Acknowledgements:

This study and the public-use data set created by it would not have been possible without the generous support of the John Randolph and Dora Haynes Foundation. The author wishes especially to thank Diane Cornwell for her strong interest in the project from its inception, and her support in realizing its completion.

Los Angeles became notorious in the 1990s as the national leader in urban social and political problems. The 1992 civil disorder and the successful passage of the anti-immigrant Proposition 187 in 1994 together seemed to signify the disintegration of the social contract. The extreme strains put on this metropolis raised many questions about whether Los Angeles could continue providing opportunities and an attractive way of life to its millions of citizens.

Answering these vital questions about the future viability of Los Angeles have been hampered by an almost complete absence of systematic knowledge about the patterns of change in the metropolis over long periods of time. Current attempts to address social and political issues such as inequalities of opportunity and neighborhood empowerment demand that we have a clear picture of how the metropolis of Los Angeles has evolved over time and in space. This is the first systematic study of race-ethnicity, socioeconomic conditions, and political behavior in Los Angeles County to cover the entire period from 1940 through the early 1990s. Using advanced Geographic Information System (GIS) techniques, this study has created a portrait of the County of Los Angeles, its 88 cities, and its 113 separate unincorporated spaces, for every census year from 1940 through 1990. Especially as we anticipate the release of the 2000 Census in the spring of 2001, we need to understand the recent and long-term trends of the Los Angeles, upon which much attention will be focused.

The critical period to examine starts with the 1940s, when Los Angeles began to grow dramatically and became vastly more diverse in terms of race and ethnicity. Asking how life chances and inter-group relations have changed in Los Angeles during the last sixty years of the 20^th century requires some very specific attention to spatial relations, as these relations have been structured through periods of official and de facto segregation. Over the period from 1940-1990, this study seeks to answer the following core research question: Over time, what has been the spatial relationship between social groups and how does spatial fragmentation affect life chances and political behavior?

The principal findings of this study give us real reasons to be concerned about the socio-political condition of the metropolis as we enter the 21^st century. The most important findings are:

1. "Segregated Diversity": Despite the dramatic growth in diversity, White isolation from other groups has increased and become entrenched in socioeconomic stratification.

Despite the dramatic growth of diversity in Los Angeles County, race-ethnic groups have remained and have even become more isolated from one another, in patterns that are entwined with socioeconomic stratification. The old boundary between Whites and Blacks has been vastly modified by a new balance of pluralities between Whites, Hispanics, Blacks, Asians, and other groups. But Whites have increasingly become spatially separated from Nonwhites, barricaded behind walls of wealth in municipal spaces far away from the center of the metropolis.

2. Class Stratification has Replaced Jim Crow to Preserve Segregation

Viewed historically and collectively, Los Angeles made an early investment in the "Whiteness" of its suburban ring. Officially sanctioned segregation prior to 1965 preserved opportunity in the fastest growing and most desirable outlying areas (especially along the Pacific Coast) for Whites only. After 1965, these areas became inaccessible to Nonwhites primarily by virtue of property values. Massive immigration and industrial restructuring in the 1970s and 1980s established a Nonwhite core that is dramatically more blue collar and has median house values below the County mean. Class stratification stepped in to replace Jim Crow. Municipal spaces have played an ambiguous role in this process. Prior to 1970, independent cities were tools of racial exclusion and nearly all diversity was contained within Los Angeles City. By 1990, fifty incorporated cities had nonwhite majorities. Half of these were very diverse. But the relative isolation of race-ethnic groups from one another has continued to grow and municipal spaces remain starkly divided between wealthy, majority-White and poor, majority-Nonwhite. The great exception is Los Angeles City, which successfully combines vastly different race-ethnic and socioeconomic class groups within one political community. These patterns highlight the need to arrest the current municipal secession movements in order to stem the trend toward greater isolation. Only political community can transcend socio-spatial segregation.

3. Space Matters: Physical Distances Have Powered Race Politics

Physical distance (measured in miles or kilometers) from the centers of Black and Hispanic population proved highly significant in explaining the race-charged voting in 1964 (repeal of the Rumford Fair Housing law), 1978 (passage of Proposition 13), and 1994 (passage of Proposition 187). So too did the type of municipal space voters lived in. Incorporated cities and cities incorporated after the Second World War were far more likely to return majorities for Propositions 14 (1964), 13 (1978) and 187 (1994). Although these location variables proved to be the strongest, the patterns in these votes also showed a remarkable continuity of hostility between Whites, Blacks, and Hispanics, and a Black-Hispanic rift evident as early as 1964.

If the trends charted in this study continue, Los Angeles County is heading for deep trouble in the 21^st century unless its citizens begin to forge political contexts for inter-group community. The socioeconomic forces creating segregated spaces are too deep to reverse in less than a few decades. Short of some unlikely totalitarian project or a surprising end to capitalist economic forces, we cannot physically relocate the millions of persons it would require to achieve a landscape that did not isolate race-ethnic groups behind class barriers. Even amid the stark divisions of the poisonous campaign of 1994, however, there was some indication of possibilities for breaking the cycle of racial polarization and segregated diversity, through alliances across race and class trench lines. Patterns throughout this study show how important it will be to maintain and expand political jurisdictions that contain the dramatically disparate race-ethnic and socioeconomic spaces.

The only immediately available solution to spatial isolation is political community. We can and should increase the opportunities for the residents of these segregated spaces to come together productively. City boundaries present a paradox: they can both divide and unify. Although municipal spaces have not, overall, mitigated an onward increase toward greater levels of segregation, a great many municipalities have become racially diverse since the 1960s, and examples abound of the capacity of Angelenos to build civic communities that transcend the traditional race-ethnic divisions. This potential is best seen in the case of the City of Los Angeles, whose boundaries snake through all possible race-ethnic and class territories of the County. Los Angeles City in fact represents the last best hope of community in Los Angeles County. It alone contains the entire spectrum of race-ethnic and socioeconomic class strata in the metropolis.

We need to move beyond the territorial politics of group division and seek political means that will create a regional, metropolitan civic community. Stemming the tide of municipal separatism will be an important first step. The results of this study clearly indicate that the county needs fewer, not more, fragmented political spaces. A counter-movement to prevent further municipal disintegration can possibly provide a valuable lesson in building a community upon the unstable foundations of segregated diversity.

Segregated Diversity? This phrase seems contradictory. "Segregation" has long spelled racial inequality and injustice in the United States, and "diversity" has emerged in the 1990s as a positive term connoting interracial/interethnic exchange, multiculturalism, and fairness. The United States is full of contradictions, however, and no less so Los Angeles. Los Angeles has two parallel traditions: the earliest being diversity and the most dominant being segregation These contradictory traditions begin with the multiracial and mostly nonwhite founders of the city in 1781, and with the self-consciously racist dreams of a white metropolis voiced by many Anglo leaders in the 1920s. During and immediately after the Second World War the interracial tensions that had been merely palpable in Los Angeles exploded into some of the most appalling and open race conflict in American history, all perpetrated by whites against nonwhites. The "internment" in 1942 of all persons of Japanese ancestry was carried out chiefly in Los Angeles, where most Issei and Nisei lived. The so-called "Zoot-Suit riots" of 1943 unleashed days of violence against Mexican-Americans. Immediately upon demobilization in 1946, a series of serious race conflicts, many involving African-Americans as targets, broke out. As the Civil Rights movement –inspired by the War in many ways—began in Los Angeles, a broad coalition of civic organizations created in 1947 the Community Relations Conference of Southern California (CRCSC). The following year the County of Los Angeles, recognizing the potential for deepening social unrest, created the Commission on Human Relations: one of the first Human Relations Commissions (HRCs) in the United States. As we argued in "The Challenge of Inter-group Relations in Los Angeles: An Historical and Comparative Evaluation of the Los Angeles City Human Relations Commission" (Ethington and West, 1998), the efforts of human relations commissions in Los Angeles have been largely ineffective in the face of large-scale social and political forces promoting inter-group animosity.

Despite the well-meaning but largely informational efforts of Los Angeles area HRCs, serious race-ethnic conflict has left its mark on Los Angeles. Two events in the 1960s showed how little progress had been made since the outset of the Second World War: the popular repeal by referendum vote of the Rumford Fair Housing Law in 1964, and the Watts Uprising in 1965. Festering tensions in the aftermath of these events was muted perhaps by the inspiring example of Los Angeles City Mayor Tom Bradley, an African-American elected by a broad coalition of whites and people of color. Although race tensions surfaced in many ways during the 1970s and 1980s, few were prepared for the ferocity of the first multi-racial "riot" in U.S. history, the so-called Civil Disorder of 1992. Bradley, shamed by the experience, retired from the political scene under a cloud of public pessimism. That deadly riot was followed just two years later by the bitter campaign for and against Proposition 187, a measure openly aimed at the region’s Latino (hereafter Hispanic) immigrants. Although ostensibly Proposition 187 attempted to deny public services only to illegal immigrants, its sponsors made no secret of their animosity toward Hispanic immigrants in general. Its passage by wide popular margins across the state was understood as a sign of intolerance and a harbinger of future conflict. The O.J. Simpson murder trial, which rapidly came to hinge on questions of systematic racism in the LAPD, seemed to focus and distill the mass public’s pent-up racial mistrust.

During most of the 1990s, the public, political leaders, and a wide range of writers and academics struggled to make sense of this gargantuan metropolis that was suddenly gaining a very ugly global reputation. Most studies have taken their cue from the prophetic writings of Mike Davis, whose City of Quartz (1990)--which seemed to predict the 1992 riot--portrayed Los Angeles as a "fortress" commanded by powerful whites who were increasingly preparing themselves for race war. Davis continued these themes in his 1998 Ecology of Fear. Davis’s work has been joined by a remarkable upsurge of serious scholarship about Los Angeles, much of it establishing, for the first time, a strong framework for understanding what lies beneath the deep tensions that have marked the past decade. Outstanding works of this type include the multiple-author volumes Ethnic Los Angeles (1996), edited by Roger Waldinger and Mehdi Bozorgmehr; Rethinking Los Angeles, edited by Michael J. Dear, H. Eric Schockman, and Greg Hise, and The City (1997), edited by Alan Scott and Edward Soja.

Two major and rather obvious dimensions were being neglected within this growing research agenda, however: space and time. Of course, many scholars have treated these dimensions in some form. Historians have narrated various aspects of the metropolis and its many cities, and sociologists have studied its present condition, while geographers have diligently mapped ethnic territories using the 1980 and 1990 censuses. Political scientists have studied the dynamics of recent elections. Until this study, however, we have had no way to understand the longer-term patterns in the social and political geography of the metropolis as a whole.

This study differs from those that have preceded it in several key respects. From the outset, the goal has been to achieve a new level of coherence in the empirical basis for interpreting the history and condition of Los Angeles County. To this end, the initial proposal to the Haynes Foundation was to build a single, robust data set for the years from 1940 through the mid-1990s, so that several dimensions of social and political experience can be analyzed by the same criteria and definitions in all of the six decades under examination. The author and his coworkers spent the first three years of the study (1996-1999) assembling the statistical database, which is now available for public distribution: Ethington, Kooistra, and DeYoung, Los Angeles County Union Census Tract Data Series, 1940-1990 (Los Angeles: USC, 2000). This data set, composed of a single table of 1,656 records (rows) and 252 variables (columns), is organized with the goal of providing maximum uniformity and comparability across all census years and census geographies. It includes aggregate data describing the characteristics of U.S. Census Tracts for race-ethnicity, occupation, education, housing, age and median household income, house value, and rent. It was assembled from two principal sources: 1) the print-published US Census tables for the years 1940, 1950 and 1960, and 2) the digitally distributed "Correspondence" data files created by the California State Department of Finance for the years 1970, 1980, and 1990. All data have been proportionally aggregated to fit the 1990 census tract geography.

This study also began with the goal of asking and answering spatial questions about this metropolis. That goal required the intensive use of Geographic Information System (GIS) technology. GIS methods were necessary both for the data assembly (1996-1999) and data analysis (1999-2000) phases of the research. In order to study spatial change between census years, data originally organized by different census tract geographies had to be proportionally fitted to a common geography. I chose the 1,656 tracts of the 1990 U.S. Census to serve as the universal geography for this study. (This choice will also maximize comparability with the 2000 Census). Election data are collected by precincts, of which there are many more than census tracts across Los Angeles County. The thousands of precinct boundaries from 1964, 1978, and 1994 were proportionally fit to the 1990 census tracts as well. For a technical methodological description, see Appendix A: User’s Guide for the Ethington, Kooistra, and DeYoung, Los Angeles County Union Census Tract Data Series, 1940-1990, Version 1.01.

Using this new data set has enabled me to go beyond some of the major limitations in studies to date. For instance, while several scholars have produced valuable collections of race-ethnic maps, none to date have contextualized these race-ethnic geographies within the socioeconomic geography of the metropolis, nor have they systematically shown the relationship of these race-ethnic geographies to the complex political geography of the region. The research questions guiding this study hypothesized a very important role for municipal boundaries in particular. Accordingly, we fitted all data not only to the 1990 U.S. Census tracts, but also to the 1994 municipal boundaries. The municipal geography of Los Angeles County is immensely complex. It is the only major metropolis in the United States with a very large proportion of its residents living in unincorporated territory (about 1 million of the 9 million residents in 1990). Further, these unincorporated areas are highly fragmented: 113 separate unincorporated spaces lie within and between the 88 incorporated municipalities of the County, for a total of 216 municipal spaces. (Constitutionally, the County itself is a "municipality," so all of these spaces are referred to as "municipal spaces." Incorporated municipal spaces are referred to as "cities." ) Analyzing data according to this complex geography even necessitated the creation of a new naming system for the otherwise undesignated 113 spaces. The best known is East Los Angeles, the historic home of the County’s Mexican-American population.

The following sections of this Report begin with a general overview of the transformation of Los Angeles County from 1940 to 1990 (Section II), and then summarize, in Sections III, IV, and V, the principal findings of the study according to research questions in three areas: 1) race-ethnic segregation, 2) municipal stratification, and 3) political polarization. I must point-out as well that this study has intentionally generated far more useful data than one researcher can possibly analyze within several years, and I do not pretend to have exhausted the possible conclusions from these data. Further, in the interest of brevity and clarity, each of the three principal sections of this Report are merely brief summaries of three major scholarly journal publications under preparation for submission by the author.

Los Angeles County grew from an official total population of 2,809,946 to 8,902,172 between the census years 1940 and 1990. As displayed in Chart 2.1, this spectacular growth was composed of several very important population shifts. The African American population (in deference to the Census terminology, hereafter "Black" and "Black Non Hispanic") population grew most rapidly during and immediately after the Second World War, due in large part to a massive migration for jobs in the war industries. After 1960, however, the relative proportion of Blacks began to decrease, countered by the rise of the Latino (hereafter "Hispanic") and Asian populations. The epochal Immigration Reform Act of 1965 opened the doors to a massive influx of immigrants from Latin American and Asia and Pacific Island nations. Chart 2.2 displays these relative group proportions, and also illustrates that between 1980 and 1990 White Non Hispanics fell from a bare majority to a minority of the overall population (41%).

Map Series 2a shows how two of the four principal race-ethnic groups have grown and changed location. As the Black population grew, it also shifted from a corridor along Central Avenue, to a new one to the west, along Crenshaw. As the Hispanic population grew, it did not shift so much as explode, from many nodes that had historicall been home to Mexicans and Mexican-Americans.

We also want to know about the general socioeconomic condition of the people of Los Angeles County. Chart 2.3 shows the proportion of persons in the two major occupational classifications, Blue Collar and White Collar (See Appendix A for technical definitions). In view of just these occupational categories, Los Angeles remained remarkable constant over this sixty-year period, which saw the rise and fall of mighty blue-collar industries in automobiles and rubber: plants that closed during the "restructuring" of the 1970s and 1980s. Although heavy consumer durable production declined, other, lighter industries rushed in to fill the void. Los Angeles has been transformed form one of the United States’ leading workshops to one if its leading sweatshops, with a huge garment manufacturing industry. Those light, small-shop industries have come to rely on immigrant laborers. Chart 2.4 shows that the proportions of Whites in each census tract, and the proportion of blue collar workers in those tracts has always been inversely correlated (as the percentage White Non Hispanic increased, the percentage blue collar decreased). The severity of this inverse correlation has, in addition, steadily grown, from a mild -.28 to a steep –.81. In other words, we are witnessing a virtual disappearance of the white working class, and a nearly complete ordering of census tracts along a spectrum of white collar White Non Hispanic tracts, to blue collar Non White tracts.

Map Series 2b shows the location of the Blue Collar population in 1960 and 1990. The remarkable lesson that emerges from these two maps is that Los Angeles County has not only retained but expanded a very large Blue Collar territory at its core. As the deepening shades of blue indicate, that core has also become much more Blue Collar by proportion. It has also become much more centrally focused.

The phenomenal growth of the Los Angeles metropolis since 1940 has steadily pushed up residential property values, as shown in Chart 2.5, which uses values indexed to the Consumer Price Index base year of 1983. Not all tracts benefited equally from this real estate boom, however. Chart 2.6 shows that the proportion White Non Hispanic and median house value have been increasingly correlated from 1940 though 1990. This positive correlation has grown from a modest .22 to a relatively steep .72. Chart 2.7 compares the correlation between all four race-ethnic groups in each census tract and the median house values of census tracts from 1940 to 1990. White the correlation has always been positive for Whites, it has been negative for the other three race-ethnic groups in all years except in 1990 for (Asians.) And while the (Asian) correlation with median house values has slowly but modestly improved, it has gotten worse for Blacks and Hispanics.

The relationship between median house values and race-ethnic groups in specific spatial locations is a very important one, which can be missed by inadequate attention to what John Logan and Harvey Molotch (1990) have termed the "political economy of place." Tests of "even" and "normal" distribution on the median house values of the 1,652 census tracts shows that tracts became more and more evenly distributed along a classic Gaussian "bell curve," a curve which became less and less skewed in an upward direction. Census tracts and municipalities, however, do not exist in abstract mathematical space. They are shapes on the ground and do not move over time. As the historical process rolls forward these shapes on the ground prosper or suffer as a function of many factors, and they show a spatial signature of the processes that shape them through time. Chart 2.8 shows how the residential prosperity generated by the long boom of Los Angeles County has been unevenly distributed among the races, living as they do in specific municipal spaces.

Although it is a complex graph, it can be easily interpreted as follows: Over time, the municipal spaces in the top quartiles of median house values became disproportionately filled with White Non Hispanics.

This preliminary survey of the overall transformation of Los Angeles County clearly indicates racially uneven development. We turn now to a spatial examination of that unevenness.

"Segregation" is the term we have used since the Civil Rights era to discuss the spatial separation of social, usually race-ethnic, groups. The Watts Riot of 1965 and its aftermath raised consciousness of the fact that Los Angeles was as segregated as the former slave states. African Americans numbered about a half-million in the 1960s, but most were concentrated in a very small district centered along Central Ave in the City of Los Angeles. Segregation was widely presumed to be a symptom of an unequal society, and several lines of attack on segregation were established. To eliminate residential segregation, fair housing laws were passed and discriminatory practices were outlawed in the courts. School busing was implemented in the Los Angeles Unified School District. Affirmative Action was implemented by public entities and many private employers.

Because so many civil rights measures were adopted by the end of the 1970s, public discussion of segregation all but disappeared in the 1980s and 1990s. But the problem not only persisted: it was radically transformed by the rise of a multiracial society resulting from the epochal Immigration Reform Act of 1965. American policymakers and scholars have been accustomed to discussing racial segregation as a question of spatial separation between just two groups: Whites and Blacks. What happens when we introduce at least two other major groups: Hispanics and Asians? What measures should we use to measure multiracial spatial relations? Led by Douglas S. Massey and Nancy A. Denton, only a handful of scholars have continued to look at these questions. Massey and Denton's numerous studies have persistently shown that measures of segregation have not diminished in U.S. metropolitan areas, despite the victories of the Civil Rights movement. While their studies are certainly of the first order of importance, they have been conducted at rather high level of abstraction, confined to tables (their principal writings contain no maps) and relatively recent time periods.

The present study of Los Angeles County sought to draw a finer picture than that provided by recent scholarship, to address in particular the nuances of the multi-racial character of Los Angeles since the 1960s, and to take into account the longer term processes at work. I measure spatial separation between four "groups": White Non Hispanic, Black Non Hispanic, Other Non Hispanic (Asian), and Hispanics. I have employed two principal measures of segregation that are well suited to regions with multiple racial groups. The Entropy (or H) Index, measures levels of diversity. For clarity, I will refer to it simply as the "Diversity Index." The Diversity Index measures the likelihood that members of multiple groups will come into contact with one another. It ranges from 0 to 1.4 and increases with the increasing balance of groups in each spatial unit (in my case tracts and municipalities). If all four groups were present in equal proportions (i.e., at 25%), this index would be at the maximum of 1.4. If only one group were present, it would equal zero.

The Diversity Index is very good for revealing the nature and extent of interracial mixing across the metropolis, but it does not tell us about the specific experiences of each distinct group. To measure those experiences, I have deployed the Isolation Index. The Isolation Index has many strengths and is well tested by sociological methodologists (Lieberson 1981). The Isolation Index measures the level of spatial isolation members of a given group experience in a metropolis, and also the probabilities of residential interaction between of members of one specific group and members of other specific groups. It has the great advantage of being asymmetrical. This means that it is sensitive to the relative sizes of different groups, showing for example that the probability of residential interaction of Blacks with Whites is different and not merely the inverse of the probability of residential interaction of Whites with Blacks.

1. "H" Entropy (or Diversity) Index

	K
H=	-Σ	Pk	Log Pk
	k=1

Where:

N=total population size

N_{k =}number of persons in Kth group

P_{k =} N_k /N

 

2. Isolation and Interaction Index

(3) Index of Dissimilarity

One of the most striking patterns in the six decade development of Los Angeles charted in this study is the growth of race-ethnic diversity in a spatial as well as a general sense. It is well known that the number of national language groups residing in Los Angeles County make this region a truly global one. The largest Mexican, Korean, and Salvadoran communities outside of those nations is a hint of this rich diversity. The nature of the data set used in this study, however, limits our view to only four groupings: White Non Hispanic, Black Non Hispanic, Other Non Hispanic (which effectively means Asian for most years of this study), and Hispanic (all races). Even so, it is quite clear that diversity has grown dramatically. Chart 3.1 shows a steady increase in the balance between proportions of the four race-ethnic groups (White Non Hispanic, Black Non Hispanic, (Asian), and Hispanic) within the tracts of the County, from 1940 to 1990.

The Diversity Index also produces a meaningful statistic for each tract, and these values, when mapped, show a remarkable and spread of diversity outward from the center of Los Angeles County from 1940 through 1990, as seen in Map Series 3A, "Diversity Index" (1940-1990). Series 3A is mapped by a procedure which averages each pixel with its "neighbors" in a ten-pixel radius. The effect is to smooth the patterns and highlight "hot spots" of high diversity and "cold spots" of low diversity. The darker shades of yellow show the most diverse, and the lighter shades show increasing homogeneity. It is important to remember that the "homogeneity" can be of any race-ethnic group. By 1990, in fact, the two most homogeneous "cold spots" were centered in Pacific Palisades on the far West Side, (homogeneous White Non Hispanic) and in East Los Angeles (homogeneous Hispanic).

Viewed in succession, it becomes clear that in the years 1940 through 1960, the most "diverse" areas in the County were highly concentrated at the center. This is because the diverse areas were coterminous with Nonwhite segregated areas. It was only possible live in "diverse" neighborhoods where people of color were strictly confined. As the walls of overt segregation began to fall in the 1960s and 70s, areas of diversity began to spread outward from that center, which itself become more and more homogeneous (first as Black and then as Hispanic). The most diverse areas began in the 1980s and 1990s to form a distinct ring around the metropolitan center, between the more homogeneous Black and Hispanic neighborhoods at the core, and the more homogeneous White neighborhoods at the periphery. These broad bands of high diversity can be understood as "Borderlands," or zones of transition and interaction between and among members of all race-ethnic groups. Most Angelenos are familiar with the visual landscape of these areas. Strip malls typically advertise their shops with up to six different languages, and it is impossible from looking along the street to tell which race-ethnic group predominates in the area. These Borderlands seem to suggest a future of "mestizaje," or willful mixture in Los Angeles. Or do they? Before we can pass judgement on that possibility, we need to examine other dimensions of segregation.

As stated above, the isolation/interaction index is an asymmetric measure of intergroup interaction, sensitive not only to the proportions but also to the relative population sizes of different the groups. Where X and Y denote different groups in a city, the isolation index measure takes the form: xP*x, and the interaction index takes two forms: (2), xP*y, and (3) yP*x. The subscript before and after the P* indicate "respectively, the group from whom and to whom interaction is directed." Thus, for a randomly selected member of Group X in a city, xP*x gives us the probability the next person that person will meet in the same census tract will also be a member of Group X. This is the measure of the group’s "isolation" from all other groups in the city. For an estimate of the probability of residential interaction with any other specific group, xP*y gives us the probability that, for a given, randomly selected individual, the next person she or he will meet will be a member of Group Y.

Chart 3.2 shows the probabilities that Whites, Blacks, (Asians), and Hispanics would primarily interact with members of their own groups in their own specific census tracts of Los Angeles County, from 1940-1990. Each group shows a different pattern. Whites were almost certain of interacting only with other Whites in their census tracts in the years 1940-1960 (90% probability or greater), but become markedly less isolated from other groups from 1970 through 1990. This is not surprising, since the White proportion of the County overall dropped to about 41% and, as we have seen, the number of highly diverse census tracts spread more widely. Blacks became more and more isolated through 1970, and then less so as their population in the metropolis declined relative to other groups. Hispanics showed the most dramatic transformation, from a very low level of isolation in 1940 (less than 10%) to almost 60%. (Asians) and Hispanics have been the least isolated overall, but have become increasingly isolated from other groups.

A glass half empty, however, is not always a glass half full. Chart 3.3 shows that although Whites were becoming less isolated from other groups from 1940 thorough 1990, those other groups had a different experience of Whites. In 1940, there was a greater than 80% probability that a Hispanic would next run into a White in his or her own neighborhood. That probability had dropped to just over 20% by 1990. Blacks were and (Asians) were similarly less likely to meet Whites in their tracts by 1990.

Chart 3.4 shows even more remarkably how Blacks have remained isolated from all three other groups. The probability of White and Hispanics residential interaction with Blacks has remained less than 10% for all years from 1940-1990. The probability of (Asian) interaction with Blacks has sometimes been slightly higher, but had fallen again to less than 10% by 1990. The probability of White residential interaction with other Whites is included in this chart for comparative reference. In Chart 3.5 we see that the probability of Black, White, and (Asian) interaction with Hispanics increased from 1940 to 1990, but only modestly. Again, the probability of White residential interaction with other Whites is included in this chart for comparative reference. Chart 3.6 shows that the probability of White, Black, and Hispanic interaction with (Asians) has been the lowest of all: less than 5% until recently, and still not above 10%.

These results seriously complicate the optimistic inference we might have drawn from the steady increase in "diversity" and the spread of borderlands over these six decades in Los Angeles County.

The most traditional measure of segregation is the Index of Dissimilarity. This familiar statistic measures unevenness of populations across spatial subareas (census tracts in this case) within a city, and is extremely easy to interpret. It tells us, simply, what percentage of a group would need to move their residential location to another part of the city in order to become evenly spread across the city. A "0" value indicates total evenness and lack of segregation, and "1.0" indicates total unevenness or maximum segregation by this measure. The chief disadvantages of this measure are a) that that it can only be used with two "groups" at a time (e.g., Whites/nonwhites; Blacks/non-Blacks, etc), and b) it is insensitive to relative population sizes. Given its widespread familiarity, however, it is not without merit, and the results can only add to our understanding of segregation in Los Angeles. Chart 3.7 shows mixed results. On the one hand, overall indices of dissimilarity for each group (measured against all other groups) have fallen slightly. Emphasis must be placed on the ‘slightly," however. It is still the case that 64% of all Blacks, 57% of all Whites, and 50% of all Hispanics, and 41% of Other Non Hispanics (Asians) in the County would need to move their place of residence to achieve an even distribution.

Map Series 3b makes visible the phenomenon of "clustering," or the tendency of groups to live in closely compact areas. The values mapped in Series 3b are the percent White Non Hispanic, shown as standard deviations around the mean. This method of mapping reveals the relative intensity of race-ethnic clustering, which clearly increased from 1940-1990. In 1940, Non-whites (displayed in blue) were clustered in central areas and a series of pockets (many of them unincorporated county spaces that were less easy for Whites to control), while Whites were far more dispersed evenly across the county. By 1990, we see the Whites are shown in darker shades of red—indicating higher concentrations, and these White clusters are unevenly concentrated around the rim of the county.

Map Series 3b also reveals something mentioned already: demographic patterns seem to coincide with municipal boundaries. The next section examines this relationship in more detail.

4. The Rise of Diversity Versus the Historical-Geographic Investment in Whiteness.

We have seen that the probability of Blacks, Hispanics, and (Asians) having residential interactions with Whites has decreased steadily, even while Los Angeles County has become more diverse. We have also seen that the wealth of the county, measured in terms of median house values, has become increasingly associated with space occupied by Whites. And we have already seen some indications that the municipal boundaries have some relationship with these patterns. Just what is the relationship between municipal spaces and "segregated diversity?" Throughout this section I shall refer to "municipal spaces" because the County is broken into 88 cities and 113 unincorporated spaces, and both types of space are significant.

The process of "incorporation," by which residents of a given territory constitute themselves as a "city" is a very important one, because once they gain that status, people fundamentally alter their relationship with one another and with others outside of that territory. Incorporated cities have the powers to establish police, fire, and other departments, to incur debts, condemn land, and design their own government to do all of these things and more. The most common use of "city-hood" has been to enhance and maintain the "quality of life" as that phrase is interpreted by each residential community. In the 19^th and first half of the 20^th centuries, leaders of cities used their powers first of all to build "growth machines," or coalitions of business, real estate, political, community, and media interests sharing the common goal of building population size and property values. Increasingly during the second half of the twentieth century, city leaders have also used their municipal powers to decrease the rate of growth and protect values already attained. They have also used their city powers to maintain racial exclusions, insofar as residents saw exclusivity as fundamental to their "quality of life."

Given this brief sketch of the special status of incorporated spaces, it should become rather obvious that these have been the most important tools with which people have invested abstract "shapes on the ground" such as invisible municipal boundaries, with concrete social and physical differences.

In Map Series 4a we can watch the gap between relatively rich and poor municipal spaces grow from 1940 through 1990. Each municipal space is shaded to show how far its median house value was from the county mean in each year. Shades from pink to dark red rise above the mean, and shades from light to dark blue fall below the mean, all in standard deviation increments. The highest house value municipalities in 1940 were centered on Beverly Hills, but the highest of all was San Marino, with house values three times higher than the county mean. In 1950 the shades of red and blue are not so dark, suggesting a moment of middle-class opportunity that we associate with the postwar low-cost housing boom. In 1960, however, the contrasts had sharpened, and a visible core of relative poverty appeared in South Central Los Angeles. This relatively poor core spread from the current Los Angeles City District 9 and adjacent City of Vernon and LA County space, to include, by 1990, LA City District 8, a large County space I call the "Athens Interzone," another County space adjacent to Huntington Park, the City of Compton, a large County space adjacent to Compton, and the City of Lynnwood. These poorest municipal spaces are neatly surrounded by another ring of below-mean house value spaces. During this same period the rich municipal spaces have become more tightly defined by the Pacific coast, but they also form an intermittent ring approximately 20 kilometers from Los Angeles City Hall. These are the wealthy suburbs of the Los Angeles metropolis. By 1990 they included, reading clockwise from the far west side: Malibu, Los Angeles City District 11, Los Angeles City District 11, Beverly Hills, La Canada Flintridge, San Marino, Arcadia, (The list would surely include Orange County municipalities in the southeast], the cities on the Palos Verdes peninsula: Rancho Park, Rolling Hills, Rolling Hills Estates, and Palos Verdes Estates, and the Marina Del Rey (L.A. County territory).

As we might expect given the indications in earlier sections of this study, these highest- house value municipalities were, with a few exceptions, also the "Whitest" in terms of race-ethnicity by 1990. Table 4.1 presents a "profile" of the 20 highest value municipal spaces in 1990. The exceptions are San Marino, with 32% "Other Non Hispanics," a category I infer to be largely Asian in this study. Rancho Park is also "less White" because of a 20% Other Non Hispanic population. I have produced in Appendix B a set of "Municipal Profiles," sorted alphabetically, and by the 60 Most and 60 Least "Diverse," "White," "Black," "Asian," and "Hispanic," for the years 1940, 1960, and 1990. Perusing these lists is very instructive. One finds that the "most diverse" municipal spaces in 1940 were, with only one small exception, actually overwhelmingly White (the exception was the future site of Marina Del Rey. The majority of 220 residents there were Asian—presumably of Japanese descent, because they disappeared by 1950—victims no doubt of the infamous Japanese "Internment" of 1942. In 1990, the "most diverse" municipal spaces were led by Carson and Gardena, each with almost equal balances between Whites, Blacks, (Asians), and Hispanics. Of the 20 "most diverse" municipal spaces in 1990, however, only one of considerable size—the City of Walnut—had median house values above the County mean, and the great majority of these spaces were below the mean.

"Least Diverse" by 1990 meant both overwhelmingly Hispanic and overwhelmingly White. The top five "least diverse" municipal spaces in 1990 were: 1) unincorporated East Los Angeles (94% Hispanic); 2) City of Maywood (93% Hispanic); 3) City of Vernon (91% Hispanic); 4) City of Huntington Park (92% Hispanic); and 5) City of Commerce (91% Hispanic). The list shifts abruptly to the City of Manhattan Beach, which was 90% White, and then wobbles back and forth between mostly Hispanic and mostly White spaces. Los Angeles City District 5 (western San Fernando Valley) is pretty far down the list (ranking 51^st among the "least diverse", but at 80% White was highly imbalanced compared with the County mean of 41% White. The most striking pattern in these profiles, however, is the way that Blacks remained by 1990 concentrated only a few municipal spaces, most of them unincorporated Los Angeles County. The percent Black falls off rapidly among these spaces, as well, reaching just 10% by the 33^rd "most Black" municipal space among all 216 spaces. By comparison, the 33^rd "most White" space was 80% White; the 33^rd "most (Asian) space was 17% (Asian), and the 33^rd "most Hispanic" space was 64% Hispanic. The largest of these "most Black" municipal spaces were Los Angeles City’s 8^th City Council District (Represented by Nate Holden at this writing), City of Compton, and City of Inglewood. Compton ranked last among incorporated cities in median house value in 1990. LA City’s District 8 was fifth from the bottom. But it is important to note that the second "most Black" space in 1990 was an enclave of relatively wealthy Blacks (median house value 1.25 times the County mean), in unincorporated LA County territory adjacent to Culver City.

In Charts 4.1 through 4.4, we return to the Isolation Index, this time by applying it to municipal spaces. Municipal spaces are usually much larger than census tracts of course, so they should provide higher probabilities of "interaction" within their borders, between members of the four principal race-ethnic groups of Whites, Blacks, (Asians,) and Hispanics. Indeed, this is true in general: we see in Charts 4.1 through 4.4 higher probabilities of interaction for Blacks with Whites, Hispanics with Whites, and Asians with Whites. More significant, however, is the way this gap closes and the Index continues to fall for all three groups across the six decades. This is remarkable, because most municipalities contain tens, or hundreds of thousands of people, and all the Isolation Index measures is the probability of a random meeting of persons across the whole space. Municipalities, we might infer from this comparison of tract- and municipal-level isolation, have been containing and diverting the growth of "diversity" so that in specific locations, the race-ethnic groups remain highly unlikely to encounter one another as neighbors.

Map Series 4b contrasts the most "White’ and most "Nonwhite" census tracts from 1940 to 1990. The heavy blue lines are the municipal boundaries, and visual inspection clearly indicates a strong correlation or "container" effect. Map Series 4c maps data that have been aggregated at the municipal level, and shows simply the growth of municipal spaces with "Nonwhite" majorities. These majority "Nonwhite" municipalities have been remarkably contiguous, and by 1990, distinctly removed from the cool clean air of the Pacific Ocean. In the 1940s, 50s, and 60s, Alameda Street was known as the "Cotton Curtain," a euphemism for the hard race boundary that divided 95% Nonwhite municipal spaces from 95% White municipal spaces. In 1990, a different version of this kind of "Edge" had emerged, this time on the West Side.

Taken as whole, the evidence presented in this study strongly indicate that the more fateful story of the growth of diversity in the Los Angeles metropolis has been a concomitant containment of that diversity within municipal boundaries, along with a clear socioeconomic stratification of census tracts and municipalities along the White-Nonwhite line. Recall especially the charts in Section II (Charts 2.4, 2.6, 2.7, and 2.8) showing the correlation between "Whiteness" of spaces and the class proportions and median house values.

Having said this, we have also seen that the City of Los Angeles is an immense exception. I have purposely avoided discussing the City of Los Angeles as a whole until now because it has been important to focus on the overall picture of municipal spaces. The City of Los Angeles has a very bizarre shape, which is a product of its aggressive annexation policies earlier in the 20^th century. It would have been absurd and useless to aggregate all of its data across its entire outline, because this would have confounded social spaces in the San Fernando Valley, for instance, with those along the "Shoestring Strip" 30 kilometers to the south, and it would have conflated the data from the poorest neighborhood of Pico Union with the richest in Pacific Palisades. Instead, I treated each of the 15 City Council Districts as separate "municipal spaces." By 1990, each of these ranked just behind the City of Long Beach in population rank, and just ahead of the Cities of Glendale, Torrance, and Pomona. The discussions above made it clear, however, that these 15 City Council Districts are among each end of the highly stratified spectrums, from richest to poorest, and from "Most white" to the "least White."

This, simply put, strongly suggests a political solution to the apparent onward march of race-ethnic isolation. It raises the question of "political community," which can be usefully explored through electoral data in the next section.

Using data on municipal growth rates in conjunction with these findings on race-ethnic and socioeconomic stratification, an historical hypothesis can be proposed, which runs as follows: During the period of greatest socioeconomic opportunity for residents of Los Angeles County, which was from 1940 to the mid-1960s, population grew fastest in the periphery of the metropolis and slowest at the center. Spaces from the West Side to San Fernando Valley to Lakewood, both incorporated and unincorporated, were acting as "growth machines" attracting jobs, defense industries, and residents. But it was exactly those years in which Jim Crow also ruled the roost in Los Angeles. Officially-approved, if not statutory, segregation kept Nonwhites in strictly segregated areas near the slow-growth, low-opportunity core. Then, beginning in 1965, the dynamic changed fundamentally. Inaugurated by the Watts Uprising of 1965 and the Immigration Reform Act of that year, not to mention the Civil Rights Act, the core began to grow faster than the periphery. Peripheral spaces that were the areas of greatest opportunity now consolidated their gains. High property values were most probably enhanced and boosted by "White flight," as the wealthier Whites bought themselves distance from what they perceived as the riotous and increasingly "foreign" core of Los Angeles. Thus, just as the race-ethnic segregation barriers fell, new socioeconomic ones rose, to insulate Whites from Nonwhites behind a wall of wealth. Taken as a huge collective project to build the Los Angeles metropolis, this process can be characterized, following an important analysis by George Lipsitz, as an historical-geographic investment in whiteness (Lipsitz, 1998).

The phenomenon of different groups of people living separately from one another is not intrinsically bad or unjust. We must assume, however, that when spatial separation is systematically linked to dramatic differences in socioeconomic status, and when these differences widen over time, that something is quite wrong with the system. Even so, it could still be the case that these contrasts, which appear dramatic to the social scientist, may go unnoticed by the residents of the metropolis themselves. Diversity and/or racial-ethnic separation may or may not have any effect on the people we are studying. The objective measurable facts of "diversity" and "segregation" are one thing, but the feelings of "mixture" or "conflict" are another. How do we measure the mass psychological component of segregation?

The registered voters of Los Angeles County have taken several opportunities over the last several decades to make clear how they felt about the spatial separation we have observed after the fact through the data analyzed in this study. Even without conducting individual interviews or consulting historical polling data about racial attitudes, we have rather direct access to the way millions of Angelenos felt about the spatial patterns of group experience. This section presents a multivariate analysis of three race-charged issue votes: 1) The repeal, by direct referendum, of the Rumford Fair Housing Law in 1964; 2) The passage of the property-tax limiting Proposition 13 in 1978; and 3) the passage of Proposition 187 in 1994, which sought to deny health and education services to undocumented immigrants.

Slightly ahead of the national drive for Civil Rights laws, the California legislature in 1963 enacted the Rumford Act, named for one of the leading African American Assembly members, W. Byron Rumford (D- Berkeley). The Rumford Act "declared racial discrimination in housing to be against public policy and forbade owners of residential property including more than four units, or owners of any publicly assisted residential property, to engage in racial discrimination in its rental or sale." Such a law ran counter to decades of real estate practices and the common sense of White property owners, who had been led to believe that the racial integration of neighborhoods guaranteed an immediate drop in property values. The California Real Estate Association, which in the 1930s had helped the Roosevelt Administration construct the Home Owner’s Loan Corporation’s "security maps" that redlined racially-mixed neighborhoods, led the attack on the Rumford law. Denouncing the new policy as "forced housing," opponents raised the necessary signatures for a referendum repeal of the law. The measure passed overwhelmingly in November of 1964, ironically in the same year the U.S. Congress passed its sweeping Civil Rights law.

Table 5.1 presents a logistic regression analysis of the vote, using each census tract as an observation point. The large number of census tracts (1,656) makes it possible to include many "predictor" variables in the model and to control for the possible overlapping effects of different variables. In asking what factors "explain" the vote, we also want to know that a racial effect is not confused with a class or distance effect, and so on. Logistic regression, chosen here because it produces easily interpreted results, is based on dichotomous variables (values of either 0 or 1). Either a tract had a certain characteristic or it did not. The "outcome variable" was defined as census tracts returning a majority vote for the repeal of the Rumford law. The "predictor variables" were defined according to the County mean for each value. Thus each tract was identified as having a White, or Black, or Hispanic population that was greater than (value of 1) the County mean, or less than the mean (value of 0). The socioeconomic status of each tract was identified by median house values greater than the mean, percent blue collar greater than the mean, and percent with college education greater than the mean. In addition to these variables defining the social condition of each tract, I constructed a unique set of "distance" variables, intended to test the hypothesis that the geometric spatial relations between neighborhoods of Los Angeles County had a significant effect on voters’ behavior, separate from the effect of their race-ethnicity or class status. These variables identify each tract in terms of its distance from the tract at the center of the Black population. These variables are dichotomous interval variables: 10-15 kilometers, 15-20 km, 20-30 km, and greater than 30 km.

The results in Table 5.1 are very striking. First, as we would expect, tracts with White populations greater than the mean were nearly six times more likely than all other tracts to return a majority vote for the repeal of Rumford, and tracts with Black populations greater than the mean were 1/10^th as likely (ten times less likely) to vote for the Repeal. Surprisingly, however, tracts with Hispanic populations greater than the mean were nearly twice as likely to return a majority for the repeal of the Rumford Fair Housing law. This unexpected result demands further investigation, but it suggests the early opening of a rift between Hispanic and Black voters that we shall see again shortly in the vote for Proposition 187. Considering the socioeconomic measures, only one was significant (variables marked "NS" for not significant had no predictive value at all on the outcome variable). The significant one was the condition of a tract having median house values above the mean for the County. Those with high house values were only half as likely to vote for the repeal of Rumford. Inferring why this was so requires some speculation, but the following possibilities are likely: a) the Rumford law was most threatening to middle- and lower-middle class Whites who had more to lose by the anticipated drop in property values; b) Wealthier Whites lived in areas that seemed less likely to experience integration; c) racial liberalism was more common among the wealthier neighborhoods. The distance variables are very interesting indeed. Each had a significant effect on the likelihood of tracts to return majorities for the repeal of Rumford in 1964, and this effect increased as the distance from the center of the Black population increased, falling again slightly at the "greater than 30 kilometers" interval. At 10 to 15 kilometers, the tracts were 1.6 times more likely to return a majority for repeal. At 15 to 20 km, tracts were almost seven times more likely to return a majority for repeal, and at the 20-30 km interval, tracts were an astounding fifteen times more likely to return a majority vote for repeal. The greater than 30 km interval fell off slightly, but still produced a huge effect: nearly twelve times more likely to return a majority for repeal.

It is important to remember that the obvious and enormous effect of geometric spatial distance has here been statistically separated from the effect of race alone or SES alone. In other words, the logistic regression allows us to say that tracts that were 20-30 km from the center of the Black population were fifteen times more likely to vote for repeal than all other tracts, even controlling for the effect of race-ethnicity and SES. These results give us our first direct evidence of the raw impact of the spatial relations between groups. The implications are very important. The Rumford law was about the policy of "open housing," but the voting to repeal that law became more determined as distance from Blacks increased. Of course, those distant tracts would be the least likely to experience integration. Hence, we seem to have a window here on a feeling of animosity that grew as the probability of contact with the other group decreased. It was the perception of a threat imposed by the group itself, and not the law, that seems to have been at work.

The story behind, and the lasting impact of, the historic passage of "Prop 13" is well known. During the 1970s the cost of living in California had increased 79%, property values had increased an average of 2% per month, and federal, state, and local taxes had increased at rates greater than the general inflation rate. While the decade of the 1970s is known for "stagflation," homeowners were only victims of tax increases in proportion to being beneficiaries of a remarkable boom in real property values. We saw in Chart 2.5 that the rate of median house value increases went up sharply during the 1970-1990 period, in constant 1983 dollars. The median value of all homes in Los Angeles County in 1970 was $60,000. In 1980 it was $100,000, and in 1990 it was close to $160,000. Although one would expect homeowners to be very happy about such appreciation, local tax increases combined with periodic re-assessments led to sudden jumps in their property tax bills of as much as 40-90%. These property taxes came to stand for a much larger discontent with liberal social spending in the 1970s, and became the focus of an attack led by Howard Jarvis and Paul Gann. Constitutional amendment initiative Proposition 13 promised to restrict property taxes to 1% of assessed value (an immediate average reduction of 57% in property taxes) It also promised to roll back property value assessments to their 1975 values, and to limit further assessments to no more than 2% per year. Proposition 13, which explicitly forbade local governments from increasing property taxes, was a radical popular measure by any standard, and its success became a historic milestone in U.S. national political trends, prefiguring in many ways the Presidential triumph of Los Angeles County’s own Ronald Reagan two years later.

Proposition 13 might abstractly be understood as purely an economic measure, but in reality it was deeply implicated in the politics of race. Conservative politicians, including the popular California Republican leaders Richard Nixon and Governor Ronald Reagan, explained the rising cost of government programs as a direct function of the Great Society welfare programs that were targeted to inner-city Blacks. By the mid-1970s this link had become commonplace. Further, 1978 was also the year that court-ordered school busing finally began implementation in the Los Angeles Unified School District—following by eight years Judge Alfred Gitelson’s 1970 ruling that the LAUSD was unconstitutionally racially segregated.

Given the focus of Proposition 13 on residential property values, and its clear popular association with race relations, we have an excellent opportunity to test the relative effect of property values and race on the vote. Table 5.2 presents the logistic regression analysis of the 1656 census tracts in Los Angeles County, using the exact same variables as in Table 5.1. Most remarkably, tracts with median house values above the mean had no effect on the likelihood that tracts would return a majority vote for Prop. 13. Instead, tracts with White populations that were greater than the county mean were nearly three times more likely to return a majority vote for Proposition 13. (We might observe here that the rates of voter registration among Hispanics in the 1970s were very low. Many were not even citizens, and those who were, and who voted, may have been markedly more conservative than other Hispanics. Further research is needed on this question). Again, tracts with Black populations higher than the mean were 1/10^th as likely (ten times less likely) to return a majority vote for Prop. 13. In 1978, the rift between Hispanics and Blacks seems to have grown: tracts with Hispanic populations greater than the mean were nearly four times more likely to return a majority vote for Prop. 13. Interestingly, tracts with Blue Collar populations greater than the mean were less than half as likely to return a majority vote. This suggests a clear "class" perception in the impact of this vote on needed (as perceived by lower-income voters) social services, regardless of race.

Even more striking than in 1964, the effect of raw spatial distance from the center of the Black population was profound. Tracts at a 10-15 km distance from the Black center were nearly four times more likely to return a majority vote for prop 13, and this figure increased with distance until the farthest interval, greater than 30 km, which was twenty-six times more likely to return a majority vote. We have a very strong indication here that Proposition 13 was powered by feelings of animosity toward the Black residents of the central core of the metropolis, and that these feelings, even controlling for the direct effect of race and class, increased as a function of spatial distance.

If the Repeal of the Rumford law and the passage of Proposition 13 took place at the height of tensions during the Civil Rights era, the passage of Proposition 187 in 1994 took place in what could be called the post-Civil Rights era. The Reagan Revolution had created new standards and popular assumptions about the value of "Liberalism," a label now clearly on the defensive. Further, the effects of Immigration Reform Act of 1965 had become fully apparent, as the Hispanic and Asian, other populations vastly outnumbered Blacks as the majority of the Nonwhite population. Indeed, in Los Angeles County, White Non Hispanics were now a 41% minority themselves. Leading a new coalition of "angry" white taxpayers, an obscure accountant named Ronald Prince authored a sweeping statutory initiative aimed at the presumed millions of undocumented "illegal" immigrants that were using expensive social services in California. The measure on the ballot read as follows: "illegal aliens are ineligible for public social services, public health care services (unless emergency under federal law), and attendance at public schools. Requires state/local agencies to report suspected illegal aliens." Because of its threatened impact on children and the infirm, and its design to convert school teachers and other public servants into adjunct immigration police, the measure was widely denounced by liberals and immigrant leaders as vindictive and racist. Indeed, the campaign for Proposition 187 was openly based on inter-group hostility in ways that surpassed the earlier propositions 14 (repeal of Rumford) and 13 (property tax limitations).

Aimed as it was primarily at Hispanic immigrants, however, analysis of this measure should give us a strong indication of whether the dynamics of race and space had changed fundamentally since what could be called the "classic" period of Black/White race relations in the 1960s and1970s. Table 5.3 strongly indicates much more continuity than contrast with the patterns we saw in the 1964 and 1978 votes. Again, tracts that had White populations greater than the mean were three times more likely to return majorities for Proposition 187. The Black-Hispanic rift continued, but this time it appears Black voters expressed frustration with growing immigrant competition for jobs and neighborhoods. Tracts with Black populations greater than the mean were two and a half times more likely to return majorities for Proposition 187, while tracts with Hispanic populations greater than the mean, predictably, were one-third as likely (three times less likely) to return a majority. A socioeconomic effect was clearly visible as well. While percent Blue Collar had no effect, it appears that wealthier areas were about half as likely (two times less likely) to return a majority vote fore Proposition 187, judging from the significant effect of both the percent college educated and the median house value variables. And again, raw physical distance had an independent effect on this vote outcome. Distances in this model are measured from the center of the Hispanic population, and as these increased, so did the probability that a tract would return a majority for Proposition 187. Tracts that were greater than 30 km from the center of Hispanic population were more than twelve times more likely to return a majority for Proposition 187.

It is useful to compare these results with the vote for Governor Pete Wilson, the Nixon protégé and former San Diego mayor who strongly backed Proposition 187. Table 5.4 shows that Wilson was even more favored by tracts with white majorities greater than the mean than was Proposition 187. And Blacks seem clearly to have split their votes: between an apparent support for Proposition 187 and a clear opposition to Pete Wilson (tracts with Black populations above the mean were ten times less likely to return majorities for Wilson). Indeed, Hispanics and Blacks seem to have found common cause in their opposition to the conservative Wilson. But Wilson had a populist edge as well. Tracts with college educated proportions above the mean were 1/5^th as likely (five times less likely) to return a majority for Pete Wilson. Distance, however, showed clear consistency with the patterns of voting for Proposition 187. Apparently, voting for Wilson was associated with the suburbs and presumably against the center city.

The voting patterns on the three race-charged issues of 1964, 1978, and 1994 present a sadly consistent pattern in many respects. In all cases the "Whiteness" of census tracts was a very strong predictor of voting. The distance of neighborhoods from the centers of the Black and Hispanic population was an even stronger predictor, however, and this finding gives us a very concrete indication that the spatial configuration of groups in historically-evolved territories has a very clear behavioral and presumably psychological correlate.

Tables 5.5 through 5.7 confirm this hypothesis by performing logistic regression analyses again on the same votes of 1964, 1978, and 1994, this time with all data re-aggregated to fit the municipal boundaries. Data configured this way allow us to test the effect of living in specific types of municipal spaces. Given the much smaller number of observations (172 in 1964 and 1978, 190 in 1994), however, these models are not as reliable as those using tracts, which numbered 1,656. Still, we have a very suggestive indication that voting behavior was structured by the kinds of local polities one lived in. In 1964 (Table 5.5), municipal spaces that were incorporated at the time of the vote were more than ten times more likely than unincorporated Los Angeles County territories to return majorities for repeal of Rumford. Race was clearly a factor. Municipal spaces with Black populations greater than the County mean were less than 1/10th as likely to return majorities for Proposition 14. The fact that the percent White variable was not significant may be an artifact of the small sample size, but it also could mean that where Whites lived (incorporated cities versus unincorporated County territory) was far more important than their simple race-ethnic identity.

The powerful effect of living in an incorporated city was apparent again in the patterns of voting for Proposition 13 in 1978 (Table 5.6). This time, both the percent White Non Hispanic above the mean and the condition of living in an incorporated city gave a municipal space a very strong likelihood of returning a majority for Proposition 13, while percent Blue Collar above the mean made spaces 1/5^th as likely to return a majority. The implication is that both race