How HUD Policy Changes Impact Reverse Mortgage Take-Up, Default Rates

The Department of Housing and Urban Development (HUD) has imposed a series of program changes for Home Equity Conversion Mortgages (HECMs) over the last few years. With new proposals already on the docket for 2016, few studies have shown the impacts of how such policy changes impact reverse mortgages—until now.

A recent research brief published by the Center for Retirement Research at Boston College summarizes the results of a recent study that estimates the effects of HECM program changes—specifically, the Financial Assessment and initial draw limitations—on both defaults and take-up of reverse mortgages.

Reverse mortgage default rates peaked in the aftermath of the financial crisis at 10% in 2013. This rise in defaults within the HECM portfolio, coupled with a negative balance in HUD’s Mutual Mortgage Insurance fund, prompted the agency to establish major changes to the HECM program.


The resulting modifications established a restriction on the amount of loan proceeds borrowers could receive in a lump sum during the first year of the HECM to 60% of the initial principal limit. HUD implemented this change in fall of 2013, less than two years before implementing more rigid underwriting standards with the implementation of the Financial Assessment in April 2015.

While the general impact of these policy changes is clear, that is, their intent should help curb the default rate by screening high-risk loan applicants and help borrowers avoid financial trouble, researchers say the magnitude of how these program changes will impact take-up of reverse mortgages remains unclear.

In their analysis, researchers relied on data from ClearPoint Credit Counseling, which provided information on households that received reverse mortgage counseling during 2006-2011. The data, which included demographic characteristics, FICO credit scores and other indicators of household financial wealth, was linked to HECM loan data from HUD, which provided details on reverse mortgage originations, withdrawals, loan terminations and defaults.

The purpose of linking the two datasets allowed researchers to analyze borrowers who ended up in default based on their financial characteristics at the time of loan origination. Of the study sample size comprising 27,894 households, 58% had a reverse mortgage.

When analyzing the linked datasets, researchers found a correlation between certain HUD reverse mortgage policy changes and their impact on reverse mortgage uptake and the probability of default.How HUD Policy Changes Impact Reverse Mortgage Take-Up, Default Rates

In the first simulation, researchers imposed the upfront withdrawal limits on all households in the sample who had reverse mortgages.

As higher withdrawal percentages are associated with higher rates of default, researchers found that imposing this restriction resulted in an 18% drop in the probability of ever defaulting. However, because a withdrawal limit may make reverse mortgage less attractive to some prospective borrowers, imposing this restriction reduced HECM take-up by 8%.

Researchers then examined a second simulation—one that does not include the withdrawal limit, but assesses the impact of using a specific underwriting threshold—using a credit score of 580 for determining whether an applicant should be approved for the reverse mortgage.

Requiring a credit score threshold of 580 allowed researchers to eliminate households that would not qualify for a reverse mortgage under this scenario. As a result, this simulation produced a 30% drop in the probability of defaulting and reduced uptake by 12%.

A third simulation, like the second one, excluded the withdrawal limit and included a credit score threshold of 580 for screening applicants’ finances. Unique to this simulation is that instead of rejecting loan applicants who fall below the credit threshold, this scenario allows them to take a reverse mortgage only if they have sufficient funds to set aside to cover future property taxes and insurance.

Compared to the other simulations, this scenario achieved a much larger drop in the default rate (37%) and a smaller drop (4%) in reverse mortgage uptake.

Lastly, the fourth simulation combined the policy changes from the first and third simulations, including HUD’s initial withdraw limit and the set-aside account for borrowers below the 580 credit score threshold. This simulation showed the largest drop in the default rate—a full 50%—and a 12% drop in take-up.

While there are lingering concerns that stricter underwriting requirements and withdrawal restrictions could ultimately make reverse mortgages less appealing to prospective borrowers, researchers expect the impact of credit-based underwriting will be small, “particularly when such standards are accompanied by a required set-aside for tax and insurance payments rather than a hard cut-off.”

“The combined impact of both types of policies could reduce take-up by 12%—primarily due to the restrictions on the initial withdrawal,” the researchers state. “However, this impact on take-up is relatively small for a rather large reduction in estimated defaults.”

Written by Jason Oliva

Join the Conversation (5)

see all

This is a professional community. Please use discretion when posting a comment.

  • Many, many thanks for this write up, and for those who conducted the research. Seeing some scientific data behind the impact of HECM changes (assuming the observation was handled with a scientific process) is encouraging. The expected reduction in uptake given the last scenario is within an acceptable margin or error to my own anecdotal expectations of reduction in origination volume due to more strict guidelines. I would be very much more interested, however, in any similar study that took today’s exact HECM requirements applied to originations over the same time period as this study. I wonder if the expected reductions in default and uptake would be the same. Does anyone have any insights?

    • David,

      Whenever a statistical sample is taken from a specific population, then the results are limited to that population at best. Since as to borrowers the sample was limited to counselees counseled solely by ClearPoint, the results as to borrowers is limited to those borrowers as well.

      What is odd is that some of the population sampled had their homes subject to the depths of the Great Housing Depression of 2008 AFTER getting a HECM and the others did not. At best the results should have been divided into into the two groups just suggested.

      It seems that few if any of the population selected will see another Great Housing Depression in their lifetime. So was the likelihood of that condition incorporated into this test and if so what was that likelihood?

      Yet the defaults being examined have had little to do with the problem we have seen with the actuarial results we have seen projected since fiscal 2009. By looping off say 40% of the value (not equity) of the homes that serve as collateral for HECMs, one would expect a terrible result on the MMI Fund even if those defaulting on taxes and insurance had not defaulted at all. While those the defaults themselves had an increased impact on the losses being shown from HECM operations they were neither material nor significant in relation to the losses reported. Remember many of these defaults had not closed that long before the defaults occurred.

      Looking at the alleged reduction of defaults and loss in business, the last scenario would have resulted in a 50% reduction in defaults but at cost to the business of 12%. It would seem the better scenario for the industry is the third one where the reduction to defaults would have been 37% but the loss in business would have only been 4%. Based on the problems with the sample and assumptions, who can say that your anecdotal and subjective results are not as reasonable as the results from this very problematic sample?

  • So here is an interesting (although not statistically reliable) phenomenon that if you impose the 60% initial limitation on disbursements, your default rate will drop by 18% of what it otherwise would have been at a cost of 8% of endorsements.

    Then the researchers acted as underwriters and permitted LESAs under their own unique conditions. If the FICO score was under 580, then we would cut down defaults by 37% and only lose 4% in endorsements by allowing these prospects to get LESAs. This test is so unrealistic, it boggles the mind. First, the population from which the borrower sample is only those who had counseling from one counseling agency. Second, how do the researchers know what percentage of those who were permitted LESAs would move forward to closing?

    There are so many theoretical problems and weaknesses with the research that discussing it is hardly worth one’s time. Once again, BCCRR shows its ineffectiveness at reliable research.

  • Moulton and company as well as our hecm industry seem oblivious to the real cause of claims against the insurance fund, a nearly 40 percent crash in housing values during the “Great Recession” which took until 2013 to reach its peak of insurance claims. A far more telling research statistic would be to project how a 40 percent crash in housing values would affect “default” and “uptake” if it were to happen again. I would hazard to guess the results would be nearly the same.

  • The statistics in the article by Jason are interesting to say the least, although, time will give us a better handle on it all.

    FA as a whole, in my opinion has put a high credibility rating on our product. I feel this at least, should out weigh the loss of business we have had because of those borrowers that don’t qualify anymore.

    In short, this new found credibility factor will give us many other professional opportunities with the financial planners, advisors, banks, credit unions, accountants attorneys and many more.

    What concerns me the most right now are the proposed changes coming up after the comment period ends, this week I believe. The final results that comes out in the end could determine where we as an industry are going.

    We also need to digest the new changes that came out in the mortgagee letter last Wednesday. All of these new events can throw these statistics out of the window, depending on the final out come. I am sure all realize this.

    This is why in my opinion, I look at these statistics as being good information but meaningless in some ways until we find out the other influencing factors that could change the course of things! We also have to hope that no more other changes surface in the mean time!!!

    John A. Smaldone

string(113) ""

Share your opinion