Approximately two years ago, the entitlement communities embarked on a joint effort to fund a study to find the answers to our questions. The entitlement cities and County staff members cooperatively prepared a Request for Proposal (RFP) for the study and actively sought bids. They also established a fair housing subcommittee of the entitlement communities' staff to make the final selection from the candidates submitting proposals. This subcommittee ultimately selected the Empirical Research Group of the UCLA School of Law, headed by Dr. Richard Sander, to conduct the study. The subcommittee has been responsible for the ongoing review of the study as it progressed, providing follow-up and communication with Dr. Sander and his research team as needed. It was agreed that the maximum funding for the study would be $50,000. The participating jurisdictions were assessed an amount proportional to their population to fund the study.

The general consensus of the fair housing subcommittee members and, we believe, the other entitlement City Staff members, is that we have a very good product and a foundation to build upon for future action. The report paints a fairly positive picture of Santa Clara County and its fair housing efforts but it is not without some constructive criticism of the jurisdictions in the county and the fair housing agencies we fund.

The purpose of this report is to assess the current extent and nature of fair housing problems in Santa Clara County, and to propose ways to improve current programs that relate to fair housing. Most of the federal Community Development Block Grant (CDBG) entitlement jurisdictions in Santa Clara County completed fair housing "assessments" in the early- and mid-1990s, but this report represents the first attempt to look comprehensively at conditions throughout the County and to use systematic data collection to evaluate several different dimensions of fair housing -- demographic change, lending patterns, residential attitudes, land use practices, and the performance of fair housing agencies. This study results from a collaboration of all the CDBG jurisdictions in Santa Clara County, who jointly funded and supervised this study in the hope that the resulting work would be broader in scope and deeper in research findings. The jurisdictions paid $50,000 to the Empirical Research Group at UCLA (ERG) to conduct the study. ERG and its principals have conducted a number of earlier fair housing studies in California, and this report builds upon and adapts techniques used in those earlier reports.

Chapters One through Six, and the two appendices, cover the substantive findings of this project. This introduction covers a few preliminaries.

Geography. The area covered by this study is Santa Clara County, and any simple mention of "the County" in this report refers to the geographic Santa Clara County. Most of the analysis and discussion in this report addresses the County as a whole. However, in a few sections of the report we look at patterns in specific cities within the County. Appendix I presents detailed demographic data for the fifteen census "places" in Santa Clara County (including its incorporated cities), and individual cities in the County have also been given feedback on the fair housing issues posed by their zoning ordinances.

Racial classifications and terminology. Racial classifications are intrinsically sensitive, political, and somewhat arbitrary. Our goal in this study is to be clear and consistent in our use of terms, and to provide a sufficient range of data and sources as to enable readers to think critically about the issues we cover. In our specific use of racial classifications, we follow closely the example used by the U.S. Census Bureau and most federal agencies. However, we simplify Census classifications in a few ways to make the presentation of data manageable. For the most part, we refer to four major racial groups: Blacks, Hispanics, Asians, and Whites. "Black" is, in this report, synonymous with African-American -- it includes all black persons living in the United States. "Hispanic" refers to any person who indicates in official forms that they are "Hispanic" or "Latino." Predominantly, this population is of Mexican, Central American, Caribbean, or South American descent. "Asian" includes the major American ethnic groups with origins in southern and eastern Asia -- Chinese, Japanese, Asian Indian, Korean, etc. -- as well as persons from nations and territories in the Pacific Ocean, including the Philippines, Fiji, American Samoa, and so on. "White" is, in this scheme, almost a residual category, including all persons who identify themselves as white in the census but do not also identify themselves as Hispanic. Mixed-race persons are allocated pro rata to all of the races with which they identify; this method is elaborated in Chapter One.

Richard Sander

UCLA

October 2002

sander@law.ucla.edu

Demographic Setting and Racial Trends in Santa Clara County

Four basic demographic facts are key to understanding fair housing patterns in Santa Clara County and its constituent cities. The first and most obvious of these is that economic change and growth, powered by the rapid development of high-tech industry, has fundamentally transformed the local economy over the past generation. Despite the current "dot-com" doldrums, the days when Santa Clara County was a mix of agricultural and bedroom communities are gone forever. By 1999, the County was home to over 940,000 jobs – over 200,000 directly in the manufacture or development of computers and computer software..

Remarkably, however, the population of Santa Clara County grew only 12.4% during the decade from 1990 to 2000, from just under 1.5 million people to just over 1.68 million people. This is the second key fact. Even though its "Silicon Valley" economy was at the heart of the greatest economic boom in United States history, and despite the fact that the County’s population had quintupled from 1950 through 1990, the County grew during the 1990s at a slightly slower rate than the United States as a whole (13.2%). In 1990 Santa Clara County was the 15^th most populous county in the nation; in 2000, it had risen one place, to 14^th. In many parts of the County, as shown in Map 1.1, population grew less than 5% during the 1990s (and in some areas even declined).

Despite the County’s relatively slow population growth in the 1990s, however, it continued to experience immigration and racial change on the same pace as earlier decades…and in some respects, at a more dramatic pace. Migration from India – fueled by strong links between Silicon Valley and the Indian computer industry – leapt during the decade; the Asian Indian population grew from nine thousand in 1990 to sixty-three thousand in 2000, a seven-fold increase that by itself accounted for a quarter of the population growth in the County. The total Asian population of the County rose from 271,000 to 436,000 from 1990 to 2000 – an absolute increase nearly as large as the County’s overall population growth. The Hispanic population also grew substantially, though less dramatically, while the White and Black populations modestly declined. The third key fact, then, is that gradual population growth occurred alongside quite rapid population change in the racial/ethnic make-up of the County.

In many of the nation’s most racially diverse cities, metropolitan diversity is accompanied by neighborhood-level segregation. In Los Angeles, for example, Blacks, Hispanics, and Cambodians are concentrated in separate regions of the County (though few areas are all-White, and some Asian groups are highly integrated). Santa Clara County, in contrast, has had fairly high levels of neighborhood integration since at least 1980, and these levels remained high as of the 2000 Census. This is the fourth key fact. Although economic segregation in the County is marked (affluence tends to steadily rise as one moves from northeast to southwest parts of the County) and though specific racial and ethnic groups have areas of concentration within the region, Santa Clara County remains one of the best examples of diversity mixed with integration in the United States (and probably the world).

The interaction of these four conditions goes a long way to explain not only fair housing conditions in Santa Clara County, but housing conditions more generally and, indeed, the region’s socioeconomic structure and much of its culture. For example:

--The interaction of rapid economic growth and modest population growth implies a strong emphasis in local government policy to limit new development. In Santa Clara County, even though only a fraction of the County is developed, the combination of federal and state-owned forests with extensive mountains limits the physical space available for expansion. On top of this, however, most jurisdictions in the County strictly limit new development, and place particularly severe restrictions on multifamily development. The expansion of business creates further competition for scarce land, and the price of land itself is, of course, extremely high. While there are good policy reasons behind many of these anti-growth policies, they naturally create conditions where housing costs steadily increase, low-income workers are gradually squeezed out of the housing market, and the affluence of the County steadily increases even while the quality of life stagnates or even declines for a significant portion of the population.

--The dramatic increase in the Asian population during the 1990s (particularly the increase in the Asian Indian population) testifies not only to the strong pressures on the labor market induced by economic growth, but also to the malleability of that labor market and the degree to which diversity is accommodated in County. Santa Clara County’s increasing diversity in the 1970s and 1980s, and its high levels of racial integration make it an unusually hospitable place for immigrants to arrive.

--Given that the Asian population rose sharply and that overall population growth was limited by public policy and natural constraints, it was inevitable that the population of other groups in Santa Clara County would either decline outright or grow much more slowly than would be expected given general patterns in California. This can best be seen by inspecting Table 1.2a , which compares the changing racial/ethnic composition of Santa Clara County with California generally.

1. Population Change in Santa Clara County

Table 1.1 shows the increase in Santa Clara County’s population over the past century. From 1900-1950, the County’s growth paralleled that of California as a whole (both increased approximately fourfold during that period). Unusual growth reached the County in the 1950s, as light industry began to enter San Jose and as other parts of the County began to be suburbanized by Bay Area commuters. Growth accelerated in the 1960s and 1970s, with much of the developable farmland in the County converted to more intensive uses; by 1980, at the threshold of the computer "boom," Santa Clara County had a diverse industrial and technological base.

Table 1.1

Growth in Population in Santa Clara County, 1900 - 2000

Decade	Population
1900	60,216
1910	83,539
1920	100,676
1930	145,118
1940	174,949
1950	290,547
1960	642,315
1970	1,064,714
1980	1,295,071
1990	1,497,577
2000	1,682,585

Ironically, population growth in the County slowed at the same time that the technology sector took off; thus, Santa Clara County's housing history over the past twenty years has been a duel between land use controls and the economic pressures of development. In both absolute and percentage terms, population growth was slower in the 1980s than in the 1970s, and slower in the 1990s than in the 1980s. As Map 1.2a shows, growth rates varied significantly across different portions of the County, tending to increase most in unincorporated portions of the County and in low-income areas. Zoning and growth restrictions appear to be the predominant factors shaping general population growth. These factors are further discussed in Chapter Five.

Tables 1.2a and 1.2b, on the following page, show the general pattern of racial change in Santa Clara County, and compares this pattern with California as a whole. The growth of the Asian population in the County is very striking. In absolute terms, the 175,000 new Asian residents in the County accounted for more than one-fifth of the statewide increase in Asians during the decade -- even though Santa Clara County accounts for only one-twentieth of the state's population. The percentage increase in Asians in Santa Clara County similarly far outdistanced state rates of growth.

Table 1.2a

Santa Clara County Racial Composition and Population Change 1990-2000


Race	1990		2000		% Change
Asian	261,466	17.5%	451,630	26.8%	72.7%
Black	56,211	3.8%	50,480	3.0%	-10.2%
Hispanic	300,966	20.1%	396,637	23.6%	31.8%
White	869,874	58.1%	763,531	45.4%	-12.2%
Other	9,060	.6%	20,307	1.2%	124%
Total	1,497,577	100%	1,682,585	100%	12.4%

Table 1.2 b

California State Racial Composition and Population Change, 1990-2000


Race	1990		2000		% Change
Asian	2,845,659	9.6%	4,037,687	11.9%	41.9%
Black	2,208,801	7.4%	2,352,081	6.9%	6.5%
Hispanic	7,436,277	25.0%	10,822,722	32.0%	45.5%
White	17,029,126	57.2%	16,167,434	47.7%	-5.1%
Other	240,158	.8%	491,723	1.5%	104.7%
Total	29,760,021	100%	33,871,648	100%	13.8%

Population changes for Hispanics, Blacks, and Whites in Santa Clara County from 1990 to 2000 are a direct result of the interaction of three factors: first, the large exogenous increase in the Asian presence; second, the restricted population growth in the County as a whole; and third, statewide patterns of demographic change. Overall, Santa Clara County grew from 1990-2000 at about the same rate as California as a whole (12.4% for the County, 13.8% for the State). However, given the much larger increase in the Asian presence, other groups in the County necessarily had a corresponding amount subtracted from their State growth rate -- 7 points lower for Whites, 14 points lower for Hispanics, and 17 points lower for Blacks. The disparity among these three growth rates corresponds directly to their relative affluence -- that is, Blacks and Hispanics have the lower median incomes and were probably thus most heavily "displaced" (in terms of expected growth rates) from the County, with Whites doing somewhat better.

2. Racial Integration of Blacks in Santa Clara County
   

With the passage of a national fair housing law in 1968, many of the systematic barriers to housing access for racial minorities fell. Real estate agents abandoned formal policies of exclusion (though much informal "steering" of minorities away from White neighborhoods continued), and banks gradually changed policies that had made it difficult for minorities to get credit and particularly difficult for them to secure mortgage loans for homes in White neighborhoods. Measured discrimination rates across the nation fell dramatically in the 1970s. Although discrimination remained a widespread, serious problem, by 1980 it was no longer ubiquitous. Smaller, more gradual declines in discrimination occurred in the 1980s and 1990s, so far as we can tell from the available systematic research. Most of this research uses "fair housing testing," a technique in which pairs of people who are similar in most ways except their race are sent to inquire about renting and homeownership opportunities, and are instructed to record their experiences. In the more recent national tests, about 4% of Black testers are falsely told that housing is unavailable, and 25-30% receive some kind of treatment that tends to discourage their pursuit of the housing.

 

 

 

 

Housing segregation in the United States is commonly measured with the "index of dissimilarity." The index is used to compare how isolated two groups of urban/suburban residents are from one another. To calculate it, we must choose a region to analyze (e.g., a metropolitan area or county), a unit of location (e.g., a "census tract," "zip code," or "block"), and two populations to compare to one another (e.g., White and Black). Suppose that we are interested in the index of dissimilarity for Whites and Blacks in Santa Clara County, by block; the index then uses census data to calculate the proportion of Blacks who would have to move to a new block to achieve the same geographic distribution as Whites across the County. A dissimilarity index of 1.00 means that 100% of one group would have to move to achieve integration with a second group (that is, the starting point is complete separation or apartheid). An index of 0 indicates complete integration.

Before the passage of fair housing laws, the Black/White dissimilarity index level in most American cities was between .85 and .90 -- extremely high levels signifying the almost complete isolation of Blacks in ghettos. In the years after 1968, hundreds of thousands of moderate- and middle-income Black families moved into White neighborhoods; but change occurred so rapidly, and was so often accompanied by White "flight," that "integration" was often a short-lived prelude to resegregation. In other words, many neighborhoods passed through rapid transition from all-White to all-Black. Genuine Black/White integration was rare in most major American cities. The index of dissimilarity, consequently, did not fall very much. The Black/White dissimilarity index for the twenty-five largest American metropolitan areas was .88 in 1960, .87 in 1970, and .81 in 1980. By the year 2000, the Black/White index had drifted down to an average of .72 in these major areas.

The story was quite different in Santa Clara County. In 1960, the City of San Jose had the lowest measured level of Black/White residential dissimilarity -- .604 -- among 208 cities studied by the leading work of the time. In the 1970s, Santa Clara County was one of a dozen metropolitan areas (among the largest one hundred in the nation) in which segregation fell sharply. The Black/White segregation index in the County fell to .48 in 1980 and .42 in 1990, when the unit of analysis is census tracts (segregation measures are higher, as we shall see, when the unit of analysis is smaller, such as blocks).

The low level of Black segregation in Santa Clara County, and in some other metropolitan areas that followed the same pattern, is tied to specific demographic conditions in the region. Other things being equal, segregation tends to be lower when the Black population is relatively small, since the low numbers make it less likely that the migration of some Blacks to a White neighborhood will lead to resegregation. Segregation is also lower in rapidly-growing metropolitan areas where existing Black neighborhoods are not bordered by tightly-knit groups of people that are resistant to change in the ethnic composition of their community. Santa Clara County, along with a number of other western urban areas such as San Diego, Seattle, and Phoenix, has all these favorable conditions, so in this sense the decline in segregation is neither unique nor mysterious. It is, nonetheless, a remarkable achievement.

Recent conditions, 1990-2000. During the 1990s, the Black/White index of dissimilarity increased for the first time in memory. The measures reported in Table 1.3 are based on blocks, not census tracts, and the smaller unit of analysis makes the numbers go up. Overall, the increases are relatively modest; the dissimilarity index for Blacks and Whites rose three points, and the dissimilarity for Blacks compared with all other races combined rose one and one-half points. Black segregation in Santa Clara County is still below general levels in California, but the decline of recent decades appears to have stopped.

Note that the dissimilarity measures in Table 1.3 are measured by block, so that, for example, if blacks and whites live in the same neighborhood but on different blocks, they still show up as segregated from one another in these measures. An index of dissimilarity calculated on the basis of census tracts (areas of about four thousand people that approximate typical notions of "neighborhood"), then the black/white index of dissimilarity for Santa Clara County is quite a bit lower: .44 in 2000. Statewide, using tracts rather than blocks lowers black/white dissimilarity much less – only to .65 in 2000. Why this difference? Probably, the difference reflects a tendency in Santa Clara County for blacks and whites to often live in the same neighborhoods but in different types of housing, or in different subclusters. For example, within similar income ranges blacks in the County are less likely to live in single-family homes than whites. Thus, in a residential neighborhood, blacks may be over-represented on a block with duplexes or small apartment buildings. The low tract-segregation index, however, suggests that blacks are very well-integrated in schools – which of course draw on neighborhoods, not blocks.

Table 1.3

Santa Clara County and State of California
Dissimilarity by Block Between Selected Race/Ethnic Comparisons

Race/Ethnic Group	Santa Clara County		California
	1990	2000	1990	2000
Black with White	.562	.591	.709	.701
Asian with White	.465	.496	.556	.587
Hispanic with White	.541	.590	.600	.619
White with all others	.458	.470	.546	.544
Black with all others	.466	.480	.647	.614
Asian with all others	.414	.440	.522	.551
Hispanic with all others	.483	.517	.542	.550

This chart contains dissimilarity indices calculated for Santa Clara and California by 2000 block. See page 1-5 for an explanation of the index of dissimilarity.

The Black population of Santa Clara County is somewhat more concentrated in the eastern half of the County than are Whites or Asians, but about as concentrated there as the Hispanic population. Two-thirds of Black households in the County are located in the City of San Jose, compared with about 50% of all households in the County. There are very few neighborhoods anywhere in the County that are more than 25% Black, and the "median" Black resident lives in a neighborhood that is less than 10% Black. But Blacks, though highly integrated, are clearly underrepresented in such predominantly single-family communities as Saratoga, Cupertino, and Los Altos. In addition to possible fair housing problems, there are two objective demographic reasons for this disparity. First, Blacks own their own homes at rates significantly below the rates one would expect given Black income (only 38% of Blacks households are owners, compared with 63% of Whites). Second, workers at the high-tech jobs in the central and western County are often imported from a national and worldwide market of computer specialists; these imported populations are predominantly White and Asian. Blacks (and Hispanics) thus are disproportionately under-represented in these jobs, and instead work predominantly in more traditional centers of the economy (e.g. light manufacturing, low-tech services), which are concentrated heavily in and around San Jose.


3. Hispanic Segregation
   

The Hispanic population of Santa Clara County is overwhelmingly Mexican in origin, as Table 1.4 demonstrates. A significant portion of the Hispanic presence has been in the County for decades, so the Hispanic population is a mixture of recent immigrants and lifelong residents. Adding to this diversity is an increase in 1990s migration from Central America which appears as "Other." As noted earlier for the Black population, Hispanics tend to be more heavily concentrated in Santa Clara County in traditional services and industries, and underrepresented in the new computer economy. The moderate levels of segregation Hispanics experience in Santa Clara County is fairly typical of levels Hispanics face throughout the nation, and appears to follow very much the classic pattern observed elsewhere -- that is, Hispanic segregation appears to result from three principal factors: lower incomes, ethnic clustering, and discrimination. Lower incomes (Hispanic median household income is about 30% lower than the median for whites and Asians) are important because they lead to Hispanic concentration in lower-income areas. (Lower wealth, regardless of ethnicity, is also important -- particularly in Santa Clara County, where it is impossible to become a homeowner in most of the northwestern part of the County without significant wealth.) Ethnic clustering is important because new Hispanic arrivals to Santa Clara County, particularly immigrants, are very likely to settle in parts of the County with a significant Hispanic presence, tapping into cultural and community networks that ease transitions. Discrimination is a significant factor, even if many landlords and realtors are not discriminatory, because the presence of some discrimination imposes search costs on Hispanics and, for many, creates important psychological barriers to seeking housing in areas dominated by other groups. The large Hispanic presence in the County generally mitigates this factor in Santa Clara County, but in the swath of western communities stretching from Los Gatos, Saratoga, Cupertino, to Los Altos and Palo Alto, the very small Hispanic presence (under 5% in most blocks) deters Hispanic buyers.


Table 1.4

Change in Hispanic Category by Detailed Ethnicity*


Ethnicity	1990	2000	% Change
Mexican	248,225	323,489	30.3
Puerto Rican	7,689	6,396	-16.8
Cuban	1,647	1,852	12.4
Other	49,552	71,664	44.6
Total	307,113	403,401	31.4

*Includes Hispanic Blacks

The measured levels of year 2000 segregation in Table 1.3 on the previous page suggest that Blacks and Hispanics have almost identical levels of segregation. This is true, when measured by the index of dissimilarity (and it's also true that levels for both groups went up a few points in the 1990s). However, the vastly larger Hispanic population makes the character of its housing segregation quite different from that facing Blacks. Since Hispanics comprise about a quarter of the county's population, the areas with higher Hispanic concentrations have very high Hispanic populations indeed. In much of eastern San Jose, [Alum Rock is a neighborhood district of San Jose] Hispanics make up more than 60% of most neighborhoods and over 80% of some. On the flip side, in the many portions of Santa Clara County where Hispanics are substantially integrated, they make up 20-40% of neighborhood populations -- thus avoiding the ethnic isolation that many Blacks perceive as accompanying their integration in Santa Clara County.

Asian Segregation

As Table 1.3 indicates, Asians are generally not as segregated as Blacks, as a group or when examined as individual nationalities. However, an inspection of the tables and maps at the end of Chapter One reveals that there are, in fact, a couple of quite distinct patterns among Asian nationalities. Koreans, Chinese, Japanese, and, to a lesser degree, Asian Indians, tend to be concentrated in the west-central County, an area centered in Cupertino and Sunnyvale. Filipinos and Vietnamese residents tend to be concentrated in the eastern portion of the County. This difference in patterns reflects a socioeconomic split as well: the first group of Asian ethnics tends to be substantially more affluent than the second group. Within these general patterns there is a fair degree of clustering within each ethnic group. While there are almost no neighborhoods in Santa Clara County where a specific Asian nationality constitutes even a majority of residents, there are identifiable neighborhoods (and presumably, identifiable ethnic institutions as well) for each of the major Asian groups. In the absence of any systematic testing data, it is hard to conclude that discrimination is not contributing to this segregation. However, research in other cities suggests that these levels of segregation are consistent with voluntary clustering.

Table 1.5a

Change in Asian Category by Detailed Ethnicity

Asian Ethnicity	1990	2000	% Change
Pacific Islanders	6,680	5,773	-13.6%
Asian Indians	20,164	66,741	231.0%
Chinese	65,027	110,632	70.1%
Filipino	61,518	76,060	23.6%
Japanese	26,516	27,257	2.8%
Korean	15,565	21,647	39.1%
Vietnamese	54,212	99,986	84.4%
Other	11,784	43,643	270.4%
Total	261,466	451,630	72.7%

Table 1.5a shows that the trend of segregation for individual Asian ethnic groups has been steadily downward over the past twenty years. Note that the trends for individual ethnic groups are uniformly downward, even though overall Asian segregation increased some during the 1990s. This is because individual Asian ethnic groups are assimilating, but many of their first stops in the assimilation process are neighborhoods with a significant Asian presence -- thus increasing overall Asian segregation. In other words, the first stage of assimilation for many Asian immigrants is a process of acculturation to the larger Asian community. Overall segregation levels for Asians are, in any event, also relatively modest. We suggest that the difference between Black and Asian levels of segregation (Black levels are higher) is due to the greater number of fair housing problems Blacks encounter from homesellers, landlords, and lenders, but this is speculative -- the differences are small enough that it is difficult to distinguish the role of income differences, preferences, and discrimination.

Table 1.5b

Population Growth and Dissimilarity Indexes with General Population

for Select Asian Nationalities


Nationality	2000 Population	2000 Dissimilarity	1990 Population	1990 Dissimilarity	1980 Population	1980 Dissimilarity
Chinese	110,632	0.37	65,924	0.37	23,929	0.39
Filipino	76,060	0.45	59,963	0.52	24,191	0.52
Japanese	27,257	0.27	27,967	0.28	21,060	0.33
Korean	21,647	0.37	15,182	0.42	6,240	0.46
Vietnamese	99,986	0.46	54,739	0.53	9,025	0.56
Pacific Islander	5,664	0.28	6,540	0.48	4,313	0.59
Asian Indian	66,741	0.39	19,675	0.41	6,775	0.55

Socioeconomic Conditions Across Racial Groups

Sociologists and economists have produced significant evidence that high levels of housing integration tend to narrow the traditional economic gaps between Blacks and Hispanics, on the one hand, and Whites and Asians on the other. This clearly appeared to be happening in Santa Clara County in the 1970s and 1980s. Income levels among Blacks in Santa Clara County, in particular, were among the highest in the nation, and at one point approached 85% of median White levels in the County. Over the past fifteen years, however, the dynamics of economic change in the County have become significantly more complex. The high rates of immigration and the great wealth creation of the computer, technology and software industries has tended to change early patterns of equalization, since new classes of "techies" and entrepreneurs have been superimposed upon the old social structure. Blacks and, to a lesser degree, Hispanics, had lower ownership rates than Whites and thus missed out to a greater degree on the incredible boom in property values in Santa Clara County. The result appears to be two competing trends: on the one hand a general rise in affluence, concentrated among college graduates in private industry (disproportionately White and Asian), while on the other hand a continuation of past trends of unusual opportunities for minority communities.

Table 1.6 provides some summary socioeconomic data for Santa Clara County, by race, for 1990 and 2000. These illustrate some of the patterns discussed in the last paragraph. Blacks in Santa Clara County not only have high incomes in absolute terms (about double the national median for blacks) but are closer to Santa Clara median for all races (the black median is 81% of the overall median) than is the case for blacks natinoally (in the U.S. as a whole, black median household income is 68% of the overall median). Nonetheless, blacks in Santa Clara County fell a bit in the 1990s in relative terms. Hispanics also fell some in relative terms, as continued immigration of mostly low-to-medium skill workers lowered the average rate of high school completion among Hispanics. Whites and Asians gained some in relative terms as very highly-educated migrants from the U.S. and overseas took many of the new, high-tech jobs.

Table 1.6

1990-2000 Summary Demographic Statistics for Santa Clara County, by Race


Variable	Overall	Asian	Black	Hispanic	White
Median household income, 1990 (adjusted to 2000 $$)	$65,430	$71,660	$53,060	$49,200	$69,450
Median household income, 2000	$74,335	$82,800	$58,920	$55,570	$80,020
% of households GT 1.01 persons per room, 1990	10.7	27.3	13.3	27.1	3.1
% of households GT 1.01 persons per room, 2000	14.3	24.6	13.8	38.9	3.2
% Persons 25 and older with high school diplomas, 1990	82.0	82.2	84.2	57.2	89.6
% Persons 25 and older with high school diplomas, 2000	83.4	84.8	88.9	55.1	93.1

Source: 1990 PUMS data from U.S. Census, and 2000 STF 3 data from U.S. Census.

The comparative data on overcrowding is also very powerful when we simultaneously look at ethnic differences and trends over time. Despite a general increase in overcrowding (from 10.7% to 14.3% of units) over the 1990s, virtually all of the change was accounted for by an increase in overcrowding in Hispanic households – which in turn was heavily influenced by an increase in the average size of Hispanic households. The overall high rate of overcrowding reflects both the large immigrant population of Santa Clara County and the very high price of housing relative to other goods in the local economy.

General Discussion of 2000 Census Data – Santa Clara County in Perspective

In May 2002, the Census released preliminary socioeconomic data from the 2000 Census for jurisdictions nationwide. Although this data is still incomplete, we have summarized in a series of tables (see Appendix I) the data, and changes since 1990, for the major jurisdictions of Santa Clara County as a whole, and for the County, State, and nation. The data provides fascinating insights into change in Santa Clara County and the ways it resembles or diverges from other regions. It should be kept in mind that the data was primarily collected in March and April 2000, and thus captures the region at the height of its economic boom. What follow are some of the most pertinent trends:

--Santa Clara County became more affluent during the 1990s. Median household income rose (in inflation-adjusted dollars) from $65,436 to $74,335, a 15% increase, with most of the rise probably occuring in the second half of the decade. Income levels in the County are 57% higher than in the State as a whole, and 77% above the national median. The poverty rate in Santa Clara County, however, remained flat in the 1990s, despite the rise in incomes and despite a national decline in poverty rates. The general impression that economic inequality in the County increased in recent years is supported by the census data.

--Although it is true that rents and home values rose rapidly in the late 1990s, the degree of fundamental change has been exaggerated. If we compare 1990 home values and rents, adjusted for inflation, with 2000 levels reported in the Census, the median value of homes in Santa Clara County as a whole rose from $390,060 in 1990 to $446,400 in 2000 -- a 14% increase -- and median monthly rents rose from $1,044 to $1,185 -- a 13% increase. In other words, housing costs rose at about the same rate that income did during the 1990s. This is reflected in the statistic (Table A.5) that the proportion of renters paying over 35% of their income in rent fell during the decade, from 31.7% to 29.6%. These statistics will astound many readers, and it is important to keep a few things in mind when using this data. First, we are comparing 1990 and 2000; since real estate values and rents fell in real terms during the early 1990s, the increases over the past few years are probably greater than the overall increase for the decade. Second, we are adjusting the data for inflation (35% overall during the 1990s) so nominal values and rents have of course increased more. Third, this data is reported by individual respondents; homeowners who have lived in their homes for a long time may underreport current values, especially in an era of rapid appreciation. Historically, however, the census numbers have proved to be an excellent indicator of housing prices.

--Housing overcrowding increased some during the 1990s. The proportion of housing units with more than one person per room rose from 10.9% in 1990 to 14.3% in 2000 -- below the statewide average, but more than twice the national rate (see Table A.5). This mostly reflects the growing immigrant population in Santa Clara County, because recent immigrants have larger families and households (average household size rose from 2.81 to 2.92 in Santa Clara during the 1990s, even while the long-term decline in the U.S. as a whole continued.

--The 2000 census data supports the view that Santa Clara's market is very tight. The total vacancy rate was 2.3% in the County, compared with 3.7% in 1990. This is much lower than the 2000 rates in California (5.8%) and the United States (9.0%) (see Table A.4). Note that these vacancy rates are literal measurements of the proportion of all housing units that were not occupied at the time of the census; thus, they include seasonal vacancies (e.g., summer homes) and units in the process of turning over, and are higher than vacancies counted by real estate agents.

--The proportion of workers using public transportation to commute to work increased in Santa Clara from 2.9% in 1990 to 3.5% in 2000 a small increase that reflects substantial local efforts to improve mass transit. It is worth noting that the County increase bucked a national decline.

Age and Race in Santa Clara County

One of the most important ways that race affects public policy in Santa Clara County is in the interaction of age and race. On the four pages following the map section are figures which map the detailed age distribution of the four major racial groups in Santa Clara County (as throughout this report, we have defined these groups to be mutually exclusive). Each of the four groups has a distinct age demography. For example, the Hispanic population (median age 26) is dramatically younger than the white population (median age 40). Although the total white population of the County is nearly twice the Hispanic population, the number of Hispanics under 10 is larger in absolute terms than the number of whites of those ages. At the other end of the spectrum, whites over the age of 80 outnumber Hispanics by a margin of 7:1. The Asian age distribution in Santa Clara County is uncharacteristic of general California patterns -- Asians usually have large families and young age distributions, similar to Hispanics; but in Santa Clara County, the large volume of young, highly-educated Asian migrants has produced an age distribution dominated by the middle ranges, from ages 25 to 40. It is notable that for all four groups, children are not the most numerous age group -- a pattern different from most of middle America, notwithstanding the "aging" of the baby boom.


4. Maps and City Tables

Appendix I presents an extensive amount of detailed data illustrating the patterns we have discussed in this chapter. The first nine pages show demographic maps of the County. The first map shows population changes (by census tract) over the 1990-2000 decade, illustrating the population stability of the west-central suburbs, and the population growth around the County periphery. The next set of maps show the detailed racial makeup of the County in the year 2000. To provide a reasonable level of detail within the confines of a standard page, each map is presented over two pages, with the first page showing northwest Santa Clara County (where 90% of the County's population is concentrated) and the second page showing southeast Santa Clara County (where nearly all of the remaining 10% lives). The series covers the four major racial groups discussed in this chapter -- Blacks, Hispanics, Asians, and Whites.

Next follow a series of tables presenting detailed demographic data from the 2000 Census for the fifteen census places in Santa Clara County, including all of its principal cities. These are intended to help jurisdictions participating in the study to evaluate how trends discussed in this chapter apply to their jurisdiction.

The Empirical Research Group at UCLA School of Law has also created a CD, available on request, that contains all of these maps and others, plus a built-in simple mapping program. The CD also contains much of the demographic data discussed in this chapter.

Fair Lending in Santa Clara County: An Analysis of Home Mortgage Data


Since 1975, the federal government has required most financial institutions to disclose information about their lending patterns. These reports, mandated by the Home Mortgage Disclosure Act, became known as HMDA data -- pronounced "Hum-da." These requirements were expanded in 1990, so that most banks, savings associations, and many "mortgage brokers" now disclose each year the racial, gender, and income characteristics of all home loan applicants, and how those applications are resolved. This makes it possible to analyze in detail the lending experiences of different groups, the lending activity in individual neighborhoods, and the performance of individual lenders.

HMDA data usually shows that Blacks, Hispanics and women are less likely to have their loan applications approved than are White men. It is misleading to infer from this alone that discrimination is occurring, however, because these groups also tend to have lower incomes and other differences that need to be "controlled" statistically to make valid comparisons. The HMDA income data does not have enough information to control for all the variables necessary to make definitive assumptions from the statistical data.

Studies that have used other data to make careful comparisons of the credit experiences of different groups tend to find that well-qualified applicants from all racial and gender groups usually are approved for the home loans they seek, but that Blacks and Hispanics are significantly more likely to be rejected. One well-known study conducted for lending in the Boston region found an 88% approval rate for Whites with solid borrowing credentials, but an approval rate of 81% for Blacks and Hispanics. Studies like this suggest that minorities fare much better with lenders today than they did in the 1960s and 1970s, when discrimination was common, but that there is still some progress to be made. Some lenders, such as Bank of America, have instituted secondary review processes for applicants more likely to encounter discrimination, and these procedures have resulted in increased loans to minorities.

1. Results for Santa Clara County

In this analysis, we have analyzed HMDA data for Santa Clara County for the years 1992 through 1999. During this period, banks and other HMDA-regulated institutions within the County processed some 180,000 applications for home mortgages. This gives us a strong database for analyzing patterns of disparate treatment. Table 2.1, below, shows the gross volume of lending activity in Santa Clara County. The sheer volume of applications from different racial groups tells us some useful things about the racial demography of the County. For example, even though the White population in the County is declining, a majority of new homebuyers are White -- thus supporting the finding in Chapter One that the White decline represents increased competition for local housing rather than a flight of the White community. In other words, Whites are still actively purchasing housing in Santa Clara County in large numbers, even though many Whites are moving away -- a sign that the sorting and replacement process going on in Santa Clara County is one driven by economics rather than race. Or, put differently, we are inferring from this data that Whites who have a good economic reason to be in Santa Clara County do not seem hesitant to buy homes here, despite the change in Santa Clara's racial makeup.

Hispanic applicant volume is less than half that of Asians, even though the Hispanic and Asian populations are comparable in size. Black applications make up less than 2% of the total, also somewhat less than their demographic presence in the County would imply.

Table 2.1

Home Loan Approval Rates by Race in Santa Clara County*


Race Group		% County Population		% House-holders		Applications		% Applicants		% Approved
Asian	26.8%		23.6 %		47,838		26.7 %		84.5 %
Black	3.0%		3.1 %		3,271		1.8 %		74.7 %
Hispanic	23.6%		15.8 %		20,333		11.3 %		74.4 %
White	45.4%		56.3 %		96,030		53.6 %		85.8 %
Other	1.2%		1.2 %		11,797		6.6 %**		78.8 %
Total	100%		100%		179,269		100 %		83.5 %

*Home loans only (includes condos and townhouses). 1992-1999 HMDA data.

Loans with "errors" removed.

** HMDA "Other" designation is overinclusive, which accounts for the

high percentage of applicants in this category.

The implication is borne out by Table 2.2, which shows Black households are much less likely to own homes than Whites, even though the Black population is generally middle class.

Table 2.2

Homeownership Rates in Santa Clara County, 1990-2000

	% of Household Owning Homes
	1990	2000
Black	33.7	38.6
Hispanic	44.6	45.5
Asian	60.6	57.4
White	63.2	66.6

The gross approval rates measured in Table 2.3 show fairly typical racial disparities -- Hispanics and Blacks are approved at lower rates than Whites and Asians -- but, as we noted earlier, this is far from dispositive evidence of discrimination. Compared to the late 1970s, indeed, the minority approval rates are relatively good. Table 2.3 takes a more sophisticated look at the disparities, using the tool of logistic regression. In a logistic regression, a number of independent variables are used to measure what effect each has "at the margin" on the likelihood of the dependent variable turning positive or negative. In this case, some of the independent variables are characteristics tied to the individual loan application -- the race, gender, and income of the borrower, and the size of the mortgage loan sought -- and other independent variables are characteristics of the census tract where the borrower is trying to buy property -- racial makeup of the tract, the affluence or poverty of the tract, the age of housing in the tract, etc. The dependent variable is the outcome of each application (approved or rejected). The regression evaluates the relative effect of each of these variables on the loan outcome.

Interpreting Table 2.3. To explain what the output in Table 2.3 means, let's examine closely the first line of results, for Blacks. The first column (Parameter) tells us what independent variable's results are reported in this line; in this case, we are examining how an application's chances of approval are affected by the applicant being Black. The second column (Comparison Group) shows us the range over which Blacks are being compared; in this case, we are comparing Blacks to Whites because we are trying to understand whether Blacks have lower approval rates than Whites when other characteristics are controlled. (If we compared Blacks to everyone else, then it would be unclear whether a result was affected by many of those in the comparison group were also experiencing discrimination.) The third column (Chi-square) is a measure of how large the statistical disparity is in Black and White outcomes. In this analysis, a Chi-square value over 4.0 means that the disparity is large enough that it is probably due to systematic factors outside those in our regression, and is a larger disparity than can be accounted for the by "control" factors or by random influences. A larger Chi-square value reflects, in part, the size of the underlying disparity (e.g., how much more likely are Blacks to be turned down for loans) as well as the size of the populations being compared (the bigger the populations compared, the more likely it is that proportionate differences are due to systematic factors, and thus the higher the Chi-square).

The fourth column in Table 2.3 (p-value) translates the Chi-square value into a formal estimate of the frequency with which a disparity as large as the measured disparity would occur, if the differences were purely the result of random fluctuations. The lower the p-value, the more important systematic influences probably are in determining outcomes. The fifth column (Odds ratio) tells us which way the systematic influences cut. A value of less than 1.0 means that, when other characteristics are controlled, the group of interest (Blacks here) is less likely to be approved for a loan than the comparison group (Whites). The lower the number, the greater the likelihood of nonapproval (though .56 is not a direct measure of the "odds" that Blacks will be approved relative to Whites). Note, for example, that the Odds ratio for Blacks is somewhat lower than for Hispanics, even though the Chi-Square value for Hispanics is higher. The difference is caused by the higher absolute

Table 2.3

Logistic Regression Predicting Loan Approvals in Santa Clara County(Exploring the Effects of Additional STF3 Demographic Variable)

Parameter	Comparison Group	Chi Square	P-value	Odds Ratio
Intercept	N/A	51.5	0.0001
Blacks	Whites	195.4	0.0001	0.56
Asians	Whites	5.1	0.02	0.96
Hispanics	Whites	507.5	0.0001	0.63
Other Races	Whites	321.5	0.0001	0.64
Males	Females	1.2	0.27	1.02
Income of applicant	(continuous)	263.4	0.0001	1
Tract of MSA Income	(continuous)	101.1	0.0001	0.99
*% Minority (Blacks + Hispanics)	(continuous)	0.29	0.59	0.93
Amount of the loan	(continuous)	8.29	0.004	1
*% Pop., age 35-44, in poverty	(continuous)	9.05	0.003	0.5
*% Male pop., age 16+, employed	(continuous)	6.06	0.014	2.39
*Median HH income	(continuous)	77.6	0.0001	1
*Median year structure built	(continuous)	55.2	0.0001	0.99
*% HS grad. or more education	(continuous)	99.1	0.0001	3.85

Data were taken from the Home Mortgage Disclosure Act for 1992-1999 (N=179,214). Overall approval rate = 83.5%

Home purchase loans only. Approval refers to loans originated in comparison to those denied – all other loan outcomes were deleted. Data with "errors" deleted.

*Data taken from 1990 STF3 census tract data

** Somers’ D = .210. Somers’ D is a fit index that reflects the concordance rate (scores closer to zero reflect poorer predictive power)

number of Hispanic applications, which makes the finding of disparity more reliable (higher Chi-square) even though the size of the disparity is greater for Blacks (lower Odds ratio).

Odds ratios are harder to interpret when the independent variable is continuous. Thus, the ratio of a census tract's income to the County's (MSA's) median income is a significant predictor of a loan's approval or rejection, but the Odds ratio is hard to interpret because one is comparing the independent variable to itself (high-income tracts compared to low-income tracts). In these cells, therefore, we simply report the direction of the influence.

Substantive results of Table 2.3. The logistic regression shows clearly that Blacks and Hispanics are significantly more likely to be rejected in applying for home loans than are relatively comparable Whites. Asians do slightly worse than Whites, but the difference is small and only marginally statistically significant. Surprisingly, women do as well as men. In our neighborhood controls, the most important finding is that the racial makeup of a subject property is not a significant predictor of the loan's approval or rejection, when we control for the affluence and socioeconomic status of the tract. These factors are certainly correlated with race, and may be indicative of some redlining -- since neighborhood factors should be very nearly irrelevant in evaluating a loan. But the non-significance of the race variable suggests that any general redlining which is occurring is probably not based on banks' screening out areas that are perceived to be predominantly minority (though this might not be true for some individual banks, especially smaller ones).

It is important to keep in mind that even this logistic regression can only tell us that "systematic factors not in our equation" are causing Blacks and Hispanics to get fewer loans. One of those systematic factors might well be discrimination, but another systematic factor is creditworthiness. The few studies that have compared actual loan files have found that credit problems are more common among Black applicants than White applicants, and studies such as Black Wealth, White Wealth (Oliver et al) have documented the much lower assets middle-class Blacks and Hispanics have compared with whites. Minorities also have more difficulty recruiting qualified co-signors for mortgage applications. The HMDA data, therefore, cannot by itself let us definitely compare the evaluation by banks of loan applications and their disposition. However, the following analysis convinces us that discrimination is playing some role in the approval gap.

Comparisons of specific banks. To analyze bank performance from another perspective, we examined, in Tables 2.4 through 2.6, the relative approval rates of the three "minority" racial groups (Asians, Blacks, and Hispanics) to Whites at the principal institutions that each of the three minority groups apply to for home mortgage loans. These tables are cruder, but easier to read than Table 2.3. Table 2.4, for example, shows the ten HMDA-reporting financial institutions in Santa Clara County that received the largest volume of applications from Asians during the 1992-99 period. The table shows the volume of applications and the loan denial rate for Asians and Whites during this period. The final column shows the "p-value" for a Chi-square comparison of rejection rates. As in Table 2.3, the p-value is the probability that the measured disparity in rejection rates would occur if only random variation was at work in the numbers.

The Asian/White comparison is a good benchmark, for as we have seen, the overall rejection rates for the two groups are similar. We see in Table 2.4 that at each institution, the two groups' denial rates are generally within a couple of points of one another (sometimes higher and sometimes lower for Asians); even though some of the results are statistically significant, as a practical matter there seems little cause for concern.

Table 2.4

Santa Clara County Denial Rates of the Top Lending Institutions
For Asians Using Whites as a Comparison

Rank	Institution	City	State	Governing Agency	Asians		Comparison to Whites		Between Race Chi-Square p-value
					Applica-tions	Denial Rate	Applica-tions	Denial Rate
1	Washington Mutual Bank F.A.	Seattle	WA	OTS	3970	8.8	6824	10.4	0.01
2	Bank of America	San Francisco	CA	OCC	3606	15.5	8697	13.3	0.002
3	Countrywide Home Loans	Calabasas	CA	HUD	2425	14.4	5697	14.2	0.78
4	Norwest Mortgage, Inc.	Des Moines	IA	FRB	1748	11.2	3431	8.6	0.003
5	World Savings and Loan Assoc.	Oakland	CA	OTS	1736	8.2	2216	7.7	0.51
6	Provident Funding Associates	Burling'e	CA	HUD	1495	9.4	1022	10.8	0.27
7	Nationsbanc Mortgage Corp.	Dallas	TX	OCC	1318	19.9	1986	17.3	0.06

 

OCC Office of the Comptroller of the Currency

FRB Federal Reserve System FDIC—Federal Deposit Insurance Corporation

FDIC Federal Reserve System

OTS Office of Thrift Supervision

NCUA National Credit Union Administration

HUD Department of Housing and Urban Development

The pattern in Table 2.5 is quite different. At five of the institutions, the difference in denial rates between Blacks and Whites is over ten points; while in four other institutions, the gap is 6.1 points or less. These differences are striking and odd. One would think that systematic differences between Blacks and Whites in terms of assets or credit records would tend to play out in similar ways across different institutions. For rejection rates to be less than 1% apart at some companies, but 17% apart at Bank of America and Washington Mutual, is very suspicious. It is also notable that the conventional banks on the list tend to have greater disparities between the races than do the mortgage companies on the list. All of this suggests the possibility of either outright discrimination by some institutions, or the adoption of underwriting criteria by some, but not all, lenders, that greatly aggravate existing racial disparities.

Table 2.5

Santa Clara County Denial Rates of the Top 10 Leading Institutions
For Blacks Using Whites as a Comparison

Rank	Institution	City	State	Governing Agency	Blacks		Comparison to Whites		Between Race Chi-Square p-value
					Applications	Denial Rate	Applications	Denial Rate
1	Bank of America	San Francisco	CA	OCC	239	30.4	8697	13.3	0.0001
2	Washington Mutual Bank, F.A.	Seattle	WA	OTS	165	27.3	6824	10.4	0.0001
3	Countrywide Home Loans	Calabasas	CA	HUD	160	25.6	5697	14.2	0.0001
4	Norwest Mortgage, INC.	Des Moines	IA	FRB	136	14.7	3431	8.6	0.014
5	Home Saving of America, F.A.	Irwindale	CA	OTS	96	26	2954	12.4	0.0001
6	Great Western Bank, A FSB	Northridge	CA	OTS	91	25.3	1862	17.4	0.06
7	The CIT Group/Sales Financing	Livingston	NJ	FRB	81	28.4	1675	22.3	0.2
8	Kaufman & Broad Mortgage Co.	Woodland Hills	CA	HUD	76	6.6	1288	5.7	0.74
9	North American Mortgage Co.	Santa Rosa	CA	HUD	70	15.7	1761	15.1	0.88
10	Downey Savings and Loan Assoc.	Newport Beach	CA	OTS	67	35.8	2014	15.8	0.0001

4. Governing Agency Codes

OCC Office of the Comptroller of the Currency

FRB Federal Reserve System

FDIC Federal Deposit Insurance Corporation

OTS Office of Thrift Supervision

NCUA National Credit Union Administration

HUD Department of Housing and Urban Development

Examining Hispanic/White differences by individual lenders (Table 2.6) shows a third pattern. Here the differences in denial rates are more consistently large (and statistically very significant), though again the range in denial rates is striking. With the exception of CIT Group, one does not find in the Hispanic/White list any rejection ratios that approach 3:1, which exist for Blacks at either Bank of America or Washington Mutual. It is thus more plausible that the Hispanic rejection rates, though high, are due to more systematic factors that are widely followed by Santa Clara County lenders. For example, lenders may be systematically underappraising older homes or demanding proof of citizenship from borrowers.

Table 2.6

Santa Clara County Denial Rates of the Top 10 Lending Institutions
for Hispanics Using Whites as a Comparison

Rank

Institution

City

State