What do the federal bureaucrats who run Fannie Mae and Freddie Mac know that private mortgage lenders don’t?

Private lenders apparently believe that when homeowners fall behind on their mortgage payments, sometimes their loan principal should be reduced. In the third quarter of 2011, 18% of homeowners who received a modification on a loan held in a bank’s portfolio saw their principal cut.

Banks are not doing charity work here. They’ve run the numbers, and they’ve concluded that under some circumstances, what makes the most sense for their own bottom lines is to grant a principal reduction.

As Columbia Business School professor Christopher Mayer said, referring to banks granting principal reductions, “I believe they like profit.”

Yes, principal forgiveness will encourage a certain percentage of borrowers to stop paying their mortgages in the hope that they, too, will get a break from their lender.

But the borrowers who get their principal reduced will be more likely to keep paying, especially if they are close to building equity in their homes again. And lenders earn an awful lot more money from a mortgage that gets paid off than they do from a foreclosure sale.

"Every time I talk to someone in the private sector, they say principal reductions are incredibly effective," Mayer told me.

The Federal Housing Finance Agency, which has been making key business decisions at Fannie and Freddie since they became wards of the state in 2008, takes a different view.

"We have been through the analytics of the underwater borrowers at Fannie and Freddie," FHFA head Edward DeMarco said during a congressional hearing in November, "and we have concluded that the use of a principal reduction within the context of a loan modification is not going to be the least-cost approach."

That explanation did not satisfy a group of House Democrats who have been calling for Fannie and Freddie to start granting principal reductions. So they pressed the FHFA to make public its analysis of the relevant data.

Last month, the agency sent to Capitol Hill a six-page document that comprises the FHFA staff’s most recent analysis of the issue. The document paints a far less definitive picture about the agency’s findings than DeMarco portrayed in his Capitol Hill testimony.

The analysis, which has gotten far less attention than it deserves, compares the expected taxpayer losses on hypothetical borrowers who would receive principal forgiveness modifications versus those who get principal forbearance.

(Under principal forbearance modifications, which Fannie and Freddie currently use, borrowers see a reduction in their monthly payments, but when they eventually pay off the loan, they are still responsible for the deferred principal.)

The agency estimates that for Freddie borrowers who qualify for the Obama administration’s loan-modification program, and who owe at least 15% more than the value of their homes, principal forgiveness would cost taxpayers about $166 million compared to forbearance.

But for Fannie borrowers in the same situation, principal forgiveness would actually save taxpayers $534 million – or three times the losses on the Freddie loans – compared with forbearance.

The FHFA also analyzed scenarios in which borrowers don’t qualify for modifications because a foreclosure sale of their homes would yield more money for the lender than a modification would. While those scenarios showed a small advantage for forbearance over forgiveness, that group of borrowers seems irrelevant; no one is proposing principal reductions for homeowners who don’t qualify for loan modifications.

Under the more important scenario – those borrowers who do qualify for modifications – the agency’s analysis found virtually no difference in the cost to U.S. taxpayers between principal forgiveness and forbearance and if anything gave a slight edge to forgiveness.

Nevertheless, DeMarco wrote in a letter to congressional Democrats, “The net result of the analysis is that forbearance achieves marginally lower losses for the taxpayer than forgiveness.”

DeMarco also stated that in order to implement principal forgiveness programs, Fannie and Freddie would have to make costly new technological investments.

"Unless there is an expectation that principal forgiveness will reduce losses," he wrote, "we cannot justify the expense of investing in major systems upgrades."

It’s a fair point, and it might be a winning argument if the rest of the agency’s analysis rested on a convincing foundation.

But the foundation is shaky. For example, the FHFA apparently did not utilize available data showing how frequently borrowers who get their principal reduced go into default. Instead, the FHFA stated that its assumptions "rely heavily on the expert judgment" of staff at Fannie, Freddie and various government agencies.

Laurie Goodman, a senior managing director at Amherst Securities and a mortgage market expert who is critical of the lack of data supporting the FHFA’s analysis, said, "They did everything in theoretical terms."

Moreover, the FHFA did not look at key subsets of borrowers where principal forgiveness might make more economic sense for taxpayers than forbearance.

As Goodman explained, loans with mortgage insurance skew the FHFA’s overall analysis in favor of principal forbearance. This is because Fannie and Freddie are going to be made whole for insured loans regardless of whether the borrower goes into default.

Consequently, it would make sense to target a principal forgiveness program at loans that don’t have mortgage insurance.

"They did not differentiate between loans with and without mortgage insurance," Goodman said of FHFA.

Moreover, the FHFA hasn’t explored the possibility of establishing a program where in return for granting a principal reduction, Fannie and Freddie get a share of any future increase in the value of the home. The benefits to taxpayers of this kind of shared-equity model should be obvious.

At this point, the FHFA may not be able to sustain its stubborn opposition to principal reductions much longer.

Democratic Reps. Elijah Cummings and John Tierney are now pushing the agency, which operates independently of the Obama administration, to provide much more information about how it reached its statistical conclusions.

The two House Democrats are also seeking answers about information they recently obtained from a former Fannie employee, who said that a pilot program for principal reductions had been terminated in 2010 by senior officials at the mortgage giant who were philosophically opposed to the idea.

DeMarco is also scheduled to appear Tuesday before the Senate Banking Committee, where he seems likely to face questions about the FHFA’s stance.

An even bigger challenge is coming from Obama’s Treasury Department, which recently announced that its mortgage-modification program will begin making bigger incentive payments for principal reductions.

An FHFA spokeswoman said Monday that the agency is now reviewing its analysis of the data in light of the Obama administration’s increased incentive payments for principal forgiveness.

In March 2011, when DeMarco was asked about principal reductions during an appearance before Congress, he responded, "We have been trying very hard to take an empirical approach to looking at this."

It’s time for his agency to start living up to that promise.

Kevin Wack is a reporter covering Capitol Hill in the Washington bureau of American Banker. The views expressed are his own.