California's Political Geography

February 2012Eric McGhee and Daniel Krimm

Summary

California has a well-earned reputation as a strongly Democratic state. But it has important, sometimes unexpected, geographic variation in ideology and opinions on specific issues. There are many places where conservative and Republican candidates might win elections in California—including places where they currently do not.

The 2012 election season promises to be unpredictable and hotly contested. Two recent electoral reforms in the state—an open "top two” primary and new electoral districts drawn by an independent redistricting commission—have shaken up the status quo and left many political observers uncertain what to expect. Likewise, voters may be asked to consider more ballot measures than usual this fall—many of them bearing important implications for the future of the state.1

To provide context for this important election, the following report summarizes the broad outlines of California’s politics and political geography. It includes information on public opinion aggregated over the past four years of the PPIC Statewide Survey, which has been conducted often enough to permit a level of detail never before possible.

Geographic Variation

Today, California looks decidedly blue. Both the state’s U.S. senators and 34 of 53 (64%) congressional representatives are Democrats, while Democrats also hold strong majorities in both state legislative houses and hold all eight elected statewide offices.2 President Obama won 61 percent of the statewide vote in 2008. In November 2011, he polled well against Republican candidate Mitt Romney—50 to 40 percent (10% undecided).3

Although California votes solidly Democratic, Californians (including non-voters) hold important elements of conservative belief in most parts of the state. On an ideological scale ranging from strong conservative to strong liberal, public opinion data show the average Californian falling in the middle and leaning slightly conservative. Statewide, Californians are divided between those who say they are willing to pay higher taxes for more government services (48%) and those who would prefer paying fewer taxes and receiving fewer services (43%). And both non-Hispanic white and black Californians are slightly more likely to say that immigrants are a burden on the state (48%) rather than a benefit (44%).4

In fact, growth in Democratic support over time has not been uniform across the state, but has had a strong geographic dimension. It is common to say that a north-south divide—with the north voting Democratic and the south voting Republican—has been replaced with an east-west, or coastal-inland divide (with the coast voting Democratic and inland voting Republican). But this shorthand tells only part of the story.

In the late 1960s and early 1970s, the Bay Area was the most Democratic region in presidential elections, and the central and southern coasts outside of Los Angeles were solidly Republican. Presidential voting is a good measure of the overall partisan complexion of a community because presidential races are high profile and tend to address the central political issues of the day. Over time, every region except inland California has shifted toward the Democrats (Figure 1).

FIGURE 1. THE BAY AREA AND LOS ANGELES HAVE BECOME MORE DEMOCRATIC THAN THE REST OF THE STATE

Insert chart, map or table title here

Today, the north-south divide has faded, but its clearest replacement may not be east versus west. Rather, presidential politics now pits Los Angeles County and the Bay Area against everywhere else.5

The following maps (Figures 2 and 3) provide some perspective by showing the vote for Obama in 2008. Figure 2 is a traditional geographic view, using what we call "places”: mostly individual counties, but sometimes groups of sparsely populated counties or subsets of heavily populated counties.6

FIGURE 2. DEMOCRATIC VOTERS DO NOT APPEAR VERY NUMEROUS IN A LAND-AREA MAP

Insert chart, map or table title here

In Figure 2, the state appears slightly Republican or closely balanced. But population is distributed very unevenly across California, and the more rural, less populated places are more likely to be Republican.

Figure 3 distorts the state’s physical geography by sizing the geographic groupings according to their relative share of the state’s population. This creates a more accurate view of Obama’s support in the state—and emphasizes the role of both the Bay Area and Los Angeles in forming the state's Democratic reputation.7

FIGURE 3. VIEWING THE STATE BY POPULATION REVEALS DEMOCRATIC DOMINANCE

Insert chart, map or table title here

Regional Variation on Social and Fiscal Issues

Even the simple distinction between the Bay Area/Los Angeles and everywhere else begins to break down when we look past partisan registration and voting and examine specific issues. Figure 4 shows how places differ not by election results or partisanship, but by opinions on social and fiscal issues.8 The social dimension combines answers to questions about abortion and gay marriage, while the fiscal dimension uses questions about the size of government and the use of spending cuts to address the state budget deficit.

We created five opinion-based groupings of our geographic places:

  • Loyal Liberal: Very liberal on both social and fiscal issues (18% of the state’s population);
  • Moderate Liberal: Moderately liberal on both social and fiscal issues (24%);
  • Conservative Liberal: Conservative on social issues and moderately liberal on fiscal issues (25%);
  • Moderate Conservative: Moderately liberal on social issues and conservative on fiscal issues (17%);
  • Committed Conservative: Conservative on both social and fiscal issues (15%).

FIGURE 4. SOCIAL AND FISCAL OPINIONS ARE NOT STRONGLY DEFINED BY GEOGRAPHIC AREAS

Insert chart, map or table title here

In contrast to Figures 2 and 3, the coastal/inland distinction is harder to see here. Only one clear regional difference appears in Figure 4: Loyal Liberal areas are found only in the Bay Area and north coast. The other categories are dispersed throughout the state, and each has at least some coastal and some inland presence. Notably, Los Angeles County contains no Loyal Liberal areas and only one that is Moderate Liberal (coastal Los Angeles). The south coast (western San Bernardino/Riverside, and Orange and San Diego Counties) is comparably mixed and includes all categories except Loyal Liberal.

These results all point to a similar conclusion: California has indeed become more Democratic, but its liberal reputation is deserved only in the Bay Area and environs. In the rest of the state, even in Los Angeles County, California is more conservative and less consistently defined by geography than conventional wisdom would sometimes suggest.

For more maps on social and fiscal issues—including abortion, gay marriage, and reducing the state budget deficit—please visit PPIC’s map room.9

Geography, Ideology, and Independents

If the state as a whole is moderate but Democratic, how do ideology and partisanship compare at the local level? Nationally, Republicans are more conservative and Democrats more liberal and at first glance, California appears similar. But as the following maps show, the extent of this self-identification depends on both party and geography.

For instance, Republicans (including independents who lean Republican) are rarely liberal anywhere (Figure 5). PPIC Statewide Survey data indicate there is no place in the state where more than 14 percent of Republicans identify themselves as liberal (responding either "liberal” or "strongly liberal” to a question about ideology). In most places, the number is 10 percent or lower. The areas with the largest share of liberal Republicans include the Bay Area, parts of Los Angeles County, Monterey/San Benito Counties, and Imperial County.

FIGURE 5. REPUBLICAN VOTERS RARELY CALL THEMSELVES LIBERAL ANYWHERE

Insert chart, map or table title here

Democrats (and independents who lean Democratic) are more ideologically diverse (Figure 6). In every part of the state, conservative Democrats make up at least 9 percent of people who identify with Democrats—and in all but three places, they make up at least 12 percent. Furthermore, the numbers of either "conservative” or "strongly conservative” Democrats are quite high in some areas: 40 percent in the eastern portion of San Bernardino County, 35 percent in Imperial County, and almost 30 percent in eastern Riverside County and much of the San Joaquin Valley.

FIGURE 6. DEMOCRATIC VOTERS ARE QUITE CONSERVATIVE IN SOME PARTS OF THE STATE

Insert chart, map or table title here

What about independents specifically? Their numbers have doubled since 1992—from 10 to 20 percent of registered voters—potentially making them a powerful new force in California politics. Although they are often called "swing voters," in reality most independents lean toward one party or the other and vote reliably that way.10 In general, the balance of these independents tilts at least slightly Democratic in almost every part of the state (Figure 7).11

This leaning pattern tends to match the broader partisan complexion of a region.12 The balance of independents tilts more Democratic in places that also show a larger Democratic advantage. The most Democratic-inclined independents can be found in parts of the Bay Area and the north coast and in coastal and central Los Angeles—places where the proportion of Democratic leaners outnumbers the proportion of Republican leaners by 25 percent or more. In contrast, the conservative southern Central Valley has about as many independents leaning Republican as leaning Democratic.

In sum, independents are more likely to reinforce the state’s political status quo than to disturb it, although they also make the state more Democratic on election day than voter registration numbers might suggest.

FIGURE 7. REGISTERED INDEPENDENTS LEAN DEMOCRATIC IN MOST PARTS OF THE STATE

Insert chart, map or table title here

Electoral Implications

California may tend to vote for Democratic presidential candidates , but many places around the state espouse views that fall to the right of the Democratic Party’s typical positions. In fact, only the Bay Area is home to extraordinarily large numbers of people who hold opinions associated with the Democratic Party.

This could signal an opportunity for Republicans. Moderate Liberal and Conservative Liberal places contain half the state’s population and seem sympathetic to many conservative positions—yet they tend to support the Democratic Party. Still, altering this status quo could prove difficult. The small number of liberal Republicans in every part of the state implies that the party's electoral coalition is ideologically solid—but that may also make the party resistant to expansion efforts. Only time will tell.

Regardless, California’s 2012 election promises to be exciting and unpredictable. Given the inherent tensions between ideology and partisanship in much of California, the political geography that emerges after the election could look very different from the current landscape.13

 
 

NOTES

We would like to thank several people outside PPIC who made this report possible. Karin McDonald and Nicole Boyle at the Statewide Database graciously offered their time, expertise, and data to help us construct election returns and other statistics below the county level. We would also like to thank Stephen Blumberg at the Centers for Disease Control and Scott Keeter at the Pew Center for the People and the Press for their help in constructing cell phone weights. Andrew Gelman and Jeffrey Lax of Columbia University provided key advice on the use of multi-level regression and post-stratification. John Ridener of the UC Berkeley Earth Sciences and Map Library assisted in acquiring and using shapefiles for area codes. John Manning at the North American Numbering Plan Administration offered valuable information about the criteria by which area codes are defined. Several experts offered helpful thoughts on the project at various points along the way: Bruce Cain, Jack Citrin, Tim Hodson, Iris Hui, Thad Kousser, Max Neiman, and Tony Quinn.

Within PPIC, we drew heavily on the wisdom and expertise of Louise Bedsworth, Abby Cook, Arabella Cureton, Bob Gleeson, Richard Greene, Hans Johnson, Jed Kolko, Dave Lesher, Kate Reber, Davin Reed, Eric Schiff, Mary Severance, Linda Strean, Lynette Ubois, and Leigh Whittier. The PPIC Statewide Survey team—Mark Baldassare, Dean Bonner, Sonja Petek, Jennifer Paluch, Jui Shrestha, and Nicole Wilcoxon—deserve special thanks for the resource of the survey itself, as well as the hard work required to create the cumulative data file from which all the survey numbers here are ultimately derived.

  1. See Dan Walters, "2012 Could Wind Up Being a Big Year in California Politics,” Sacramento Bee, January 2, 2012 (www.sacbee.com/2012/01/02/4157178/dan-walters-2012-could-wind-up.html).
  2. These are governor, lieutenant governor, secretary of state, attorney general, treasurer, controller, insurance commissioner, and superintendent of public instruction.
  3. Field Poll, November 14–27, 2011.
  4. The state’s ideological scale is based on combined PPIC Statewide Survey data from 2008 through 2010. Preferences for government services and taxes are based on data from combined PPIC Statewide Surveys in January and May 2008; January, May, and September 2009; and March 2010. The immigrant benefit/burden question is based on PPIC Statewide Survey data from March, June, and December 2007; March and August 2008; and March and September 2010, combined.
  5. The pattern is broadly the same for both gubernatorial elections and party registration. For similar evidence of polariza-tion in California and a discussion of its origins, see Iris Hui, "Growing Geographic Polarization and the Perpetuation of the Electoral Disconnect” (PhD dissertation, University of California, Berkeley, 2010). The Bay Area/north coast consists of Alameda, Contra Costa, Humboldt, Lake, Marin, Mendocino, Napa, San Francisco, San Mateo, Santa Clara, Santa Cruz, Solano, and Sonoma Counties. The central and south coasts consist of Monterey, Orange, San Benito, San Diego, San Luis Obispo, Santa Barbara, and Ventura Counties. Los Angeles is just Los Angeles County, while the inland consists of all counties not otherwise assigned. For further information, contact the authors.
  6. For the sake of the survey analysis we change the standard county map by combining counties where survey data are scarce and dividing counties where survey data are plentiful. Details regarding these divisions and combinations are available from the authors upon request.
  7. This map was created in ArcMAP with an add-on utility that creates Density Equalizing Cartograms using the methodology developed by Mark Newman and Michael Gastner at the University of Michigan (Michael T. Gastner and M. E. J. Newman, "Diffusion-based method for producing density-equalizing maps", Proceedings of the NAS, May 18, 2004, vol. 101, no. 20, pp 7499-7504). Density Equalizing Cartograms change the shape of map polygons so that their size is based upon another attribute such as population. The size and shape of the polygons are changed, sometimes dramatically, but their original neighbors remain neighbors, and no new neighbors or new gaps are added.
  8. To increase the reliability of our estimates, we move beyond the simple demographic weights that the PPIC Statewide Survey always employs to a more sophisticated technique called multilevel regression and post-stratification (MRP). MRP first uses a multilevel model to assign a predicted opinion to each respondent based on a weighted average of two general factors: the respondent’s demographic characteristics, along with the region and political voting history of the place he or she lives; and the average point of view of the respondent’s place independent of these contextual variables. The larger the sample size in a given place, the larger the weight on the place’s unique point of view, and the closer the resulting values get to the raw unweighted data. These predictions are then weighted by census demographics in the manner typical of any PPIC Statewide Survey. MRP thus preserves as much unique information about a place as possible, and where sampling error is large, it asks the estimates to conform with expectations from other politically relevant data. We conducted extensive validation analysis to ensure that the model was working properly and not introducing any bias in the results. Please contact the authors for more details on the method, its application to this study, and the validation analysis.
  9. These issues include the fiscal and social questions from which the two-index categories in Figure 4 are derived.
  10. From 2008 to 2010, roughly 60 percent of independents interviewed by the PPIC Statewide Survey said they lean toward one party or the other. In the 2008 presidential, 2010 gubernatorial, and 2010 U.S. Senate races, independent leaners in the survey voted with at least as much loyalty as respondents who identified openly, but weakly, with a party. For a more detailed argument about the loyalty of independent leaners, see Bruce Keith et al., The Myth of the Independent Voter (University of California Press,1992).
  11. The denominator for these calculations is all independents, making these numbers similar to the party balance numbers reported earlier.
  12. Independent voters who lean toward one party or the other have been included in Figures 5 and 6, so the consistent conservatism of Republicans seen in Figure 5 incorporates the potential moderating effect of independent voters.
  13. A full set of numeric tables for measures cited in this paper and other data are available on request.