At least since hte Great Depression, federal and state governments have regularly intervened in the mortgage markets during economic crises through household debt relief and foreclosure prevention policies. During the current crisis, the largest such intervention took the form of the Home Affordable Modification Program, or HAMP.

This program, unveiled in early 2009 by the Obama administration, allocated about $50 billion to provide sizeable financial incentives to mortgage servicers relative to their normal compensation. These incentive payments were meant to facilitate renegotiation on mortgages, which could include actions such as reducing balances and interest rates of loans at risk of foreclosure.

The program’s goal was to alleviate several perceived barriers to mortgage renegotiation, such as the private market’s inability to internalize negative externalities imposed by excessive foreclosures and the institutional frictions to renegotiation of securitized mortgages resulting from separation of ownership and control of such loans.

Policymakers estimated that the program would result in permanent loan modifications of 3 million to 4 million households by the original program end date in December 2012.

Since its inception, HAMP’s impact and effectiveness has been the object of heated political debate, with most arguments guided by anecdotes. In order to inform this debate, we undertook the first formal systematic study of HAMP to understand how such programs work and which aspects facilitate or impede their effectiveness.

We argue below that this exercise provides important lessons for both the future design of mortgage finance and the assessment of other recent initiatives, such as the Home Affordable Refinancing Program known as HARP and QE3, the third round of quantitative easing approved by the Federal Reserve.

To answer the question of what HAMP did, we used unique loan-level data from the U.S. Treasury in our paper. These data cover more than 30 million U.S. residential mortgage loans (about 60% of the market) and contain information on loan and borrower attributes, including precise information on loan payments, the type of loan renegotiation taken (e.g., principal reduction), whether the renegotiation was undertaken under HAMP or outside of it, and the servicer responsible for the mortgage.

The biggest obstacle in evaluating the impact of the program is estimating hypothetical outcomes, such as servicers’ renegotiation activity in the absence of HAMP. We construct such estimates by examining the renegotiation activity of servicers with respect to mortgages that did not meet the eligibility criteria for the program. Comparing this activity with that on qualified mortgages helps us quantify the effects of the program.

Using this approach, we estimate that HAMP fell significantly short of its objective. We find that renegotiations induced by the program will reach just about one-third (about 1.2 million) of its targeted 3 million to 4 million indebted households by the original end date of December 2012. As a result of these modifications, the program will have achieved a modest reduction in both foreclosures (at most 800,000 overall) and house price declines. We found that the program had no effect on other variables that could have been impacted by household debt reduction such as durable consumption — possibly because detecting consumption responses at the household level would have required a much larger take-up rate than what actually occurred under the program.

Interestingly, our analysis does not support the view held by some administration officials that HAMP boosted proprietary modifications. Two pieces of evidence debunking this notion are particularly worth mentioning. First, servicers who did relatively few proprietary modifications before HAMP continued to do few during the program. This seems inconsistent with the argument above, since under that view one would have expected to see more proprietary modifications. Moreover, our evidence suggests that HAMP had an adverse impact on the effectiveness of proprietary modifications, as measured by the rate at which borrowers re-default subsequent to receiving a modification.

SERVICER WEAKNESSES

The answer to why HAMP failed to meet its target largely lies in servicer quality. In particular, we find great discrepancy in the response to HAMP between the mortgage servicers, with some servicers modifying loans under the program at more than twice the rate of others. In fact, we find that the low renegotiation intensity of a few large bank servicers explains much of the low overall response.

The muted response of these servicers cannot be accounted by differences in contract, borrower, or regional characteristics of mortgages across servicers. Instead, their low renegotiation activity — which is also observed before the program — reflects their preexisting organizational capabilities. Servicers with lower renegotiation activity had existing organizational design that was much less conducive to
mortgage modification.

These organizations had smaller, less trained, and overloaded servicing staff, and servicing call centers that were unable to efficiently handle a large number of calls. One of the most important differences among mortgage servicers is their ability to deal with distressed loans. Some servicers have expertise in dealing with nonperforming mortgages, while others only specialize in efficient processing of checks from borrowers.

These differences across servicers are economically important. About 75% of loans are in the hands of servicers with an organizational design less conducive to renegotiations. This makes some sense because during “normal” times, it may have been efficient to just specialize in mass processing of checks from borrowers. However, had these low-response mortgage servicers modified loans at the same pace as their more active counterparts, HAMP would have induced about 800,000 more modifications. Hence, instead of about 1.2 million, the program would have resulted in about 2 million modifications. That is still well short of the program goal, but it is a big difference.

LESSONS EMERGE

Two related lessons emerge from our work. One is that the organizational capability of servicers could have played a very important role in determining the rate of modifications and foreclosures during a financial crisis — and such capability takes a long time to build.

Large servicing platforms suited for the mass processing of payments were ill-equipped to deal with distressed borrowers. Moreover, given that the incentive payments were quite substantial, policies that rely on organizations to quickly change their operations — by altering their focus from processing checks to loan renegotiation — face significant hurdles.

A more fruitful direction for policymakers would be to design policies to alleviate frictions that prevent the reallocation of resources between intermediaries. For example, in the case of HAMP, it may have been productive for the program to allow the transfer of distressed mortgages from inefficient servicers to those more capable of conducting many renegotiations.

One way to address this issue in the future is to borrow some ideas from the commercial real estate market. In the commercial mortgage-backed securities market, performing loans are handled by primary servicers.

Upon the occurrence of certain specified adverse events, primarily serious distress, the administration of the loan is transferred to a special servicer. The special servicer — similar to a “SWAT” team — is an organization designed to efficiently handle distressed loans, including their potential renegotiation.

Of course, if implemented in the residential sector, the provisions for these transfers would have to be included in the contractual agreements with servicers and understood by investors at the time of purchase.

The second, broader lesson that emerges is that policies relying on the voluntary participation of intermediaries need to recognize that certain organizations may be better equipped than others to implement a given initiative. Such policies are not limited only to HAMP but also apply to other initiatives undertaken by the administration in response to the foreclosure crisis. For example, an effective HARP and QE3 initiative requires significant refinancing activity by intermediaries.

Our work suggests that such policies may also face significant hurdles due to limited organizational capabilities of some financial intermediaries. Therefore, future policies should consider provisions that alleviate implementation hurdles due to intermediary heterogeneity.