Homeownership determinants for Chinese Americans: assimilation, ethnic concentration, and nativity

Chinese homeownership rates in the Los Angeles CMSA adjusted by socioeconomic and housing market characteristics are on average 18 percentage points higher than those of native white households Painter et al. (2003). This finding runs contrary to most of immigration literature, which suggests that immigrants usually lag behind the host society in measures of economic well-being. This study focus on two additional factors, which most economic studies of homeownership choice ignore, that may play a role in helping Chinese households achieve high homeownership in ways that other immigrant group do not. The results of this analysis find that the high homeownership rates cannot be explained by the English skills of households. On the other hand, the cultural influence of home-owning peers may have partially contributed to the higher homeownership of Chinese households. While living in ethnic Chinese communities lowers homeownership rates, in general, it helps improve the likelihood that Chinese immigrants to own a home. Finally, we find that there is great diversity among Chinese subgroups with respect to their likelihood of owning a home, but very little diversity with respect to the education and income level of Chinese households.

Assimilation, Ethnic Concentration, and Nativity Homeownership Determinants for Chinese Americans: Gary Painter,* Lihong Yang,** and Zhou Yu*** * Lusk Center for Real Estate, University of Southern California Los Angeles, CA 90089-0626 or [e-mail]. ** Lusk Center for Real Estate, University of Southern California Los Angeles, CA 90089-0626 or [e-mail]. *** Lusk Center for Real Estate, University of Southern California Los Angeles, CA 90089­0626 or [e-mail]. 1 Chinese homeownership rates in the Los Angeles CMSA adjusted by socioeconomic and housing market characteristics are on average 18 percentage points higher than those of native white households Painter et al. (2003). This finding runs contrary to most of immigration literature, which suggests that immigrants usually lag behind the host society in measures of economic well-being. This study focus on two additional factors, which most economic studies of homeownership choice ignore, that may play a role in helping Chinese households achieve high homeownership in ways that other immigrant group do not. The results of this analysis find that the high homeownership rates cannot be explained by the English skills of households. On the other hand, the cultural influence of home-owning peers may have partially contributed to the higher homeownership of Chinese households. While living in ethnic Chinese communities lowers homeownership rates, in general, it helps improve the likelihood that Chinese immigrants to own a home. Finally, we find that there is great diversity among Chinese subgroups with respect to their likelihood of owning a home, but very little diversity with respect to the education and income level of Chinese households.In recent years, issues of the importance of and access to homeownership have triggered substantial academic research and policy debate. The research on access to homeownership (see, e.g., Coulson 1999; Gyourko and Linneman 1996) is, in part, motivated by sizable and persistent gaps in homeownership attainment between white and minority households/ While the U.S. homeownership rate rose perceptibly over recent years to a record 67.1 percent in mid-2000, Simmons (2001) reports that the longstanding white-minority homeownership gap of about 28 percentage points was little changed. These facts are particularly troubling given recent research that has demonstrated positive socioeconomic and community impacts associated homeownership (see, e.g., Green and White 1997; Haurin, Parcel and Haurin 2002). Population changes in the past decade signal a large increase in non-white population. Results from the 2000 Census in the United States suggest that Latino populations have increased by 58 percent and that Asian populations have increased by about 76 percent over the past decade" Within the Asian minority group, Chinese immigrants now well exceed two million - the largest Asian immigrant group in the United States (U.S. Census Bureau 2002). More recently, Mainland China has become the second largest immigrant sending country next to Mexico (Office of Policy and Planning 2002). These changing demographics have the potential to create an adverse impact on overall homeownership rates, because ethnic minorities have homeownership rates that are much below that of white, non-Hispanic households (see, e.g., Alba and Logan 1992; Coulson 1999; Krivo 1995; Painter, Gabriel and Myers 2001). Thus, it is critical to understand the factors that lead immigrants to own homes, and to investigate under what conditions successful transition into homeownership has been accomplished. 2 Unlike evidence shown in Bianchi, Reynolds and Spain (1982) and Wachter and Megbolugbe (1992) concerning homeownership gaps between African Americans and whites, recent studies by Coulson (1999) and Painter, Gabriel, and Myers (2001) suggest that the national gaps between whites and Latinos and Asians are largely due to income differences, residence in high cost metropolitan areas, and the high mobility rates among recent immigrants. Two recent studies-Painter et. al.(2001) and Painter, Yang, Yu(2003)-of large gateway metropolitan areas suggest that there exists very little differential between Asians and whites in the likelihood of homeownership. Further, Painter et. al. (2003) found that Chinese households were much more likely to own homes than whites and Asians other than Chinese. This finding would be surprising to proponents of the traditional assimilation literature (see, e.g., Alba and Nee 1997; Gordon 1964), as that theory is oriented to explaining the process of immigrant adaptation into society as a function of catching up to status of native white households. Recent Chinese immigrants, rather than climbing up the socioeconomic ladder over time, may have achieved a socioeconomic status comparable to that of native-born whites soon after arrival. While higher Chinese homeownership rates are surprising, Alba, Logan, and Zhang (2002) propose a conception of emerging ethnic communities among immigrants that may provide insights into the mechanism for why Chinese homeownership rates are so high. This notion of immigrants choosing to live together even with elevated socioeconomic status may suggest that these groups have unique socioeconomic ties and are opting for close access to ethnic resources rather than immersing into white majority neighborhoods through spatial assimilation. Further, if groups are choosing to live together, and these groups have both high initial preferences for homeownership and large ethnic resources for sharing, there could be peer 3 influence among Chinese households to buy homes that could greatly reinforce homeownership attainment among these households (Glaeser, Sacerdote and Scheinkman 1996; Manski 1995). The aim of this paper is to better understand the role that assimilation, ethnic concentration, and nativity play on the homeownership rates of Chinese households. We proceed by testing four hypotheses using 1990 Public Use Microdata (PUMS) data from the Census. First, we examine the role of language proficiency, as an indicator of assimilation, in enabling households to own a home. We follow the procedure in Alba and Logan (1992) and Myers and Lee (1998). Second, we examine the role of peer influence to own homes among Chinese households in communities with a large presence of Chinese and relatively high Chinese homeownership rates. Next, we examine the role of different nativities among Chinese households to examine if households coming from places with different national wealth possess different likelihoods of owning a home. Finally, we examine the role of socioeconomic status and time since first immigration and its interaction with nativity. This enables us to isolate the factors that may have led Chinese households to be so successful in achieving high homeownership rates. Past Research and Theory Assimilation Traditionally, research on homeownership choice has not focused on immigrants, and therefore has neglected investigating factors that may play an important role in determining the likelihood that an immigrant will own a home. Two potentially important influences on immigrant's homeownership are the extent to which an immigrant household has been assimilated into the 4 host society, and the extent to which peers influence their decision making and potentially provide information or resources to the household. As revealed in Gordon (1964) and Alba and Nee (1997), assimilation theory describes a straight- line process of adaptation and acculturation, leading immigrants to a state of structural integration into the host society. These theories were based primarily on the experience of earlier European immigration. Many scholars have challenged the validity of this "melting pot" theory of assimilation in the contemporary context. Immigrants after 1965, mostly coming from Asia and Latin America, have been characterized by their drastic diversity of socioeconomic backgrounds and national origins. Instead of convergence over time and forming a unified group, Portes (1995) shows that many recent immigrants have experienced different adaptation processes, and sometimes even shown a pattern of perpetual ethnic differences. As indicated in Zhou (1997a), the literature has recognized that the importance of contextual effects in immigrants' assimilation, which underpins socioeconomic stratification in the host society. The high homeownership attainment of Chinese immigrants may be an additional case for the notion of "segmented assimilation" (Portes and Zhou 1993; Zhou 1997b). Assimilation is manifested in many socioeconomic characteristics, such as the cultural norms, beliefs, and behavior patterns. In practice, English proficiency has been widely used as an indicator of the assimilation process (see, e.g., Alba and Logan 1992; Krivo 1995; Myers and Lee 1998). Assimilation theory suggests that immigrants with higher English language ability are able to adapt better to the host society. English language ability is also a necessary skill for communicating with other people and negotiating the transactions necessary for purchasing a 5 home. Consequently, English language ability should be positively associated with spatial assimilation, socioeconomic well being, and homeownership attainment (Alba and others 1999; Carliner 2000; Fong and Kumiko 2000; Krivo 1995). Ethnic Concentration While many recent studies have revealed ambiguities in the assimilation theory with respect to the socioeconomic outcomes of assimilation, few have focused on the role of ethnic concentration on homeownership attainment. On the one hand, forced concentration may restrict housing availability for certain ethnic groups, as discovered in Flippen (2001), Galster (1987), White (1987), and Massey, Gross and Shibuya(1994). Toussaint-Comeau and Rhine (2000) find that residence in an area dominated by ethnic enclave leads to lower homeownership rates in that area.iii Concentration has also been shown to lead to inferior labor market outcomes (see, e.g., Borjas 1998; 2000) and less socioeconomic mobility (see, e.g., Alba, Logan and Stults 2000; Allen and Turner 1996; Massey and Denton 1987). On the other hand, concentration may provide better access to ethnic resources, such as an informal financial system, credit sharing, and neighborhood support as revealed in Fong and Gulia (2000) and Schoeni, McCarthy, and Vernez (1996). Further, contextual effects, as discussed by Manski (1995) and others, may arise among ethnic groups that reinforce strong initial preferences to own homes. Haurin, Dietz, and Weinberg (2003) provide a thorough review of the effects of neighborhood homeownership rates, through social interactions, on the residents of a community and an adjacent community. Through feedback mechanisms or social multiplier effects, neighborhood context can strengthen mutually supported social behavior such as religion, poverty, or homeownership attainment. This conception of contextual effects 6 presents additional explanations of peer influences on homeownership attainment, particularly within immigrant communities whose residents share cultural preferences and social network. Peer influence could thus lead to higher homeownership rates than native white populations among some groups, if these initial preferences for homeownership were particularly strong. In other words, residing in immigrant communities may bolster homeownership rates of immigrants, if the community is particularly predisposed to homeownership. Chinese Subgroups A final issue relevant for understanding the homeownership rates of Chinese households concerns the heterogeneous background of Chinese immigrants in the U.S. The Chinese population in the U.S. has experienced significant changes in composition with respect to national origin and reasons for immigration. For example, while immigrants from Mainland China chose to come to the U.S. mainly because of economic prospects, most Chinese immigrants born in Vietnam were forced to leave their county and came as refugees after the Vietnam War (Chang 1999). Meanwhile, recent immigration from Taiwan and Hong Kong/Macau was largely due to concerns over economic security and heavily influenced by the relationship with Mainland China, as shown in Li (1998), Tseng (1995), and Wachman (1994). While it is a rather recent phenomenon to have a large number of immigrants directly from Mainland China, immigration from Hong Kong/Macau and Taiwan began soon after the passage of the immigration reform law in 1965, as revealed in Ng (1998) and Skeldon (1995). Economically, Chinese immigrants from Vietnam and Mainland China are relatively impoverished, while those from Taiwan and Hong Kong/Macau have higher educations and income before their immigration (Tseng 2000; Zhou 1992). While Chinese immigrants come 7 from regions with different socio-economic and political conditions, Chang (1999) and Takaki (1994) show that they also possess a common cultural identity. Natural questions arise as to what extent that heterogeneity exists in homeownership attainment across different Chinese groups and, how such differences fare in comparison to the homeownership rates of native-born whites. The varied experiences of Chinese immigrants provide a unique setting for research on immigration, as they represent a group of people with cultural connections yet economically and politically diverse backgrounds. Bifurcations are evident in recent Chinese immigrants, who are clustered at both ends of the socioeconomic spectrum (see, e.g., Chang 1988; Cheng and Yang 1996; Li 1998; Zhou and Gatewood 2000). Despite wide acknowledgement of internal diversity of Asian immigrants (White, Biddlecom and Guo 1993), the existing body of work has largely treated Chinese immigrants as an internally coherent group. This research will be able to highlight the differences that exist with respect to one socioeconomic outcome, namely homeownership, and the role played by different nativities - Mainland China, Taiwan, and Hong Kong/Macau, different initial socioeconomic status, and different timing of immigration. Data This analysis primarily uses data from the 5% Public Use Microdata Sample (PUMS) file of the 1990 decennial Census.iv The 1980 5% PUMS data and the recently released Census 2000 5% PUMS data will also be used to compare trends in Chinese homeownership rates from 1980 to 2000. The analysis focuses on Chinese households in Los Angeles Consolidated Metropolitan Statistical Area (CMSA), which comprises four Primary Metropolitan Statistical Areas (PMSA). 8 The four PMSAs include Los Angeles-Long Beach PMSA, Anaheim-Santa Ana PMSA, Riverside-San Bernardino PMSA, and Oxnard-Ventura PMSA. The Los Angeles region has the largest Asian population among all metropolitan areas, and it is also a gateway metropolitan area for Chinese immigrants (Waldinger and Bozorgmehr 1996). Although the primary focus of the analysis is on the Los Angeles CMSA due to sufficient sample size of Chinese sub-populations in the region, we also verify our results using the New York, San Francisco, and Washington DC metropolitan areas. The sample includes all households that either own or rent their primary residence, excluding persons who reside in group quarters. The samples are also limited to those householders that are aged between 18 and 64. Based on place of birth, Chinese immigrants are divided into four groups, which are those born in Taiwan, Hong Kong/Macau and Macau, Mainland China, and other places. In addition, U.S.-born Chinese, Asian other than Chinese, and non-Hispanic white households are included to provide a useful benchmark/ Even though this categorization is not precise in all cases, the place of birth provides a consistent grouping scheme based on the information available" Table 1 shows the homeownership rate for all ethnic groups in 1990 for both the full sample of households, and the sample of recent moversvii Asians and whites have similar homeownership rates in 1990 (57 and 61 percent, respectively), while Latinos and African-Americans have substantially lower homeownership rates (43 and 37 percent, respectively). The Chinese subgroup has substantially higher homeownership rates at 68 percent. Table 2 displays the rates for each Chinese subgroup based on the place of birth and extends the time frame from 1980- 9 2000. The overall rate of Chinese homeownership has remained between 61 and 68 percent over the period. Within the Chinese subgroup, the nativities with the highest homeownership rates are Taiwan and Mainland China, with over 70 percent homeownership in 1990. The Chinese nativities with the lowest homeownership rates - approximately 50 percent - are households of Chinese ancestry from Vietnam and other parts of Southeast Asia. Tables 1 and 2 about here To determine how much of the differences in homeownership rates can be explained by socioeconomic characteristics and other factors, we include demographic factors (race-ethnicity, age group, marital status, number of persons in the household, number of workers in the household, migration origin, and migration history), economic factors (income, education level of the head of household), and variables to capture local housing market conditions (housing price and rent levels)viii These variables enable the researcher to capture factors that influence homeownership choice based on the user cost of homeownership and factors related to preferences of households correlated with demographic characteristics such as the life cycle (see, e.g., Myers, Megbolugbe and Lee 1998; Skaburskis 1996). Instead of simply including household income, we use measures of permanent and transitory household income. Using the method of Goodman and Kawai (1982), permanent income is the predicted value of a regression of household income on a set of demographic and human capital characteristics™ Transitory income is calculated as the residual of observed household income and predicted income. In addition, a household's dividend and interest income is included to partially capture wealth requirements that are needed to meet mortgage downpayment requirements. Since no direct measures of wealth are available in these data, education attainment of the head of household can 10 also be considered as a proxy to indicate the future earning potential of households (Gyourko and Linneman 1996). Three final variables that are important in the immigration literature (see, e.g., Krivo 1995; Myers and Lee 1998) -- immigrant status, length of stay, and migration origin - are also included. Appendix I reports the mean values of all independent variables used in the study. Rather than discuss all of the differences in detail, we focus on some of the larger differences concerning income and immigrant status in Figures 1 and 2. Figure 1 presents the difference in permanent incomes by groups. U.S. born Chinese and whites have the highest household income while those from other places or Taiwan have the lowest household income, suggesting that the higher homeownership rate of Chinese households is not due to higher incomes.x Figure 2 reveals that there had been substantial growth in Chinese population in Los Angeles during the 1980s. For instance, Taiwanese immigrants increased by almost two-fold during the 1980s, while immigrants from Mainland China grew by almost fifty thousand. New immigrants who came in the last 10 years have contributed to most of the growth in the Chinese population. Figures 1 and 2 about here As shown in Figure 2, Chinese subgroups are also diverse with respect to immigrant status. This study is particularly concerned with the implications of ethnic concentration and assimilation on Chinese high homeownership rates. In addition to the independent variables discussed above, this paper uses two additional sets of variables. The first set measures English proficiency and its interaction variable with the Chinese ethnicity categorical variable. Using the self-assessment of English proficiency in the PUMS data, we divide household heads into two groups: high or 11 low English proficiencyxi The group with high English proficiency is further divided by whether or not only English is spoken exclusively at home. This leads to three categorical variables: English spoken well and exclusively at home, English spoken well but not exclusively at home, and English spoken poorly, which is used as the omitted category in model estimation. This characterization of language ability and its interaction with Chinese ethnicity enables a test of the independent effect of language, and its relative importance for Chinese households. The second set of variables is designed to examine whether the homeownership rates of any individual Chinese household is in part influenced by the concentration of Chinese households in the same area. Two different measures are used. It could be the case that cultural factors influence Chinese preferences for homeownership as indicated in Brown and Pannell (2000) and Zhou (1992), and thus concentration as measured by percent of an area that is comprised by Chinese households can increase the attractiveness of certain areas. This variable would also capture if homeownership is more attractive to Chinese households in areas with other Chinese. A second conception of peer influence emphasizes the residence of Chinese homeowners. This peer influence variable is constructed by multiplying the difference in homeownership rates between Chinese and white households in the same area by the categorical variable for Chinese ethnicity. It is presumed that higher values of peer influence exist if a household is surrounded by greater numbers of Chinese homeowners. The underlying white household homeownership rate is included to control for some of the housing market characteristics that are unobserved in the data. 12 The delineation of each area is primarily based on the Public Use Micro Area (PUMA) in the PUMS data. We create the variable for the percent share of Chinese households in the area at individual PUMA level. When there are not sufficient Chinese households residing in some PUMA's, we aggregate surrounding PUMA's based on spatial adjacency and similar socioeconomic conditions. For the method that focuses on Chinese homeowners, the Los Angeles metropolitan area consists of 27 aggregated areasxii Appendix II shows the corresponding municipalities in each area. Figure 3 shows that for most areas, Chinese homeownership rates are at least 10 percentage points higher than whites. Figure 3 about here Model Most recent studies of homeownership evaluate cumulative attainment of homeownership (tenure status) among a sample of existing households (see, e.g., Alba and Logan 1992; Coulson 1999; Gyourko and Linneman 1996). The cumulative approach has been justified by the view that homeownership is a long-term decision based as much upon anticipated future needs as on present needs (Edin and Englund 1991). However, among households who are ages 45 or older, cumulative attainment of homeownership may largely reflect the lagged effects of past choices. A second approach assumes that homeownership decisions of recent movers more closely reflect equilibrium conditions and avoid that lagged effect (see, e.g., Boehm, Herzog and Schlottmann 1991; Ihlanfeldt 1981)xiii A key drawback to basing analysis of homeownership choice on a sample of recent movers concerns possible sample selection bias. Renters and those less settled, such as immigrants, are 13 over-represented in a sample of recent movers. For that reason, estimates of the determinants of homeownership choice could be biased. Although Census data do not report the tenure status of households prior to their move, one can estimate a model of their likelihood of entering the mover sample. To address possible sample selection bias, this study uses a Heckman-style correction described by Painter (2000).xiv The homeownership choice model with correction for selection bias is adapted from Van de Ven and Van Pragg (1981), in which both the selection equation and the tenure choice equation have binary dependant variables. The selection equation uses a probit model with the choice to move as the dependent variable with controls for socioeconomic factors that may affect the moving propensity of households. Homeownership choice is assumed to be observed only if a household moves. It is assumed that the error terms in both models are jointly normally distributed with correlation coefficient p. The resulting model is estimated using a maximum-likelihood procedure to obtain the parameters of each equation and the correlation between each choice.xv Results While Tables 1 and 2 highlight how much greater Chinese homeownership rates are than whites or other Asians, we now consider how large these gaps would be after controlling for a variety of socioeconomic and local characteristics. Four model specifications are presented in Table 3. The first includes Chinese and Asians other than Chinese separately, with white households as the reference groups. The second, third and fourth specifications add two sets of additional variables based on English proficiency and population and homeownership concentration of Chinese households. Overall, being married, higher education, higher income, lower housing 14 prices, higher rents, and moving within Los Angeles all lead to higher homeownership rates. The impact of these variables is consistent across model specifications. In addition, the correlation coefficients between the tenure choice equation and the mobility equation are significant in all of three model specifications, suggesting the importance of controlling for mobility explicitly in estimating the model. Table 3 about here The results in Table 3 (Model I) mirror the results in Painter et al. (2003) Chinese homeownership rates are significantly above those of whites and other Asians. The difference translates into a marginal difference in the probability of homeownership of 18 percentage points. The English proficiency variables and their interactions with the Chinese ethnicity variable are introduced in Model II of Table 3. As is common in the literature (see, e.g., Alba and Logan 1992; Krivo 1995), better English skills increase the likelihood of owning a home. For Chinese households, however, it appears that speaking multiple languages in the home gives them an advantage in the housing market as evidenced by the insignificant coefficient on the Chinese interaction term, while speaking English exclusively at home effectively cancels out the effect of good English skills. The final two columns in Table 3 (Models III and IV) test for the role of both concentration of households and Chinese homeowners. The results are consistent, and provide suggestive evidence as to the importance of concentration. Areas with higher Chinese concentrations are associated with lower homeownership rates in the general population. However, these areas are strongly associated with higher Chinese homeownership rates. This finding is similar to other 15 recent work (see, e.g., Borjas 2002; Toussaint-Comeau and Rhine 2000), which has found that immigrant homeownership rates increase in areas of high ethnic concentrations. We further find that in the model results including concentration of Chinese homeowners that the unexplained higher homeownership rates of Chinese households falls by 30 percent, suggesting that peer influence may play a role in reinforcing Chinese preferences for homeownership. Decomposing the Chinese into Five Subgroups We next investigate how the determinants of homeownership attainment and the role of English proficiency and concentration across different Chinese subgroups because of the research (see, e.g., Brown and Pannell 2000; Chang 1999) that suggests that Chinese immigrants have diverse immigration experiences. The first model specification in Table 4 demonstrates that all five Chinese subgroups have higher homeownership likelihoods relative to whites when controlling other household and housing market characteristics. Among Chinese groups, those households from Taiwan have the highest homeownership likelihood, followed by households from Mainland China and Hong Kong/Macau. These adjusted differences are different from the unadjusted differences reported in Table 2. The calculated difference relative to whites in marginal probabilities is 30 percentage points for households from Taiwan, 16 percentage points for households born in Mainland China or Hong Kong/Macau, 11 percentage points for U.S. born Chinese, and 8 percentage points for Chinese households born in other regions. Table 4 about here The importance of English proficiency (Table 4:Model II), while important for Asians in general, differs significantly in each Chinese subgroup. For Chinese households from Taiwan, English 16 proficiency does not contribute to homeownership attainment at all. For those from Mainland China, speaking English only in the home does not increase homeownership, while having good English skills, but not speaking English in the home is only moderately associated with higher homeownership rates. On the other hand, English skills are consistently more important for U.S. born Chinese and Chinese from other regions of Asia. As in Table 3, higher concentrations of Chinese households and homeowners implies lower homeownership for the full population of Asians and whites (Table 4: Model III and IV), but is positively or insignificantly related to homeownership for all Chinese sub-groups. After controlling for both the language variables and peer influence variables together, the unexplained homeownership differences between the whites and Chinese households from Hong Kong/Macau, the U.S., and other regions, as indicated by the categorical variables, are not significant. On the other hand, unexplained homeownership differentials between households from Taiwan and Mainland China remain at 25 and 17 percentage points, respectively. There are some differences, however, with respect to the importance of living in areas with either high concentrations of Chinese households or concentrations of Chinese homeowners by subgroup. Taiwanese households are much more likely to buy a home if there are greater numbers of Chinese homeowners, while households from the mainland are equally influenced by the presence of any Chinese households. Some of these differences may be related to the location preferences by households who choose to live in certain areas when buying a home. It may also suggest that while all Chinese households are more likely to own in areas with greater 17 numbers of Chinese households, the mechanism by which peers influence each other may be quite different for different groups. Results from Other Metropolitan Areas To confirm these research findings from Los Angeles, we extend our analysis into three additional metropolitan areas - San Francisco-Oakland-San Jose, CA CMSA, New York- Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA, and Washington, DC-MD-VA, MSA. Although the Chinese population is not as large as it is in Los Angeles, we are able to examine the extent to which the high homeownership of Chinese is unique to Los Angeles, and whether the homeownership differences across sub Chinese groups are also observed in other metropolitan areas. We are unable to test for evidence of concentration because there is not sufficient sample size of Chinese households across PUMAs. As reported in the Appendix 3, the high level of homeownership is also evident in other three metropolitan areas. Asians, other than Chinese, have either the same or slightly higher level of homeownership attainment than their white counterpart. Again, as observed in Los Angeles, while speaking English well in general elevates one's homeownership probability, Chinese households that speaking English only at home appear to be at a disadvantage relative to those that speak multiple languages in the home. Next, we investigated the differences across sub Chinese groups in their homeownership attainment in San Francisco and New York, but not in the Washington DC metropolitan area, due to insufficient sample size in the Chinese population. As is evidenced in the Appendix 4, Taiwanese households in San Francisco and New York have the same high homeownership 18 propensity as they do in Los Angeles. Meanwhile, Chinese who were born in the U.S., Mainland China, and Hong Kong also indicate a strong propensity for homeownership in San Francisco, but these effects are somewhat smaller in New York. Consistent with the results observed in Los Angeles, speaking English well improves homeownership in general. However, this effect is lessened for these Chinese subgroups. Tests for Bifurcations Finally, we examine whether the findings reported above are consistent for different education groups, different income groups, and different immigration groups, and over time (Table 5). First, we separate the whole sample into two sub-samples based on the level of education attainment of household head as a means of distinguishing between alternative hypotheses for why Chinese homeownership rates are higher than that of whites. It would be expected that highly educated Chinese households would only differ from whites due to preferences for homeownership, as highly educated households would have fewer unmeasured wealth differences that might constrain lower educated households from obtaining a down payment. In the "high education" sample, the household heads have college degree or higher, and the "low education" sample contains households with high school diploma or less. Table 5 about here Contrary to expectation, there is actually a higher unexplained probability of Chinese homeownership among the highly educated group when compared to the lower educated group. The difference is 20 percentage points for the highly educated group in comparison to 11 percentage points for the low education group. Part of this difference can be attributed to the variable that captures peer influence, which is slightly more important for the low education 19 sample. When we decomposed the sample into higher than median and lower than median income, the results mirror the decomposition by educational attainment. Since most Chinese Americans are immigrants, we also tested whether there were differences in unexplained homeownership based on the period of immigration. In this test, we group the Chinese households coming to the United States after 1980 census as "new immigrants." Surprisingly, English proficiency has less impact on Chinese new immigrants than their counterparts who had arrived many years before, but the new immigrants are much more affected by peer influence than the others. In addition, both the new and old Chinese immigrants have similar unexplained homeownership rates when compared to whites in their specific group, which is contrary to other Asians which are less likely to own in the "new immigrants" group but are slightly more likely to own in the "old immigrants" group. One possible explanation for these results is the large number of Chinese immigrants during 1980s from Taiwan and Hong Kong/Macau that went through rapid economic growth in that period. Some of the new immigrants accumulated wealth in their origination areas before immigration, which enabled them to meet any downpayment constraints easier than similar households born in the United States. At the same time, because the advantage of Chinese households is consistent across samples, it is unlikely that unmeasured wealth can be the whole story. Next, we considered how robust the results on Chinese homeownership are across the decades from 1980 to 2000. Because of data limitations, we cannot test the relative importance of peer influence over the decades. Comparing these three time periods, the predicted homeownership rate gap between Chinese and white households broadened for all Chinese subgroups during 20 1980s, and went away for most Chinese subgroups except for those from Hong Kong/Macau and Taiwan during 1990s. Thus, the group of immigrants that arrived in the 1980s may have had unusually high levels of unmeasured wealth, but in every period Chinese homeownership rates are higher than native populations. Finally, we considered an alternative measure of concentration that has been used in the literature. We created an Entropy (or Diversity) index based on Theil (1967) for Chinese households in an area.xvi Consistent with all the robustness checks discussed, this categorization of Chinese concentration also plays an important role in determining homeownership among Chinese households. The relationship largely mimics the results that were obtained using the simple household concentration measure. Conclusion Only recently has research begun to explore homeownership choice among immigrant groups despite the fact that nearly all of the population growth in the United States comes from immigrants and the children of immigrants. Some of this research (see, e.g., Painter, Yang and Yu 2003) has found that Asian homeownership rates are similar to that of whites and Chinese homeownership rates are substantially higher than other ethnic groups. In this research, we have tested a variety of hypotheses regarding the determinants of Chinese homeownership rates. We focused on the relative importance of English proficiency and ethnic concentration in Chinese homeownership attainment as two additional factors that may influence homeownership rates in ways that traditional homeownership analyses have neglected. We further examined the differences in the rate that Chinese subgroups achieve homeownership to see if these households have consistent patterns despite different nativity. 21 We found that language proficiency, while important for in the general population of Asians, was less important for Chinese households, and for Taiwanese and households from Mainland China, language had no effect. This finding suggests that for homeownership attainment, language assimilation is not as important for Chinese immigrants as it is for other immigrant groups (see, e.g., Alba and Logan 1992; Krivo 1995). The positive impacts of English that were found were largely confined to the group that spoke English well, but not in the home. This finding is consistent with the research by Feliciano (2001) and Portes and Zhou (1993), which found maintaining multicultural or bilingual status are positively associated with educational success and better socioeconomic outcomes. On the other hand, we found suggestive evidence that the presence of ethnic Chinese communities can partially explain why Chinese groups own homes at higher rates than the native white population. Whether this is due to the presences of shared resources or to peer influence, this research design cannot distinguish, but future research is needed to sort out the relative importance of each. Finally, we found significant differences based on the nativity of Chinese households. Groups from Taiwan had the highest unexplained rates (25 percentage points), which may be partially due to initial wealth that is unmeasured in the data. At the same time, households from Mainland China and Vietnam without high initial levels of wealth had significantly higher, adjusted rates of homeownership than white households (17 and 8 percentage points, respectively). These households may be accessing informal networks of wealth, but this is again unmeasured in the data. After controlling for socioeconomic factors, language ability, and ethnic concentration, 22 only households from Taiwan and Mainland China had significantly higher homeownership rates when compared to whites or other Asians. Finally, we found no difference in a household's likelihood of owning a home due to differential education levels or "newness" of immigration, relative to whites. One important topic for future research is to simultaneously model the location choices of immigrant households with their homeownership choice. In this study, we estimated the factors that influence homeownership choice conditional on a household's choice of location. While there are controls for location characteristics such as the price of housing, rents, and the concentration of Chinese households in an area, there is a relatively new literature such as Gabriel and Painter (2003) and Deng, Ross, and Wachter (2003) that suggests that dual consideration of location choice and homeownership choice can yield important insights into how households make these decisions. To the extent that households make both decisions simultaneously, future research will investigate how sensitive the results of this study are to the endogeneity of location choice. In this analysis the endogeneity of location choice may bias up the estimated impact of Chinese homeownership concentration, but because this was shown to be important across a variety of sub-samples (Table 5), it is unlikely to eliminate the importance of this factor. A topic not discussed in this research is whether higher homeownership observed among Chinese households is a good thing for the community and the individual households. While higher homeownership may have neighborhood benefits (see, e.g., Green and White 1997; Rohe, Van Zandt and McCarthy 2000), it has been suggested that homeownership may have negative 23 effects on households under certain circumstances. For low-income households, owning one's home causes the household to be more vulnerable to the idiosyncratic risk of the real estate market, because housing usually comprises the largest portion in their investment portfolio as indicated in Rohe, Van Zandt, and McCarthy (2000). More importantly, homeownership may limit household mobility, since the cost associated with moving for homeowners is much higher than renters as shown in Quigley and Weinberg (1977). This restriction in mobility may hamper the labor market opportunities of households (see, e.g., Boehm 1981; Green and Hendershott 2001). This is likely to be particularly important for new immigrants who are usually less settled and more responsive to the shifts in the labor market as shown in Borjas (2001). Future research should explore the impact that elevated homeownership plays on these other outcomes. 24 The authors thank the Lusk Center for Real Estate for providing funding for this project. Zhou Yu appreciates the financial support of the HUD Doctoral Dissertation Research Grant (H- 21390SG). Points of views or opinions in this paper are those of the authors and do not necessarily represent the official position or policies of the grantors. 25 References Alba, R., and J. R. Logan. 1992. Assimilation and Stratification in the Homeownership Patterns of Racial and Ethnic Groups. International Migration Review 26(4): 1314-1341. Alba, R., and V. Nee. 1997. Rethinking Assimilation Theory for a New Era of Immigration. International Migration Review 31(4): 826-874. Alba, R. D., J. R. Logan, and B. J. Stults. 2000. The Changing Neighborhood Contexts of the Immigrant Metropolis. Social Forces 79(2): 587-621. Alba, R. D., J. R. Logan, B. J. Stults, G. Marzan, and W. Zhang. 1999. Immigrant Groups in the Suburbs: A Reexamination of Suburbanization and Spatial Assimilation. American Sociological Review 64(3): 446-460. Allen, J. P., and E. J. Turner. 1996. Spatial Patterns of Immigrant Assimilation. Professional Geographer 48(2): 140-155. Bianchi, S. M., F. Reynolds, and D. Spain. 1982. Racial Inequality in Housing: An Examination of Recent Trends. Demography 19(1): 37-51. Boehm, T. P. 1981. Tenure Choice and Expected Mobility: A Synthesis. Journal of Urban Economics 10: 375-389. Boehm, T. P., H. W. Herzog, and A. M. Schlottmann. 1991. Intra-Urban Mobility, Migration, and Tenure Choice. Review of Economics and Statistics 73(1): 59-68. Borjas, G. 1998. To Ghetto or Not to Ghetto: Ethnicity and Residential Segregation. Journal of Urban Economics 44(2): 228-253. ------. 2001. Does Immigration Grease the Wheels of the Labor Market? Brookings Papers on Economic Activity 2001(1): 69-134. Borjas, G. J. 2000. Ethnic Enclaves and Assimilation. Swedish Economic Policy Review 7: 89­122. ------. 2002. Homeownership in the Immigrant Population. Journal of Urban Economics 52(3): 448-476. Brown, C. L., and C. W. Pannell. 2000. The Chinese in America. J. O. McKee. editor. Ethnicity in Contemporary America: A Geographical Appraisal. Rowman & Littlefield Publishers: Lanham, MD. Carliner, G. 2000. The Language Ability of U.S. Immigrants: Assimilation and Cohort Effects. International Migration Review 34(1): 158-182. Chang, G. H. 1999. Diaspora and Immigration - Case Studies - Writing the History of Chinese Immigrants to America. The South Atlantic Quarterly 98(1): 135-143. Chang, M. S. H. 1988. From Marginality to Bimodality: Immigration, Education, and Occupational Change of Chinese Americans, 1940-1980, Unpublished doctoral dissertation, Stanford University, Stanford. Cheng, L., and P. Q. Yang. 1996. The "Model Minority" Deconstructed. R. Waldinger, and M. Bozorgmehr. editors. Ethnic Los Angeles. Russell Sage Foundation: New York. Coulson, N. E. 1999. Why Are Hispanic- and Asian-American Home-Ownership Rates So Low?: Immigration and Other Factors. Journal of Urban Economics 45(2): 209-227. Deng, Y., S. L. Ross, and S. M. Wachter. 2003. Racial Differences in Homeownership: The Effect of Residential Location. Regional Science and Urban Economics. Edin, P.-A., and P. Englund. 1991. Moving Costs and Housing Demand: Are Recent Movers Really in Equilibrium? Journal of Public Economics 44: 299-320. 26 Feliciano, C. 2001. The Benefits of Biculturalism: Exposure to Immigrant Culture and Dropping out of School among Asian and Latino Youths. Social Science Quarterly 82(4): 865-880. Flippen, C. A. 2001. Racial and Ethnic Inequality in Homeownership and Housing Equity. Sociological Quarterly 42(2): 121-149. Fong, E., and M. Gulia. 2000. Neighborhood Change within the Canadian Ethnic Mosaic, 1986­1991. Population Research and Policy Review 19(2): 155-177. Fong, E., and S. Kumiko. 2000. Suburbanization and Home Ownership: The Spatial Assimilation Process in U.S. Metropolitan Areas. Sociological Perspectives 43(1): 137-157. Gabriel, S., and G. Painter. 2003. Pathways to Homeownership: An Analysis of the Residential Location and Homeownership Choices of Black Households in Los Angeles. The Journal of Real Estate Finance and Economics 27(1): 87-109. Galster, G. C. 1987. Residential Segregation and Interracial Economic Disparities: A Simultaneous-Equations Approach. Journal of Urban Economics 21: 22-44. Glaeser, E. L., B. Sacerdote, and J. A. Scheinkman. 1996. Crime and Social Interactions. The Quarterly Journal of Economics 111(2): 507-548. Goodman, A. C., and M. Kawai. 1982. Permanent Income, Hedonic Prices, and Demand for Housing: New Evidence. Journal of Urban Economics 12: 214-237. Gordon, M. M. 1964. Assimilation in American Life: The Role of Race, Religion, and National Origins. Oxford University Press: New York. Green, R. K., and P. H. Hendershott. 2001. Home-Ownership and Unemployment in the U.S. Urban Studies 38(9): 1509-1520. Green, R. K., and M. J. White. 1997. Measuring the Benefits of Homeowning: Effects on Children. Journal of Urban Economics 41(3): 441- 461. Gyourko, J., and P. Linneman. 1996. Analysis of the Changing Influences on Traditional Households' Ownership Patterns. Journal of Urban Economics 39(3): 318-341. Haurin, D. R., R. D. Dietz, and B. A. Weinberg. 2003. The Impact of Neighborhood Homeownership Rates: A Review of the Theoretical and Empirical Literature. Journal of Housing Research 13(2): 119-151. Haurin, D. R., T. L. Parcel, and R. J. Haurin. 2002. Does Homeownership Affect Child Outcomes? Real Estate Economics 30(4): 635-666. Ihlanfeldt, K. R. 1981. An Empirical Investigation of Alternative Approaches to Estimation of the Equilibrium Demand for Housing. Journal of Urban Economics 9: 97-105. Krivo, L. J. 1995. Immigrant Characteristics and Hispanic-Anglo Housing Inequality. Demography 32(4): 599-615. Li, W. 1998. Anatomy of a New Ethnic Settlement: The Chinese Ethnoburb in Los Angeles. Urban Studies 35(3): 479-502. Logan, J. R., R. D. Alba, and W. Zhang. 2002. Immigrant Enclaves and Ethnic Communities in New York and Los Angeles. American Sociological Review 67(2): 299-322. Maddala, G. S. 1983. Limited-Dependent and Qualitative Variables in Econometrics. Econometric Society Monographs in Quantitative Economics. Cambridge University Press: Cambridge. Manski, C. F. 1995. Identification Problems in the Social Sciences. Harvard University Press: Cambridge, Mass. Massey, D. S., and N. A. Denton. 1987. Trends in the Residential Segregation of Blacks, Hispanics, and Asians: 1970-1980. American Sociological Review 52(6): 802-825. 27 Massey, D. S., A. B. Gross, and K. Shibuya. 1994. Migration, Segregation, and the Geographic Concentration of Poverty. American Sociological Review 59(3): 425-446. Myers, D., and S. W. Lee. 1998. Immigrant Trajectories into Homeownership: A Temporal Analysis of Residential Assimilation. International Migration Review 32(Fall): 593-625. Myers, D., I. Megbolugbe, and S. W. Lee. 1998. Cohort Estimation of Homeownership Attainment among Native-Born and Immigrant Populations. Journal of Housing Research 9(2): 237-269. Ng, F. 1998. The Taiwanese Americans. Greenwood Press: Westport: CT. Office of Policy and Planning, S. D. 2002. Annual Report: Legal Immigration, Fiscal Year 2000: Washington D.C., U.S. Department of Justice, Immigration and Naturalization Service. Painter, G. 2000. Tenure Choice with Sample Selection: Differences among Alternative Samples. Journal of Housing Economics 9(3): 197-213. Painter, G., S. A. Gabriel, and D. Myers. 2001. Race, Immigrant Status, and Housing Tenure Choice. Journal of Urban Economics 49(1): 150-167. Painter, G., L. Yang, and Z. Yu. 2003. Heterogeneity in Asian American Homeownership: The Impact of Household Endowments and Immigrant Status. Urban Studies 40(3): 505-530. Portes, A. 1995. Children of Immigrants: Segmented Assimilation and Its Determinants. A. Portes. editor. The Economic Sociology of Immigration. Russell Sage Foundation: New York. Portes, A., and M. Zhou. 1993. The New Second Generation: Segmented Assimilation and Its Variants. Annals of the American Academy of Political and Social Science 530: 74-96. Quigley, J. M., and D. H. Weinberg. 1977. Intra-Urban Residential Mobility: A Review and Synthesis. International Regional Science Review 2: 41-66. Rohe, W. M., S. Van Zandt, and G. McCarthy. 2000. The Social Benefits and Costs of Homeownership: A Critical Assessment of the Research. Working Paper, v. no. 00-01. Research Institute for Housing America: Washington D C. Schoeni, R., K. F. McCarthy, and G. Vernez. 1996. The Mixed Economic Progress of Immigrants: Santa Monica, CA, Rand Corporation, p. 121. Simmons, P. A. 2001. Changes in Minority Homeownership During the 1990s. Fannie Mae Foundation Census Notes 07. Skaburskis, A. 1996. Race and Tenure in Toronto. Urban Studies 33(2): 223-252. Skeldon, R. 1995. Emigration from Hong Kong : Tendencies and Impacts. Hong Kong Series. Chinese University Press: Sha Tin, N. T., Hong Kong. Takaki, R. T. 1994. From Different Shores: Perspectives on Race and Ethnicity in America. Oxford University Press: New York. Theil, H. 1967. Economics and Information Theory. Studies in Mathematical and Managerial Economics ; V. 7. North-Holland Pub. Co.: Amsterdam. Toussaint-Comeau, M., and S. L. W. Rhine. 2000. Ethnic Immigrant Enclaves and Homeownership: A Case Study of an Urban Hispanic Community. Consumer and Community Affairs Policy Studies, v. 2000-6. Federal Reserve Bank of Chicago: Chicago. Tseng, Y.-F. 1995. Beyond "Little Taipei": The Development of Taiwanese Immigrant Businesses in Los Angeles. International Migration Review 29(1): 33-59. ------. 2000. The Mobility of Entrepreneurs and Capital: Taiwanese Capital-Linked Migration. International Migration 38(2): 143-168. 28 U.S. Census Bureau. 2002. The Asian Population: 2000, Census 2000 Brief, U.S. Census Bureau, Demographic Surveys Division. Van de Ven, W. P. M. M., and B. M. S. Van Pragg. 1981. The Demand for Deductibles in Private Health Insurance: A Probit Model with Sample Selection. Journal of Econometrics 17: 229-252. Wachman, A. 1994. Taiwan : National Identity and Democratization. Taiwan in the Modern World. M.E. Sharpe: Armonk, N.Y. Wachter, S. M., and I. Megbolugbe. 1992. Racial and Ethnic Disparities in Homeownership. Housing Policy Debate 3(2): 333-370. Waldinger, R. D., and M. Bozorgmehr. 1996. Ethnic Los Angeles. Russell Sage Foundation: New York. White, M. J. 1987. American Neighborhoods and Residential Differentiation, v. 72. Russell Sage Foundation: New York. White, M. J., A. E. Biddlecom, and S. Guo. 1993. Immigration, Naturalization, and Residential Assimilation among Asian Americans in 1980. Social Forces 72(1): 93-117. Zhou, M. 1992. Chinatown: The Socioeconomic Potential of an Urban Enclave. Temple University Press: Philadelphia. ------. 1997a. Growing up American: The Challenge Confronting Immigrant Children and Children of Immigrants. Annual Review of Sociology 23: 63-95. ------. 1997b. Segmented Assimilation: Issues, Controversies, and Recent Research on the New Second Generation. International Migration Review 31(4): 975-1008. Zhou, M., and J. V. Gatewood. 2000. Contemporary Asian America: A Multidisciplinary Reader. New York University Press: New York. 29 Notes: i References to white refer to persons of white and non-Hispanic origin. ii References to Latino refer to persons of Hispanic origin, who may be of any race. A person is counted as Asian if he/she chose Asian as the race option in the Census 2000. 111 They also find that effect is somewhat mitigated if the household is Latino. iv The sample size is much larger than comparable data available from the American Housing Survey (AHS) or the Current Population Survey (CPS). Because of the relatively small share of Chinese in the total population, the sample size is critical for this analysis, enabling detailed analysis of Chinese subgroups. The PUMS data file provides detailed information about both the housing units and individuals who reside in, which is sufficiently numerous to identify separate marginal effects for each of the groups studied here. In addition, these datasets do not have specific information on migration histories and detailed race categories among Asians. v Taiwanese immigrants are defined as those persons who were born in Taiwan and chose either Chinese or Taiwanese as their race on the U.S. Census form. Tseng (1995) argues that, by relying on birthplace in the 1990 Census, one would underestimate Taiwanese immigrants. She suggests that country of last residence is a better way to define Taiwanese immigrants. Unfortunately, the U.S. Census does not provide such information. In addition, less than 15 percent of all the residents in Taiwan were in-migrants from Mainland China after the Second World War. The vast majority of people who were born in Taiwan were decedents of the "local people" who moved from Mainland China centuries ago (Ng 1998). Thus, very few persons are likely to have been born in Mainland China and emigrated from Taiwan. The characteristics of immigrants who were born in Taiwan should be representative of that of Taiwanese. Moreover, Taiwanese identity is socially constructed and deeply rooted in socioeconomic and political evolution of the island. It is also not immediately clear whether the majority of those who were born in Mainland China and later immigrated to the United States would consider themselves as Taiwanese. vi A typical example is that some Chinese moved from the mainland to Taiwan in 1949 and then immigrated to the United States. In terms of initial household endowments, those persons may be more like the Taiwan 30 subgroup given the income differences between Mainland China and Taiwan. But using place of birth may still include them into the mainland Chinese subgroup. v11 The sample of "Recent Movers" is defined in the Census as those households who moved in the past 5 years. viii This paper uses PUMA as the geographical unit of local housing market. The information regarding the housing price and rent is based on this unit. Housing price is measured as the 25th percentile home price and rent as the median rent in one PUMA. The use of these proxies follows Gyourko and Linneman (1996). ix Results of these household income regressions are available upon request. x Taiwanese have a higher average wage income per worker than those from Mainland China. However, Taiwanese in general have far fewer workers per household than those from the mainland. Therefore, aggregate household income of Taiwanese is lower than those from the mainland. xi We categorized household heads as high English proficiency if their self-assessments of English speaking ability were "very well" or "well". Those with self-assessment results of "not well" or "not at all" are identified as low English proficiency. xii To avoid the misleading values in Chinese homeownership rate due to too few Chinese observations, we require that each PUMA must have at least 100 Chinese household observations to be an individual area, or it will be combined with other PUMAs to meet the standard. xiii Cohort analysis may be employed as an alternative to cumulative attainment in static, cross-sectional samples (Myers, Megbolugbe and Lee 1998). A focus on mover households achieves dynamic analysis of cross-sectional samples by a means different than the cohort method. xiv This modeling procedure has been applied in two recent papers, Painter et. al. (2001) and Painter et. al. (2003). The results are particularly sensitive with respect to the age profile and immigrant length of stay. xv Formally, the log likelihood function that is estimated is the following, yi=1 yi=0 L = ^ ln[02(Xi P, Zi y, p)] +^ ln[02(-Xi P, Zi y, p)] +^ ln[1 - Ot(Zi y)] i&S i&S i£S where S is the set of observations for which OWNi is observed, Mi is the standard cumulative normal and M2 is the cumulative bivariate normal distribution function. The selection model is identified by both functional 31 form assumptions and by the addition of Duncan's Occupational code index not included in the homeownership choice equation. xvi The formula used here is: (percentage of Chinese*ln (percentage of Chinese) + percentage of all other groups* ln (percentage of all other groups))/(-1.386). Theoretically, the index will be zero when there are no Chinese households or when all households are Chinese in an area. The index has its maximum value of 0.5 when the Chinese share of total households is 0.5. Most PUMAs have a small number of Chinese households, usually lower than 10%. The maximum level is about 24% of the total households in the PUMA. Therefore, the Entropy index in our study becomes larger as the share of Chinese households increases. Detailed estimation results using the Entropy index are available upon request. 32 Table 1. Homeownership rates by race and birthplace, Los Angles CMSA, 1980 and 1990. Percentage 1980 1990 Full Sample Movers Only Full Sample Movers Only All 54.5 40.8 56.1 43.0 White 58.2 43.8 61.4 47.6 Black 39.0 22.7 37.3 21.8 Latino 44.2 32.0 43.1 30.1 Asian (all) 52.6 44.9 57.3 49.6 Chinese 61.1 56.6 68.2 64.1 Non-Chinese Asian 50.6 42.3 54.2 45.5 No. of Observations 92,033 56,106 163,657 93,974 Note: The samples are limited to those householders that are aged between 18 and 64. The homeownership rate in one ethnic group is the ratio of homeowners to the total households within that group. Source: 5% Public Use Microdata Samples of the US Census, 1980 and 1990.Table 2. Homeownership rates by race and birthplace, Los Angeles CMSA, 1980-2000. Percentage 1980 1990 2000 White 58.2 61.5 66.1 Chinese 61.4 68.2 64.7 Who were born in Mainland China 68.7 70.7 62.3 Taiwan 58.6 75.0 71.7 Hong Kong/Macau 55.0 63.7 71.4 U.S. 61.0 67.4 60.5 Other places 46.5 51.4 58.4 Note: Householders are aged between 18 and 65. Source: 5% Public Use Microdata Samples of the US Census, 1980, 1990, 2000Table 3. Estimation results for White-Asian sample with all Chinese as one group. Model I Model II Model III Model IV Variable Coeff. t -Stats Coeff. t -Stats Coeff. t -Stats Coeff. t -Stats Ethnicity-Chinese 0.602 ** 17.64 0.714 ** 10.15 0.834 ** 10.52 0.458 ** 6.09 Ethnicity-Other Asians 0.001 0.04 0.024 0.98 0.056 * 2.27 0.025 1.03 Speak English only at home - - 0.256 ** 5.83 0.246 ** 5.59 0.258 ** 5.85 Interaction with Chinese - - -0.355 ** -3.33 -0.454 ** -4.17 -0.507 ** -4.60 English good but not the only language at home - - 0.282 ** 7.09 0.275 ** 6.89 0.282 ** 7.06 Interaction with Chinese - - -0.044 -0.55 -0.073 -0.92 -0.127 -1.58 Homeownership rate differences - - - - - - -0.637 ** -8.96 Interaction with Chinese - - - - - - 3.459 ** 9.54 Chinese concentration by PUMA - - - - -2.095 ** -11.54 - - Interaction with Chinese - - - - 0.943 ** 3.10 - - Correlation Coefficient (rho) 0.503 ** 126.3 0.503 ** 120.2 0.202 ** 11.51 0.203 ** 11.26 Log Likelihood -107,995 -107,956 -107,833 -107,846 Pseudo-R2 0.196 0.196 0.197 0.197 Number of Observations 124,626 124,626 124,626 124,626 *: significant at 5% confidence level; **: significant at 1% confidence level Note: Each model specification also includes certain control variables, such as household characteristics (age, education, income, marital status, family structure, etc.), housing market conditions (housing price and market rent), migration and immigration status, which are discussed in the paper. Pseudo-R2 is calculated by the forumla 1 - L1/L0, where L1 = the estimated log-likelihood in the estimated equation and L0 = n(nm)*ln(p(nm))+n(m,o)ln(p(m,o))+n(m,r)ln(p(m,r)) (Maddala, 1983). Here, n(nm) = number of non-movers p(nm) = percentage of non-movers in the full sample n(m,o) = number of owners in the movers sample p(m,o) = percentage of owners in the movers sample times percentage of movers in the full sample n(m,r) = number of renters in the movers sample p(m,r) = percentage of renters in the movers sample times percentage of movers in the full sampleTable 4. Estimation results for White-Asian sample with different chinese subgroups. Model I Model II Model III Model IV Variable Coeff. t -Stats Coeff. t -Stats Coeff. t -Stats Coeff. t -Stats Subgroup-Mainland 0.542 ** 9.17 0.713 ** 7.03 0.697 ** 5.67 0.575 ** 5.46 Subgroup-Taiwan 1.042 ** 16.63 1.429 ** 11.12 1.691 ** 11.49 0.860 ** 5.53 Subgroup-Hong Kong/Macau 0.525 ** 5.30 0.437 1.36 0.301 0.85 0.395 1.21 Subgroup-U.S. born 0.372 ** 4.29 0.052 0.07 0.274 0.35 -0.115 -0.16 Subgroup-Other 0.273 ** 4.50 0.118 1.05 0.280 2.13 -0.092 -0.77 Ethnicity-Other Asians 0.007 0.23 0.024 0.96 0.056 2.24 0.021 0.86 Speaking English only at home - - 0.251 ** 5.70 0.239 ** 5.41 0.253 ** 5.87 Interaction-Mainland - - -0.260 -0.62 -0.276 -0.65 -0.329 -0.81 Interaction-Taiwan - - -1.194 ** -4.53 -1.376 ** -5.09 -1.190 ** -4.62 Interaction-Hong Kong/Macau - - -0.472 -0.99 -0.393 -0.81 -0.476 -1.02 Interaction-U.S. born - - 0.376 0.53 0.200 0.25 0.254 0.34 Interaction-Other - - 0.086 0.37 -0.012 -0.05 -0.084 -0.35 English good but not the only language at home - - 0.281 ** 7.03 0.272 ** 6.79 0.280 ** 7.14 Interaction-Mainland - - -0.167 -1.37 -0.169 -1.36 -0.239 -1.90 Interaction-Taiwan - - -0.443 ** -3.07 -0.485 ** -3.28 -0.444 ** -2.99 Interaction-Hong Kong/Macau - - 0.140 0.41 0.160 0.47 0.109 0.33 Interaction-U.S. born - - 0.213 0.29 0.030 0.04 0.142 0.19 Interaction-Other - - 0.326 2.46 0.282 2.08 0.234 1.73 Homeownership rate differences - - - - - - -0.600 ** -8.54 Interaction-Mainland - - - - - - 1.988 ** 3.03 Interaction-Taiwan - - - - - - 5.202 ** 6.65 Interaction-Hong Kong/Macau - - - - - - 0.476 0.44 Interaction-U.S. born - - - - - - 2.236 * 2.14 Interaction-Other - - - - - - 3.284 ** 4.70 Chinese concentration by PMUA - - - - -2.069 ** -11.42 - - Interaction-Mainland - - - - 1.868 ** 3.92 - - Interaction-Taiwan - - - - -0.324 -0.59 - - Interaction-Hong Kong/Macau - - - - 2.564 ** 3.41 - - Interaction-U.S. born - - - - 0.357 0.30 - - Interaction-Other - - - - 0.676 1.37 - - Correlation Coefficient (rho) 0.504 ** 113.3 0.289 ** 130.0 0.201 ** 11.91 0.201 ** 11.79 Log Likelihood -107,944 -107,882 -107,767 -107,791 Pseudo-RA2 0.196 0.197 0.197 0.197 Number of Observations 124,626 124,626 124,626 124,626 *: significant at 5% confidence level; **: significant at 1% confidence level Note: See note on Table 3 regarding control variables. Table 5. Estimation results for different education, income and immigrant groups. College Degree or Higher High School Diploma or Less Variable Coeff. t -Stats Coeff. t -Stats Ethnicity-Chinese 0.750 ** 5.01 0.377 ** 4.23 Ethnicity-Other Asians 0.070 * 2.14 0.037 0.99 Speak English only at home 0.250 ** 3.51 0.297 ** 5.20 Interaction with Chinese -0.75 ** -4.45 -0.463 * -2.30 English good but not the only language at home 0.214 ** 3.17 0.383 ** 7.68 Interaction with Chinese -0.310 * -2.10 -0.109 -1.03 Homeownership rate differences -0.600 ** -5.62 -0.628 ** -6.57 Interaction with Chinese 3.195 ** 6.40 3.675 ** 6.90 High Income Low Income Variable Coeff. t -Stats Coeff. t -Stats Ethnicity-Chinese 0.635 ** 3.95 0.384 ** 4.52 Ethnicity-Other Asians 0.153 ** 3.96 -0.060 -1.86 Speak English only at home 0.257 ** 2.98 0.170 ** 3.31 Interaction with Chinese -0.370 -1.71 -0.590 ** -4.39 English good but not the only language at home 0.192 * 2.33 0.260 ** 5.70 Interaction with Chinese -0.172 -1.02 -0.107 -1.16 Homeownership rate differences -0.630 ** -5.59 -0.596 ** -6.63 Interaction with Chinese 2.559 ** 3.72 3.647 ** 8.59 New Immigrants (<= 10 yrs) Old Immigrants (> 10 yrs) Variable Coeff. t -Stats Coeff. t -Stats Ethnicity-Chinese 0.423 ** 4.39 0.435 ** 2.57 Ethnicity-Other Asians -0.170 ** -3.63 0.046 0.98 Speak English only at home 0.149 1.90 0.248 * 2.52 Interaction with Chinese -0.879 ** -2.94 -0.274 -1.16 English good but not the only language at home 0.221 ** 4.21 0.377 ** 4.27 Interaction with Chinese -0.087 -0.92 -0.090 -0.52 Homeownership rate differences 0.213 0.84 -0.022 -0.09 Interaction with Chinese 3.473 ** 6.62 1.385 * 2.22 *: significant at 5% confidence level; **: significant at 1% confidence level Note: See note on Table 3 regarding control variables. Figure 1. Permanent household income by race and birthplace, Los Angeles CMSA, 1980-1990. 70 T 60 - ______ 50 - ______ _____ ______ ___ 40 - -------- -------- 30 - 20 - 10 - White Chinese born in Taiwan □ 1980 Mainland China Hong Kong and Macau □ 1990 U.S. Other places 0 Note: The vertical axis shows the mean value of permanent income in 1000s. All dollar figures are in 1989 dollars. Source: 5% Public Use Microdata Samples of the US Census, 1980, 1990.1980 1990 80,000 i 80,000 Figure 2. Foreign-born Chinese by immigrant status and birthplace, Los Angeles CMSA, 1980-1990. Taiwan □ Over 20 Years □ 5-10 Years Mainland Hong Kong Other and Macau places □ 10-20 Years □ Arrived in Last 5 Years 60,000 40,000 20,000 Taiwan Mainland Hong Other Kong and places Macau □ Over 20 Years □ 5-10 Years □ 10-20 Years □ Arrived in Last 5 Years Note: Chinese immigrants from other places refer to foreign-born Chinese who were not born in Taiwan, mainland China, Hong Kong, or Macau. 0 Source: 5% Public Use Microdata Samples of the US Census, 1980, 1990.Appendix 1. Variable Summary Statistics in Full and Movers-Only Sample in the Los Angeles CMSA LA Full Sample LA Movers Sample Variable Mean Std Dev. Mean Std Dev. Ownership Rate 0.611 0.487 0.479 0.500 Age 18-24 0.047 0.211 0.074 0.262 Age 25-34 0.261 0.439 0.372 0.483 Age 35-44 0.292 0.454 0.302 0.459 Age 45-54 0.219 0.414 0.159 0.366 Age 55-64 0.182 0.386 0.093 0.290 Not Married, Male Head Of Household 0.191 0.393 0.226 0.419 Not Married, Female Head 0.217 0.412 0.226 0.418 No High School Diploma 0.093 0.290 0.090 0.286 High School Dip. W/ College 0.445 0.497 0.442 0.497 College Degree or Better 0.462 0.499 0.468 0.499 Number Of People In Household 2.782 1.484 2.736 1.486 Number Of Workers In Household 1.723 0.903 1.675 0.839 Permanent Income (1000s) 57.82 22.57 54.28 21.87 Transitory Income (1000s) 0.000 39.80 -0.357 38.20 Dividend Income (1000s) 2.782 9.430 1.916 7.677 The 25th Percentile Housing Price (log) 12.09 0.444 12.07 0.458 Puma Median Rent (log) 6.499 0.202 6.494 0.209 Ethnicity-White 0.875 0.331 0.858 0.349 Ethnicity-Chinese 0.032 0.176 0.037 0.189 Ethnicity-Mainland 0.010 0.099 0.010 0.098 Ethnicity-Taiwan 0.008 0.091 0.011 0.105 Ethnicity-Hongkong and Macau 0.003 0.055 0.004 0.062 Ethnicity-Native 0.004 0.065 0.004 0.062 Ethnicity-Other Chinese 0.007 0.082 0.008 0.092 Ethnicity-Other Asian 0.093 0.291 0.105 0.306 Speaking English only at home 0.816 0.387 0.793 0.405 English good but not only language at home 0.154 0.361 0.170 0.375 Homeownership Rate Difference 0.121 0.086 0.118 0.088 Chinese concentration by PUMA 0.025 0.044 0.024 0.044 Moved From Within Same State(s) 0.051 0.221 0.089 0.284 Moved From Within U.S 0.087 0.282 0.151 0.358 Moved From A Foreign Country 0.034 0.181 0.059 0.235 Immigrant 0.185 0.388 0.207 0.405 Came To U.S. In The Past 5 Yrs. 0.033 0.178 0.054 0.226 Came To U.S 5-10 Years Ago 0.042 0.202 0.057 0.231 Came To U.S 10-15 Years Ago 0.037 0.189 0.041 0.199 Came To U.S 15-20 Years Ago 0.020 0.140 0.019 0.136 Came To U.S 20-30 Years Ago 0.029 0.168 0.022 0.148 Came To U.S More Than 30 Years Ago 0.023 0.149 0.014 0.115 Number of Observations 124,626 72,061Appendix 2. Cities and counties in each area. Group Cities or Counties 1 Burbank,Santa Clarita,Lancaster, Palmdale,North Hollywood,Semi Valley 2 Glendale,Pasadena,La Canada 3 Calabasas,Malibu,Santa Monica,Brentwood 4 Van Nuys,Northbridge,Encino 5 Pomona,Azusa,Baldwin Park,Irwindale,La Verne,Claremont 6 Covina,West Covina,City of Industry 7 Diamond Bar,La Habra Heights 8 Whittier 9 El Monte,Arcadia,San Marino,San Gabriel,Temple City 10 Montery Park,Rosemead 11 Alhambra,South Pasadena 12 Eagle Rock 13 Beverly Hills,Hollywood 14 Downtown LA,Westlake 15 East LA,Vernon,South Gate,Lynwood,Compton,Downey,Gardena,Hawthorn 16 Venice,Westwood 17 Torrance,El Segundo--Manhatton--Redondo Beaches,Palos Verdes--Rolling Hill Estates 18 Carson,Lakewood,Bell Flower,Harbor City, Long Beach 19 Santa Fe Springs,La Mirada,Cerritos,Artesia 20 Ventura County 21 Riverside County 22 San Bernadino County 23 Santa Ana,Westminster,Garden Grove 24 Laguna Beach,Laguna Niguel,San Juan Capistrano, La Habra,Yoba Linda 25 Irvine city, Tustin City 26 Buena Park, Huntington Beach,Newport Beach,Costa Mesa 27 Anaheim,Fullerton Note: The designation of these spatial areas is based on PUMA unit from the Census PUMS data. When one PUMA does not have enough Chinese household observations, spatially adjacent PUMAs with similar social and economic characteristics are be grouped together.Appendix 3. Estimation results for White-Asian sample with all Chinese as one group by metropolitan areas. Variable San Francisco CMSA New York CMSA Washington DC MSA Coeff. t -Stats Coeff. t -Stats Coeff. t -Stats Ethnicity-Chinese 0.905 ** 10.17 0.727 ** 9.86 1.280 ** 5.26 Ethnicity-Other Asians 0.118 ** 3.29 0.084 * 2.46 0.413 ** 5.99 Speak English only at home 0.245 ** 3.43 0.219 ** 4.94 -0.052 -0.50 Interaction with Chinese -0.409 ** -3.72 -0.426 ** -3.16 -0.823 ** -2.64 English good but not the only language at home 0.268 ** 4.01 0.131 ** 3.15 -0.134 -1.33 Interaction with Chinese -0.014 -0.15 -0.032 -0.38 -0.483 -1.85 Correlation Coefficient (rho) 0.260 ** 6.60 0.480 ** 8.80 0.110 0.70 Log Likelihood -50,656 -109,669 -27,245 Number of Observations 59,705 146,306 32,746 *: significant at 5% confidence level; **: significant at 1% confidence level Note: Each model specification also includes certain control variables, such as household characteristics (age, education, income, marital status, family structure, etc.), housing market conditions (housing price and market rent), migration and immigration status, which are discussed in the paper.Appendix 4. Estimation results for White-Asian sample with different Chinese subgroups by metropolitan areas. San Francisco CMSA New York CMSA Variable Coeff. t -Stats Coeff. t -Stats Subgroup-Mainland 1.014 ** 10.37 0.654 ** 7.80 Subgroup-Taiwan 1.641 ** 6.56 1.641 ** 8.44 Subgroup-Hong Kong/Macau 1.094 ** 4.81 0.668 * 2.49 Subgroup-U.S. born 1.533 ** 3.43 -0.002 0.00 Subgroup-Other 0.051 0.30 0.330 1.74 Ethnicity-Other Asians 0.110 ** 3.03 0.092 ** 2.67 Speaking English only at home 0.266 ** 3.68 0.214 ** 4.80 Interaction-Mainland -0.845 ** -3.23 -0.661 -1.41 Interaction-Taiwan -1.338 ** -3.22 -2.051 ** -4.05 Interaction-Hong Kong/Macau -0.925 ** -2.81 -0.182 -0.34 Interaction-U.S. born -0.950 * -2.09 0.254 0.49 Interaction-Other -0.047 -0.17 0.178 0.50 English good but not the only language at home 0.278 ** 4.15 0.132 ** 3.17 Interaction-Mainland -0.223 -1.82 0.000 0.00 Interaction-Taiwan -0.411 -1.55 -0.662 ** -3.13 Interaction-Hong Kong/Macau -0.151 -0.61 0.088 0.30 Interaction-U.S. born -0.747 -1.62 0.276 0.53 Interaction-Other 0.527 ** 2.74 0.133 0.60 Correlation Coefficient (rho) 0.261 ** 6.64 0.435 ** 8.29 Log Likelihood -50,581 -109,625 Number of Observations 59,705 146,306 *: significant at 5% confidence level; **: significant at 1% confidence level Note: See note on Appendix 3 regarding control variables.