The FHA home mortgage insurance program was reformed in 1990 in order to put the MMI Fund on an actuarially sound basis. Those reforms have substantially increased the net worth of the fund. The latest actuarial study of the fund reports that its net worth is well above the level mandated in the legislation. There are, however, several reasons for treating that finding with caution. The study does not include the full data for fiscal year 1999, and the technical analysis is not directly comparable with previous actuarial studies. The 1990 legislation was complicated and the policy debate was vigorous and intense. Before doing anything else, Congress should make sure the estimated surplus net worth in the MMI Fund is really there.
If the MMI Fund is in solid financial shape, FHA should resume paying distributive shares to homeowners who pay off their mortgage. This is what Congress anticipated when it decided in 1990 to suspend payment of distributive shares. Congress explicitly rejected the idea of permanently terminating distributive shares. Congress expected the Secretary of HUD to resume paying distributive shares when the MMI Fund was actuarially sound again.
Paying distributive shares is also the fair thing to do. The MMI Fund's actuarial position depends on whether FHA-insured homeowners pay off their mortgages or default on them. The 1990 reforms were based on the best judgment that could be made at the time as to how many homeowners would be in each category. If the MMI Fund's position is stronger than Congress mandated, that's because more homeowners have paid off their mortgage than anyone expected 10 years ago. Those homeowners have paid more in insurance premiums than in hindsight they needed to. They deserve a rebate of their excess premiums. They have earned it.
The Homeowners Rebate Act is an appropriate way to resume paying distributive shares. It protects the MMI Fund and the taxpayer. If the MMI Fund continues to build net worth in the future, it would also be appropriate to lower the mortgage insurance premium. That will help FHA achieve its historic mission of promoting homeownership. Young families will find it a little easier to buy their first home.
Any excess reserves in the MMI Fund should not be spent on housing subsidy programs. They are not general tax revenues, they are payments into a mutual insurance program. Whatever the merits of the housing programs, those programs should be enacted through the normal authorization and appropriations process.
Thank you for inviting me to testify before the Subcommittee today to discuss the proposed Homeowners Rebate Act of 2000, introduced as H.R. 4785 by Chairman Lazio in the House and as S. 2914 by yourself and Chairman Gramm in the Senate.
I should begin by stating that, although I am Director of Urban Policy Studies and a Senior Fellow at the Hudson Institute, a nonprofit policy research institution, I do not speak for the Hudson Institute or any other group. I am here to speak only for myself, based on my experience at HUD and my professional activities. I am an economist, specializing in housing and urban issues throughout my professional career. I have served at HUD in three different administrations over the last 30 years, most recently as Assistant Secretary for Policy Development and Research under President Bush and Secretary Kemp from 1989 to 1993. In that capacity, I helped to design the 1990 FHA reform legislation that protected the Mutual Mortgage Insurance (MMI) Fund from threatened insolvency and put in place the policies that have created the present positive net worth position. At that time I testified about FHA before this subcommittee and your counterpart subcommittee in the House, and I'm pleased to have the opportunity to appear before you this morning on the same subject.
In my testimony today, I will first describe the position of the MMI Fund when Congress and the Administration found it necessary to reform the home mortgage insurance program. Then I will discuss the current actuarial position of the Fund. The second part of my testimony concerns the issue of distributive shares. I will explain the issue and the reasons for suspending payment of distributive shares in 1990. With this background, I will comment on the on the Homeowners Rebate Act of 2000, which proposes to resume the payment of distributive shares to FHA-insured homeowners. I'll end with a few thoughts on appropriate and inappropriate policy options.
The FHA home mortgage program has been financially self-sufficient since it was enacted in the 1930s. It has been widely praised as a social experiment that worked, helping millions of young families become homeowners without cost to the taxpayers. This positive image was rudely jarred in 1989, on the heels of the savings and loan crisis. Policymakers in both the executive and legislative branches became concerned that the MMI Fund might be insolvent. Secretary Jack Kemp therefore commissioned the accounting firm of Price Waterhouse to conduct the first independent actuarial study of the Fund. Price Waterhouse issued its report in June 1990. It concluded that the Fund was solvent but not sound. It had a net worth of less than $3 billion, about 1 percent of the total amount of FHA mortgage insurance in force. This was a significant decline in the course of a decade. Price Waterhouse calculated that in 1979 the Fund had a present value of over 5 percent of the total amount of insurance in force. It projected that the Fund would lose money on each year's book of business from 1980 through 1989. It further projected that the Fund would lose about $200 million on each year's book of business in the future. FHA was clearly underpricing its mortgage insurance, and had been doing so for quite some time.
Price Waterhouse concluded that the MMI Fund's net worth should be at least 1.25 percent, and said further that "the 1.25 percent is not a desired ratio, but a minimum ratio." That meant the Fund should have a net worth of $3.75 billion at the end of fiscal year 1990, about $1 billion more than it was expected to have.
The Price Waterhouse study was released at a hearing before ths subcommittee in June 1990. That hearing launched a vigorous and intense debate on FHA reform during the summer and fall. The Senate and House passed very different reform proposals, but eventually managed to resolve the differences and include FHA reform in the Cranston Gonzalez National Affordable Housing Act. It is important to emphasize that there was agreement throughout the policy debates - between the Administration and Congress, between the House and Senate, between Republicans and Democrats - on the net worth targets. After considering the findings in the Price Waterhouse study, the Administration and Congress agreed that FHA should achieve the minimum net worth ratio of 1.25 percent within 24 months. They also established a higher target net worth ratio of 2 percent, to be achieved by the end of the 1990s.
The FHA reform legislation required HUD to provide for an annual independent actuarial study of the MMI Fund. The latest actuarial study, for fiscal year 1999, estimates that the Fund has a substantially higher net worth than mandated by the Cranston-Gonzalez National Affordable Housing Act. The actuarial study estimates a net worth of 3.66% as of the end of fiscal year 1999, and projects a net worth of 3.81% at the end of the current fiscal year. This is welcome information, certainly, but there are several reasons to treat it with caution.
The 1999 actuarial study is the first one conducted by the accounting firm of Deloitte Touche; all of the previous studies back to the first one were conducted by Price Waterhouse. There are some notable differences between them. As of the end of fiscal year1998, Price Waterhouse estimated that the net worth would be 3.06% at the end of fiscal year 1999. That's a large difference, and it isn't fully accounted for in the study. Deloitte Touche calculates that two-thirds of the difference is due to changes in economic assumption and differences between the projected and actual experience of the MMI Fund in fiscal year 1999. That leaves one-third to be accounted for. Deloitte Touche is using a different econometric model than Price Waterhouse; that may be part or even all of the explanation, but Deloitte Touche isn't able to compare the models to see.
The Deloitte Touche actuarial study does not include the full 1999 book of business. It includes only the first nine months of the year. All the previous studies used data for the full year. That may not matter, but there is no necessary reason to produce an actuarial study based on nine months' data. The actuarial study states that Deloitte Touche plans to update the analysis "during the next two months (December 1999 and January 2000)," but to my knowledge has not made public any subsequent analysis.
The 1998 Price Waterhouse actuarial study calculated a slight decline in the MMI Fund's net worth, from 2.81% at the end of fiscal year 1997 to 2.71% at the end of fiscal year 1998. That wasn't a large change, but it was a change in the wrong direction and it was surprising. The economic expansion during 1998 should have caused an increase in the Fund's net worth, consistent with other years of economic expansion. One contributing factor is that FHA default rates in 1998 turned out to be higher than projected. That happened again in 1999. The Deloitte Touche actuarial study reports slightly higher defaults for fiscal year 1999 than Price Waterhouse projected a year before, but the MMI fund's net worth is estimated to be higher in 1999, rather than lower. (Defaults are reported by Deloitte Touche only for 30 year fixed rate mortgages, but that is the lion's share of FHA's business.)
None of these points imply that the Deloitte Touche actuarial study is wrong. Data for the last three months of fiscal year 1999 may not have any appreciable effect on the Fund's net worth. Also, defaults are not the only factor affecting the Fund's net worth; they may be offset by other factors. I know from working with Price Waterhouse 10 years ago that these actuarial studies require complicated and sophistical econometric analysis. We should not read too much into the differences between their first study and the last one by Price Waterhouse.
But they do suggest caution. It would be prudent to see if the reserves of the MMI Fund continue to increase, before deciding to do anything with any amount in excess of the appropriate amount of capital. It would be useful to have two full year actuarial studies from Deloitte Touche before making major policy decisions based on the first partial year study.
Putting the MMI Fund on the road to actuarial soundness was hard work 10 years ago. It was hard work for HUD senior management and hard work for the housing subcommittees of Congress. It was hard work for the mortgage bankers, the realtors, and the builders. I can't imagine that anyone wants to go through that process again. Before Congress decides what to do with any "surplus" net worth, I urge you to make very sure that the surplus is really there.
The FHA single-family insurance program was established on the principle of mutuality. This was done because it is not possible to forecast mortgage defaults and prepayments very accurately. (Certainly it was extremely hard to forecast them in the late 1930s.) Both defaults and prepayments depend on economic conditions. If interest rates fall, prepayments rise. If there is a recession, defaults rise. The timing of economic fluctuations is also relevant. Defaults are especially likely in the first few years after the mortgage is issued; then gradually default becomes less likely as the homeowner builds up some equity in the house. Once that happens, the homeowner is not very likely to default during a recession, even if the owner loses his or her job and decides it's necessary to move; there is still equity in the home, and the owner will be better off selling it rather than defaulting on the mortgage.
This means that the mortgage insurance premium may turn out to be too high or too low. If it is too high - if the economy performs better than mortgage insurers and homeowners expect
when the mortgage is issued - then one way to deal with the excess of premium income over claims is to return the excess premium income to the policyholders. The payments are called "distributive shares" and the type of insurance is known as participating or mutual insurance. The same principle of mutuality is applied to life insurance; there are some mutual life insurance companies.
FHA began to pay distributive shares in 1943. They were paid to homeowners on the basis of the year in which they bought their home and took out their FHA-insured mortgage (known as the endorsement year). If there was a surplus of premium income over claims among all mortgages insured in fiscal year 1960, for example, FHA paid a distributive share to a mortgagor when the mortgage was prepaid (for example, when the house was sold) or when the mortgage was paid off at maturity, in 20 or 30 years.
Through 1989, the MMI Fund had returned almost $1.5 billion in distributive shares to mortgagors. FHA had declared that distributive shares would be paid on mortgages underwritten through 1978. FHA's practice was to declare that distributive shares would be paid when it estimated that income would exceed claims for the mortgages insured in a particular endorsement year
The Price Waterhouse findings raised questions about distributive shares. Price Waterhouse's calculation of net worth did not factor in the payment of the distributive shares that FHA had already announced. If FHA continued to pay distributive shares on mortgages insured between 1961 and 1978, standard accounting practices required that the Fund's estimated net worth should be reduced by $1.5 billion, because $1.5 billion was the present value of the distributive shares that FHA was committed to pay. The policy question was whether to continue paying distributive shares on endorsement years through 1978 even though the Fund as a whole was not in sound condition, or whether to renege on a commitment by the federal government.
The Administration's original reform proposal, as stated by Secretary Kemp before this subcommittee in June 1990, was to continue paying distributive shares on past books of business through 1978, but to terminate the practice for the future. The subcommittee preferred to suspend the payment of distributive shares, for past endorsement years as well as for the future, until the 2.0 percent capital target was achieved. That was incorporated in the legislation passed by the Senate. The House subcommittee and the House itself subsequently accepted the same approach, without much debate. The actual language in the Cranston-Gonzalez National Affordable Housing Act was drafted in the House. It states that the Secretary can determine whether to issue distributive shares (Section 331). If the Secretary determines that the MMI Fund is not maintaining an adequate capital ratio, the Secretary may not issue distributive shares (Section 332).
Thus in 1990 Congress made an explicit decision to suspend payment of distributive shares until the MMI Fund was solvent. Congress rejected a proposal by the Administration to terminate payment of distributive shares on future books of business. It left open the possibility of paying distributive shares on those future books of business, depending on the financial condition of the Fund.
If the Deloitte Touche actuarial study is correct, then FHA is now in a position to being paying distributive shares again, in accordance with the Cranston-Gonzalez National Affordable Housing Act. That is what Congress committed FHA to do, when it passed the reform legislation 10 years ago. Congress suspended payment of distributive shares because the MMI Fund was in an unsound actuarial position. That problem has been resolved. FHA has done better than anyone expected. The projections of FHA's claims experience - the best judgment that anyone could make at the time - turned out to be too pessimistic. More homeowners either prepaid their mortgage or continued to make their monthly payments. As a result, they have paid more in insurance premiums than they needed to. They deserve a rebate of the unnecessarily large premium. They have earned it.
Resuming payment of distributive shares is appropriate for another reason. If at a later date the MMI Fund's net worth starts to decline again, then the HUD Secretary can once again suspend payment. That would be entirely consistent with the FHA reform that Congress enacted 10 years ago. It gives HUD the flexibility to protect the fund - a first line of protection, buying time for more fundamental legislative changes if they become necessary.
The Homeowners Rebate Act includes several protections for the MMI Fund.
(1) By setting a capital target of 3 percent before the Secretary can issue distributive shares, the bill ensures that the fund will not be depleted by an unexpected economic downturn.
(2) Also, the bill changes the basis on which distributive shares would be paid. They would be paid on all mortgages that are terminated when the fund has a net worth above 3 percent, regardless of the origination year. That is more reasonable than the past practice of declaring that distributive shares will be paid on individual books of business, regardless of when the mortgages are actually terminated. The solvency and soundness of the fund depend on the mortgage payments of all insured homeowners, not those who bought their homes in a particular year. There are not separate insurance funds for each year's book of business.
For these reasons, I think the Homeowners Rebate Act is a very good way to resume the payment of distributive shares, if the MMI Fund is in an actuarially sound position with sufficient reserves to protect taxpayers and homeowners.
The most important part of the FHA reform in 1990 was an increase in the mortgage insurance premium. Each book of business since those reforms went into effect in fiscal year 1992 has contributed to the MMI Fund's growing net worth. The premiums that homeowners are paying exceed the losses from defaults by a substantial amount, so that net worth is rising faster than insurance in force. If that continues in the future, then Congress would be in a position to reduce the mortgage insurance premium for homebuyers. That will help FHA achieve its historic mission of promoting homeownership. Young families will find it a little easier to buy their first home.
There is recent precedent for this. In 1996 President Clinton reduced the premium by 25 basis points, on the grounds that the MMI Fund was going to achieve the mandated net worth ratio of 2 percent by the year 2000. If in fact the capital ratio is now between 3 and 4 percent, and likely to keep rising, there is better justification for cutting the premium now than there was four years ago.
I know that there have been proposals to spend the MMI Fund's reserves on various housing programs. I haven't looked at the specific proposals, because I think it is fundamentally wrong to treat the reserves as if they were general tax revenues paid by all taxpayers to support the activities of the federal government. The mortgage insurance premium would become a tax on young families to support housing subsidy programs or other housing purposes. That's not the way to promote homeownership. Whatever the merits of the housing programs, they should be enacted through the normal authorization and appropriations process.
Distributive shares were suspended in 1990 because the MMI Fund was not actuarially sound. If the fund's position is as strong as the latest actuarial studies indicate, then it would be consistent with the law to resume paying distributive shares. The Homeowners Rebate Act is a good way to resume payment. It protects the fund and the taxpayers at the same time that it rewards homeowners who have been making their mortgage payments regularly over the years. I believe the public would be well served if Congress passed the Homeowners Rebate Act.
Home | Menu | Links | Info | Chairman's Page