Study Shows Apparent Disparities in Lending to Women


new study has revealed some apparent differences in the success rate of female and
male mortgage applicants

The Woodstock Institute, a non-profit research organization, conducted a
study based on Home Mortgage Disclosure Act (HMDA) data in the six county Chicago
region and found that applications for either purchasing or refinancing a home
were less likely to be approved if submitted by a female.  The same was true of applications submitted
by a female with a male cosigner when compared to applications from a male
applicant with a female cosigner.

would expect to see no significant difference in the origination rates for
male-headed joint applications and female-headed joint applications, since the
backgrounds of both borrowers on joint applications are considered by mortgage
lenders,” said Spencer Cowan, vice president of research at Woodstock
Institute. “The fact that there are such large disparities raises troubling
questions about potentially discriminatory underwriting.”

Institute says that the wealth gap between men and women is well documented;
one study showed that single women between the ages of 18 and 64 have a median
wealth that is 49 percent of that for single men and 12 percent of the median
for couples.  The differences for women
of color are even greater.  That same
study suggests that the wealth gap would persist even if the gap in income
between men and women and between whites and persons of color did not exist. 

homeownership is the most common means of wealth creation, women are
disadvantaged here as well.   They are
three times as likely to be custodial parents with greater expenses and less
opportunity to save.  That and their
lower income means they are less likely to have a significant down payment,
income to qualify for more valuable properties (and thus greater appreciation)
and more likely to have received subprime loans during the housing bubble, and
this made sustainable homeownership more difficult and, by increasing the cost
of homeownership, reduced the wealth-building effects.

study examines women’s access to mortgages following the housing collapse to
see if they now have substantially equal access to mortgages in the absence of
the subprime products that had been disproportionately marketed to them.  The Institute examined data on applications both
those originated and denied for conventional, FHA and VA loans between $20,000
and $800,000 intended as first liens on one-to-four family units.  Applicants had incomes between $20,000 and
$999,000.  The study included 58,870
applications for home purchases and 198,522 applications for refinancing.

for loan-to-value ratio, the Institute found that female mortgage applicants
were about 8 percent less likely overall to
have home purchase mortgages
originated than were male
applicants.  Those disparities, however, may reflect the influence of demographic characteristics which vary systematically between
female and male applicants.  For example, 77 percent of female applicants for
purchase mortgages had no co-applicant versus only 52
percent of male applicants; 71 percent of female
for refinance mortgages had no co-applicant versus
32 percent for male
The more likely presence of children in single-women headed households may also put female applicants with no
co-applicants in a significantly different family and financial situation
male applicants with no co-applicants.

To control for those potentially confounding demographic factors, the
study ran an analysis on a subset of the data
consisting of joint applications with a male applicant with a female co-applicant or a female
applicant with a male
In this subset, there were
21,432 purchase applications
and 111,304 refinance
applications.  Female-headed
joint applications
overall were 24
percent less likely to have
purchase mortgages originated, and 39
percent less likely to
have refinance mortgages originated, than were male-headed joint applications. The
disparities applied across all racial categories with the largest found in African American female- headed
joint applications which were
34 percent less
likely to
be originated than were African American male-headed joint applications for
purchase and 44 percent less
likely for refinancing. Latino female-headed
joint applications
were 19 percent less
likely to
have purchase mortgages
originated and 29
percent less
likely to
have refinance mortgages
than were
joint applications.

The Institute
references a Department of Housing and Urban Development disparate impact analysis which found that female applicants for mortgages,
purchase and refinance, are significantly less likely to
have loans originated than are male applicants.  Even when
analysis is limited to
with either a female
and male co-applicant or male applicant with a
female co-applicant, reducing the probability that the applicant is
a single parent, the disparity remains and appliances among all racial and
ethnic categories. 
The Institute says such a consistent set of findings
showing a significant disparity in women’s
access to mortgage
credit is troubling and
warrants further investigation.

The Woodstock Institute study admits a
in that the HMDA data does not contain key categories of information such
as credit scores, overall debt-to-income ratio or appraised values and says
without those data points, it is not possible
to determine
whether female-headed joint applications
are denied while male-headed joint applications with
similar qualifications are approved.  It
urges the Consumer Financial
Protection Bureau
(CFPB) to expeditiously finalize enhancements it is permitted under Dodd-Frank to make to the
HMDA data to include at a minimum the applicant’s credit score,
debt-to-income ratio, and the appraised value of the property. 
That additional data could show more
clearly whether there was disparate impact or some legally defensible justification for the disparities shown in its
study.  “If disparities for women persist with those additional data, additional fair
lending training,
and enforcement will be needed,”
the study concludes.

Leave a Reply