The 2009 Home Affordable Modification Program (HAMP) was supposed to help about 3 to 4 million homeowners in underwater or otherwise troubled mortgages by offering substantial financial incentives for loan modifications.
The program was well-designed and attractive in theory: applicants had to meet screening criteria to ensure that the most severe cases were prioritized, and all stakeholders faced substantial financial incentives for each successful modification. And, the policy’s architects attested, modifying mortgages in lieu of foreclosing would stave off further housing price declines, mitigate investor losses, keep families in their homes, and ensure cities didn’t lose tax revenue to support crucial services, all while preventing larger losses to banks in the form of the cost of foreclosures. It might even stimulate consumer consumption in the most heavily affected regions of the country.
Yet four years later, only about one third of the 4 million homeowners estimated to potentially benefit from HAMP during its initial lifespan (through December 2012) have received modifications through the program.
Professor Tomasz Piskorski looked at what went right and wrong with HAMP, working with a team of colleagues that included Sumit Agarwal and Gene Amromin of National University in Singapore, Itzhak Ben-David of Ohio State, Souphala Chomsisengphet of the Office of the Comptroller of Currency, and Amit Seru of the University of Chicago.
To trace HAMP’s performance, the researchers analyzed comprehensive loan-level data from the Office of the Comptroller of Currency and the US Treasury, which included information on more than 30 million residential mortgages, loan repayment amounts, servicers, delinquency rates, and principal and interest rate reductions (including whether such adjustment of terms was a result of HAMP or otherwise). All in all, the data represent about $6 trillion in mortgages.
To get at the underlying question of why the program underperformed, the researchers first looked at what HAMP did accomplish. The biggest obstacle in evaluating the impact of the program was estimating hypothetical outcomes, such as servicers’ renegotiation activity had HAMP not been put in place. They estimated this by examining the renegotiation activity of servicers of mortgages that did not meet the eligibility criteria for HAMP. Comparing non-HAMP renegotiation activity with renegotiation on HAMP-qualified mortgages helped the researchers quantify the effects of the program.
They found that HAMP fell significantly short of its objective: net renegotiations induced by the program have reached just about one-third (about 1.2 million) of its targeted three to four million indebted households by the original end date (December 2012). As a result of these modifications, the program will have achieved only a modest reduction in both foreclosures and house price declines.
So why did the program fail? The researchers’ analysis points toward servicer quality. In particular, they find a large discrepancy in the response to HAMP among servicers, with some modifying loans under the program at more than twice the rate of others. They also show that renegotiation intensity of a few large bank servicers explains much of the low overall response. The muted response of these servicers cannot be accounted for by differences in contract, borrower, or the regional characteristics of mortgages across servicers. Instead, their low renegotiation activity — which was also low prior to the implementation of HAMP — reflects their preexisting organizational capabilities. Servicers with lower renegotiation activity had existing organizational designs that were 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 servicers is their ability to deal with distressed loans: some servicers have expertise in dealing with non-performing mortgages, while others only specialize in efficient processing of checks from borrowers.
These differences across servicers are economically important. Piskorski and his co-authors find that about 75 percent of loans are in the hands of servicers with organizational designs that were less conducive to renegotiations. This makes some sense because during “normal” times, it may have been efficient to just specialize in the mass processing of checks from borrowers. However, had these low-response 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 two million modifications. That is still well short of the program goal of three to four million, but it is a big difference.
Finally, the researchers also explore what the effects of debt relief programs such as HAMP would be if they were implemented more intensively. To do so, they exploit regional variation in the share of loans in the hands of more capable servicers. Their findings suggest that debt relief programs, when used with sufficient intensity, may have a meaningful impact on foreclosure rates and house prices. However, their results also suggest that such programs targeted at distressed borrowers may not necessarily result in a sizeable increase in non-durable or durable consumption, at least in the near term.
Why is that the case? It appears that distressed borrowers use additional resources from debt reduction to pay down their other debt in lieu of spending on new consumer goods. These findings are consistent with arguments that the large accumulation of household debt prior to the crisis is an important factor limiting household consumption.
The analysis offers a lesson for policy makers and servicers: any policy intervention in response to foreclosure or other crises may face similar organizational hurdles and should be accounted for in its design. In the case of HAMP, for instance, modification rates might have been higher if the program had allowed servicing for modifications to be shifted to servicers that were better prepared to review, negotiate, and process modifications.
“Such a policy could be modeled on the existing framework in the commercial real estate market,” Piskorski says. Such provisions could be included in servicers’ contracts, which investors in mortgage-backed securities would be informed of at the time of purchase. “There, when pre-defined ‘adverse events’ occur, the loan is transferred to a special servicer, designed to efficiently handle distressed loans, including renegotiation.”
Tomasz Piskorski is the Edward S. Gordon Associate Professor of Real Estate and Finance at Columbia Business School.
This work was made possible in part due to the generous research support of the Paul Milstein Center for Real Estate.
Read the Research
"Policy Intervention in Debt Renegotiation: Evidence from the Home Affordable Modification Program"