Database Reference
In-Depth Information
Recent mortgage pricing studies consider gender and racial discrimination in con-
sumer credit (Edelberg, 2007), such as credit cards and education loans, in private
firm credit (Albareto & Mistrulli, 2011; Blanchard et al., 2008; Blanchflower et al.,
2003; Cavalluzzo et al., 2002; Muravyev et al., 2009), in subprime home loans (Bo-
cian et al., 2008; Reid & Laderman, 2009), in household credit (Weller, 2008). Using
survey data, (P. Cheng et al., 2009) found that women pay higher rates because they
do not search for best-rate loans as much as men do.
Mortgage default
studies adopt the percentage of mortgage defaults as a mea-
sure of discrimination. Intuitively, if different default rates are observed for equally
creditworthy groups that differ in some discrimination ground, this is considered
prima facie
evidence of discrimination. Recent contributions on the subject include
(C. L. Brown & Simpson, 2010; S. Chan et al., 2010; Yezer, 2010). A discussion
of the limitations of data on mortgage defaults, including unobserved variables and
sample-selection bias, can be found in (Turner & Skidmore, 1999, Chapter 5).
Discrimination in mortgage rejection and pricing has often occurred indirectly,
through the practice of
redlining
(Hillier, 2003),(Turner & Skidmore, 1999, Chapter
4), which consists of denying credit or of applying higher interest rates to people
living in some specific neighborhood. The use of geographic attributes may hide
(intentionally or not) the fact that such a neighborhood is populated mainly by peo-
ple of a specific race or minority. US cities, in particular, show a very high racial
divide. The percentage of individuals of a protected group in a neighborhood is
often used as a measure of the level of segregation (James & Tauber, 1985; Rear-
don & Firebaugh, 2002). Empirical works combine HMDA data with census data
(Silverman, 2005; E. Blank et al., 2005; Blackburn & Vermilyea, 2007; Ding et
al., 2008; Ezeala-Harrison et al., 2008; Wyly et al., 2008; Rugh & Massey, 2010;
Squires et al., 2009; Vicki et al., 2009; G. A. Dymski et al., 2011) to test for such a
form of indirect discrimination. As an alternative, (Campbell et al., 2008) use pro-
prietary data on unsecured debt. Other studies on redlining use house market data
(Aalbers, 2007; Ezeala-Harrison et al., 2008), consumer credit card data (Brevoort,
2011; Cohen-Cole, 2009), and insurance data (Ong & Stoll, 2007; Ross & Tootell,
2004).
Finally, in the related context of consumer markets,
price discrimination
is the
practice of a retailer, wholesaler, or manufacturer of selling the same product, with
the same marginal cost, at different prices based on buyers' willingness to pay
(Armstrong, 2006). Differential pricing discriminating racial minorities has been
observed in the car sales market (Ayres, 1995; Ayres & Siegelman, 1995; Goldberg,
1996).
6.6
Knowledge Discovery Perspective
The issue of discrimination analysis has been considered from a knowledge discov-
ery, also known as data mining, perspective along two directions: discrimination
discovery and prevention.
Search WWH ::
Custom Search