Since the release last fall of the Home Mortgage Data Act figures that showed apparent bias against minorities in mortgage lending, a central aspect of the ensuing debate has been whether the Federal National Mortgage Association and the Federal Home Loan Mortgage Corporation are to blame because of their underwriting guidelines. Ann B. Schnare, senior vice president and director of ICF Inc. and manager of the firm's Housing and Real Estate Group, commented on a paper presented at Fannie Mae's recent annual housing conferences that examined the issue. The paper was presented by Glenn B. Canner, senior economist in the Division of Research and Economics at the Federal Reserve Board, and Stuart A. Gabriel, associate professor of finance and business economics at Stanford University. Following are excerpts of her comments. Additional excerpts will appear in the June 15 issue of The Mortgage Marketplace.
Canner and Gabriel document what community groups have been claiming for a relatively long time - namely, that loan applications are less likely to be accepted if the applicant is a member of a minority group or if the property is located in a minority neighborhood. While the data do not make it possible to control for a number of important factors - most notably, the applicant's wealth and credit rating - the findings are, at least, highly provocative.
Documenting the extent to which the secondary market contributes to these outcomes is more problematic. Portfolio lenders are rapidly becoming a thing of the past. Since the majority of lenders now sell their mortgages to the secondary market, either at origination or after holding them for some period of time, they should theoretically be willing to originate any loan that Freddie Mac or Fannie Mae is willing to buy. If you followed this line of reasoning to its extreme, you could conclude that the policies of the secondary market are largely responsible for the patterns revealed by the HMDA data.
However, the reality is not so simple. Since the secondary market does not originate loans, Freddie Mac and Fannie Mae can only buy the product that lenders have to sell. If lenders avoid or discriminate against certain groups or types of neighborhoods - for whatever reason and no matter how irrational discrimination may appear from a profit-maximization point of view - the dearth of such loans will ultimately affect the composition of Freddie Mac's and Fannie Mae's purchases. Separating out the impact of the primary and secondary market is thus exceedingly difficult and may be, to a large degree, an example of the "chicken or the egg" debate.
Data presented in the Canner-Gabriel study shed relatively little light on this issue. The distribution of conventional originations by primary lenders is compared with the distribution of purchases by Fannie Mae and Freddie Mac in table 1. In particular, the proportion of loan originations and sales that involved low- and moderate-income borrowers or neighborhoods, where such borrowers and neighborhoods have been further stratified by ethnicity and race, are compared.
Some lenders have claimed that they are forced to portfolio their CRA loans rather than sell them to an unreceptive secondary market. While the data lend some support to this hypothesis, the evidence is not compelling. For example, 14% of all conventional originations are made to low- and moderate-income borrowers; however, such loans accounts for only about 12% of all Freddie Mac and Fannie Mae purchases. Likewise, 11% of all conventional originations went to lower income and moderate-income neighborhoods, compared with 9% for Fannie Mae and about 12% for Freddie Mac.
Thus, while there are differences in the distribution of originations and sales, the differences are relatively small, a pattern that is at least consistent with the notion that Freddie Mac and Fannie Mae tend to buy what lenders have to sell. However, such statistics will mask the role of the secondary market if lenders avoid making loans to minority borrowers or neighborhood because they believe they cannot be sold. Unfortunately, it is difficult to accept or reject this hypothesis based on the data presented in table 1. ... The distributions of loans in the primary and secondary markets are mirror images of each other, making it difficult to distinguish between cause and effect.
As an aside, however, it is interesting to note that Freddie Mac appears to do as well or marginally better than Fannie Mae with respect to its purchases of low- and moderate-income loans. I cite this finding not to embarrass our host, but rather, to highlight the importance of mainstream lending operations in community lending initiatives.
Most observers would agree that to date. Fannie Mae has been significantly more active than Freddie Mac in its affordable housing initiatives. Yet the impact of such special programs does not show up in the aggregate data. Apparently, the volume of lending that has been generated through such activities is dwarfed by the volume of lending that occurs through the regular operations of the primary and secondary markets.