*Names in bold indicate Presenter
A likely reason why Asian American borrowers have received less research and regulatory attention is the belief that members of this group are, on average, more economically successful than other minority groups in the US. Earlier studies of discrimination in lending, for example, rarely found any instance where Asian Americans would be denied mortgage loans more often than non-Hispanic whites. As credit distributions for Asian Americans mirror those of non-Hispanic whites, not much evidence of pricing differentials appeared.
A reliance on mean outcomes, however, does not reflect the heterogeneity of Asian American experiences. Because of differences across key factors such as time since immigration and country of origin, many believe that the Asian American economic distribution is bi-modal, with clusters at the high- and low-ends. The positive experiences of a relatively small number of elite Asians bias average statistics upwards, whereas the bottom of the distribution suffers disadvantages comparable to the similarly situated members of other minority groups. Some advocates have claimed that the focus on mean outcomes, as well as social constructions such as the model minority myth, have led to Asians having an overstated perception of success and a resultant lack of social service targeting.
In this paper, we examine the experiences of Asians in US mortgage markets. We test for differences in the price level (as measured by mortgage annual percentage rates) and approval rates compared with other racial and ethnic groups, while controlling for key economic factors such as credit worthiness or type of loan product. As well, we analyze differences based on the racial composition of borrowers’ neighborhoods. We pay particular attention to the question of whether outcomes differ at different points in the distribution of various economic measures, using matching and quantile regression methods. The primary data source for this paper is a unique sample of private lender mortgage records, and we supplement these data with data available from HMDA and Census.
Full Paper: