The Federal Housing Administration (FHA) guarantees mortgages made by approved lenders using its Mutual Mortgage Insurance (MMI) Fund, which pays for approved claims. An annual report to Congress on the MMI Fund is an important barometer of the fiscal health of the FHA’s mortgage programs.
Separating the forward and reverse mortgage businesses will better serve both programs, which (respectively) finance approximately half of all first-time home purchases and provide seniors the opportunity to tap into home equity and age in place. The programs have different risk characteristics and missions. Lumping them together distorts their financial performance, interfering with policymakers’ ability to make sound decisions.
The report shows that the MMI Fund, which supports $1.23 trillion of insurance-in-force, had a capital ratio of 2.09 percent, just above the 2 percent statutory requirement.
But the MMI Fund provides insurance for two different programs:
- The forward program: $1.15 trillion in current insurance-in-force (8 million mortgages), a capital ratio of 3.33 percent, and an economic net worth of $38.4 billion
- The HECM program: $73 billion in current insurance-in-force (413,000 mortgages), a capital ratio of -19.84 percent, and an economic net worth of -$14.5 billion
The forward program, with 93.5 percent of the total insurance-in-force, has a capital ratio well over the statutory minimum. The reverse program, with 6.5 percent of the total insurance-in-force, shows the opposite, and its volatility makes it difficult to model. This puts the dependability of the reverse mortgage estimate in doubt.
It’s hard to accurately estimate the financial health of the reverse mortgage program
The figure below shows the capital ratios for the MMI Fund and each of the two programs since 2012.
The capital ratio on the forward fund has ranged from -1.34 percent to 3.33 percent, with predictable year-to-year growth of 0.5 to 1 percent following the Great Recession. The capital ratio on the HECM fund, on the other hand, ranged from -19.84 percent to 3.07 percent, with year-to-year variations of 8 to 14 percent.
This volatility stems principally from two issues that make HECMs hard to model: the assumption-driven nature of this product, including the impact of interest rates and home price appreciation, and the significant improvements to the program made between 2013 and 2015, which limit the amount of performance data available for modeling recent vintages.
Many unknowns make HECM modeling difficult and unreliable
Analysts must make assumptions about several unknowns to analyze potential losses for the HECM program. When borrowers over age 62 tap into home equity with a HECM, they receive cash from an FHA-approved lender. The FHA does not make the loan, but it insures the loan if the borrower defaults.
Unlike a traditional forward mortgage, the balance owed by the borrower on a reverse mortgage increases over time because the borrower does not make monthly payments. The loan is typically repaid when the homeowner (or their estate) sells the property. To estimate losses, analysts must assume future interest rates, home prices, and the time the borrower will be in the home.
Analysts must also answer the following:
- How long will the loan last?
- For how much will the home sell?
- Will price appreciation on homes owned by HECM borrowers match area price appreciation?
- What will be the prevailing interest rate in the coming years? (Cash flow from the HECM will be discounted at this rate, and this rate will be used to calculate the growth in the balance of any adjustable-rate mortgages.)
The variability of interest rates raises one of the biggest challenges to making accurate HECM performance predictions. Because a HECM borrower takes out a long-term loan secured by the property’s value, and many of these loans have a fixed interest rate, the value of that loan falls as rates rise, and rates fall as the HECM loan value rises.
Interest rate changes explain most of the MMI fluctuations in 2014 and 2015. In 2016, there was a change in assumptions as to the rate of home price appreciation for HECM borrowers. The analysis extrapolated recent lower home price appreciation over the life of the loan that may be 30 years or more. In 2017, a different model was used, which explains why the program continued to deteriorate in a year of robust home price appreciation. Overall, the modeling assumptions used are conservative and likely overstate the fund’s negative net worth.
Current HECM estimates do not account for recent risk reduction efforts
Several rounds of significant program changes are not reflected in the current HECM estimate. Since late 2013, the maximum amount that could be withdrawn (the so-called principal limit factor) was reduced substantially for every age bracket. Moreover, the proportion of this that a borrower could take out in the first year has been reduced from 100 to 60 percent of this limit, unless the up-front amount was being used to pay off a mortgage secured by the home.
Since early 2015, HECM borrowers must also conduct a financial assessment of their ability to meet their obligations, including property taxes and home insurance. These changes reduce HECM risks and will likely cut losses significantly by as much as 32 percent.
The annual report does not incorporate these data on defaults because it uses claims data that lags by several years. We would expect the estimates of the economic value of HECMs to improve in future years as the claim data get updated to incorporate these program changes.
Moreover, early in 2017, the FHA made further reductions to the maximum amount that could be taken out, which will positively affect the economic value of future vintages.
Removing reverse mortgages from the MMI Fund can help address this problem
Congress could take a positive step by removing the HECM portfolio from the MMI Fund. HECM estimates are heavily assumption driven and do not reflect recent program changes. Lumping the HECM and the forward products together distort decisionmaking for both programs.
Given the importance of HECMs in allowing seniors to age in place, the pre-2016 vintages should also be removed from the HECM MMI Fund, as the current HECM program has been dramatically improved, and past deficiencies shouldn’t hinder it going forward.
We also encourage the FHA to make data on HECM performance more available, so techniques for modeling performance can be improved.