Study Finds Largest Servicers Hampered HAMP
Pre-HAMP experience with loan modifications may have been key to the outcome of the Home Affordable Modification Program (HAMP) according to a paper published recently by the Federal Reserve Bank of Chicago, and the Office of Comptroller of the Currency (OCC) on The Social Science Research Network. The internal structure of some servicers may have reduced the number of HAMP modifications by as much as 70 percent.
"Policy Intervention in Debt Renegotiation: Evidence from the Home Affordable Modification Program" written by a consortium of six academics from several universities looked at the HAMP program to determine whether government intervention in the functioning of mortgage markets is effective. Proponents argue that efforts such as household debt relief and foreclosure prevention initiatives prevent excessive foreclosures that could lead to deadweight losses for borrowers and lenders and negative externalities for the society. They also say that such policies help reduce high levels of household debt that may distort consumption and investment decisions. Critics argue that these policies risk generating moral hazard problems that are likely to raise the cost of credit in the long run and have undesirable redistributional consequences. There, however, have been few studies to explore either of these contentions.
This research was conducted using OCC Mortgage Metrics data which includes origination and servicing information on 34 million mortgages, about 64 percent of U.S. residential loans. About 11 percent of the loans are bank held, the remainder is sold to investors through Fannie Mae, Freddie Mac, and the private market.
The research team studied loans over the period July 2008 through December 2010. HAMP was implemented in March 2009 so the study included data spanning nine months of the mortgage crisis before HAMP and 21 months of the program period. To provide a control group for the study the investigators used two strategies; the first was to assume that properties that were investor owned would have been eligible for HAMP were they owner occupied and the second was to study loans with balances above $729,750 which were by definition ineligible for HAMP.
The paper has two objectives: first to evaluate how HAMP affected various margins related to renegotiation decisions by servicers - both those done under HAMP and outside of it - and the impact of HAMP on broader outcomes such a house prices. Second, the paper documents and exploits the significant heterogeneity in program response across intermediaries (servicers) to allow an understanding of their role in implementing the program.
HAMP appears to have increased the modifications offered to eligible loans but the analysis also looked at whether HAMP modifications substituted for modifications which might have otherwise been done outside of the program. They found no evidence of such a decline and found that HAMP led to an increase in the annual rate of permanent modifications of about 0.7 percent or about 1.2 additional permanent modifications over the study period, significantly short of the three to four million households targeted for intervention.
Researchers did find that private modifications offered after the program was introduced were less aggressive. In particular, the use of more aggressive tools like rate reduction declined by 11 percent, term extensions by 9 percent, and principal reduction by 2 percent. At the same time the use of less aggressive tools like capitalization of unpaid interest increased - in that instance by close to 10 percent. These changes however were offset by the higher effectiveness of HAMP modifications. For example, rate reductions through HAMP were offered at a rate 55 percent higher than through private permanent modifications.
These results could simply reflect a channeling of more promising loans to HAMP because of higher incentive payments for successful modifications. Alternatively, servicers may have redirected some resources to HAMP, adversely affecting of private modifications. Regardless of the reasons, there was clearly an adverse impact by HAMP on the effectiveness of private modifications in the treatment group.
The analysis then examined the impact of HAMP on the goal it was designed to affect, the rate at which loans are foreclosed. They found that the program resulted in a moderate decrease in the rate of completed foreclosures in the treatment group, a differential 0.48 percent decrease in the annual foreclosure rate or about 800,000 fewer foreclosures over the duration of the program (i.e. through December 2012). This was substantially lower than the program target. The authors make no projections about how many of these foreclosures were permanently prevented.
The study found no evidence that HAMP had any impact on external outcomes such as house prices in regions more exposed to the program or on auto sales, credit card loans, employment, or other externalities, suggesting that the program "generated limited spillovers, at least in the near term."
There was also little evidence that HAMP may have induced strategic defaults. There was only a small relative increase in the delinquency rate in the treatment group in the pre-HAMP period - about 0.027 percent each quarter. The authors say these results may provide guidance for future programs, in particular since HAMP guidelines required borrowers to provide extensive documentation on their economic hardship and provided additional compensation to servicers for modifying loans that were current.
A key finding was the substantial differences in the way servicers handled the program and the large affect on outcomes. Some servicers modified loans at a rate more than four times the rate of others and this cannot be accounted for by differences in contract, borrower, or regional characteristics of mortgages. There were also significant differences in the conversion rate from trial to permanent modification status across servicers. In sum, the study found a strong positive relationship between the intensity of renegotiations done by a servicer in the pre-HAMP period and the rate of permanent modifications done through HAMP. The authors suggest that, had large servicers performed at the same rate as the others, there would have been about 70 percent more permanent modifications put in place.
So why did some servicers perform better pre-program and carry that behavior forward? The investigators looked at relationships among several variables and found the number of full-time servicing staff to be positively correlated with the intensity of renegotiations in the pre-program period. They also found those larger staffs were also assigned fewer loans per member to service. The companies with a higher intensity of modifications also devoted more time to staff training.
The study also looked at the servicers' call centers using as proxies the percentage of dropped calls and the average hold time per call. Servicers who were more efficient in handling phone queries also conducted more renegotiations.
The authors suggest that a reallocation of resources might have improved implementation of the program and specifically suggest transferring distressed mortgages to designated servicers as is done in commercial real estate. They also suggest that while extensive screening for eligibility and the incentives offered to servicers may have prevented large-scale strategic defaults, they may also have stalled the pace of the program.