noncommercial signs or banners except for the protection of public health or safety or if the posting violates a local, state, or federal law. Over the years, the legislature has also appointed several committees and working groups to review state policies toward common interest developments. In 1983, the assembly convened a select committee to streamline previous codes. Their efforts culminated in the 1985 DavisStirling Common Interest Development Act (California Civil Code §§ 1350 et seq.), the main body of law governing CIDs in California. Since its adoption, the act has been amended more than 40 times. As a result, it is a cumbersome piece of legislation that has been criticized as complicated and hard to understand, contradictory, uneven in its coverage, difficult to enforce and, offering weak protection for individual rights (French, 2000). In 1999, the legislature authorized the California Law Revision Commission to undertake a multiyear study of common interest law, including the Davis-Stirling Act and other key statutes. The objectives of the study are: • to set a clear, consistent, and unified policy with regard to the formation and management of CIDs and the transaction of real property interests located within them; 15 • to clarify the law, eliminate unnecessary or obsolete provisions, consolidate existing statutes in one place in the codes; and, • to determine to what extent common interest housing developments should be subject to regulation (California Law Revision Commission, 2002b, p. 3). Thus far, the commission has recommended instituting requirements for procedural fairness in homeowner association rulemaking and decisionmaking, as well as a nonsubstantive reorganization of the DavisStirling Act. Assemblywoman Patricia Bates introduced implementing legislation for these changes (AB 512) in the 2003–2004 session. The commission has also recommended improving existing alternative dispute resolution requirements and requiring that common interest development associations provide internal dispute resolution mechanisms. Several emerging policy issues also have yet to be addressed by the state legislature. A major issue for CID residents is “double taxation,” or their responsibility for local property taxes even if they are consuming private alternatives to public services. Some state and local governments have agreed to property tax rebates for CID residents subject to certain public access requirements. For example, New Jersey’s Municipal Services Act requires that local governments reimburse CID homeowner associations for street sweeping, garbage collection, snow plowing, and street lights provided that the CID accepts public dedication of its roads or streets (New Jersey Statutes, § 40:67-23-3). Similar arrangements exist in Houston, Texas; Kansas City, Missouri; and Montgomery County, Maryland (Dilger, 1992).5 A related concern is that local real property taxes are deductible under the federal Internal Revenue Code but homeowner assessments are not. According to Executive Council of Home Owners (ECHO) lobbyist Robyn Boyer Stewart, “It is only a matter of time before the tax-and-equity bomb blows. . . . The politician ____________ 5Where a CID is located exclusively on private property, local governments may not provide routine services such as police patrols, trash collection, and animal control because of litigation concerns unless they have negotiated prior agreements with the CID (Advisory Commission on Intergovernmental Relations, 1989). 16 who manages to capture this constituency, speak to its needs, and offer it a voice will be amply rewarded” (Kennedy, 1995, p. 775). Nonresidents, on the other hand, are concerned about the social consequences of common interest developments. Critics of CIDs argue that these communities “compete with cities for the affluent, siphoning off their tax dollars, their expertise and participation, and their sense of identification with the community” (McKenzie, 1994, p. 23). The chapters to follow provide an empirical basis for this debate by analyzing the demographic, socioeconomic, and political features of common interest developments in California. 17 3. What Do California CIDs Look Like? This chapter surveys common interest developments in California. It highlights the diversity among CIDs, presenting breakdowns of the main attributes of these developments by type and over time. It concludes with an examination of the financial resources of common interest developments and their relative position among local governments in California. Note on Data Sources The chapters above noted the lack of previous empirical studies on CIDs because of data limitations. For example, the federal government only recently began collecting information on community associations in the 2001 American Housing Survey. Various states, however, do collect this information as part of their regulatory oversight.1 In California, developers of subdivided land consisting of five or more lots or units must file a public report with the Department of Real Estate (DRE) before offering units for sale (California’s Subdivided Lands Act, California Business and Professions Code §§ 11000-1200). The report must include the location and size of the development as well as the governing documents and proposed budget of the homeowner association (Roland, 1998, p. 17). Each addition or material change to the development requires an amendment to the initial DRE report. Although not required by law, most homeowner associations also incorporate as nonprofit mutual benefit organizations.2 The Davis____________ 1Florida, Virginia, Maryland, and Nevada have reporting requirements similar to those in California. 2Unincorporated homeowner associations do not receive the tax benefits or liability protections of incorporated CIDs, although they are required to follow most of the same laws (Lave-Johnston and Johnston-Dodds, 2002). 19 Stirling Act requires that incorporated associations file articles of incorporation with the Secretary of State (SOS), along with a statement identifying the corporation as a CID, the location of the association's business or corporate office, and the name and address of the association's managing agent. A separate statement of “principal business activity,” listing three officers, their addresses, and an agent's name and address for service of process, must be filed within 90 days of incorporation and every two years thereafter. Information from these two sources is available through HOA-Info, a proprietary database compiled by Levy and Company, CPAs of Oakland, California. These data were supplemented with other sources, such as homeowner association websites and directories as well as property tax assessment, deed transfer, and mortgage records. The result is the most comprehensive existing database on common interest developments in California.3 A Large and Diverse Group There are currently over 36,000 common interest developments— including planned developments, condominiums, and cooperatives—in California. These communities include over three million CID housing units, equivalent to approximately one-quarter of the state’s housing stock (Table 3.1). The prevalence of common interest developments varies substantially by region. The South Coast has the most developments and CID units in the state, whereas the Sacramento metropolitan area and San Diego region exhibit the highest concentration of common interest developments as a share of all housing units (42 percent). Roughly one-third of all CIDs in California are PDs and two-thirds are condominiums and cooperatives (Table 3.2). These proportions are nearly reversed when one considers units rather than developments. About 61 percent of all California CID units are in planned developments, and 39 percent are in condominiums or cooperatives. ____________ 3There are remaining gaps in the data. Specifically, the number of units is missing for 22 percent of all records, acreage is missing for 31 percent, and fee and revenues are missing for 61 percent. 20 Table 3.1 Geographic Distribution of Common Interest Developments Region Sacramento Metro San Diego Inland Empire Sierras Far North South Coast Bay Area Central Coast San Joaquin Valley Total No. of CIDs 1,325 4,772 2,250 312 663 16,311 7,769 1,459 1,321 36,182 CID Units 301,329 450,256 450,584 31,308 106,504 956,613 491,523 77,731 144,822 3,010,670 % of Total Housing Units 42 42 38 32 22 21 19 16 13 25 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: The counties included in each region are as follows: Far North: Butte, Colusa, Del Norte, Glenn, Humboldt, Lake, Lassen, Mendocino, Modoc, Nevada, Plumas, Shasta, Sierra, Siskiyou, Sutter, Tehama, Trinity, and Yuba. Sacramento Metro: El Dorado, Placer, Sacramento, and Yolo. Sierras: Alpine, Amador, Calaveras, Inyo, Mariposa, Mono, and Tuolumne. Bay Area: Alameda, Contra Costa, Marin, Napa, San Francisco, San Mateo, Santa Clara, Solano, and Sonoma. San Joaquin Valley: Fresno, Kern, Kings, Madera, Merced, San Joaquin, Stanislaus, and Tulare. Central Coast: Monterey, San Benito, San Luis Obispo, Santa Barbara, and Santa Cruz. Inland Empire: Riverside and San Bernardino. South Coast: Los Angeles, Orange, and Ventura. San Diego: Imperial and San Diego. Common interest developments have grown exponentially in California since the early 1960s. As in the rest of the nation, planned developments are the fastest-growing type of CID, increasing by 84 percent since 1985 (Figure 3.1). Although planned developments are on average larger than condominiums and cooperatives, there is considerable overlap in the size distributions of these two types of CID (Figure 3.2). Planned 21 Table 3.2 Types of Common Interest Developments in California Type Planned developments % of total Condominiums and cooperatives % of total Total Developments 9,313 33 19,310 67 28,623 Housing Units 1,495,913 61 938,857 39 2,434,770 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Figures do not include developments of unknown or “other” types of associations in business parks, mobile homes, lofts, timeshares, roads, airports, and docks. 2,500 Number of units (1000s) 2,000 1,500 1,000 500 0 1962 1970 1975 1980 1985 1990 1995 1999 Planned developments Cooperatives Condominiums SOURCE: HOA-Info (2002). NOTE: Figure does not include CIDs of unknown type. Figure 3.1—Growth of Common Interest Developments in California, by Type developments can include as few as three units and condominiums can encompass as many as 27,000 units at build-out. The average planned development includes 168 units and the average condominium or cooperative has 52 units. There is a trend toward smaller common interest developments over time, as evidenced in Table 3.3. However, recent years have witnessed a slight up-tick in the average size of CIDs, as measured by both the number of units and total acreage. 22 Number of developments 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0 1–5 6–10 11– 16– 21– 26– 51– 76– 101– 151– 201– 301– 501– 1,001+ 15 20 25 50 75 100 150 200 300 500 1,000 Number of units Planned developments Condominiums and cooperatives Source: HOA-Info (2002). Figure 3.2—Size Distribution of Common Interest Developments in California, by Type Table 3.3 Average Size of Common Interest Developments, by Year of Construction Year Built Pre-1940 1940–1949 1950–1959 1960–1969 1970–1979 1980–1989 1990–1995 1996–2002 Planned Developments Units Acres 350 135 190 14 155 40 111 28 76 14 47 11 39 10 60 15 Condominiums and Cooperatives Units Acres 11 1 11 1 24 1 36 2 24 1 15 1 14 1 18 1 SOURCE: HOA-Info (2002). NOTE: Cells include the median number of units or acres for each type of CID. 23 The dataset used in this study does not include information on the services provided by each common interest development. One proxy measure of service activity is revenue collected by the homeowner association through monthly assessments per unit. (This measure does not include funds raised through special assessments for nonroutine expenses.)4 As with other CID characteristics, monthly assessments exhibit wide variation, ranging from merely nominal amounts to thousands of dollars per unit per month. Three-quarters of all reporting planned developments charged at least $68 per unit per month in 2002, with a median monthly assessment of $112. Similarly, three-quarters of all condominiums and cooperatives charged monthly fees of at least $138, for a median of $186 per unit per month (Table 3.4).5 Monthly fees translate into varying annual revenues for homeowner associations (Figure 3.3). Total association revenues were estimated at $6.3 billion in 2003 (HOA-Info, 2003). Thus, homeowner associations are a major presence in California. Subsequent chapters will focus on planned developments—the CIDs that most resemble local governments—and their consequences for neighborhood diversity and segregation as well as civic engagement. Table 3.4 Distribution of Monthly Fees of Common Interest Developments Planned developments Condominiums or cooperatives 5 $25 $81 SOURCE: HOA-Info (2002). 25 $68 $139 Percentile 50 $112 $186 75 $168 $249 95 $313 $443 ____________ 4The HOA-Info database includes assessments reported by the developer to the California Department of Real Estate before offering units for sale. These figures were supplemented with current financial data, where available, or inflated to current dollars assuming a 4 percent annual inflation rate. 5Anecdotal evidence suggests that condominium fees are higher because they generally offer more common areas and amenities. 24 3,000 2,500 Number of developments 2,000 1,500 1,000 500 0 0–25 26–50 51–75 76–100 101– 201– 200 300 Thousands of dollars 301– 400 401– 1,000+ 500 Planned developments Condominiums and cooperatives SOURCE: HOA-Info (2002). Figure 3.3—Revenues of Common Interest Developments in California 25 4. Focus on Planned Developments The remainder of this report focuses on planned developments. This is the most prevalent and fastest-growing type of common interest development housing in both California and the nation. These CIDs also bear the strongest resemblance to traditional local government and attract the most scrutiny for their potential policy consequences. This chapter presents a portrait of planned developments in California.1 It first surveys the regional distribution of these communities and the types of places in which they are located. The chapter then compares the demographic and socioeconomic characteristics of planned developments to characteristics of other neighborhoods. It concludes with a detailed examination of the racial and ethnic composition of PDs and their contribution to neighborhood diversity and metropolitan area segregation. Geographic Distribution Planned developments are primarily concentrated in urbanized regions of the state, including the South Coast, Bay Area, and San Diego region (Table 4.1). They represent particularly large shares of the housing stock in fast-growing regions such as the Inland Empire and Sacramento metropolitan area as well as in certain less urbanized areas, such as the Sierras and Far North. ____________ 1The analysis presented in this chapter and the remainder of this report relies on the results of a geographic matching procedure described in Appendix A. Comparisons between planned developments and other neighborhoods are based on census block groups with and without PDs. Census block groups are the smallest level of geography at which socioeconomic variables from the U.S. Census are publicly available. There were 22,133 block groups in California in 2000, of which about 16 percent had at least one planned development in 2002. 27 Table 4.1 Planned Developments, by Region Region Bay Area Central Coast Far North Inland Empire Sacramento Metro San Diego San Joaquin Valley Sierras South Coast Total No. of Planned Developments 2,475 722 316 907 646 960 648 96 2,530 9,300 % Single- % Population Population % of All Family Growth, per Square Housing Housing 1990–2000 Mile, 2000 10 17 13 980 9 13 12 121 15 22 11 26 17 24 26 119 15 21 21 353 17 29 13 353 10 14 20 121 26 34 16 9 10 18 10 1,959 12 19 14 217 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). Approximately two-thirds of planned developments are in suburbs, or in metropolitan areas but beyond the central city (Table 4.2). Contrary to the notion that homeowner associations in PDs are substitutes for traditional local government, over three-quarters of PDs are within incorporated cities. Most of these cities were incorporated before 1950 and have populations of less than 100,000 (Table 4.3). Demographic and Socioeconomic Characteristics Planned developments differ from other neighborhoods along a number of dimensions.2 Table 4.4 indicates that, on average, planned developments have significantly higher percentages of the population who are white, in both central city and suburban areas (60 and 65 percent, respectively).3 Correspondingly, they have lower proportions who are black (4 percent in central cities and 3 percent in suburban ____________ 2Comparisons between planned developments and other neighborhoods refer to census block groups with and without PDs. See Appendix A for details on the procedure used to match PD locations to census geography. 3All reported demographic and socioeconomic information is from the 2000 Census unless otherwise noted. Racial and ethnic groups include only individuals who identify themselves as belonging to one race. The terms “white,” “black,” and “Asian” in the text refer to non-Hispanic members of these groups. 28 Table 4.2 Planned Developments, by Central City/Suburban Location and City Incorporation Status Incorporated city Unincorporated area Total Central City 1,648 — 1,648 Suburb 2,550 1,088 3,638 Nonmetropolitan Area 65 161 226 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). Total 4,263 1,249 5,512 Table 4.3 Planned Develoments, by City Size and City Incorporation Date City Population Less than 1,500 1,500–1,999 2,000–2,499 2,500–4,999 5,000–9,999 10,000–19,999 20,000–24,999 25,000–49,999 50,000–99,999 100,000–249,999 250,000–499,999 500,000–999,999 1,000,000–2,499,999 2,500,000–4,999,999 Total Before 1950 2 4 6 69 110 45 355 823 704 257 170 152 131 2,828 City Incorporation Date 1950– 1960– 1970– 1980– 1990– 1959 1969 1979 1989 2000 Total 2 13 4 4 7 11 2 26 8 19 2 18 2 118 23 28 34 16 18 229 6 20 10 31 4 116 173 63 119 124 21 855 169 65 12 83 43 1,195 151 58 55 20 988 257 170 152 131 538 267 232 292 90 4,247 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). areas) and Hispanic (20 and 19 percent, respectively). Differences in Asian population shares are not significant between planned developments and other neighborhoods in central cities. They are significant, but in the opposite direction for the suburbs. That is, planned developments tend to have larger Asian populations in these areas. 29 Table 4.4 Selected Demographic Characteristics of California Planned Developments (in percent) Characteristic Black Hispanic Asian White Ages 0 to 19 Ages 20 to 39 Ages 40 to 64 Age 65 and up Central City Planned Non-PD Developments Neighborhoods 4.2 9.8** 20.1 33.5** 11.7 12.3 60.2 40.8** 25.1 28.3** 30.2 33.2** 30.7 27.6** 14.0 10.9** Suburb Planned Non-PD Developments neighborhoods 2.7 5.1** 18.8 32.6** 10.3 9.2** 64.6 49.5** 27.0 30.3** 26.1 28.7** 33.6 29.6** 13.3 11.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. **Denotes statistically significant difference at the 1 percent level. Planned developments also exhibit significant differences in age structure compared to other neighborhoods. Although a few well-known developments cater primarily to seniors (e.g., Leisure World, incorporated as Laguna Woods in 1999), planned developments as a group are more heterogeneous with respect to age. On average, planned developments include more individuals age 65 and older than other neighborhoods (differences of 2 to 3 percentage points, depending on location). However, they also include more residents ages 40 to 64 (differences of 3 and 4 percentage points for central cities and suburbs, respectively). They have correspondingly smaller population shares of younger persons and also slightly fewer married couples with children than other neighborhoods (Table 4.5). There are major socioeconomic differences between planned developments and other neighborhoods as well (Table 4.5). Planned developments had significantly higher median annual household incomes in 1999, by about $12,000 in central cities and $14,000 in suburban areas. They also tend to have more households in the highest income categories and fewer households in the lowest categories. For example, 30 Table 4.5 Selected Socioeconomic Characteristics of California Planned Developments Central City Suburb Planned Non-PD Planned Non-PD Characteristic Developments Neighborhoods Developments Neighborhoods Median household income in 1999, $ 58,687.8 46,688.3** 67,045.9 53,297.9** Household income, % Less than $35,000 32.9 42.9** 26.8 35.6** $35,000–$49,999 14.2 15.0** 13.7 15.5** $50,000–$74,999 19.1 17.5** 19.8 20.1 $75,000–$99,999 12.4 10.0** 13.5 12.0** $100,000–$199,999 16.6 11.6** 19.5 13.5** $200,000 and up 4.9 3.0** 6.6 3.3** % Married couples with children 30.9 31.9** 34.4 36.4* % with no college 28.0 40.4** 26.5 38.8** % in managerial and professional occupations 42.6 33.8** 42.3 32.4** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. nearly 22 percent of PD households in central cities earned more than $100,000 in 1999, and 26 percent in suburban areas earned at least this amount (compared to 15 and 17 percent for non-PD neighborhoods). Individuals residing in planned developments are also more likely to have attended college (differences of roughly 12 percentage points in both central city and suburban areas) and to work in managerial and professional occupations (differences of 9 to 10 percentage points in each area). In addition, there are several distinguishing features of planned development housing itself. Homes in these communities are generally newer and of a higher median value ($280,000 in central cities and $324,000 in suburban areas). They also are more likely to be singlefamily structures and owner-occupied (Table 4.6). 31 Table 4.6 Selected Housing Characteristics of California Planned Developments Characteristic Housing density (units per square mile) % single-family units % owner-occupied units Median year built Median home value Central City Suburb Planned Non-PD Planned Non-PD Developments Neighborhoods Developments Neighborhoods 2,853.9 64.8 59.8 1,968.3 280,189.8 4,806.2** 60.2** 49.5** 1,947.8** 233,449.4** 1,839.3 75.1 69.7 1,971.9 323,518.1 2,482.0** 73.7* 63.3** 1,960.2** 231,384.0** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Table includes sample means for each subgroup. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. Many observed demographic differences between planned developments and other neighborhoods persist even after controlling for housing characteristics. Notably, planned developments have fewer blacks and Hispanics, holding constant the median home value and the decade when the majority of units were built. Differences between planned developments and other neighborhoods narrow at higher housing market values, however (Tables 4.7 and 4.8). In sum, planned developments conform in many ways to the popular image of “the secession of the successful.” Residents of these communities are on average white, older, and better educated, and earn higher incomes. Nevertheless, PDs are also more heterogeneous than their image as exclusive luxury enclaves would suggest. Although they house more of the highest-income Californians, planned developments also include many residents in the middle- to higher-income categories (those with annual earnings of $35,000 to $49,999, $50,000 to $74,999, and $75,000 to $99,999). In addition to the age 65 and older category, there are sizable population shares in the younger age groups. Differences in the racial composition of planned developments also diminish somewhat after controlling for housing characteristics such as year of construction and median value. 32 Table 4.7 Percentage Black, by Planned Development Status and Housing Characteristics 33 Year of Construction Before 1960 1960–1969 1970–1979 1980–1989 1990–1999 Low Non-PD Planned Neighborhoods Developments 11.1 6.2** 7.5 3.4** 6.1 4.2** 8.1 4.9** 7.5 5.8 Median Home Value Medium Non-PD Neighborhoods Planned Developments 8.8 3.7** 5.8 3.5 4.7 3.0** 5.5 4.2* 5.4 4.0* High Non-PD Planned Neighborhoods Developments 4.1 2.3** 2.2 1.6 2.6 1.7** 2.7 2.0 3.2 2.3* SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTES: “Low” home values are the first through 32nd percentile for the state ($149,100) in the 2000 Census; “medium” home values are the 33rd through 65th percentile for the state ($149,200 to $244,700); and “high” home values are the 66th through 100th percentile for the state ($244,800 and above). *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. Table 4.8 Percentage Hispanic, by Planned Development Status and Housing Characteristics 34 Year of Construction Before 1960 1960–1969 1970–1979 1980–1989 1990–1999 Low Non-PD Planned Neighborhoods Developments 49.5 34.5** 38.2 26.3** 31.4 23.1** 36.9 29.7** 37.7 29.0** Median Home Value Medium Non-PD Neighborhoods Planned Developments 44.1 33.0** 32.0 24.0** 26.0 20.7** 23.9 16.7** 22.4 16.1** High Non-PD Planned Neighborhoods Developments 15.8 13.6** 14.5 11.7* 14.3 11.3** 11.9 10.5 15.2 12.0** SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTES: “Low” home values are the first through 32nd percentile for the state ($149,100) in the 2000 Census; “medium” home values are the 33rd through 65th percentile for the state ($149,200 to $244,700); and “high” home values are the 66th through 100th percentile for the state ($244,800 and above). *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. Diversity and Segregation This section traces the consequences of the different racial and ethnic composition of planned developments for neighborhood diversity and overall metropolitan area segregation. It relies on the entropy index of diversity, which measures how evenly the population of any given area is distributed across the major racial and ethnic groups.4 The index reaches a maximum of 100 when all groups are equally represented and a minimum of zero when only one group is present (see Appendix B for a fuller discussion of the index). Planned developments are often less racially and ethnically diverse than other neighborhoods, even after controlling for whether the neighborhood is in a suburb or a central city (Table 4.9). At the most extreme, in the Sacramento area, suburban planned developments have an average diversity score of 43, compared to 55 for other neighborhoods. On the other hand, in some metropolitan areas, planned developments are more diverse than or about the same as other neighborhoods (e.g., Los Angeles–Riverside–Orange County). A general definition of residential segregation is the extent to which individuals belonging to different groups live separately from one another (Massey and Denton, 1988). The entropy index of segregation relates the diversity of an area as a whole to the average diversity of its constituent parts (see Appendix B). It describes, for example, how different the racial and ethnic composition of a city or metropolitan area is from its neighborhoods or, equivalently, how neighborhoods within a given city differ from each other. The index ranges from a minimum of 0, when all neighborhoods have the same racial and ethnic composition and hence the same diversity, to a maximum of 100, under complete segregation or where the diversity of each neighborhood is 0. An attractive property of the entropy index is that it can be decomposed into “between” and “within” components (see Appendix B). The “within” component is the portion of segregation attributable to ____________ 4This report considers the four largest racial and ethnic groups in California— Hispanic and non-Hispanic white, black, Asian and Pacific Islander—as well as an “other” category encompassing individuals who do not fit any of the above categories, those of more than one race, and American Indians. 35 Table 4.9 Average Diversity Scores of California Neighborhoods, by Metropolitan Area, Planned Development, and Central City/Suburban Status Central City Suburb Planned Non-PD Planned Non-PD Metropolitan Area Developments Neighborhoods Developments Neighborhoods Bakersfield 61 61 43 47 Chico–Paradise 36 39 32 43 Fresno 63 65 45 46 Los Angeles–Riverside– Orange County 57 56 55 54 Merced 63 68 56 57 Modesto 58 60 51 53 Redding 33 35 25 30 Sacramento–Yolo 70 74 43 55 Salinas 50 53 53 47 San Diego 56 59 50 57 San Francisco–Oakland– San Jose 62 65 57 60 San Luis Obispo– Atascadero–Paso Robles 41 46 39 41 Santa Barbara–Santa Maria–Lompoc 51 50 41 51 Stockton-Lodi 68 71 59 59 Visalia–Tulare–Porterville 52 56 53 41 Yuba City 67 59 61 59 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Each table entry is the population-weighted average diversity of census block groups in that category. racial and ethnic differences among neighborhoods of a given type, for example, central cities or suburbs. The “between” component is the remaining portion of segregation, which cannot be reduced without individuals relocating. By extension, central city and suburban segregation can be decomposed into differences among planned developments, as well as between planned developments and other neighborhoods. The latter component is the portion of overall segregation that would be eliminated if individuals moved out of planned developments and into other suburban or central city neighborhoods. Results from this decomposition reveal that the contribution of planned developments to overall metropolitan area segregation in 36 California is actually quite small (Table 4.10). In suburbs, differences in the racial and ethnic composition between planned developments and other neighborhoods explain only 2 percent of overall segregation. In central cities, the share attributable to planned developments is even smaller at 1 percent. (Appendix B includes decompositions of segregation indices by metropolitan area.) This finding is perhaps unsurprising because of the small proportion of the population currently living in planned developments. Recall that decompositions of the entropy measure of segregation reflect the relative population shares in each geographic component. Although planned developments represent more than 40 percent of new single-family homes, they include less than 20 percent of the existing stock of these homes and 12 percent of total units (HOA-Info, 2002; U.S. Census Bureau, 2002a). In light of racial and ethnic differences between planned developments and other neighborhoods presented in this chapter, it is conceivable that as these communities proliferate, they could exert a greater effect on residential segregation over time. Table 4.10 Decomposition of Statewide Metropolitan Racial Segregation Component Between central cities and suburbs Central cities Between planned developments and other neighborhoods Within planned developments Within other neighborhoods Suburbs Between planned developments and other neighborhoods Within planned developments Within other neighborhoods % Share 2 43 1 4 37 56 2 9 44 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Numbers may not sum to 100 because of rounding. 37 5. Voting Behavior This chapter investigates the voting behavior of Californians living in planned developments. It begins with a look at linkages between socalled “private governments” and political participation, as measured by voter turnout, registration, and party affiliation. It then considers differences between planned developments and other neighborhoods in “yes” votes on statewide ballot propositions. Finally, it examines how robust these differences are even after controlling for other factors that might influence voting behavior. Many social commentators have expressed concerns about residents of common interest developments opting out of the public sphere. Perhaps most famously, Robert Reich has referred to the growth of CIDs as the “secession of the successful.” Similarly, Charles Murray (1991, p. 26) has expressed concern about the democratizing potential of CIDs to allow “10 or 20 percent of the population… to bypass social institutions it doesn’t like in ways that only the top fraction of 1 percent used to be able to do.” One argument is that the homogeneity of these communities reduces the incentives, or controversies, that draw citizens into civic affairs (e.g., Putnam, 2000; Oliver, 2001).1 Alternatively, if planned developments are substitutes for traditional local government, then residents of these developments may perceive less benefit to taking part in elections and other aspects of public life (e.g., Blakely and Snyder, 1997). These concerns have taken on added meaning in light of recent evidence for declining “social capital” in the United States, as documented in Robert Putnam’s Bowling Alone. Analogous to physical or human capital, social capital refers to the stock of social norms, trust, and civic engagement that enhances both individual productivity and ____________ 1See Alesina and La Ferrara (2000) for an opposing view, i.e., that heterogeneity depresses participation in voluntary associations. 39 aggregate welfare. Indicators include voter turnout and other forms of political participation, membership in voluntary associations, and surveys regarding trust in government. Turnout in U.S. presidential elections has fallen from roughly 63 percent of the voting age population in 1960 to 51 percent in 2000 (U.S. Federal Elections Commission, 2000). In California, voter turnout has fallen from roughly 50 percent in nonpresidential statewide elections during the 1970s to 36 percent in November 2002 (California Secretary of State, 2002). Moreover, as in the rest of the nation, few Californians are involved in political parties, elections, or campaigns (e.g., Baldassare, 2000, p. 34; Verba, Scholzman, and Brady, 1995). These trends could have lasting consequences to the extent that government policies tend to reflect the preferences of those who participate in public life. There are several reasons, however, to believe that residents of planned developments will be more rather than less politically active. As seen in the previous chapter, residents of these communities tend to be older, better educated, and more affluent—factors that are all associated with higher rates of voter turnout (e.g., Rosenstone and Hansen, 1993). Beyond these factors, homeowners in general are more likely to take part in elections and other forms of civic participation because of an interest in preserving neighborhood quality and property values (DiPasquale and Glaeser, 1999).2 According to one gated community resident interviewed by Blakely and Snyder (1997, p. 72), for instance, The reason for the activism in Rancho Mirage and Indian Wells is the residents are rich and used to having their voices heard, and they are retired and have time. Also, the gated community has a stronger investment in expressing their needs because they identify with that community. Neighborhoods would have drawn that kind of allegiance in the past. There has been a sense of belonging and of shared destiny between an individual and where they live throughout history which has been lost. The gated community replaces that. I’m wondering if they aren’t a market response to that desire for belonging. . . . The gate and the homogeneity reinforces a feeling of shared destiny and [shared interest in] property values . ____________ 2Homeowners may also self-select into these communities because of an interest in preserving property values, for example, through more stringent land use and other regulation (Helsley and Strange, 2000). See Appendix C for a discussion of the selection issue. 40 This motivation further suggests that members of planned developments may vote differently from other Californians. In particular, in keeping with the “double taxation” issue discussed at the end of Chapter 2, they may favor public spending that complements private expenditures (e.g., police protection) while opposing that which is duplicative or redistributive (e.g., infrastructure and welfare). This chapter tests the above hypotheses using voting and registration information from the 2000 statewide general election.3 Results indicate that members of planned developments do not exhibit significantly different voting behavior once relevant demographic characteristics are taken into account. These findings undermine the popular view that socalled “private governments” crowd out political participation or influence the direction of voting behavior on individual ballot propositions, at least in statewide general elections. Voter Turnout, Registration, and Party Affiliation Planned developments exhibit higher rates of voter turnout and registration than other California neighborhoods (Table 5.1). Differences in turnout between planned developments and other neighborhoods are 6 percentage points, on average, whereas those in voter registration are even higher at 8 percentage points. Planned developments are more politically conservative than other neighborhoods, with higher Republican Party registration (by about 9 percentage points) and lower Democratic Party affiliation (by roughly 8 percentage points). (These differences are statistically significant at the 1 percent level). Voting on Statewide Ballot Propositions Differences in political ideology are apparent not only in the party affiliation of planned development residents but also in their patterns of voting on statewide ballot propositions (Table 5.2). In the 2000 statewide general election, residents of planned developments were more likely to support measures related to privatization, such as school ____________ 3These data were matched to census levels of geography by the Statewide Database at UC Berkeley’s Institute of Governmental Studies. See Statewide Database (2002). 41 Table 5.1 Voter Turnout and Party Affiliation, by Planned Development Status (in percent) Voter turnout Registered Democrat Republican Independent Other Planned Developments 74.1 70.7 40.7 39.8 14.4 5.1 Other Neighborhoods 68.4 62.8 49.1 31.2 14.2 5.5 Difference 5.7** 7.8** –8.5** 8.6** 0.2* –0.4** SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Rates of voter registration are expressed as a percentage of the voting age population. All other percentages are of registered voters. Independent voters are those who decline to state a party affiliation. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. vouchers (Proposition 38) and legislation enabling state and local governments to contract with private firms for public works projects (Proposition 35). They also voted in favor of fiscally conservative measures such as a two-thirds majority vote requirement for regulatory fees (Proposition 37) and against a measure lowering the voter approval requirement for local school bonds from two-thirds to 55 percent (Proposition 39). A Closer Look at Voting Behavior A key question is whether discrepancies in voter turnout and behavior persist after controlling for the demographic and socioeconomic characteristics of planned developments. As discussed in Chapter 4, these communities tend to differ from other neighborhoods along a number of dimensions. Table 5.3 summarizes these differences and their predicted effects on voter turnout based on previous studies. Taking account of just one factor—the proportion of the population that is non-white—narrows the gap in voter turnout considerably, 42 Table 5.2 Approval Rates for Statewide Ballot Propositions, by Planned Development Status (in percent) Prop. Planned Other No. Purpose Developments Neighborhoods Difference 32 $500,000,000 bond issue for farm and home aid to California veterans 66 69 –3** 33 Allows California legislature to participate in the Public Employees' Retirement System 37 41 –3** 34 Limits individual campaign contributions; establishes voluntary spending limits; expands public disclosure requirements 60 61 –1** 35 Provides that state and local governments may contract with private entities for engineering and architectural services 58 53 5** 36 Requires probation and treatment for conviction of possession, use, transportation, or being under influence of controlled substances 60 62 –3** 37 Requires two-thirds vote of state legislature, or either majority or two- thirds of local electorate, to impose any activity fees to monitor, study, or mitigate effects of that activity 49 47 2** 38 Authorizes state-funded school vouchers of $4,000 per pupil per year for private and religious schools 31 28 3** 39 Authorizes bonds for repair, construction, or replacement of school facilities, classrooms, if approved by a 55 percent local vote 52 56 –5** SOURCES: HOA-Info (2002); Statewide Database (2000). NOTE: Differences may not be precise because of rounding. **Denotes statistically significant difference at the 1 percent level. although differences remain statistically significant at the 1 percent level (Table 5.4). More strikingly, the gap in voter registration between planned developments and other neighborhoods is all but eliminated after controlling for the proportion of the population that is non-white (Table 5.5). 43 Table 5.3 Summary of Demographic and Socioeconomic Differences Between Planned Developments and Other Neighborhoods and Predicted Effects on Voter Turnout (in percent) Characteristic White Income above $75,000 With college degree Ages 40 to 64 Age 65 and up Homeowner Moved in last 5 years Planned Other Predicted Effect Developments Neighborhoods Difference on Turnout 64 47 17** Increases 37 26 11** Increases 21 15 6** Increases 33 29 4** Increases 14 11 2** Increases 66 57 9** Increases 49 48 1** Decreases SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: Differences may not be precise because of rounding. **Denotes statistically significant difference at the 1 percent level. Table 5.4 Voter Turnout, by Planned Development Status and Racial and Ethnic Diversity (in percent) Percentage Non-White 0–25 25–50 50–75 75–100 Planned Developments 79 74 69 63 Other Neighborhoods Difference 77 2** 72 2** 66 2** 60 3** SOURCES: HOA-Info (2002); Statewide Database (2000). NOTE: Differences may not be precise because of rounding. **Denotes statistically significant difference at the 1 percent level. Regression analysis enables us to control for many demographic and socioeconomic characteristics simultaneously, thereby parsing out the effect of planned developments alone, all else being equal. Figure 5.1 summarizes the estimated effect of planned developments on voting behavior, holding constant other neighborhood features such as age structure, levels of educational attainment, income mix, racial and ethnic composition, homeownership, and urban location.4 ____________ 4Appendix C includes detailed regression results and a discussion of modeling issues. 44 Table 5.5 Voter Registration, by Planned Development Status and Racial and Ethnic Diversity (in percent) Percentage Non-White 0–25 25–50 50–75 75–100 Planned Developments 81 72 59 46 Other Neighborhoods Difference 80 1 71 1 59 0 47 –1* SOURCES: HOA-Info (2002); Statewide Database (2000). NOTE: Differences may not be precise because of rounding. *Denotes statistically significant difference at the 5 percent level. 10 8 6 4 2 0 –2 –4 –6 –8 –10 Voter turnout Percent registered Percent Republican Percent Democrat Percent Independent Raw difference in means Controlled difference in means Percentage points SOURCES: HOA-Info (2002); Statewide Database (2000). Figure 5.1—Estimated Effects of Planned Developments on Voter Turnout, Registration, and Party Affiliation 45 Differences in voter turnout, registration, and party affiliation between planned development and other neighborhoods are reduced substantially after controlling for these influences. Results from the regression model indicate that planned developments have rates of voter turnout that are higher by two-thirds of 1 percentage point; this is in contrast to the 6 percentage points detected in the uncontrolled difference. More dramatically, differences in voter registration rates switch from a positive 8 percentage points to a negative 0.77 percentage points. In other words, planned developments exhibit slightly lower rates of voter registration than we would expect given residents’ socioeconomic status. Differences in party affiliation, although muted, are still apparent. Residents of planned developments are more likely to register with the Republican Party, for an average difference of approximately 2 percentage points, with a nearly offsetting reduction in Democratic Party affiliation. Regressions seeking to explain voting on statewide ballot propositions hold constant Republican Party registration in addition to the demographic and socioeconomic neighborhood attributes described above. Again, the gap in voting behavior between planned developments and other areas narrows considerably (Figure 5.2). Differences in percentage voting in favor of “conservative” measures—such as school vouchers (Proposition 38), state and local government contracting out for public works projects (Proposition 35), and a two-thirds vote requirement for fees (Proposition 37)—drop to 1 percentage point or less. In contrast to the uncontrolled differences presented above, planned developments are associated with slightly higher approval rates for measures to lower the vote requirement for local school bonds (Proposition 39) and to issue a bond for veterans (Proposition 32). In general, these differences, although statistically significant, are quite small. Taken together, they suggest that planned developments do not strongly affect either political participation or patterns of voting behavior in statewide general elections. Residents of planned developments vote similarly to other Californians of comparable age, income, education, race, and ethnicity. 46 Percentage points 6 4 2 0 –2 –4 –6 Proposition 32 Proposition 35 Proposition 37 Proposition 38 Proposition 39 Raw difference in means Controlled difference in means SOURCES: HOA-Info (2002); Statewide Database (2000). Figure 5.2—Estimated Effects of Planned Developments on Votes for Statewide Ballot Propositions 47 6. Conclusions Common interest developments are a dominant source of new housing in California. This report is one of the first empirical analyses of these communities. It presents a comprehensive overview of common interest developments in general, including their origins, institutional features, and powers and responsibilities and compares them to traditional local government. It also surveys the features of common interest development housing in California, documenting trends in the location, size, and financial resources of these communities. The report focuses on one particular type of common interest development, the planned development. This is the fastest-growing and most prevalent CID as well as the most direct form of potential competition with traditional local government. The study relies on a variety of data sources to describe these communities and discern their potential consequences for residential segregation and civic engagement. Results indicate that concerns regarding negative social consequences of planned developments are, for the most part, unwarranted. Planned developments are less diverse than other neighborhoods, yet their contribution to area segregation is minor. Similarly, planned developments do not exhibit differences in political participation or voting behavior once relevant demographic and socioeconomic factors are taken into consideration. There are several important caveats to bear in mind, however. As recent trends in the growth of the population and planned developments continue, it is conceivable that the effects of these communities on residential segregation may become more pronounced. Although planned developments currently represent only 12 percent of all housing units, they constitute roughly 30 percent of new housing starts and over 40 percent of new single-family home sales (HOA-Info, 2002; U.S. Census Bureau, 1994, 1995, 1996, 1998, 1999, 2002b; Construction Industry Research Board, 2000). 49 Local authorities may wish to consider such potential costs before encouraging this form of development to offset the costs of additional public services and infrastructure. Nevertheless, eliminating planned developments altogether would not necessarily reduce segregation even by the modest amounts suggested in Chapter 4. These developments are but one way that individuals can sort themselves based on desires for either enhanced services or racial and economic homogeneity. In addition, voting behavior in statewide general elections is only one form of political participation, albeit one that is more tractable for empirical analysis. Other important indicators of civic engagement include joining in voluntary associations, contacting elected officials, and attending public hearings or meetings. An important area for future research is the level of involvement within common interest developments themselves through boards of directors and committees. Only this information will illuminate the total effect of planned developments on participation in public life. Finally, a consistent and important finding throughout the report is the considerable heterogeneity among common interest developments in general and planned developments specifically. This finding in particular militates against any “one-size-fits-all” policy solution. 50 Appendix A Data and Methods This appendix describes the procedure used to match locations of planned developments to U.S. Census Bureau geography. It also explores the potential consequences of the limitations of this procedure for results presented in the report. In some cases, it presents results separately based on alternative sample definitions. Planned development addresses in the HOA-Info database were geocoded—matched to latitude and longitude coordinates using geographic information systems (GIS). Matching was successful for about 8,100, or 94 percent, of the approximately 8,600 planned developments for which address information was reported.1 These coordinates were then used to link each property to census levels of geography (e.g., blocks, block groups, and tracts) and political jurisdictions (e.g., cities, counties) (Figure A.1). Geographic matches were based on the reported street address of the homeowner association president. Properties reporting only a manager address were excluded from the sample, as were any properties reporting anomalous addresses that could not be reliably distinguished from off-site management companies.2 This procedure yielded roughly 6,200, or 66 percent, of all planned developments that existed in California as of December 2002. ____________ 1Of the 9,318 planned developments in the dataset, address information was missing for 667 observations. Searches of homeowner association websites and directories as well as public tax assessment, deed transfer, and mortgage records yielded locations for several missing PDs. These developments were statistically indistinguishable from planned developments that did report address information on all observable dimensions except size (i.e., number of units and acreage), most likely reflecting the procedure used to locate missing addresses rather than systematic reasons for reporting or not reporting to the California Department of Real Estate or Secretary of State. 2These include multiple records with identical addresses except for homeowner associations that are affiliated with one another (e.g., separate building phases of the same development). Unique observations in areas that were zoned predominantly for commercial or industrial use according to county assessor data were also purged from the analysis. 51 Figure A.1—Example of Geocoding Procedure This sample differs somewhat from the population of planned developments as a whole. Notably, they are significantly smaller and older than other planned developments (Table A.1). These differences may reflect the tendency for larger and newer developments to report offsite management companies as their primary address; these would be excluded from the sample. Efforts to obtain address information from large property management companies were unsuccessful because of confidentiality concerns. Table A.2 shows that excluding these developments from the study is unlikely to bias results toward zero because they constitute a small proportion of total housing. In other words, they do not constitute a large enough share of all units to contaminate the “control group.” 52 Table A.1 Selected Characteristics of Planned Developments in California Characteristic 5th 25th Minimum Percentile Percentile Median Study Sample Units Acres Year founded Monthly fees ($) Annual revenues 3 0 1924 0 62 6 1 1968 21 6,908 16 42 3 10 1978 1984 65 113 19,022 48,947 Remaining PDs Units Acres Year founded Monthly fees ($) Annual revenues 2 0 1929 0 86 10 1 1971 30 13,617 39 7 1980 71 45,714 84 17 1987 110 105,356 SOURCE: HOA-Info (2002). 75th 95th Percentile Percentile Maximum 108 33 1990 168 132,011 460 30,000 250 11,000 1999 2002 310 4,006 591,784 35,235,192 170 662 11,335 50 329 5,716 1993 2000 2002 167 317 1,654 240,995 1,072,115 13,645,714 Nevertheless, the exclusion of these PDs could limit the external validity or generalizability of this study. A second issue is matching planned developments to census levels of geography. Ideally, we would like to assume that census block groups and planned developments are co-extensive and to use block groups to compare planned developments with other neighborhoods. However, planned development and block group boundaries do not perfectly coincide. The average block group contains roughly 550 housing units, whereas the average planned development includes 150 units. In addition, planned developments may span multiple block groups. These concerns are somewhat mitigated by the Census Bureau requirements that census block groups be as homogeneous as possible and that they reflect existing neighborhood boundaries (U.S. Census Bureau, 2000). Nevertheless, we replicated the analysis in Chapter 4 using a narrower definition of what constitutes a planned development 53 Table A.2 Regional Distribution of Planned Development Units, by Sample Definition Region Bay Area Central Coast North Coast and Mountains Inland Empire Sacramento Metro San Diego San Joaquin Valley Sierras South Coast Total Study Sample 159,501 6 40,444 8 57,946 12 111,695 9 73,926 10 88,866 8 79,259 7 21,587 22 216,719 5 849,943 7 Remaining PDs 106,307 4 5,388 1 16,971 3 93,894 8 35,183 5 98,134 9 31,167 3 3,493 4 245,324 5 635,861 5 Total Housing Units 2,552,402 492,256 488,006 1,186,043 714,981 1,084,040 1,107,260 97,456 4,492,105 12,214,549 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTE: The numbers in the second row for each region are percentages. block group.3 Findings were not substantially different in sign or in magnitude. Results are available from the author upon request. ____________ 3Specifically, to avoid mistakenly characterizing block groups as planned developments, we restricted the sample to block groups where PDs constituted a majority (60 percent) of the housing stock. Many of these block groups included more planned development units than total housing units, suggesting a PD that spans multiple block groups. To address this second type of error, we coded all block groups within a census tract as planned developments if the “spillover” from the initial PD constituted more than 70 percent of all housing units in the tract. 54 Appendix B Measuring Diversity and Segregation This appendix reviews the entropy measures of diversity and segregation. It also presents results for the decomposition of segregation measures by metropolitan area (Table B.1). The entropy index of diversity for a city or metropolitan area is E = n ∑ Qr r =1 log ⎛ ⎝⎜ 1 Qr ⎞ ⎠⎟ (B.1) where Qr is the proportion of the total population in group r, and n is the number of groups. Analogously, for a neighborhood the index is defined as Ei = n ∑ qri r =1 ⎛ log⎝⎜ 1 q ri ⎞ ⎠⎟ (B.2) where qri is the proportion of individuals belonging to racial or ethnic group r in neighborhood i. Conceptually, the index expresses the level of uncertainty as to the racial or ethnic identity of any individual selected at random from the area as a whole. It reaches its maximum of log(n) when all groups are equally represented in the population and its minimum of 0 when there is only one group present. Index values can be normalized to range between 0 and 100 by using the number of groups, n, as the base of the logarithm and multiplying by 100. The entropy index of segregation relates the diversity of a city or metropolitan area as a whole to the population-weighted average diversity of its constituent neighborhoods: k ∑ [ti /T (E − Ei )] H = i =1 E (B.3) 55 Table B.1 Decomposition of Segregation, by Planned Development and Central City/Suburban Status 56 Metropolitan Area Bakersfield Chico–Paradise Fresno Los Angeles–Riverside– Orange County Merced Modesto Redding Sacramento–Yolo Salinas San Diego Between Central Cities and Suburbs (1) 0.84 3 0.15 1 1.67 7 Total Central Cities (2) 6.14 25 4.59 35 10.39 43 Within Central Cities Between PDs and Within Within Non-PDs PDs Non-PDs (3) (4) (5) 0.58 0.77 4.79 23 19 0.06 0.79 3.74 06 28 0.72 1.08 8.59 35 36 0.43 11.24 1 37 0.97 5.11 6 33 0.86 7.23 6 54 0.98 2.58 14 37 2.9 5.45 17 31 1.25 9.93 4 35 0.93 13.68 4 59 0.42 0.6 12 0.34 0.11 21 0.32 0.68 25 0.06 0.65 19 0.18 0.73 14 0.06 3.8 0 13 0.55 1.46 26 10.22 34 4.66 30 6.23 47 1.87 27 4.54 26 6.07 21 11.67 51 Total Suburbs (6) 17.66 72 8.46 64 11.86 50 18.59 61 9.47 61 5.2 39 3.48 50 9.14 52 17.18 61 8.42 37 Within Suburbs Between PDs and Within Within Non-PDs PDs Non-PDs (7) (8) (9) 1.54 1.32 14.8 65 60 0.39 0.50 7.57 34 57 0.8 0.8 10.26 33 43 0.45 2.57 18 0.14 0.25 12 0.24 0.19 21 0.12 0.14 22 0.59 1.57 39 0.92 4.62 3 16 0.59 1.6 37 15.57 51 9.08 58 4.77 36 3.22 46 6.98 40 11.64 41 6.23 27 Total (10) 24.4 100 13.19 100 23.94 100 30.25 100 15.54 100 13.3 100 7.03 100 17.5 100 28.36 100 23.02 100 Table B.1 (continued) 57 Metropolitan Area San Francisco– Oakland–San Jose San Luis Obispo– Atascadero–Paso Robles Santa Barbara–Santa Maria–Lompoc Stockton–Lodi Visalia–Tulare– Porterville Yuba City Total Between Central Cities and Suburbs Total Central Cities 0.94 12.22 4 51 0.08 4.2 1 29 2.78 11.68 13 54 2.45 10.98 13 60 2.26 8.70 11 42 0.18 3.35 2 28 0.45 12.14 2 43 Within Central Cities Between PDs and Within Within Non-PDs PDs Non-PDs 0.16 1.91 18 10.15 42 0.15 0.83 16 3.22 22 0.37 2.98 2 14 0.89 1.1 56 8.33 39 8.99 50 0.64 1.64 38 0.13 0.35 13 0.42 1.23 14 6.42 31 2.87 24 10.49 37 Total Suburbs 10.81 45 10.37 71 7.15 33 4.72 26 9.68 47 8.31 70 15.92 56 Within Suburbs Between PDs and Within Within Non-PDs PDs Non-PDs Total 0.23 3.05 1 13 7.53 23.97 31 100 0.42 3.9 3 27 6.05 14.65 41 100 0.36 2.22 2 10 0.07 0.49 03 4.57 21.6 21 100 4.16 18.15 23 100 0.31 0.55 23 0.2 0.36 23 0.62 2.65 29 8.82 43 7.75 66 12.65 44 20.64 100 11.83 100 28.51 100 SOURCES: HOA-Info (2002); U.S. Census Bureau (2002a). NOTES: The first entry in each column represents the “between” or “within” component of overall metropolitan segregation. The second entry is the percentage share of metropolitan segregation attributable to this component. Columns 1, 2, and 6 sum to column 10. Columns 3, 4, and 5 sum to column 2, and Columns 7, 8, and 9 sum to column 6. Index values reported in this table are multiplied by 100. where T and ti are the respective area and subunit populations and k is the number of subunits. Since the entropy of an area as a whole (i.e., E ) can never be less than the average entropy of its parts (i.e., ∑ik=1(ti /T )Ei )), the segregation index necessarily ranges between 0 and 1. At 0, the racial composition of each subunit is identical to that of the area as a whole, whereas at 1 each subunit contains only one group. An attractive feature of the entropy index of segregation is that it may be decomposed into segregation between and within groups. The general formula for any decomposition is H = ∑ p∈P ⎛ ⎝⎜ Tp T ⎞ ⎠⎟ ⎛ ⎝⎜ Ep E ⎞ ⎠⎟ H p (B.4) where p represents racial or geographic components of the decomposition (Reardon et al., 2000, p. 354). For instance, the segregation of a metropolitan area comprising central city and suburban block groups can be written as H = H c ,s + ⎛⎝⎜TTc ⎞ ⎠⎟ ⎛ ⎝⎜ Ec E ⎞ ⎠⎟ H c + ⎛⎝⎜TTs ⎞ ⎠⎟ ⎛ ⎝⎜ Es E ⎞ ⎠⎟ H s (B.5) where Tc and Ts, Ec and Es, and Hc and Hs are, respectively, the total populations, entropy measures of diversity, and entropy measures of segregation for city and suburban block groups (Reardon et al., 2000, p. 355). The first term, Hc,s, indicates the portion of overall segregation attributable to differences between the central city and the suburbs. That is H c ,s = 1 − ⎛⎝⎜Tc E c +Ts TE E s ⎞ ⎠⎟ . (B.6) The second and third terms represent portions of segregation attributable to differences among central city and suburban subunits. These latter terms can be further decomposed into “between” and “within” components with respect to planned development status. Specifically, each “within” term in Equation (B.5) above includes 58 H PD,non−PD + ⎛⎝⎜TTPD ⎞⎛ ⎠⎟ ⎝⎜ E PD E ⎞ ⎠⎟ H PD + ⎛⎝⎜T noTn − PD ⎞⎛ ⎠⎟ ⎝⎜ E non−PD E ⎞ ⎠⎟ H non − PD (B.7) where the first term is segregation between planned developments and other block groups defined analogously to Equation (B.6), the second term represents the share of segregation among planned developments, and the third term represents the share of segregation among non-PD block groups. 59 Appendix C Technical Considerations and Regression Results A methodological complication in any study of neighborhood effects on political participation is the potential endogeneity of residential choice. That is, individuals may choose to live in a planned development because of observable or unobservable factors that are also related to voting behavior. As a result, apparent relationships between planned developments and voter turnout may be spurious. This selection bias can work in either direction. For example, individuals may sort into a planned development based on a taste for involvement in both public and private government. Alternatively, they may move into a PD to opt out of the public sphere without sacrificing collective goods and services. In either case, estimates of the effect of living in a planned development will be both biased and inconsistent. Gordon (2003) explores this possibility using several specifications to control for potential selection bias. Methods include least squares regression with block-group fixed effects, a Heckman selection correction model, and a propensity-score approach. Results from these models are not substantially different from the least squares regression with demographic covariates presented in this report.1 The remainder of this appendix includes detailed results from the multivariate analysis underlying the charts in Chapter 5. ____________ 1Gordon (2003) also includes results of regressions based on a more restricted sample definition as in Appendix A. 61 Table C.1 Estimated Effects of Planned Developments on Voter Turnout, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 5.700 (0.166)** 68.374 (0.079)** 22,055 0.04 Ordinary Least Squares 0.661 (0.107)** 0.093 (0.012)** 0.173 (0.007)** –0.118 (0.004)** –0.053 (0.004)** –0.123 (0.003)** –0.160 (0.014)** –0.049 (0.016)** 0.023 (0.010)* 0.086 (0.011)** 0.068 (0.004)** –0.043 (0.005)** 0.110 (0.009)** 0.127 (0.010)** 0.108 (0.009)** 0.059 (0.009)** –1.696 (0.226)** 63.193 (0.855)** 21,802 0.63 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 62 Table C.2 Estimated Effects of Planned Developments on Percentage Registered, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 7.827 (0.328)** 62.833 (0.148)** 21,039 0.02 Ordinary Least Squares –0.765 (0.206)** 0.083 (0.022)** 0.162 (0.014)** 0.077 (0.007)** –0.266 (0.008)** –0.345 (0.006)** –0.168 (0.024)** –0.108 (0.028)** 0.176 (0.019)** 0.379 (0.021)** 0.150 (0.007)** –0.186 (0.009)** 0.010 (0.017) –0.006 (0.022) –0.009 (0.015) 0.050 (0.018)** 4.507 (0.440)** 60.758 (1.572)** 20,874 0.66 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 63 Table C.3 Estimated Effects of Planned Developments on Percentage Registered Democrat, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means –8.460 (0.220)** 49.116 (0.107)** 22,060 0.05 Ordinary Least Squares –1.807 (0.175)** 0.136 (0.014)** 0.095 (0.010)** 0.547 (0.004)** 0.243 (0.006)** 0.055 (0.006)** –0.023 (0.016) –0.061 (0.018)** –0.272 (0.014)** 0.077 (0.015)** –0.065 (0.004)** –0.144 (0.006)** –0.015 (0.011) –0.049 (0.013)** –0.100 (0.012)** –0.143 (0.017)** 5.243 (0.265)** 46.224 (1.116)** 21,807 0.60 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 64 Table C.4 Estimated Effects of Planned Developments on Percentage Registered Republican, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 8.604 (0.252)** 31.223 (0.111)** 22,060 0.04 Ordinary Least Squares 2.031 (0.202)** –0.126 (0.016)** 0.044 (0.010)** –0.476 (0.005)** –0.185 (0.006)** –0.160 (0.006)** 0.026 (0.017) 0.061 (0.018)** 0.284 (0.015)** –0.153 (0.018)** 0.152 (0.005)** 0.134 (0.007)** 0.007 (0.013) 0.053 (0.014)** 0.116 (0.013)** 0.192 (0.019)** –4.899 (0.333)** 26.724 (1.219)** 21,807 0.54 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 65 Table C.5 Estimated Effects of Planned Developments on Percentage Registered Independent, 2000 Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and older Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (= 1 if in an urbanized area) Constant No. of observations R2 Difference in Means 0.211 (0.079)** 14.165 (0.037)** 22,060 0.00 Ordinary Least Squares –0.148 (0.061)* –0.013 (0.006)* –0.070 (0.004)** –0.070 (0.002)** –0.023 (0.002)** 0.134 (0.003)** –0.017 (0.008)* –0.002 (0.008) –0.002 (0.006) 0.061 (0.007)** –0.065 (0.002)** 0.011 (0.003)** 0.027 (0.004)** 0.018 (0.005)** 0.019 (0.005)** 0.007 (0.006) –0.077 (0.109) 16.855 (0.479)** 21,807 0.45 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 66 Table C.6 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 32, 2000 Difference Ordinary in Means Least Squares Planned development –2.752 0.314 = 1 if > 1 PD (0.096)** (0.066)** = 0 otherwise Percentage Ages 40–64 0.004 (0.007) Age 65 and older 0.096 (0.005)** Non–Hispanic black 0.140 (0.003)** Hispanic 0.019 (0.002)** Non–Hispanic Asian –0.050 (0.003)** With less than high school 0.012 (0.010) High school diploma –0.007 (0.011) With some college –0.011 (0.007) With college degree –0.021 (0.007)** Homeowner –0.016 (0.002)** Moved in last 5 years –0.006 (0.003) With income $50,000–$74,999 –0.016 (0.005)** With income $75,000–$99,999 –0.027 (0.006)** With income $100,000–$199,999 –0.048 (0.006)** With income $200,000 and up 0.011 (0.007) Urban (= 1 if in an urbanized area) 0.678 (0.128)** Percent registered Republican –0.208 (0.003)** Constant 68.993 75.398 (0.048)** (0.567)** Observations R2 22,036 0.03 21,801 0.64 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 32 is a $500,000,000 bond issue for farm and home aid to California veterans. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 67 Table C.7 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 35, 2000 Difference Ordinary in Means Least Squares Planned development 5.272 1.068 = 1 if > 1 PD (0.124)** (0.077)** = 0 otherwise Percentage Ages 40–64 0.000 (0.007) Age 65 and older –0.039 (0.005)** Non–Hispanic black –0.143 (0.003)** Hispanic –0.001 (0.003) Non–Hispanic Asian –0.025 (0.003)** With less than high school –0.041 (0.008)** High school diploma –0.027 (0.009)** With some college 0.013 (0.006)* With college degree 0.052 (0.007)** Homeowner –0.046 (0.002)** Moved in last 5 years 0.055 (0.003)** With income $50,000–$74,999 0.017 (0.005)** With income $75,000–$99,999 0.051 (0.007)** With income $100,000–$199,999 0.100 (0.005)** With income $200,000 and up 0.161 (0.008)** Urban (= 1 if in an urbanized area) –0.546 (0.140)** Percent registered Republican 0.287 (0.003)** Constant 52.509 42.972 (0.058)** (0.508)** No. of observations R2 22,032 0.06 21,799 0.70 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 35 allows state and local governments to contract with private entities for engineering and architectural services. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 68 Table C.8 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 37, 2000 Difference Ordinary in Means Least Squares Planned development 2.331 0.167 = 1 if > 1 PD (0.141)** (0.078)* = 0 otherwise Percentage Ages 40–64 0.013 (0.008) Age 65 and up –0.021 (0.005)** Non–Hispanic black –0.042 (0.004)** Hispanic 0.097 (0.003)** Non–Hispanic Asian 0.128 (0.003)** With less than high school 0.108 (0.010)** High school diploma 0.135 (0.010)** With some college 0.089 (0.007)** With college degree –0.075 (0.009)** Homeowner 0.009 (0.003)** Moved in last 5 years 0.034 (0.003)** With income $50,000–$74,999 –0.002 (0.007) With income $75,000–$99,999 0.003 (0.007) With income $100,000–$199,999 –0.022 (0.005)** With income $200,000 and up –0.066 (0.007)** Urban (= 1 if in an urbanized area) –0.553 (0.138)** Percent registered Republican 0.503 (0.004)** Constant 46.991 21.736 (0.066)** (0.598)** No. of observations R2 22,029 0.01 21,800 0.72 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 37 requires a vote of the state legislature or local electorate to impose fees to monitor, study, or mitigate an activity. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 69 Table C.9 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 38, 2000 Difference in Means Ordinary Least Squares Planned development = 1 if > 1 PD = 0 otherwise Percentage Ages 40–64 Age 65 and up Non–Hispanic black Hispanic Non–Hispanic Asian With less than high school High school diploma With some college With college degree Homeowner Moved in last 5 years With income $50,000–$74,999 With income $75,000–$99,999 With income $100,000–$199,999 With income $200,000 and up Urban (=1 if in an urbanized area) Percent registered Republican Constant No. of observations R2 2.662 (0.113)** 28.282 (0.052)** 22,036 0.020 0.112 (0.071) 0.020 (0.007)** –0.033 (0.005)** –0.027 (0.003)** 0.024 (0.003)** 0.069 (0.003)** 0.062 (0.008)** 0.087 (0.011)** 0.064 (0.007)** –0.043 (0.006)** –0.018 (0.002)** 0.029 (0.003)** 0.003 (0.005) 0.004 (0.005) –0.022 (0.007)** 0.045 (0.007)** –1.84 (0.132)** 0.379 (0.003)** 13.645 (0.532)** 21,801 0.670 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 38 authorizes state-funded school vouchers of $4,000 per pupil per year for private and religious schools. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 70 Table C.10 Estimated Effects of Planned Developments on “Yes” Votes for Proposition 39, 2000 Difference Ordinary in Means Least Squares Planned development –4.536 0.931 = 1 if > 1 PD (0.173)** (0.099)** = 0 otherwise Percentage Ages 40–64 –0.041 (0.011)** Age 65 and up 0.028 (0.006)** Non–Hispanic black 0.156 (0.004)** Hispanic 0.065 (0.004)** Non–Hispanic Asian –0.011 (0.003)** With less than high school –0.072 (0.011)** High school diploma –0.175 (0.018)** With some college –0.109 (0.010)** With college degree 0.079 (0.009)** Homeowner –0.058 (0.003)** Moved in last 5 years 0.036 (0.004)** With income $50,000–$74,999 –0.031 (0.008)** With income $75,000–$99,999 –0.039 (0.008)** With income $100,000–$199,999 –0.015 (0.009) With income $200,000 and up 0.115 (0.009)** Urban (=1 if in an urbanized area) 2.770 (0.179)** Percent registered Republican –0.473 (0.004)** Constant 56.307 73.201 (0.089)** (0.787)** No. of observations R2 22,035 0.02 21,801 0.78 SOURCES: HOA-Info (2002); Statewide Database (2000). NOTES: Robust standard errors are in parentheses. Percentages are calculated over individuals except for income, which is a proportion of all households. Proposition 39 authorizes bonds for school facilities, classrooms if approved by a 55 percent local vote. *Denotes statistically significant difference at the 5 percent level. **Denotes statistically significant difference at the 1 percent level. 71 References Advisory Commission on Intergovernmental Relations, Residential Community Associations: Private Governments in the Intergovernmental System? Washington, D.C., May 1989. Alesina, Alberto, and Eliana La Ferrara, “Participation in Heterogeneous Communities,” Quarterly Journal of Economics, Vol. 115, No. 3, August 2000, pp. 847–904. Baldassare, Mark, California in the New Millennium, University of California Press, Berkeley, California, and the Public Policy Institute of California, San Francisco, California, 2000. Barton, Stephen E., and Carol J. Silverman, “History and Structure of the Common Interest Community,” in Stephen E. Barton and Carol J. Silverman, eds., Common Interest Communities: Private Governments and the Public Interest, Institute of Governmental Studies Press, Berkeley, California, 1994. Barzel, Yoram, and Tim R. Sass, “The Allocation of Resources by Voting,” Quarterly Journal of Economics, Vol. 16, August, 1990, pp. 745–771. Black’s Law Dictionary, 5th ed., West Publishing Company, St. Paul, Minnesota, 1979. Blakely, Edward J., and Mary Gail Snyder, Fortress America: Gated Communities in the United States, Brookings Institution Press, Washington, D.C., 1997. California Law Revision Commission, “Common Interest Development Law: Scope of Project (Background Study),” Staff Memorandum 2001-19, Palo Alto, California, January 22, 2001. 73 California Law Revision Commission, “Nonjudicial Dispute Resolution Under CID Law: Alternative Dispute Resolution (Additional Information),” Second Supplement to Memorandum 2002-10, Palo Alto, California, 2002a. California Law Revision Commission, “Alternative Dispute Resolution in Common Interest Developments,” Tentative Recommendation, Palo Alto, California, 2002b. California Secretary of State, Statement of Vote, 2002, available at http://ss.ca.gov/elections/sov/2002_general/contents.htm. Community Associations Institute, Data on U.S. Community Associations, 2003, available at http://www.caionline.org/about/facts.cfm. Construction Industry Research Board, Characteristics of New Homes Sold: California and Selected Counties, 1982–1999, Burbank, California, 2000. Devaney, John F., Tracking the American Dream: 50 Years of Housing History from the Census Bureau, 1940–1990, U.S. Department of Commerce, Washington, D.C., 1994. Dilger, Robert J., Neighborhood Politics: Residential Community Associations in American Governance, New York University Press, New York, 1992. DiPasquale, Denise, and Edward L. Glaeser, “Incentives and Social Capital: Are Homeowners Better Citizens?” Journal for Urban Economics, Vol. 45, March 1999, pp. 354–384. Egan, Timothy, “Many Seek Security in Private Communities,” New York Times, September 3, 1995, pp. A1, A22. Ellickson, Robert C., “New Institutions for Old Neighborhoods,” Duke Law Journal, Vol. 48, No. 1, 1998, pp. 75–110. Foldvary, Fred, Public Goods and Private Communities: The Market Provision of Social Services, Edward Elgar, Brookfield, Vermont, 1994. 74 Foldvary, Fred E., “Private Communities and Community Associations,” in David T. Beito, Peter Gordon, and Alex Tabarrok, eds., The Voluntary City: Restoring Urban Life in Crisis Times, Michigan University Press and The Independent Institute, Ann Arbor, Michigan, 2002. French, Susan F., Scope of Laws Affecting Common Interest Developments, background study prepared for the California Law Revision Commission, November 2000. Gordon, Tracy M., “Crowd Out or Crowd In? The Effects of Common Interest Developments on Political Participation in California,” The Annals of Regional Science, Vol. 37, No. 2, May 2003, pp. 203–233. “Government by the Nice for the Nice,” The Economist, July 25, 1992, pp. A25–A27. Helsley, Robert W., and William C. Strange, “Social Interactions and the Institutions of Local Government,” American Economic Review, Vol. 90, No. 5, 2000, pp. 1477–1490. HOA-Info Information Services Division, HOA-Info, Oakland, California, 2002. HOA-Info Information Services Division, California Community Association Statistics, Oakland, California, 2003. Kennedy, David J., “Note: Residential Associations as State Actors: Regulating the Impact of Gated Communities on Nonmembers,” Yale Law Journal, Vol. 105, December 1995, pp. 761–793. Kress, Carl B., “Comment: Beyond Nahrstedt: Reviewing Restrictions Governing Life in a Property Owner Association,” UCLA Law Review, Vol. 42, 1995, pp. 837–883. Lave-Johnston, Julia, and Kimberly Johnston-Dodds, Common Interest Developments: Housing at Risk? California Research Bureau, Sacramento, California, 2002. Massey, Douglas S., and Nancy Denton, “The Dimensions of Racial Segregation,” Social Forces, Vol. 67, 1988, pp. 281–315. 75 McKenzie, Evan, Privatopia, Yale University Press, New Haven, Connecticut, 1994. Murray, C., “The Shape of Things to Come,” National Review, Vol. 43, 1991, pp. 26–31. Nelson, Robert H., “Privatizing the Neighborhood: A Proposal to Replace Zoning with Private Collective Property Rights to Existing Neighborhoods,” George Mason Law Review, Vol. 7, No. 4, Summer 1999, pp. 827–880. Oliver, Eric J., Democracy in Suburbia, Princeton University Press, Princeton, New Jersey, 2001. Pack, Janet R., “BIDs, DIDs, SIDs, and SADs: Private Governments in Urban America,” Brookings Review, Fall 1992, pp. 18–21. Putnam, Robert D., Bowling Alone: The Collapse and Revival of American Community, Simon and Schuster, New York, 2000. Reardon, Sean F., John T. Yun, and Tamela M. Eitle, “The Changing Structure of School Segregation: Measurement and Evidence of Multiracial Metropolitan-Area School Segregation, 1989–1995,” Demography, Vol. 37, No. 3, August 2000, pp. 351–364. Reich, Robert B., “Secession of the Successful,” New York Times, January 20, 1991, Section 6, p. 16. Roland, Helen E., Residential Common Interest Developments: An Overview, California Research Bureau, Sacramento, California, March 1998. Rosenstone, S. J., and J. M. Hansen, Mobilization, Participation, and Democracy in America, Macmillan, New York, 1993. Senate Housing and Community Development Committee, Hearing, “Common Interest Developments (CID’s): Incorporation and Registration,” AB 643 Assembly Bill-Bill Analysis, July 1, 2002. Statewide Database, Statewide Database, 2000, available at http:// swdb.berkeley.edu/info/PDRDocument.txt. 76 Statewide Database, Documentation, 2002, available at http:// swdb.berkeley.edu/data/data.html. Theil, Henri, Statistical Decomposition Analysis, Amsterdam: NorthHolland Publishing Co., 1972. Tiebout, Charles M., “A Pure Theory of Local Public Expenditures,” Journal of Political Economy, Vol. 64, October 1956, pp. 416–424. Tocqueville, Alexis de, Democracy in America, in J. P. Mayer, ed., Anchor, Garden City, New Jersey, 1969 (1835), pp. 513–517. Treese, Clifford J., Community Associations Factbook, Community Associations Institute Research Foundation, Alexandria, Virginia, 1999. U.S. Census Bureau, Census of Housing, General Housing Characteristics California, U.S. Department of Commerce, Washington, D.C., 1970, 1980, and 1990. U.S. Census Bureau, Geographic Areas Reference Manual, U.S. Department of Commerce, Washington, D.C., 2000. U.S. Census Bureau, Census 2000 Summary File 3—California, Washington, D.C., 2002a. U.S. Census Bureau, Statistical Abstract, U.S. Department of Commerce, Washington, D.C., 1994, 1995, 1996, 1998, 1999, 2002b. U.S. Federal Elections Commission, National Voter Turnout in Federal Elections: 1960–1996, 2000, available at http://www.fec.gov/pages/ htmlto5.htm. Verba, S., K. L. Scholzman, and H. E. Brady, Voice and Equality: Civic Voluntarism in American Politics, Harvard University Press, Cambridge, Massachusetts, 1995. Worley, William S., J. C. Nichols and the Shaping of a Kansas City: Innovation in Planned Residential Communities, University of Missouri Press, Columbia, Missouri, 1990. 77 About the Author TRACY GORDON Tracy Gordon is a research fellow at the Public Policy Institute of California, where her research interests include state and local public finance and political economy. She has authored several journal articles and conference papers on topics relating to local government, including common interest developments and their effects on residential segregation and voting behavior, the local initiative process, and local property taxes. She holds a B.A. in human biology from Stanford University and a Ph.D. in public policy with a concurrent M.A. in economics from the University of California, Berkeley. 79 Related PPIC Publications Who Pays for Development Fees and Exactions? Marla Dresch and Steven M. Sheffrin The Ties That Bind: Changing Demographics and Civic Engagement in California S. Karthick Ramakrishnan and Mark Baldassare “Who’s Your Neighbor? Residential Segregation and Diversity in California” California Counts: Population Trends and Profiles Volume 4, Number 1, August 2002 Juan Onésimo Sandoval, Hans P. Johnson, and Sonya M.Tafoya “California’s Newest Neighborhoods” California Counts: Population Trends and Profiles Volume 5, Number 1, August 2003 Hans P. Johnson and Joseph M. Hayes 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 81" ["post_date_gmt"]=> string(19) "2017-05-20 09:36:22" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_304tgr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:36:22" ["post_modified_gmt"]=> string(19) "2017-05-20 09:36:22" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_304TGR.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) }