Dueling views on home lending discrimination

WASHINGTON -- July 28, 2006 -- A consumer advocacy group, after analyzing loan information provided to the government, says there is clear evidence of discrimination against black and Hispanic borrowers.

A law firm that represents mortgage lenders, using much of the same data, says its study proves that there is no evidence of discrimination in lending.

The lawyers call the consumer advocates' views "skewed," while the advocates think the lawyers' findings are "limited."

The studies
A May report by the Center for Responsible Lending, based in part on information from the Home Mortgage Disclosure Act, said when it comes to most types of high-priced home loans, usually called subprime loans, black and Hispanic borrowers were more than 30 percent more likely to receive expensive subprime home loans than their white peers with similar finances.

CRL spokeswoman Sharon Reuss thinks a leading cause might be yield-spread premiums, which are monetary incentives paid to mortgage brokers who get increased rates on subprime loans.

"It could be that borrowers of color are relying on brokers who, because of yield-spread premiums, are delivering more expensive loans than what the borrowers qualify for," Reuss says. "Perhaps more expensive subprime lenders are marketing to African-American and Latino communities while these same communities are being insufficiently served by lower-cost lenders."

The day after the Center for Lending announced its findings, Traiger & Hinckley LLP, which represents lenders, ignited a controversy when it reported that the new 2005 loan information from the same law, the Home Mortgage Disclosure Act, shows that the 172 percent increase in the amount of high-priced loans, during 2004 to 2005, has more to do with borrowers becoming greater credit risks than with racial discrimination.

First, the firm claims, borrowers are taking out second mortgages on their homes, reducing value and increasing their chances of defaulting. And second, more borrowers are using a greater amount of their incomes to repay their mortgage debts, which makes the loans riskier.

It also reported that while the volume of high-priced loans increased, the average amounts that home buyers of different races and ethnicities received showed only small differences from 2004 to 2005, and that lenders had increased their services to minorities.

The firm says that a substantial effort is being made to serve minority home buyers.

"There are not only more loans to minorities but a greater proportion of lending to minorities. We shouldn't lose sight that lending to minorities has increased," says Warren Traiger, partner at Traiger & Hinckley LLP.

Both studies analyzed pricing information from the Home Mortgage Disclosure Act. This law requires most metropolitan mortgage lenders to gather information about their lending activities and submit the data to the government. Lenders are also instructed to provide information about prices that exceed certain thresholds. These are high-priced loans.

The methodology
Traiger & Hinckley, in its study, "Fair Lending Indications of the 2005 Home Mortgage Disclosure Act Data," looked at 2005 loan information from the Home Mortgage Disclosure Act, or HMDA, while the Center for Responsible Lending looked at the information from 2004.

More specifically, the law firm focused on the high-priced loans from the HMDA, looking at 1.6 million mortgages from a list of the top 10 lenders, according to American Banker.

The Center for Responsible Lending looked at 50,000 mortgages in its study called "Unfair Lending: The Effect of Race and Ethnicity on the Price of Subprime Mortgages." In addition to using the HMDA data, the center filled in gaps of missing loan information by using another loan resource, the LoanPerformance subprime asset-backed securities database, which the study says accounts for 87 percent of the U.S. subprime loan originations in 2004. Using these two sources they were able to get the borrower, loan and property characteristics.

Researchers also utilized public information from the Federal Reserve, the Office of Federal Housing Enterprise Oversight and the Census Bureau.

The Center for Responsible Lending calls its study a "first of its kind," saying that by merging the two sources they were able to control the variables of credit scores, loan-to-value ratios and other underwriting factors. They used a statistical technique called "regression analysis" to isolate the impact of race and ethnicity on loan pricing from the effects of borrower risk and economic conditions.

The opponents duel
Traiger says he thinks the report from the Center for Responsible Lending is a useful study. "Well done, given inherent limitations. Normally you would take a sample of all HMDA lending. They weren't able to do that," he says.

He says the report is a little skewed because the researchers only looked at loans that security issuers deemed "subprime."

"It's not based on a random sample," he says. "The Fed went through the HMDA data and identified 200 lenders and instructed regulators to look to see if there was discrimination. The CRL study is moot because the federal regulators are on this issue."

Kevin Mukri, spokesman for the Office of Comptroller of the Currency, says these federal examinations are part of the annual process.

"Banks and lenders provide HMDA data," he says. "The data goes to the Federal Reserve. They sort the numbers and send the appropriate findings to the appropriate bank regulators. The Office for the Comptroller of the Currency, for instance, will conduct our own review of the aggregate numbers from HMDA. We look into what the numbers would indicate. We don't have a time limit for our review, but we'll do whatever's necessary."

Reuss says the Traiger & Hinckley study is very limited in scope and did not control any variables.

"Comparing apples to oranges. The Traiger & Hinckley and Center for Responsible Lending studies each asked different questions, and Center for Responsible Lending looked at more data, which allowed Center for Responsible Lending to control for credit risk factors to isolate the impact of race and ethnicity on the probability of being in the higher-rate pool," says Reuss.

"The Traiger & Hinckley study looked only at HMDA data from 10 large lenders, which did not permit them to control for traditional credit-risk price factors such as FICO scores and specific loan-to-value.

"The Center for Responsible Lending study used regression analysis to control for credit-risk characteristics, and also controlled for geography. Traiger & Hinckley did not control for any variables."

Traiger says consumer advocacy groups should "hold their fire, pending the results of the examinations undertaken by the banking regulators as a result of the Federal Reserve's review of the 2004 HMDA data.

"Only these examinations, which review the actual loan file, including a borrower's credit history and loan details, can definitively establish if discrimination took place," he says.

The purpose
The Center for Responsible Lending's study was intended to influence lenders and lawmakers. It called on them to create pricing standards, make lenders and brokers responsible for providing loans that are right for consumers, expanding HMDA's disclosure requirements, providing enough resources to fully enforce the fair lending laws and creating policy incentives that lead the market to better serve blacks and Hispanics.

Traiger says his firm regularly analyzes mortgage lending data for its clients and that the report was undertaken to make a contribution to the public debate about what the HMDA data means.