In the 1990 study, Price Waterhouse allowed the FHA ceiling loan amount to increase to 95 percent of each state's median house price. This indexing of FHA loan limits provides a higher ceiling in many metropolitan areas thereby increasing FHA's business in states with high home prices. By increasing FHA's volume of business, premium income would rise, thus improving the MMIF's cash position, according to the Price Waterhouse analysis. More importantly, Price Waterhouse determined that the increased ceiling would provide larger cash reserves for the MMIF for two reasons. First, the fund would receive greater premium income because it would insure more mortgages. Second, Price Waterhouse, after reviewing FHA and private mortgage claim (default) and non-claim (prepayment) experience, found that the default and loss rates would be lower for higher balance loans. For example, FHA experience over the last ten years indicates that larger loans tend to show lower default rates and experience lower percentage losses when they do default. This was true at all Loan-to-Value (LTV) ratios, i.e., holding LTV constant, lower balance loans have a higher default rate than higher balance loans. The report indicated that even under severe economic scenarios, the MMIF would be better off if the mortgage loan limit were raised. NAR strongly urges that the FHA loan limit be raised to 95 percent of the area median sales price with a national ceiling of 85 percent of the Fannie Mae and Freddie Mac conforming loan limit. In conjunction with this, for loans above $50,000 we would support establishing a minimum downpayment requirement of 3 percent of the first $25,000 of the loan amount, 5 percent of the loan amount between $25,000 and $125,000 and 10 percent of the amount over $125,000. For loans below $50,000 the current 3 percent downpayment requirement would remain. The 10 percent minimum downpayment requirement on the balance above $125,000 would assure the quality of large loans from high-cost areas. In past Congressional debates on raising the loan limit, higher balance loans were characterized has possessing a greater contingent liability for the MMIF than lower balance loans, and therefore might erode the reserves of the MMIF in the future. Data on the experience of both FHA- and conventionally-insured mortgages discussed above, however, demonstrates that this concern is both unnecessary, and incorrect. Prudence, however, requires entertaining the prospect that the future claims experience of higher balance loans will diverge from historical experience. It is for this reason that NAR suggests the above downpayment requirement. NAR believes that raising the mortgage limit and increasing the downpayment requirement will improve the financial position of the MMIF by broadening the usage of the program in terms of serving borrowers in more areas and also in terms of the range of borrowers who would buy homes that could qualify under the program. Structural Changes Needed to Increase Homeownership Opportunities We need to reemphasize that the effect of HUD's regulatory implementation of NAHA and the Omnibus Budget Reconciliation Act of 1990 has been to severely weaken the future viability of the MMI program. Simply reversing the changes to the FHA single-family program implemented in 1991, however, will not return the MMIF to solvency; a more systematic approach must also be made. The NATIONAL ASSOCIATION OF REALTORS® believes that changes to current FHA programs could broaden homeownership alternatives for borrowers who are displaced from the current 203(b) insurance under the MMIF, but could qualify under other FHA insurance programs if certain barriers were removed. These changes would require congressional action; however, they all would shore up the financial condition of the MMIF. Broadening the appeal for alternatives to the MMIF programs will provide homeownership opportunities to those households displaced from the single family mortgage insurance program. Three specific FHA housing programs can be identified as alternatives to the Section 203(b); (1) the Section 221(d)(2) single-family program insured through the General Insurance Fund; (2) the Section 223(e) single- and multi-family insurance programs insured through the Special Risk Insurance Fund; (3) the Section 234(c) condominium home mortgage insurance program insured through the General Insurance Fund. Both the General Insurance (GI) and Special Risk Insurance (SRI) Funds are intended to require subsidies from Congress. There has never been a requirement that these funds be actuarially sound, but rather they are directly funded through Congressional appropriation. Broadening the scope of these alternative programs will accomplish two major goals: (1) improve the quality of loans entering the MMIF, and thus preserving the financial viability of the fund; and (2) accommodating displaced households explicitly within FHA housing programs. A. B. Section 221(d)(2) single-family mortgage insurance program. Raising the loan limit on Section 221(d)(2) from $31,000 ($36,000 in high-cost areas) to 30 percent of the FHLMC/FNMA conforming mortgage limit ($202,300 in 1992) would broaden the appeal for the Section 221(d)(2) program. This program provides mortgage insurance for loans used to finance the purchase of low cost one-to-four family dwellings for low or moderate income families or families displaced by governmental action (urban renewal, code enforcement, etc.) or as a result of a circumstance declared by the President to be a major disaster. The limits for highcost areas are available only in those areas in which construction costs cannot be maintained within the standard limits. This determination is made by HUD, and limits are established separately for each area. Indexing the maximum loan limit to a proportion of the conforming conventional loan limit would remove the principal barrier to this program, which is the low maximum mortgage limit. Section 223(e) program. The Section 223(e) program provides insurance to areas defined as older, economically declining metropolitan areas. Expanding the number of areas that would be eligible for these loans would provide an alternative for many current MMIF borrowers who more appropriately should be insured pursuant to Section 223(e). HUD insures lenders against loss on mortgage loans under other sections of the National Housing Act, e.g., Section 221(d)(2) or Section 221(d)(3). C. The determination to use Section 223(e) is at the discretion of HUD. These loans may be used to finance the purchase, repair, rehabilitation, and construction of housing in older, declining urban areas where conditions are such that certain normal eligibility requirements for mortgage insurance under a particular program cannot be met. The property must be an acceptable risk giving consideration to the need for providing adequate housing for low- and moderate-income families. Section 234(c) Condominium Units Program. The purpose of the Section 234(c) Additionally, an environmental assessment required for a proposed project under The requirements on units in a project converted from rental housing were instituted during the 1970s and have outlived their public policy needs. Relaxing these requirements would open up homeownership opportunities in many areas of the country. Similarly, the owner-occupancy requirements of this program necessitate time-consuming verification for lenders and significantly lower lenders desire to make these loans. HUD needs to simplify this process to expand program activity. Price Waterhouse FY 1990 Actuarial Status Report and FY 1991 FHA Financial Audit HUD has recently released two significant reports which relate to the Mutual Mortgage Insurance Fund. On March 19, HUD Secretary Jack Kemp released the annual MMIF Actuarial Status report, as required by NAHA. This actuarial analysis was performed by Price Waterhouse. On March 27, HUD released the FHA Audit of Fiscal Year 1991, Financial Statements. This year's audit was released by HUD's Inspector General for Audit, and was performed by Price Waterhouse. Both of these releases portray a troubled FHA Mutual Mortgage Insurance Fund, and the NATIONAL ASSOCIATION OF REALTORS® welcomes the opportunity to comment on these reports. NAR wishes to convey our concerns with both the modeling assumptions as well as the economic assumptions utilized by Price Waterhouse in its model of the economic value of the MMIF. We feel that the combination of both model specification error (in particular the claims model), as well as the excessively pessimistic economic and house price appreciation projections, significantly over-estimate the default risk of current and future "books of business" in the singlefamily mortgage insurance programs. This results in a much more negative assessment of the economic value of the MMIF than we feel is warranted. Proving this contention, however, has been difficult. In a Freedom of Information Act (FOIA) request submitted to HUD by NAR in May 1991, our request for access to the data provided to Price Waterhouse was denied, because HUD said it "did not maintain a copy and cannot recreate the database." An earlier FOIA request by the National Association of Home Builders for the Price Waterhouse model was denied because the Department said it "does not possess either the model or additional data assumptions." To date, access to the model and the data used to estimate the model has been denied to interested industry participants. This lack of access has only raised NAR's uneasiness with the sensitivity of this model to its assumptions. The most recent Price Waterhouse status report illustrates our concern of the "sensitivity" of the Price Waterhouse model to its economic assumptions. Since the original Price Waterhouse assessment, the new report indicates that the economic value of the Fund declined by $5.8 billion dollars, from a positive $2.6 billion to a negative $2.7 billion, with nearly half of this decline attributive to a change in house price appreciation and economic assumptions alone. We are stunned that in using only one more year of data, this model can generate results this divergent in its bottom-line assessment of economic value. We are even more skeptical that HUD relied so heavily on an untested and unproven econometric "model" to make drastic and dramatic changes to such a highly utilized and socially important housing program. As we previously mentioned, the original Price Waterhouse report assumed that previous FHA homebuyers with more savings than the typical FHA homebuyers would, when faced with the new "risk-based" premium structure, increase their downpayment and remain within the FHA program. However, as was demonstrated in Table I, adverse selection has occurred and these homebuyers have opted instead for the conventional mortgage market. This has left the risk profile (as measured by average LTV) of new FHA insurance as high if not higher than it was pre-reform. Price Waterhouse did not do any real world modeling of consumer behavior when faced with a price increase, and simply assumed that this business would be retained. It was not, and the future viability of the MMIF was not enhanced. The Fiscal Year 1991 Financial Audit of FHA also paints a troubling picture of the current cash performance of the MMIF. The most recent financial audit of the FHA reported that the MMI Fund incurred a operating loss of $31 million during Fiscal Year 1991 after reporting a profit of $66 million in FY 1990 and $11 million for FY 1989. The FY 1991 audit also reported that Government Equity in the MMI Fund declined from last year's audit, decreasing from $924 million in FY 1990 to $871 million in FY 1991. NAR would be even more concerned with this report if we were not convinced that Price Waterhouse is incorrectly crediting unearned premium in its financial statements, resulting in an overly low estimate of government equity. The NATIONAL ASSOCIATION OF REALTORS® strongly urges consideration of legislative language to require HUD to revamp its refund schedule of the upfront mortgage premium for the single-family mortgage insurance program to reflect the claims experience in the single-family program during the 1980s. This refund schedule, which was created in 1984 and never updated, significantly over-refunds premium to FHA borrowers who prepay their mortgage, dramatically reducing the earned reserves of the MMIF. Our main concern is that both HUD and Price Waterhouse continue to fail to address this issue. Even the Price Waterhouse FY 1990 Actuarial status report indicates that the upfront premium is not being recognized commensurate with the risk it represents in the early years of a mortgage. In Appendix B of the Price Waterhouse Actuarial Review dated March 1992, the Price Waterhouse model is used to estimate the appropriate refund of the unearned premium. After making this estimate, Price Waterhouse compared their estimate to the HUD refund schedule implemented in Mortgagee Letter 84-9 and found that their estimate of what should be refunded was approximately three-quarters of what HUD was actually refunding, based on Mortgagee Letter 84-9. The consequence of using an incorrect premium recognition schedule has been amply documented by GAO. NAR strongly believes that the estimated loss of $31 million in FY 1991 would have actually resulted in a profit if the premium had been recognized correctly. FHA Multifamily Programs FHA's principal multifamily rental housing programs for moderate-income families during the 1980s were the Section 221(d)(3) and Section 221(d)(4). FHA provided mortgage insurance to finance rental or cooperative multifamily housing for moderate-income households, including projects designated for the elderly. Under Section 221(d)(3), HUD could insure 100% of total project costs for nonprofit and cooperative mortgagors, but only up to 90 percent under Section 221(d)(4), the multifamily coinsurance program, irrespective of the type of mortgagor. Problems within the FHA coinsurance program have been well publicized. Loss reserves in the FHA General Insurance Fund (GI) were increased substantially because of the defaults in the Fund's multifamily coinsurance programs starting in 1988. Substantially all FHA coinsurance loans were sold through the issuance of mortgage-backed securities fully guaranteed by GNMA. Beginning in September of 1988, GNMA declared a number of major FHA lender/coinsurers in default of its obligations for GNMA securities backed by coinsured loans, causing FHA to assume full insurance responsibility and bear 100 percent of all insurance losses. Price Waterhouse, in the 1989 FHA Fiscal Year Financial Audit, identified poor underwriting, undercapitalization of coinsurance endorsers, poor economic conditions, and excessive concentration of risk among too few lenders, as the reasons for the failure of the FHA coinsurance program. In the audit, Price Waterhouse identified many weaknesses with HUD's control and monitoring of the FHA programs, and has yet to provide a non-qualified audit of FHA. What is unclear, however, is whether FHA's multifamily insurance programs fell on hard times due to the design of the FHA programs, poor evaluation of the risks inherent to insuring low- and moderate-income multifamily projects, or HUD's inability to properly manage these programs. HUD has indicated that its full insurance programs will continue to be available for creditworthy projects. HUD created its Delegated Processing Program (DPP) to enhance its ability to deliver full insurance commitments. Under DPP, HUD assigned to private lenders certain processing functions while retaining final authority over underwriting decisions. In February 1991, HUD issued a request for proposals (RFP) from lenders interested in becoming FHA-approved |