GAO HOME OWNERSHIP Limiting Mortgage Assistance Provided to Owners With High- Income Growth a 111111 142314 ~;Ao,‘l2~:Irl:I~-!~o-l 17 Resources, Community, and Economic Development Division B-211608 September 26,199O The Honorable Lloyd Bentsen Chairman, Committee on Finance United States Senate The Honorable Dan Rostenkowski Chairman, Committee on Ways and Means House of Representatives As required by section 4006(i) of Public Law 100-647,we are reporting on the recapture provisions for qualified mortgage bond and mortgage credit certificate assistance.The report discusseshow the recapture mechanism is likely to affect assistedhome owners and whether it may be expected to limit the home ownership subsidy provided to householdswith rapid income growth. Also, the report discussesalternatives to the recapture mechanism that might limit the subsidy provided to those with rapid income growth. Chapters 3 and 4 contain several matters for consideration by the Congress. We are sending copies of this report to the Senateand House committees responsible for tax and housing matters; the Secretary of the Treasury; the Secretary of Housing and Urban Development; the Commissioner of Internal Revenue;the Director, Office of Managementand Budget; and other interested parties. We will also make copies available to others upon request. This report was prepared under the direction of John M. Ols, Jr., Director, Housing and Community Development Issues,who may be reached at (202) 276-6526.Other major contributors to this report are listed in appendix IX. Y J.Assistant Dexter Peach Comptroller General ./ Executive Summary The Congressenacted legislation in 1988 to recapture someor all of the Purpose subsidy from first-time home owners who receive qualified mortgage bond assistance.As stated in the conferencereport, the purpose of the recapture mechanismwas to retrieve the subsidy from owners who experience rapid income increasesafter they purchased their assisted home and, as a result, did not need the subsidy to remain home owners. The Congressdelayed the effective date of the recapture provisions until 1991 to enable GAOto report on whether the enacted provisions will accomplish the act’s purpose and to identify alternative approaches. State and local government agenciesassist first-time home buyers by Background using qualified mortgage bond proceeds(often called mortgage revenue bonds) to provide fixed-payment mortgage loans at below-market interest rates. These loans lower owners’ monthly payments. Assisted buyers must meet income, home purchase price, and other eligibility standards. The recapture provisions were imposed to recover someor all of the assistanceif the owner’s income increasedbeyond certain levels. Assisted owners pay the recapture amount as part of their federal income tax for the year in which the home is sold. The recapture amount is basedon an assumedinterest rate reduction of 126 percentagepoints per year during the first 6 years of ownership. It increaseson a pro -- rata basis from $0 at the start of the first year to a maximum of 6.26 percent of the mortgage amount for a home sold after 6 years. The recapture amount is then phased out during years 6 to 10, after which no recap- ture amount is due. Also, the recapture amount may be reduced to $0 if the household’s income has not increased by certain amounts, as deter- mined through a complex income test. In any case,the recapture amount cannot exceedSOpercent of the gain on the sale. While the recapture provisions will recover someor all of the subsidy Results in Brief received by certain assistedhome owners, it doesnot treat all owners with income increasesequitably. For example, the 1.26 percentagepoint recapture rate may be larger than the interest-rate reduction owners received. Moreover, becausethe recapture amount is computed only after the subsidy is discontinued, someowners will continue to receive the assistanceeven after their income has risen to a level where they could make unassisted housing payments. Finally, assistedloans can be Page2 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Executive Summary made to those who could afford market-rate conventional loans with comparable terms, and therefore do not need mortgage bond assistance. Various ways exist to more closely tailor the amount recaptured to the interest-rate reduction received by the owner. A more fundamental changewould be to discontinue the assistancewhen the owner could afford to remain an owner without the reduced-interest-rate loan. Finally, agenciescould limit this assistanceto those who could not ini- tially afford a market-rate conventional loan with comparable terms. GAO’s Analysis The recapture formula estimates whether assistedowners would still meet income eligibility standards when they sell their home. If income is less than the estimated eligibility standard, the recapture amount is reduced or eliminated. The income increasesneededto pay any or the maximum recapture amount in any year depend on (1) median income in the area, (2) how the owner’s income compareswith the maximum income eligibility limits when the home was purchased, and (3) changes in household size. GAOdeveloped five hypothetical recapture situations. Under the most conservative one, income would have to increaseby about 18 percent for the household to pay the full recapture amount in the first year, and an average of 11 percent per year if the home was sold after the second year. The income increasesneededto pay the full recapture amount fur- ther decline from an averageof 9 percent per year after year 3, to an average of 6 percent per year near the end of year 10. For the other cases,the income increasesneededto pay the full recapture amount were almost always higher, and most were significantly higher than the most conservative case. Recapture Not Tailored to Many assistedowners will compute their recapture amount using a Assistance Received higher rate than the interest-rate reduction they receive. Others will use a lower rate. These situations occur becausemost assistedowners will have a mortgage interest reduction that is different from the recapture formula’s assumed 1.25 percentagepoint reduction. For example, in a 1988 study, GAOfound that one-quarter of the sampled buyers received interest rate reductions of three-quarters of a percentagepoint or less.* ‘Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GAO/ - 8_111, Mar. 28,1988). Page3 GAO/RCED-90-117 Recapturingthe Qualified MortgageBondSubsidy JCxecutlve Summary I For an $80,000 mortgage in which the owner received a l-percent reduc- tion in the mortgage interest rate, the owner could pay as much as $1,000 more in recapture than the benefit received. Subsidy Continues Even Becausethe recapture amount is phased out from years 6 through 10, When Income Increases owners with large increasesin income and/or those who receive assis- tance the longest have their recapture amounts reduced and eventually Significantly eliminated. This reduction occurs even if owners could afford unassisted housing payments. Also, a recapture amount that is phased out may induce assistedowners to stay in their home longer to avoid or reduce the recapture. As a result, the subsidy period is extended even if the owner could afford unassisted housing payments. Initially, Many Assisted Buyers can receive mortgage bond assistanceeven if they may have Owners May Have Been been able to afford housing payments on a market-rate loan of compa- rable terms without it. Using two standard mortgage qualification tests, Able to Afford Unassisted GAOfound that about half of the assistedhouseholdsmay have been Housing Payments able to afford the samehome with a market-rate fixed-rate mortgage. Limiting assistanceto only those who could not afford housing pay- ments on a conventional loan of comparable terms would be a better use of mortgage bond funds, Alternative to the An alternative to the recapture mechanism would be to increase an Recapture Mechanism owner’s monthly payment for principal and interest (up to the level determined by the market interest rate at the time the loan was made) when the owner’s income has increased sufficiently to afford those pay- ments. One advantage of this approach is that it halts assistancewhen it is no longer needed,rather than when the home is sold. However, this approach would require housing agenciesor their agents to periodically reexamine owners’ income and would make the program administration more complex and costly. The enacted recapture mechanism is preferable to no recapture mecha- Matters for nism at all becausesomeor all of the assistancecould be recovered from Congressional at least someof those who probably could have waited a short time and Consideration purchased a home without it. However, it is a relatively ineffective way to identify and recapture benefits from those who do not need continued assistancebecausethe recapture requirements are triggered solely by an owner’s decision to move. Page4 GAO/RCED-B&117 Recapturingthe QualifiedMortgageBondSubsidy ;. ExecutiveSummary If the Congressdecidesto changethe recapture provisions, it could either (1) keep the current mechanism with somerevision or (2) substi- tute a different approach.2Regarding the former, the Congressmay wish to (1) tailor the recapture amount more closely to the interest-rate reduction received and (2) eliminate the phaseout after year five so that those owners who continue to benefit from the assistancehave someor all of it recaptured. If the Congressdecidesto replace the recapture approach, it may wish to enact an alternative that ends assistancewhen owners can afford unassisted housing payments. The proposals discussedabove do not keep those buyers who could afford housing payments with a market-rate loan of comparable terms from initially receiving assistance.The Congressmay wish to require that housing agenciesinitially limit assistanceto buyers who could not otherwise obtain a market-rate loan with comparable terms without qualified mortgage bond assistance. The Association of Local Housing Finance Agencies,National Council of Agency Comments State Housing Agencies,and Department of the Treasury commentedon a draft of this report. The Association and the Council disagreed with GAO'S alternatives to the current recapture mechanism. They cited addi- tional administrative complexity of these approacheswith little, if any, perceived gain. GAO agreesthat administrative requirements would increase somewhat. The added requirements should be weighed against better achieving the stated congressionalgoals for the recapture mechanism. The Association and the Council also stated that an affordability test was not needed,Their disagreementscenter on their belief that current eligibility standards for income and home purchase price are sufficient and that GAO'S analyses are flawed. GAO conducted an additional anal- ysis and continues to believe that about half of the assistedbuyers may have been able to otherwise afford a market-rate loan. Therefore, GAO continues to believe that a conventional affordability test is needed. Treasury did not express support or opposition to GAO'S proposals but emphasizedthe complexity and administrative effort neededto imple- ment them. 2GAO's March 1988 report questioned the continued use of qualified mortgage bonds to assist first- time buyers. This report should not be considered as a change in that position. Page6 GAO/RCED-90-117 Recapturingthe QuallfledMortgageBondSubsidy contents , Executive Summary 2 Chapter 1 10 Introduction How QMB and MCC Programs Operate 10 Serving Lower-Income Households 11 Tightened Eligibility Standards and a Subsidy Recapture 12 Objectives,Scope,and Methodology 13 Chapter 2 16 Many Owners May How the Recapture Amount Is Calculated Combined Effect of Several Factors Determines the 16 17 Pay a Reduced Recapture Amount Due Recapture Amount State and Local Recapture MechanismsAre Apparently 24 Limited summary 24 Chapter 3 26 Changesto the Increasing Equitable Treatment in the Recapture Provisions 26 Recapture Provisions Subsidy Allowed to Continue Despite Owners’ Significant 31 to Promote More Income Increases Equitable Treatment A Prospective Affordability Test Is Needed 33 Conclusions 36 of Owners Matters for Consideration by the Congress 36 Agency Comments 36 Chapter 4 39 Alternatives That Substituting a Roll-Over Approach for the Recapture 39 Mechanism Would Better Identify Income-BasedVariable Payment Loans as an Alternative 42 and Deny Subsidy to to the Recapture Mechanism Those With Rapid Conclusions 43 Matter for Consideration by the Congress 43 Income Growth Agency Comments 44 Appendixes Appendix I: Example of How to Compute the Recapture 46 Amount Y Appendix II: Income IncreasesNeededto Pay the 47 Recapture Amount for an Area With a $26,000 Median Income Page6 GAO/RCED-B&117 Recapturingthe QualifiedMortgageBondSubsidy \ Cmtmta Appendix III: Recapture Mechanismsin Federal, State, 61 and Local Housing Programs Appendix IV: Expected Recapture Effects on Program 66 Participation and Tax Administration Matters Appendix V: Approach Used in Applying a Total-Debt- 69 Expense-to-IncomeTest to Assisted Buyer Data Base Appendix VI: CommentsFrom the Association of Local 62 Housing Finance Agencies Appendix VII: CommentsFrom the National Council of 72 State Housing Agencies Appendix VIII: CommentsFrom the Department of the 91 Treasury Appendix IX: Major Contributors to This Report 98 Related GAO Products 100 Tables Table 2.1: Summary of Recapture Requirements 17 Table 2.2: Dominant Recapture Provision Over Time for 21 HouseholdsA and B Table 3.1: Effect of Uniform Recapture Liability on 28 Assisted Owners Receiving Different Interest Rate Reductions Table 3.2: Appreciation Rates of Existing Single-Family 29 Homesfor SelectedMetropolitan Areas - 1981-88 Table IV. 1: Example of Present Value of QMB Assistance 66 With and Without Recapture Provisions Table V. 1: Comparison of QMB-AssistedBuyers to 60 Household Attributes in the Survey ResearchCenter Survey Figures Figure 2.1: Expected Recapture Amount for Hypothetical 19 HouseholdsWith Relatively Large Income Increases Figure 2.2: Average Annual Income IncreasesNeededfor 22 HouseholdsA and B to Pay the Maximum or Any Recapture Amount Figure 3.1: Alternatives to the Recapture Phaseout 31 Figure II. 1: Average Annual Increase in Income Neededto 48 Pay the Recapture Amount for Household C Y Page7 GAO/lWJD~117 Recaptuhtgthe QualifiedMortgageBondSubsidy Contents , Figure 11.2:Average Annual Increase in Income Neededto 49 Pay the Recapture Amount for Household D Figure 11.3:Average Annual Increase in Income Neededto 60 Pay the Recapture Amount for Household E Abbreviations FRlHA Farmers Home Administration GAO General Accounting Office HUD Department of Housing and Urban Development MC!C Mortgage credit certificate &MB Qualified mortgage bond Page8 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Page9 GAO/RCED-90-117 Recapturingthe Qualif¶edMortgageBondSubsidy Chapter 1 I Introduction A popular form of assistancefor first-time home buyers is a below- market interest rate mortgage loan financed through tax-exempt quali- fied mortgage bonds (QMB).’ A related form of assistanceis the mortgage credit certificate (MCC), which allows first-time buyers to take a credit on their federal income tax for a portion of mortgage interest paid. The Congresshas restricted the eligibility of those receiving QMB loans and MCCS to prevent upper-income households from receiving them. In 1988, the Congressfurther restricted home buyer eligibility require- ments. It also imposed a tax that, under certain circumstances,will recapture someor all of the assistanceprovided. However, the Congress deferred the effective date of the recapture provisions for loans to those made on or after January 1,1991, and required us to assessthe mecha- nism and alternatives that would accomplish its purpose. The issuer of QMBS, usually a state or local government housing finance How QMB and MCC agency, issuesthese bonds when it believes that sufficient demand by Programs Operate home buyers exists for the resulting mortgage funds. Home buyers are attracted to QMB loans becausethe interest rates are lower than those available on conventional loans.2Typically, the agency issuesbonds when it believes that the resulting &MB mortgage interest rates will be l- l/2 to 2 percentagepoints below the conventional mortgage rate for fixed-rate loans. Depending on the movement of conventional mortgage interest rates between the time the bonds are issued and the time the loans are made, the spread may be higher or lower than the agency attempted to achieve. The reduced payment from the below-market &MB loan benefits the buyer.” However, the effect of this benefit is somewhat reduced since buyers often pay one-time fees to the housing agency to help defray pro- gram coststhat a borrower in the conventional market would not pay. The prospective home buyer executesa purchase contract on a home and applies for assistedfinancing through a mortgage lender acting for the housing agency.The lender makes the fixed-rate QMB loan after determining that the purchaser has not had an ownership interest in a *These bonds are often called “mortgage revenue bonds.” ‘QMB funds may also be used for other kinds of loans, such as home improvement loans; however, home mortgage loans are by far the most prevalent form of assistance. 3We found the median reduction to be 1.44 percentage points from January 1983 to June 1987, or about $40 per month after taxes. See p. 38 of Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GA--B&l 11, Mar. 28, 1988). Page10 GAO/RCED-90-117 Recapturingthe QualUkl MortgageBondSubsidy chapter 1 Introduction principal residencefor the 3 previous years (“first-time buyer”), meets income and home purchase price requirements, shows repayment ability at the below-market rate set by the housing agency,is a good credit risk, and meets other qualifying requirements. MCCS, on the other hand, entitle buyers to take a credit against their fed- eral income tax liability (i.e., a dollar-for-dollar reduction in tax lia- bility) for a portion of their mortgage interest paid during the year.4The credit may be from 10 to 60 percent of the interest paid. In taking the credit, home owners must reduce the amount of the home mortgage interest deduction (i.e., a reduction from gross income) taken on their year-end income tax returns by the amount of the credit. If the home owner’s federal income tax liability is less than the amount of the credit, the balance of the credit is lost to the owner in that year (a “nonrefund- able” credit). However, the owner can carry forward any unused por- tion of the credit for 3 years. Loans with MU33 are made through a participating mortgage lender at the market interest rate. The lender checks to make sure that the buyer can afford the home and qualifies the buyer for both the mortgage loan and the certificate. Lenders may or may not use the expected credit to qualify the buyer. Buyers also pay a fee to help defray the cost of the program. The Mortgage Subsidy Bond Tax Act of 1980 imposed the first statutory Serving Lower-Income restrictions on the ability of state and local governments to issue tax- Households exempt bonds for financing home mortgages.Generally, these restric- tions limited (1) the dollar volume of bonds that could be issued; (2) those assistedto first-time buyers; and (3) the purchase price of the assistedhome to no more than 90 percent of the area average.In drafting the 1980 act, the Congress,while expecting that lower-income householdsbe the primary beneficiaries, largely allowed the bond issuers to determine the proportion of proceedsthat would be used for lower-income households. The Congressrecognizedthat qualified mortgage bond programs operate in housing markets that differ considerably in affordability, and the 4QMBs are one type of *called “private activity bonds.” The Internal Revenue Code (the Code) limits the total amount of private activity bonds that can be issued each year. In lieu of issuing bonds, the agency may convert some or all of its unused bond issuance authority to authority to issue a certain volume of MCCs. Page11 GAO/RCED-O-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter1 , Introduction broad eligibility standards that it enacted gave state and local govern- ments wide flexibility in structuring their programs. Through legislative changesin 1982 and 1986, this flexibility has remained, although the Congresshas attempted to tighten eligibility requirements by adding a requirement that, generally, assistedbuyers’ household income cannot exceed 116 percent of area median income. With the Technical and MiscellaneousRevenueAct of 1988 (P.L. lOO- Tightened Eligibility 647, Nov. 10, 1988), the Congresstightened eligibility requirements once Standards and a again in responseto evidencethat the (1) program was serving a higher Subsidy Recapture proportion of single-personhouseholdsthan the first-time home buyer market servesgenerally and (2) income and purchase price eligibility requirements did not reflect the significant differences in which first homes are affordable in different housing markets.6The act added a household size adjustment to the income eligibility standard. Under the act, household income for household sizesof one or two persons cannot exceed 100 percent of the area median income. The income eligibility standard for householdsof three or more persons remains at the 1986 level of 116 percent of area median income. In addition, on a formula basis, the income eligibility standards are adjusted upward or down- ward in areas of high or low housing costs,respectively. The act also imposes a tax on assistedhouseholdsthat is designedto recapture all or a portion of the expected subsidy from someQMB- or Moo-assistedowners. The recapture tax was added becausethe Congress believed that “in those &MB- and MCC-assisted households where income has risen rapidly since acquisition, the special subsidy provided by the program was not necessary in order to become or remain a homeowner.“0 6Miscellaneous Revenue Act of 1988, Report of the Committee on Ways and Means, House of Repre- sentatives, U.S. Congress. Report 160-o-796, July 26, 1988, pp. 434-6. See also (GAO/RCEDS&111, Mar.28, 1988) and Home Ownership: Targeting Assistance to Buyers Through Qualified Mortgage Bonds (GAO/RCEr, June 27,1988). eTechnica1 and Miscellaneous Revenue Act of 1988, Conference Report to Accompany H.R. 4333. House ot Hepresentatives, U.S. Congress. Report 100-1104, Vol. II, Oct. 21, 1988, p. 86. The House legislative report noted that while home ownership provides the owner both consumption benefits and investment benefits (through house price appreciation), the recapture mechanism was intended to return the consumption assistance received through the reduced-interest QMB loan or MCC tax credit. See Miscellaneous Revenue Act of 1988, p. 436. Page12 GAO/RCED-SO-117 Recapturingthe QualifiedMortgageBondSubsidy chaptm 1 Intmduction While the revised eligibility provisions under the 1988 act becameeffec- tive for bonds issued (and elections not to issue bonds in favor of MCCS) after December31,1988, the Congressdelayed the effective date for the recapture provisions until January 1,199l. In the interim, the act requires that we study the enacted recapture mechanism and alterna- tives to it. Objectives, Scope,and Section 4006(i) of the 1988 act requires that we report on the enacted provisions and any alternatives that would accomplish the purposes of Methodology the section. Accordingly, we studied the formula contained in the enacted legislation and alternatives in terms of how they would be likely to achieve the conferencereport’s stated objective of identifying those with rapid income increasesand recapturing the assistanceprovided by the QMB loan or MCC.' These alternatives include potential modifications to the legislated provisions and substitution of different approachesfor achieving the Congress’goal. We also assessedwhether the recapture mechanismor other legislative provisions would prevent those who could afford a market-rate loan from initially receiving a QMBloan or MCC. To determine the recapture requirements for QMB loans and MCCS, we reviewed the pertinent sections of the Technical and MiscellaneousRev- enue Act of 1988 and the legislative reports leading to it. The Internal RevenueService, which has primary regulatory responsibility, told us that the Service plans to have taxpayer guidance completed by December1990. To determine how two major federal home ownership programs struc- ture their recapture mechanisms,we reviewed laws, regulations, and program documents, and met with officials at the Department of Housing and Urban Development (HUD) for its section 236 program and the U.S. Department of Agriculture for the Farmers Home Administra- tion’s (FmHA)section 602 program. We also obtained information on the extent that recapture mechanisms are used by state and local housing agenciesand whether potential overlap with the federal recapture mechanism may exist. For state agency practices, we obtained a summary of a 1989 survey of member 70ur March 1988 report concluded that the Qh%B-financingmechanism is costly, serves mostly buyers who could afford homes otherwise, and does little to increase affordability for low- and moderate- income buyers (see (GAO/RCED-S%111, Mar. 28,1988)). This report primarily limits its scope to recapture and related issues. Pagela GAO/RCJ3D-90-117 Recapturingthe Qualif’iedMortgage BondSubsidy Chapter1 Introduction agenciesperformed by the National Council of State Housing Agencies. To learn about local agency recapture practices, we mailed a survey instrument or telephoned 128 local housing agenciesthat are members of the Association of Local Housing Finance Agencies. We reviewed economicand housing literature related to possible alterna- tive recapture mechanismsand topics related to the behaviors of first- time buyers. We performed our work between January and August 1989. We sent a draft of this report to the Association of Local Housing Finance Agencies,National Council of State Housing Agencies,and Department of the Treasury for their review and comment. The Associa- tion and the Council disagreed with most aspectsof our draft report. Treasury did not express support or opposition to GAO'S proposals but suggestedadditional discussion about the complexity and administrative effort neededto implement the proposals. Their comments are repro- duced in appendixes VI through VIII, respectively. Our evaluation of their comments is set out at the end of chapters 3 and 4 and appendixes VI through VIII. On the basis of those comments,we performed an addi- tional test on the extent to which QMB loan recipients may have been able to purchase the samehome with a market-rate loan (seech. 3 and wp. VI. Page14 GAO/NED-90-l17Recapturingthe QualifiedMortgageBondSubsidy Chapter 2 Many Owners May Pay a Reduced Recapture Amount The recapture provisions are a set of complex formulas that are applied when the owner disposesof the assisted home.’ The provisions are designedto recoup the assistedowner’s assumedinterest rate reduction if the owner sells the home during the first 6 years of home ownership and if income has increasedbeyond established limits. After 6 years, the recapture amount is gradually phased out until, at the end of the tenth year, no recapture amount is due, The increasesin income neededto pay the maximum, or even any, recapture amount during this period vary by year and household situa- tion. However, in the initial years the increasesneededare substantial, suggestingthat many assistedowners may pay a reduced recapture amount or, perhaps, none at all. And, those who sell their assistedhome after 6 years will have the recapture amount phased out. Therefore, we would expect that many owners, including those whose income increased rapidly after they purchased the assistedhome, will pay less in recapture than the benefit that they received. Somestate and local housing agenciesimpose forms of recapture that are independent of the federal recapture. However, the incidence of these recaptures is apparently limited. Therefore, few owners may be subject to multiple recaptures. The recapture provisions impose a potential federal income tax liability How the Recapture on those whose income grows after they purchase the assistedhome. Amount Is Calculated The recapture amount is due for the tax year in which the assistedhome is sold. The maximum recapture amount is 6.26 percent of the owner’s original mortgage amount (from QMB loans or MCCS) if the home is sold 6 years after it is purchased. The recapture amount is reduced to reflect periods of time shorter or longer than 6 years (the holding period adjust- ment). In addition, adjustments for (1) gain on sale of the assistedhome and (2) changesin household income may reduce or eliminate the amount. Holding Period Adjustment The recapture amount increaseson a --pro rata basis from $0 to 6.26 per- cent of the assistedmortgage amount over the first 6 years, according to the number of months from date of purchase until the home is sold, and Y similarly decreasesevenly during years 6 to 10. For example, after 1 ‘Dispositions are generally sales, exchanges, or gifts. We refer to dispositions as “sales” throughout the report. Page15 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy , Chapter2 , ManyOwnen May Paya Reduced RecaptureAmount year the maximum recapture tax is 1.26 percent of the assistedloan amount (6.26 x 12/60 months); after 2 years the maximum tax is 2.6 percent of the assistedloan amount, and so on, up to 6 years.2 Gain-On-SaleAdjustment The recapture amount cannot exceed60 percent of the gain realized on the sale of the home. The gain is computed by subtracting the basis of the home (purchase price of the home plus cost of improvements and certain purchase costs) from the selling price, less certain selling expenses.For example, if the gain on the sale of the home was $6,000, the amount recaptured could not exceed$2,600. Income Adjustment The general nature of the income adjustment is to estimate, at time of sale, whether the assistedowner would still meet the income eligibility limits for a QMB loan or MCC at that time. If the assistedowner’s house- hold income at time of sale is higher than the estimated eligibility limits, then no reductions to the recapture amount (becauseof income) apply. If the owner’s household income is less than the estimated eligibility limit, then the recapture amount is reduced to a lesser amount, as described below. The income adjustment provides that the recapture amount is reduced by 2 percent for every $100 by which the assistedowner’s income (adjusted gross income plus any tax-exempt interest) is less than the estimated income eligibility limit. The estimated income eligibility limit is the highest household income level that would have met the Code’s income eligibility limit in the area where the buyer lived when the loan was made, plus $6,000. The highest household income level is deter- mined after two adjustments. The first adjustment is to reflect the mortgagor’s family size on the date of sale. Generally, the income eligibility limit is 116 percent of area median income for householdsof three or more persons and 100 percent of area median income for householdsof one or two persons.3The secondadjustment is to increasethe highest household income level by 6 ZTax writers estimated that the typical after-tax benefit (reduction in mortgage interest rate) seen by assisted buyers was 1.26 percent per year. This estimate was based on a typical before-tax reduction in our March 1988 report. See p. 38 of Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GA& 31f the seller’s household size changes, say from a two-person household at time of purchase to a three-person household at time of sale, then the qualifying income dollar amount is based on the 116 percent standard, rather than the loo-percent eligibility standard, and vice versa. Page10 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter2 ManyOwneraMay Paya Reduced RecaptureAmount percent (simple interest) for each full year the assistedloan is out- standing. The S-percentincrease is a proxy for the amount that income eligibility limits might increaseeach year. The Code also contains several exceptions and special provisions to the computation. Table 2.1 summarizes the primary recapture requirements, and appendix I provides a detailed example of a recapture calculation. Table 2.1: Summary of Recapture Requlrement8 Condition Provlrrion To whom recapture applies 7; T,rs who receive QMB loans or MCCs starting January 1, !i When recapture is due tfh;i;zein which the owner sells or otherwise disposes of How much is recaptured The maximum amount is 6.25 percent of the assisted mortaaae amount. Reductions and to limitations The recapture amount increases from $0 the maximum amount over the first 5 years of ownership, and then decreases to $0 after 10 years. The recapture amount cannot exceed 50 percent of the gain realized on sale. The recapture amount is reduced if household income is less than the estimated income eligibility standard, subject to household size and income change adjustments. How recapture is paid Recapture is part of the mortgagor’s individual income tax liability. Other provisions The housing agency must inform the mortgagor of the potential recapture amount within 90 days of settlement. Home improvement loans are excepted from recapture. Other requirements, including limited exceptions to those subject to recapture, and refinancing were also enacted. Source: Section 4005(g) of P.L. 100647, IRC Stn. 143(m). This section examines how the reductions and limitations affect the Combined Effect of recapture amount due on the basis of a hypothetical example. The Several Factors income test and the gain-on-saletest may reduce the recapture amount Determines the due for householdsthat have relatively large increasesin income during the first 3 to 4 years. After the fifth year, the holding period adjustment Recapture Amount reduces and eventually eliminates the amount due for all assisted Due owners. Figure 2.1 shows the maximum recapture amount that would be col- lected over a lo-year period and comparesthat amount with the amount that would be collected for two householdsthat have relatively large Page17 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter2 I ManyOwnersMayPaya Reduced RecaptureAmount increasesin income. For this comparison, we constructed two hypothet- ical assistedhouseholds.Both householdslive in an area with a median income of $34,000. Both householdspurchase $94,600 homes and obtain assistedloans of $85,000.4Both householdshave three members and both experiencerelatively large annual increasesin income of 10 per- cent ann~ally.~Both households’ homes have a 7-percent appreciation rate, which is about equal to the nationwide average for existing single- family homes for the 1979-88period? 4The numbers chosen are illustrative of recapture provisions only. Incomes and house prices vary widely across the nation. For this example, the area median income chosen is about equal to HUD’s estimate of the expected 1989 nationwide household median income and its estimate of statewide household median income for Ohio, Washington, and Wyoming. 6These increases are larger than the &percent increase in the qualifying income dollar amount built into the recapture formula income test. They are also larger than the nationwide average increase in nominal income for all home owners over the last 10 years that we computed from Appendix Table I of the Joint Center for Housing Studies’ The State of the Nation’s Housing, 1989. 6National Association of Realtors’ data show an average annual appreciation of 6.7 percent on the sale of existing homes nationwide from 1979-88. To compute the net gain, we assumed that selling expenses, such as realtors’ commissions, transfer taxes, and other selling expenses, were 8 percent of the sales price. Page18 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter2 ManyOwnenMay Paya Reduced RecaptureAmount Figure 2.1: Expected Recapture Amount for Hypothetlcal Howehold, Wlth Relatively Large Income Increaw8 5500 Amount Recrptumd (Dollars) 1,‘; : ‘Z 4400 i , f’ ,i’ i 1’ I’ / aaoo / : / I’ /’ : /I’ I’ 2200 1’ P ,,’ : 1100 , ,’/’ r’ ,/ r’ i r’ , / z 0 ” / ---a---- 0 1 2 a 4 5 9 7 8 9 10 Yearn - Household A (115% of areamedian income) --I.- Household B (100% of areamedan income) m Maximum Amount From Holding Period Adjuabnent Note: After the Household A and B lines intersect the maximum tax line, the amount due in this example is determined by the holding period adjustment. The recapture amount is based on an $85,000 assisted loan amount. For other details, see text. The householdsdiffer only in their income at the time that they purchase their assistedhomes. Household A has the maximum income to be eligible to participate, $39,100 (116 percent of the area median income) and Household B’s income equals the area median, or $34,000. As the thicker solid line in figure 2.1 shows, the recapture amount due before the income and gain-on-saleadjustments reachesits apex at the end of the fifth year and decreasessymmetrically through the tenth year of ownership. This line also shows that, for a household that obtains an $86,000 loan, the maximum recapture amount rises from $0 when the home is purchased, to $6,313 at the end of year 6, and then declines at the samerate until the end of year 10, after which no amount is due. Page19 GAO/RCED-99-117 Recapturiugthe QuaUfledMortgageBondSubsidy Chapter2 ManyOwnersMay Paya Reduced RecaptureAmount Different Recapture Figure 2.1 also shows the consolidated effects of the income and gain- Provisions Dominate Over on-saletests and the holding period adjustment on HouseholdsA and B, both of which bought homesunder identical terms and have relatively Time large annual increasesin income (10 percent annually). Figure 2.1 shows that, becauseof the income and gain-on-saletests, even those with lo- percent income increasesand averagegains on sale are likely to pay less than the maximum recapture amount if they sell their home before the end of the third year in the caseof Household A and before the end of the fifth year, for Household B.7For example, Household A’s income must rise by 18 percent in the first year to pay the full recapture amount.sSincewe assumeda lesser increase of 10 percent, the income adjustment would reduce the recapture amount due. Also, since we assumedthat selling expenseswould be 8 percent of the salesprice and the home appreciated 7 percent in the first year, no gain on sale would result in the first year, and, hence,no recapture amount would be due even if income had risen by 18 percent or more. In addition, figure 2.1 shows that the recapture amount due dependson the difference between household income at the time of sale and the esti- mated income eligibility standard under the recapture formula. That is, for determining the recapture amount, an assistedowner’s increase in income is measured against the maximum qualifying income. The lower the income of the household when buying the assistedhome in relation to the income eligibility standard, the more income has to rise to be sub- ject to the maximum or any recapture amount. Therefore, the income increasethat Household B must have before it pays any or the max- imum recapture amount is greater than that for Household A since (1) both are tested against the 116-percenteligibility amount of $39,100 and (2) Household B has a lower starting amount.0 Table 2.2 shows in dollar values the effect of the holding period, income, and gain-on-saleprovisions in determining the recapture amount due for HouseholdsA and B. For most years, the holding period adjustment determines the recapture amount due. ‘Two major federal home ownership programs, HUD’s section 236 and FmHA’s section 602 programs for low- and moderate-income buyers, have recapture provisions. Neither recapture provision makes ah adjustment for changes in the assisted owner’s income in calculating the recapture amount due. (See app. III.) ‘($39,100 x 1.06) + $6,000 = $46,066, which is 18 percent greater than the $39,100 income eligibility limit. ‘The results would have been the same if Household B was a one- or two-person household when it bought the assisted home and a three-person or larger household when it sold it since the qualifying income dollar amount is calculated on the basis of household size at sale. Page20 GAO/RCED-W-117 RecaptwAng the QuallfledMortgageBondSubsidy Chapter2 ManyOwnersMayPaya Reduced &xaptureAnlouut Table 2.2: Domlnant Recapture Provirlon Over Tlmo for Households A and B Maximum amount based Hourehold A Hourehold B on holding Amow; d; Amouum;d; End of period Dominant Dominant year ad)urtment provis Bon8 provision provis Pon8 provision 1 $1,063 $0 GOS $0 GOS,INC 2 2,125 1,628 INC 0 INC 3 3,168 3,188 HPP 184 INC 4 4,250 4,250 HPP 2,430 INC 5 5,313 5,313 HPP 5,313 HPP 6 4,250 4,250 HPP 4,250 HPP 7 3,188 3,188 HPP 3,188 HPP 8 2,125 2,125 HPP 2,125 HPP 9 1,063 1,063 HPP 1,063 HPP 10 0 0 HPP 0 HPP $E%olciing period provision INC - income provision GOS - gain-on-sale provision After the fifth year, the holding period adjustment reducesthe recap- ture amount to levels below the assumedbenefit seenby the assisted owner. The act’s legislative history doesnot indicate why this phaseout was enacted. Income Increases Needed Figure 2.2 shows the annual averagepercentageincreasesin income by Owners to Pay the neededin any year to pay both the maximum recapture amount and any recapture amount. For Household A, the increaseneededto pay the Maximum or Any maximum recapture amount declines from 18 percent for a home sold at Recapture the end of the first year to just over 6 percent for a home sold before the end of the tenth year. The declining increasesneededto pay the recap- ture amount are due to the declining influence over time of the $6,000 amount added to the adjusted qualifying income, since adjusted quali- fying income is increasedby 6 percent (simple) each year. If the $6,000 amount was eliminated from the formula, income increasesneededto pay maximum recapture would be constant, that is, 6 percent in any year. Page21 GAO/RCED&L117 Recapturhgthe QualifiedMortgageBondSubsidy chapter 2 Many OwneraMay Paya Reduced RecaptureAmount Figure 2.2: Average Annual Income Increare8 Needed for Houwholds A and 40 Average Pmconl Incroau in Income B to Pay the Maxlmum or Any Recapture Amount 35 ’ 30 0 1 2 3 4 6 6 7 8 9 10 YOtMO - Full recapture amount for household A - - - - Any recapture amount for household A m Full recapture amount for household B n n mn Any recapture amount for household B Note: Assumes an area median income of $34,000. When it bought the assisted home, Household A’s income was 115 percent of area median income, or $39,100. Household B’s income matched the area median. See text for details. For Household B, the annual income increase neededto pay the max- imum recapture amount ranges from about 36 percent for a home sold after the first year to almost 9 percent for a home sold just before the end of year 10. The results for Household B are larger than for House- hold A because(1) HouseholdsA and B are both tested against the 116- percent eligibility amount of $39,100 and (2) Household B has a lower starting point than Household A. Of course, for both HouseholdsA and B, the recapture amount computed just before the end of year 10 will be negligible becauseof the holding period adjustment. For Household A, the income increaseneededto pay any recapture amount is slightly more than 6 percent per year. Again, the result is different for Household B with the average annual income increases neededto pay any recapture amount decreasingfrom 21 percent at the end of the first year to about 7 percent near the end of the tenth year. Page22 GAO/RCED-!30-117 Recapturingthe QualifiedMortgageBondSubsidy chapter 2 ManyOwnersMayPaya Reduced RecaptureAmount But, the reasoning is the sameas for the determination of income increasesneededto pay the maximum recapture amount. The results that we present here depend upon a number of factors. While the samegeneral results will hold for several different assump- tions, the dollar values derived depend on initial values, such as the area median income. The lower the initial income, the larger the percentage increase in income that must occur to pay the maximum recapture amount. Again, this is the result of the $6,000 constant.1oSeeappendix II for an example of recapture amount calculations using lower area median income levels. Many Assisted Owners In our March 1988 report, we found that 66 percent of the assisted Could Pay Less Than the buyers in our sample had incomes at or below area median income.ll Also, as discussedabove, those whose incomes at time of purchase were Maximum Recapture equal to the area median must have quite large income increasesbefore Amount they pay the maximum, and, in somecases,any recapture amount. Therefore, many assistedbuyers will probably pay less than the max- imum recapture amount. Figure 2.1 and table 2.2 showed how the recap- ture amount would be reduced by the income test for owners whose income was at 100 percent of the area median income level, rather than the maximum of 116 percent of area median income. Appendix II takes this analysis one step further by comparing income increasesneededto pay the maximum or any recapture amount when the owner’s income at time of purchase was 80,100, and 116 percent of area median income, and it uses a lower area income. *‘For example, assume median incomes of $24,000 for Area X and $34,000 for Area Y. Household income in Area X would have to rise by 21 percent to reach $29,000 (Area X median income plus $6,000) and by only 16 percent to reach $39,000 in Area Y (Area Y median income plus $6,000). “Based on 136,000 loans made during the January 1983 through June 1987 period. See (GAO/ RCED88-111, Mar. 28,1988), p. 81. Page23 GAO/RCED-90-117 Rec~~pturing the Qualified MortgageBondSubsidy Ch8ptar2 MauyOwnemMayPaya Reduced RecaptureAmount Several state and local governments impose someform of recapture on State and Local buyers whom they assist. However, these are limited, and any overlap Recapture that may exist between them and a recapture of QMB loan or MCC assis- Mechanisms Are tance is apparently quite small. Apparently Limited A 1989 survey by the National Council of State Housing Agenciesof its national membership showed that eight states had someform of recap- ture mechanism as part of one or more of their home ownership pro- grams. None of the eight members reported that it applies recapture to its QMB loans. Rather, the recapture mechanismsappear to be used only with special programs of limited duration or availability. In 1989, we surveyed 128 membersof the Association of Local Housing Finance Agencies for the samepurpose. Becausenot all local housing agenciesbelong to the association,our survey may not cover all recap- ture approachesin use. Twenty of the responding agenciesindicated that they employed a recapture or recapture-like mechanism.12In a number of cases,these mechanismswere limited to a single development or small set of properties rather than applying to the agencies’housing programs in general. In only a few casesdid the respondents indicate that the recapture mechanismswere applied to QMB loans and/or MCCS, although survey results did not provide details on these situations. Consequently, the home owner, in rare instances,may be subject to mul- tiple recaptures. A multiple recapture could be avoided if state and local governments recaptured only the enhancementsthey add to the QMB loans or MCCS (such as closing cost or down payment assistance). Becausethe federal recapture provisions are effective for loans made starting January 1, 1991, none of the home owners currently receiving state or local assistanceand subject to a state or local recapture provi- sion will be affected by the federally enacted recapture provisions. A description of the state and local recapture mechanismsidentified by the two surveys, as well as descriptions of recapture of subsidies in HUD'S and F&A'S home ownership programs, are contained in appendix III. The recapture provisions are a complex set of formulas that impose a Summary potential tax liability on assistedowners whose income grows after they Y 12About 91 percent of the 128 members surveyed responded. Also, one of the respondents wss not an Association member, but returned our survey when another agency sent the questionnaire to it. Page24 GAO/RCED-90-117 Recapturingthe Qualifbd MortgageBondSubsidy purchase the assistedhome. Many of those who sell their homes during the first 6 years may have the recapture amount reduced from the max- imum amount in any year becauseof the income test and the gain-on- sale test. During years 6 to 10, the legislated recapture formula reduces the amount to be recaptured and the recapture amount becomes$0 at the end of the tenth year. Although somestate and local agenciesimpose someform of recapture, the number of owners subject to state and local recapture is apparently quite limited. Therefore, few assistedowners may be subject to a mul- tiple recapture when the QMB loan and MCC recapture provisions become effective starting January 1, 1991. Page 26 GAO/RCED-90-117 Recapturing the Qualifkl Mortgage Bond Subsidy , Chapter 3 Changesto the &capture Provisions to Promotk More Equitable Treament of Owners The recapture provisions will identify and impose a potential recapture payment on certain householdswhose incomesincrease after they pur- chasedtheir assistedhomes.However, several features treat assisted owners differently. For example, the recapture provisions can impose the samerecapture amount on householdsthat receive different amounts of assistance. Further, the recapture provisions do not addresstwo fundamental weaknessesin how QMB loan and MCC assistanceis provided: (1) assis- tance can be provided to those who initially could afford market-rate housing payments without &MB loan or MCC assistanceand (2) once obtained, the assistanceis continued regardlessof the owner’s ability to support unassisted housing payments. Several revisions could be made readily to the current recapture mecha- nism to treat assistedowners more similarly and to ensure that only those who cannot support unassisted housing payments, initially, receive assistance.However, discontinuing assistancewhen the owner could support unassisted housing payments would require a funda- mental changein how assistanceis delivered. Equitable treatment between taxpayers is one goal of tax policy. Two Increasing muitable basic conceptsof equity are the “benefits principle” and the “ability-to- Treatment in the pay principle.” Under the benefits principle, equity is fostered when Recapture Provisions those (1) who benefit more pay more in taxes than those who benefit less and (2) with equal benefits are taxed at the samelevel, all else being equal. However, the recapture legislation contains provisions that treat differently situated assistedhouseholds similarly and similarly situated householdsdifferently. As a result, assistedowners may be treated inequitably. Under the ability-to-pay principle, equity is fostered when those with greater financial resourcesare taxed more than those with fewer resources.As applied to the recapture provisions, the income adjust- ment provision is generally consistent with this principle, while the gain-on-saleprovision is not. Page26 GAWBcEDI)o-117 Becspturinethe QualifiedMortgageBondSubsidy Chapter 8 Cbangeato the RecapturfProvlsionato PromoteMoreEquitableTreatment of Ownen Single Recapture Rate Not The recapture formula is not fully consistent with the benefits principle Fully Consistent With the since it uses a single rate that is equal to the presumed benefit, rather than the actual benefit received, as the basis for the recapture amount1 Benefits Principle In one respect, this approach satisfies equity principles in that assisted owners who have larger loans are subject to a larger recapture amount than those with smaller loans, all else being equal. However, this approach raises other equity considerations because many assistedowners will receive larger or smaller interest-rate reduc- tions from QMB loans than the yearly after-tax 1.26 percentagepoint amount imbedded in the recapture provisions. For example, in March 1988 we reported that, while the median before-tax reduction was 1.44 percentagepoints, one-fourth of the assistedbuyers we studied received reductions of about three-quarters of a percentagepoint or less from the conventional interest rate.2Similarly, the MCC credit may give a larger or smaller benefit than the after-tax surrogate 1.26 percentagepoint figure, becauseof either the nonrefundability of the credit or the credit rate the MCC bears (see ch. 1). The results of the disparity between the uniform recapture rate and the actual benefit received, all else being equal, is quantified in table 3.1 for three levels of interest-rate reductions.3For example, an assistedowner with an $80,000 QMB loan would be liable for a maximum of $6,000 in recapture amount on the basis of the legislated formula, after 6 years. However, if the owner received an after-tax benefit of 1 percent, then the owner would pay $1,000 more in recapture amount than the benefit received, Similarly, an owner who received an after-tax benefit of 1.S percentagepoints would pay $1,000 less in recapture amount than the benefit received. ‘Home ownership typically conveys several benefits, such as increased equity as a result of house priceappreciation. For this discussion, we use the term “benefits” to mean the interest-rate reduction from a QMB loan and the tax credit received from an MCC by the assisted owner. 2Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GAO/ 88-l 11, Mar. 28, 1988), p. 38. 3This exampleserves as an illustration, only. Therefore, it does not consider the time value of money from the time the assistance is received to the payment of the recapture amount. Page27 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter3 Changesto the RecaptureProvisionato PromoteMoreEquitableTreatment of Ownera Table 3.1: Effect of Uniform Recapture Llabillty on Assisted Owners Recelvlng Additional amount recaptured (not Different Interest Rate Reduction8 reca tured) based on an after-tax Amount of assisted Recapture Pnterest rate reduction of mortgage amount* 1% 1.25% 13% $40,000 $2,500 $500 $0 ($500) 8Qooo 3,750 750 0 (750) 80,000 5,000 1,000 0 (1,0001 @Assumes the home is sold at the end of 5 years and no other reductions apply To increase equity of treatment between assistedowners, the recapture provisions could be changedto basethe recapture amount on the assistedowner’s monthly payment reduction. For example, the partici- pating lender could document the before-tax monthly payment reduc- tion due to the QMBloan at the time the loan is closed.This document could state the difference in monthly principal and interest payments between a comparable conventional loan and the QMBloan of the same size. At time of sale, the maximum amount to be recaptured would be the dollar amount of the reduction (say, $60 per month) times the number of months that the assistedloan is held. For a loan with a MCC, the assistedowner could tally the credits taken from prior years’ tax returns and use that amount as the basis for the recapture amount. As with the current recapture provisions, the maximum recapture amount could be capped or phased out, if so desired.4 Moves for Other Than Householdsmove for a variety of reasons,including the desire to live in Housing ReasonsAre Less a larger home or reside in another neighborhood. Here, the assisted owner’s choice is purely voluntary. Other situations, such as moving for Consistent With the employment purposes, are also voluntary, but are not dictated by a Ability-To-Pay Principle changein housing choices.However, under the first situation, an assistedowner may be more likely to time the move to minimize the recapture amount (or accumulate the financial resourcesto pay it), all else being equal. This may not be the situation when the move is made for reasonsother than housing choice. From this standpoint, the recapture provisions may not be fully consis- tent with the ability-to-pay principle since householdsthat move for other than housing reasonsmay not have set aside financial resources 41ncommenting on a draft of this report, Treasury suggested that, in theory, the appropriate policy would be to recapture the benefit only for the years in which the assisted owner’s income exceeded maximum qualifying income. Treasury also acknowledged that this would require constant moni- toring of owners’ incomes, which would increase administrative and compliance costs. Page28 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chaptara changeeto the ReeaptlmProvMousto PromoteMoreEquitableTreatment of Ownen for the changein ownership. However, in this casethe decision to move and pay a recapture amount is still voluntary. This situation is not easily remedied. It would not be equitable to forgive the recapture amount for a voluntary move. Alternatives that do not present this problem are presented in chapter 4. Gain-On-SaleProvisions The gain-on-saleprovision fulfills a purpose becauseit leaves a portion Are Not Always Consistent of the home owner’s gain on sale available for the purchase of a replace- ment home. However, the provision is not always consistent with the With the Ability-To-Pay ability-to-pay principle when householdswith rapid income increases Principle escapepaying recapture becausethey live in areas of low or no housing price appreciation. Under law, the recapture amount cannot exceed60 percent of the gain on the sale of the assistedhome. In effect, this limita- tion leaves someof the equity buildup due to appreciation with the assistedowner, which then can be applied toward the purchase of a replacement home.6However, different areas of the country and dif- ferent segmentsof a local housing market experience different rates of appreciation (seetable 3.2 for examples). Equity considerations occur becausethe gain-on-saleprovisions may reduce or eliminate the recap- ture amount for an assistedowner with a rapid increase in income if the home does not appreciate. Another household with similar income increaseswould pay a greater recapture amount if it had a greater gain on sale. fable 3.2: Appreclatlon Rater of Exirting Single-Family Homes for Selected Median sales price Metropolitan Areas - 1981-86 (Percent appreciation from 1981) Metropolitan area 1981 1984 1986 1988 $93,000 $101,200 $132,500 Washington, DC $88,300 (5.3) (14.6) (50.1) 65,100 68,300 75,700 Birmingham, Ala. 59,200 (10.0) (15.4) (27.9) 68,200 69,900 74,500 Milwaukee, Wis. 64,500 (5.7) (8.4) (15.5) 63,900 63,000 56,200 Oklahoma City, Okla. 54,100 (18.1) (16.5) (3.9) 73,500 70,900 64,700 Baton Rouge, La. 69,600 (5.6) (1.9) (-7.0) Source: GAO analysis of National Association of Realtors data “HUD’s section 236 and FmHA’s section 602 home ownership programs also limit recapture amounts based on gain on sale (see app. III). Page29 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter3 Changesto the RecaptureProvleionsto PromoteMoreEquitableTreatment of Owners Similar assistedowners would pay different recapture amounts depending on whether their homes appreciated markedly or had low appreciation. In addition, two assistedowners in the samemarket area, one with high income growth and the other with lower income growth, could pay a small recapture amount in areas of low appreciation. This meansthat those who experience rapid increasesin income (but who live in a market area with little or no appreciation) not only pay a reduced recapture amount, but may have less need for the appreciation equity if they chooseto buy a replacement home in the samemarket area, in contrast with an assistedhousehold in higher appreciation mar- kets experiencing a smaller income growth. By allowing owners to retain someequity for a replacement house purchase, the gain-on-saleprovision fulfills a purpose: it does not force these owners to return to renter status becauseof a large recapture amount. As such, the above equity considerations are not easily reme- died, Alternatives to the current recapture provisions that do not create this equity consideration are discussedin the next chapter. Phaseout Provisions Not Although the legislative history is silent on the intent of the phaseout of Consistent With the the recapture from years 6 to 10, it is an explicit part of the recapture formula. However, the phaseout is not consistent with the benefits prin- Benefits Principle ciple since those who receive greater total benefits over time are subject to a lesser recapture amount than those who receive the same annual benefit, but over a shorter period. As shown in table 2.2, the phaseout overrides the income test. Thus, those owners who benefit the most from a lengthy subsidy period and/or who have large increasesin income over this period have their recapture amount reduced and ulti- mately eliminated. The recapture formula could be revised to increaseequity of treatment between two groups of assistedowners who have rapid increasesin income but, in one case,stay in the assistedhome for 6 years or less, and, in the other, more than 6 years. Figure 3.1 shows three alterna- tives. First, the provisions could allow the recapture amount to increase (after 6 years and until the home was sold) at the enacted recapture rate. Second,the recapture amount could increase after a certain period (say, 6 years) but at a somewhat reduced rate. And third, the law could establish a plateau from which the recapture amount would neither increase nor decrease. Page30 GAO/RCED-W117 Recapturingthe QualifiedMortgageBondSubsidy chapter 3 Changesto the RecaptureProvisionsto PromoteMore JSquitable Treatment of Owners Figure 3.1: AlternatIves to the Recapture Phalreout low0 Amountto k RocaDtuNd 8ooo 6ooo 4oao a00 A’ ,./’ 0 0 1 2 3 4 5 6 7 8 9 10 Ym - fbte remains atnatant owr life of loan 1111 Recapture Increases at a Lower Rate Enacted Recapture Rate nnnn Recapture Amount WI Plateau The recapture provisions are imposed after the assistedowner sells the Subsidy Allowed to home. This approach is consistent with the general framework of QMB Continue Despite loans and MCCS,where the continued need for the assistanceis not peri- Owners’ Significant odically determined. It is also easier and less costly to administer since the amount is computed once, as opposedto periodic testing. Income Increases However, someowners’ incomes will increaseto the point that they do not need continued assistanceto remain home owners. The current recapture provisions do not addressthis consideration. In this regard, the recapture provisions (1) test for the increasesin income only after the assistedhome owner has sold the assistedhome; and (2) may cause the owner to stay in the home longer and thus extend the length of time the subsidy is provided. Page31 GAO/RCED-30-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter 3 (r Changesto the RecaptureProvisionato PromoteMore EquitableTreatment of Owners Subsidy Continues Since the recapture mechanism tests for income growth only at sale, QMB Regardlessof Household’s and MCC subsidiescontinue as long as the assistedowner owns the home, or for the life of the mortgage. Thus, those who experience rapid Income Growth increasesin income can continue to receive the subsidy when the income increasesmake continued participation unnecessary.Further, unless prospective participants believe that they will be worse off financially with a QMB loan or MCC subject to recapture than with a comparable con- ventional loan, they are likely to participate in the program even when faced with a recapture amount.6 While the QMBS are outstanding, the government loses substantial rev- enue from foregone tax revenues.’ Similarly, for MCCS, every dollar in tax credit taken is a dollar of tax revenue foregone. A different approach for testing whether income is sufficient to afford the home without continuing the subsidy would be one in which a home owner’s income is periodically tested to determine continued need for assistance with housing payments. This approach is discussedin chapter 4. Recapture May Extend Another result of the recapture mechanism is that assistedowners may Subsidy Duration for stay in their homes longer (or sell earlier) to minimize the recapture tax. Potential home buyers often do not know with certainty how long they Some,But Not for Others will reside in a housing unit, whether it is subsidized or not. The poten- tial recipient of a QMB loan or MCC with the current recapture mechanism has a relatively small probability of a very low subsidy if the recipient moves within 6 years and a larger probability of a greater subsidy if the recipient stays in the home longer.* As a result, the potential recipient is presented with a “no lose” situation, since the present value of the sub- sidy is greater than the expected present value of the recapture amount. Even though the potential recapture amount increasesduring the first 6 years, we would not expect that the recapture amount would create an “In more formal terms, home buyers would choose to participate if they believe that the net present value of their stream of benefits will be greater than zero. The net present benefit to the home buyer is the sum of the annual savings from the reduced interest rate mortgage over a conventional mort- gage, less the anticipated recapture tax, and less additional costs borne by the borrower that the conventional buyer would not incur (participation fees), discounted to reflect the time value of money. ‘Several studies have estimated that an annual federal tax expenditure of between $20 million to $30 million occurs for every $1 billion in QMBs issued, or between $160 million and $200 million in pre- sent value terms. See (GAO/RCED-88-111, Mar. 28,1988), pp. 60-62. *For a general discussion of economic behavior with risk and uncertainty present, see Milton Friedman, Price Theory, 2nd. ed. (Chicago: Aldine Publishing, 1976), pp. 80-84. Page32 GAO/RCED-!IO-117 Recaptwiugthe Qualified MortgageBondSubsidy Chapter3 Changeato the l&capture Provih9ione to PromoteMoreEquitableTreatment of Ownera incentive for owners to move during that period becauseowners who expected the benefit from the QMB loan or MCCto be less than the recap- ture amount (in present value terms) would not be likely to participate in the program. Alternatively, a substantial recapture amount might be expected to motivate assisted owners to remain in their homes for more than 6 years to reduce or eliminate the amount due.” This situation could causesomehouseholdsto receive the subsidy longer than they would have without the enacted recapture mechanism.The reduced mobility would increase federal tax expenditures. Alternatives that do not have these effects on moving behavior are dis- cussedin chapter 4. Additional considerations regarding the recapture provisions’ expected effects on home owner participation are contained in appendix V. A Prospective In establishing the recapture provisions, the Congressattempted to iden- tify those assistedowners (1) whose incomes will increase rapidly so Affordability Test Is that they could have waited a short period and bought a home without Needed assistanceor (2) who no longer need the assistanceto remain home owners. However, the recapture provisions do not attempt to identify a third group: those who do not need a QMB loan or MCCto afford the assisted home, but are eligible to receive it under Internal RevenueCode eligibility requirements. The Code does not require that QMB loan and MCC assistancebe limited to buyers who cannot afford their first home without this assistance. In March 1988, we reported that only 2 of the 26 agencieswe contacted attempted to limit assistanceto those who could not afford a similarly priced first home without a reduced-interest rate QMB loan. Using housing agency records for over 149,000 buyers who received QMB loans from 1983 through 1987, we estimated that 66 percent of the assisted buyers probably could have bought the samehome at the sametime with a conventional fixed-rate loan of the samesize with comparable terms. The report indicated that even more probably could have afforded the samehome at the sametime if they had chosena conven- tional adjustable-rate loan, which offers lower initial interest rates than *For findings that indicate that two out of three first-time buyers remain in their housing units for over 6 years, see John Simonson, “Prospective First-Time Homebuyers: Salient Characteristics and Their Implications,” National Association of Realtors (Feb. 1939), p. 32. Page33 GAO/RCEJJ-90417 Recaptdng the Qualif¶edMortgageBond Subsidy I chapter 3 Cbangea to the Recapture Provisions to Promoti More JCquitable Treatment of Owners fixed-rate loans. We based this result on the prevailing industry housing-expense-to-incomequalifying test in making these analyses.‘” For this report we expanded our affordability test by adding the second standard qualifying test, the total-debt-expense-to-incometest. This test measureswhether total debt expense(housing expenseand consumer installment payments, such as automobile loans) exceedsa certain per- cent of household income. At the time of our prior review, the 36 per- cent of income could typically be devoted to debt expense.Becausethis secondtest acts as an additional screen,it would be expected that some who probably would have qualified under the first test, might be screenedout by the secondtest. That is, their added debt expensemight make them an unacceptable mortgage credit risk. For this report, we ran the total debt test on those who passedthe housing expensetest. We found that the total-debt-expense-to-income test had little effect on the proportion of assisted householdsthat may have been able to afford the sameprice home at the sametime with a market-rate, fixed-rate loan, The proportion of householdsthat may have been able to afford the same home with a market-rate, fixed-rate loan decreasedslightly, from 66 percent to 61 percent of the over 149,000 householdswe reviewed.” As such, we continue to believe that most assisted buyers may have been able to buy a home conventionally. A conventional affordability test could better target QMB loans or MCC assistanceto those who need assistancewith housing payments to buy their first home. To determine whether buyers qualifying for &MB loans could afford the samehouse with a market-rate loan, the lender could run the samequalifying tests using mortgage interest payments at the market rate. If prospective buyers could afford a market-rate loan, then they would not be eligible for a bond-assistedloan. Lenders would con- tinue to use the prospective borrower’s credit and employment histories in the sameway that they do for QMB loans and market-rate loans to determine the expected risk of the loan. For MCCS, the processwould be similar. Sincethe buyer receives a market-rate loan, the lender would determine whether it would lend the same amount of money to the buyer if the MCX were not available. “‘See (GAO/RCEDSS-111, Mar. 28, 1988), pp. 32-34,46, and 66-68. See also Home Ownershi : Targeting Assistance to Buyers Through Qualified Mortgage Bonds (GAO/RC@ne 27, 1988), pp. 16-1’7.We recognize that some owners prefer fixed-rate over adjustable-rate loans to elimi- nate interest rate variations “To estimate consumer debt, we used data from a 1987 Federal Reserve System study. See app. V for a description of how we performed this analysis. Page 34 GAO/RCELNO-117 Recapturing the Qualified Mortgage Bond Subsidy chapter 3 changerto the BecaptllreProvlsloluJto PromoteMoreRquitableTreatment of Owner9 Conclusions The legislated recapture formula is consistent with the general design of QMBand MCCassistancesince the assistanceis provided until the home is sold and the recapture is imposed only after the home is sold. Although the rationale for many of the provisions are not contained in the act’s legislative history, the provisions are specific mathematical formulas set out in law. However, the effect of the provisions raises considerations of equitable treatment between assistedowners since the provisions treat differently situated householdssimilarly and similarly situated house- holds differently. These differences in treatment are reflected in the (1) uniform versus observedbenefit rate, (2) gain-on-saleprovision, and (3) phaseout of the recapture amount from years 6 to 10. In particular, the uniform (1.25 percentagepoint) recapture rate will recapture more ben- efit than someowners received and less than others received. Also, the phaseout of the recapture amount reduces and eliminates the recapture amount due even though the total benefit received by the owner increaseseach year. Further, becausethe provisions are triggered only when the home is sold, the recapture provisions (1) allow the subsidy to continue when it is not neededto support housing payments, if the assistedowner decides not to move, and (2) may act as an incentive not to move as soon as may have occurred if the recapture was not present. While the recapture provisions retrospectively attempt to identify owners who no longer need the assistance,no mechanism exists to pre- vent agenciesfrom providing QMBloans and MCCSto those who do not need it to afford their first home. A conventional affordability test made at the time a prospective buyer applies for a &MBloan or MCCwould better ensure that only those who need the assistancereceive it. The current recapture mechanism is preferable to none at all because someor all of the subsidy will be recaptured from at least someowners who probably could have waited a short time and bought a home without assistancedue to rapid growth in household income. Overall, however, the current mechanism is a relatively ineffective way to iden- tify and recapture benefits from those who do not continue to need assistancewith housing payments becausethe recapture processis trig- gered solely by an assistedowner’s decision to move. Without altering its basic form, the recapture provisions could be improved if the changes discussedin this chapter were made. Other improvements, such as con- tinuation of the subsidy when it is no longer needed,would require more fundamental changesto the existing framework. Page35 GAO/RCELMtO-117 Recapturingthe QualifiedMortgageBondSubsidy chapter 3 Changesto the RecaptureProvisionsto PromoteMoreEquitableTreatment of Owners Alternatives to the recapture provisions may better achieve the purpose of denying the subsidy to those who no longer need it and increase equity of treatment between assistedowners, although at an increased administrative cost. These alternatives are the subject of chapter 4. If the Congressdesires to retain the existing recapture framework, it Matters for could improve equity of treatment between owners by providing that Consideration by the the recapture amount be basedon the actual benefit received, rather Congress than a uniform rate. For the QMB program, this could be accomplished by basing the rate of recapture on the actual before-tax housing payment reduction realized by the owner over the period of ownership. For the MCC program, the recapture could be based on the amount of the tax credit actually taken. Also, if the Congressdecidesto retain the existing recapture framework and to have the recapture amount better reflect the total amount of ben- efit received from owners with rapid income growth, it may wish to eliminate the phaseout of the recapture amount after year 6. If, how- ever, the Congresswishes to limit overall recapture liability, it could provide for a lesser recapture rate for years 6 and beyond, or cap the recapture amount at the amount determined after year 6, subject to other current adjustments. Further, if the Congresswants to improve the targeting of those who initially receive QMB loans and MCCS, it could require that they be pro- vided only to those who cannot obtain a market-rate loan with compa- rable terms. The Association of Local Housing Finance Agencies,National Council of Agency Comments State Housing Agencies,and Department of the Treasury provided com- ments on a draft of this report (see app. VI through VIII, respectively). The Association and the Council objected to basing the recapture rate on the actual interest rate reduction received by the assistedowner, rather than the uniform 1.26 percentagepoint amount. Both groups thought our proposal would needlessly encumber the program and that the single rate was easier to understand and administer. Treasury com- mented that our suggestedapproach was generally appropriate as a matter of fairness. However, as a matter of tax policy, Treasury sug- gestedthat the fairness objective would have to be balanced against associatedadministrative and compliance costs.We agreethat a single rate is easier to administer. However, our concern is that the assisted Page36 GAO/RCEJHO-117 Recapturingthe Qualified MortgageBondSubsidy Chapter3 RecaptureProvisioneto Changes to the PromoteMoreEquitableTreatment of Owners owner not be placed in a position of potentially paying more in recap- ture than the benefit received. We also do not believe that the administrative requirements for the assistedowner or lender would be as burdensomeas suggested.While we did not quantify the added costs involved in tailoring the recapture formula more to the benefit received, we expect that the additional cost per home owner would be small. The additional effort would comprise two steps. First, the lender would determine the monthly principal and interest payment for a fixed-rate market-rate loan of comparable terms. Second,the lender would determine the difference between the housing payment for the market-rate loan and the QMB loan. The results of this one-time computation would be provided to the assistedowner and retained by the lender or the housing agency. In deciding whether to make this revision to law, the Congressmust weigh the effort required in performing these calculations against the benefit of refining the recapture mechanism. Similarly, the Association and the Council objected to our analysis of the effects of the recapture phaseout provision. The Association believed that our proposal would discourageowners from remaining in their home, apparently becausethe recapture would not be phased out. Our discussionpoints out that the enacted phase-out provision allows the recapture provision to be reduced to as far as $0 even though the house- hold (1) continues to receive the benefit from a reduced interest loan for the life of the loan and (2) has income increasesthat meet or exceed those in the income test. From a standpoint of achieving the Congress’ stated goals (as set out in the conferencereport), we believe it is appro- priate to recapture at least someof the benefit from those who continue to receive it and meet income increasethresholds. The Council said that the phaseout would have little effect on those with rapid increasesin income becausethey (1) will move out as soon as they can afford to and (2) are better off selling earlier, rather than later, to take advantage of the provision that adds $6,000 to adjusted quali- fying income.12The Council ignores the fact that the substantial transac- tion costs involved in selling a home creates significant disincentives to selling within the first few years of ownership. Treasury had no specific comments on our phaseout discussion. “As discussed in ch. 2, in the first few years of ownership, the $6,000 amount adds significantly to the amount income must increase before the full recapture amount is due. Page37 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Chapter3 Changesto the RecaptureProvieIonstn PromoteMoreEquitableTreatment of Ownere The Association and the Council strongly disagreed with our analysis of whether assistedowners probably could have afforded the samehouse without assistance.Both groups said that if the secondstandard mort- gagequalifying test (total-debt-expense-to-incometest) was applied, our results would have been substantially different. As a result, we per- formed that test and found little changein our results. Finally, both the Association and the Council stated that an affordability test was not needed.In its comments,however, the Council stated that it agreeswith the objective of discouraging those from get- ting a QMB loan when conventional financing is a reasonablealternative. Treasury also agreedthat providing assistanceto those who could otherwise afford a market-rate loan with comparable terms is not con- sistent with the goals of the QMB and MCC programs. On the basis of our affordability test in this report, we maintain that this sort of screening mechanismis neededas a way of demonstrating that potential first-time home buyers need QMB loans and MCCS to purchase their homes. Page88 GAO/RCEBf%117 Recapturhgthe QuaUfiedMortgageBondSubsidy Alternatives That Would Better Idefy and Deny Subsidy to Those With Rapid Income Growth Chapter 3 discussedimprovements that could be made within the frame- work of the enacted recapture provisions to provide more equitable treatment between sets of assistedowners who experienced rapid income growth. As chapter 3 also indicated, the recapture approach does not identify and suspendthe subsidy to those with rising incomes becauseit is triggered only by the recipient’s decision to move. Mortgage loans that discontinue assistancebecauseof pre-established affordability criteria (roll-over loans) and periodically increase housing payments as long as income increases(income-basedvariable payment mortgage loans) are income-driven alternatives to the current recapture approach. These alternatives improve equity of treatment between assistedowners and are not based on an assistedowner’s decision to move. However, these approachesare most likely more complex and costly to administer than the recapture provisions. A roll-over mortgage loan is a mechanism that continues or ends assis- Substituting a Roll- tance on the basis of periodic examination of current income. If the Over Approach for the owner’s income rises to a point where the owner could afford a compa- Recapture Mechanism rable market-rate loan, the owner is required to refinance the assisted loan with a market-rate mortgage loan. If income growth is insufficient for the owner to afford a market-rate loan, then assistanceis continued with no changein monthly payment. The roll-over decision points could come at any time. Roll-over mortgageswith 3- and 5-year intervals are common in the Canadian private mortgage market.’ This approach dif- fers from the recapture approach discussedin the preceding chapters in that the roll-over approach doesnot require that assistancebe paid back. Since a roll-over decision is driven by income growth, this approach is consistent with the ability-to-pay principle. Similarly, it does not create inequities cited in chapter 3, making it consistent with the benefits prin- ciple. A roll-over mortgage does not require a voluntary move to end assistanceduring the lifetime of the assistedmortgage and, therefore, more efficiently continues assistanceto those who require assistanceto meet monthly payments, while halting assistanceto those who no longer need it. Recipients who do not experience the income growth required for the roll-over to occur are not discouraged from moving as much as if ‘For a further discussion of the roll-over mechanism, see David Ling and Mark Smith, “Another Look at Mortgage Revenue Bonds,” Journal of Policy Analysis and Management, Vol. 7, No. 3 (Spring 1988), pp. 662-4. Page 39 GAO/RCED-99-117 &capturingthe Qualified Mortgage Bond Subsidy Chapter4 AlternativesThnt WoUrdBetter Identify and DenySubsidyto ThoseWith Rapid IncomeGrowth they were subject to a recapture. They lose the ongoing subsidy if they move, but they do not becomesubject to the recapture provisions2 The current recapture provisions and roll-over loans introduce adminis- trative costs and complexities. The recapture mechanism adds to reporting requirements by assistedowners and others and may increase general Internal RevenueService coststo the extent that the Service audits recapture tax reporting (see app. IV). The use of roll-over loans is likely to incur costs since they require routine examinations that would probably be performed by the mortgage servicer acting for the housing agency.3 Refinancing and Other A loan that “rolls over” to a market-rate loan requires determining what Costs of a Roll-Over Loan the market rate should be: either the prevailing rate for similar loans at the time the loan was made or the rate at the time the roll-over occurs. Could Be Avoided Also, converting to an unsubsidized conventional loan will entail refi- nancing costs, which are often substantial. Last, periodic examinations of income will require increased administrative activities by the mort- gageservicers and increasethe costs of the program. The easiestmethod, administratively, for determining the market rate and refinancing costs for home owners with QMB loans is to require them to incur refinancing costs and find a market-rate mortgage in the cur- rent mortgage market. This method, however, subjects recipients to one certain and a possible secondcost. First, refinancing chargescan be sub- stantial. Second,if mortgage interests rates in the private market have risen since the time of home acquisition, recipients could becomeworse off than if they had never participated since owners have to pay a higher mortgage interest rates than if they had never participated in the QMBprO$Jram. The FIMA section 602 home ownership program handles this situation by establishing a “market rate” when the household becomesa recipient. As the assistedhousehold’s income rises, housing payments increase up to the amount determined by the “market rate.” However, since the loan *Both the roll-over loan and the current recapture provisions treat assisted owners who have little or no income growth similarly. In the former, assistance is continued until the home is disposed of; in the latter, no recapture amount is imposed either during the life of the loan or when it is discontinued. 3Roll-over loans and a recapture could be combined but would entail higher governmental administra- tive costs and a greater home owner burden than if only one approach was used. The two combined, however, both limit the duration of assistance and impose a recapture amount on recipients with potentially the greatest ability to pay. Page 40 GAO/RCED-90-117 l&capturing the QualifiedMortgageBondSubsidy . Chapter4 Alt.mnativeeThat WouldBetter IdentLfyand DenySubsidyto ThorreWith Rapid IncomeGrowth is not converted to a conventional loan, no refinancing costs are gener- ated. A similar mechanism could be set in place for QMBloans to avoid interest-rate risk and refinancing costs4 The MCCprogram doesnot create this tradeoff. The recipient already holds a private mortgage instrument made at market-rate interest. The MCCcan be cancelled without triggering a changein the mortgage interest rate or incurring refinancing costs.Therefore, a roll-over mech- anism would involve the ending of the mortgage credit rather than shifting to a new private mortgage instrument.6 Finally, periodic reexamination of income leads to increased administra- tive costs.These costs could be paid from participation fees, bond pro- ceedsthat might otherwise be used to make mortgages,or from other (IIOIPQMB or MCC-related) sourcesavtilable to somehousing agencies.We did not attempt to quantify the likely cost increasesthat would result from the roll-over mechanism. However, a Mortgage Bankers Associa- tion of America official estimated that a lender’s activities relating to roll-over recertification would add about one-eighth of a percentage point to the lender’s servicing fee (or about $100 on an $80,000 mort- gageloan). If roll-over approach replaces the recapture mechanism, the reexamination period could be set at a point, such as every 3 or 6 years rather than annually, to lessenthe burden but still test for sustained income increases. Shallow Subsidy Reduces While the roll-over approach halts the subsidy when it is no longer Roll-Over Impact neededto afford housing payments on a comparable market-rate loan, the impact is small becausethe original subsidy is relatively small. As discussedearlier in this report and in our March 1988 report, QMBloans provide relatively shahow subsidies-typically a 1.44-percentagepoint reduction in the mortgage interest rate (or about $40 per month, after taxes).” In contrast, HUD and FmHA home ownership programs reduce mortgage interest rates to as low as 1 percent. Clearly, a roll-over mech- anism that operated in a deep subsidy program would have a greater 4An income-averaging approach could be used so that increased payments do not result from tempo- rary increases in household income. 6Although implementing thii alternative would be much simpler for a MCC loan than for a QMB loan, QMBs form the overwhelming majority of the assistance provided. 8ee Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GAO-W-1 11, Mar. 28, T9W, PP~11 and 66-90. %ee p. 38 of (GAO/RCED-88-111, Mar. 28,1988). Page41 GAO/RCED-f#117 Recapturingthe QualifiedMortgageBondSubsidy Chapter4 AkemativeaThat WouldBetter IdentiQ and DenySubaldyto ThoseWith Rapid IncomeGrowth potential impact than a similar roll-over for a QMBloan and MCC, which typically provide a much smaller amount of assistance. A secondincome-basedalternative to the recapture provisions is an Income-BasedVariable income-basedvariable payment loan. In this type of loan, the owner’s Payment Loans as an payments for principal, interest, taxes, and insurance are a fixed per- Alternative to the centageof household income. As income increases,so do monthly pay- ments. While the roll-over mortgage caps the mortgage interest rate at Recapture Mechanism the market rate, the income-basedvariable payment loan allows the assistedowner’s interest rate to rise above the market rate as income continues to grow. Mortgage payments increase as long as income does. For QMB loans, this mechanismwould motivate recipients to leave the program when their monthly payments for principal and interest becomegreater than payments for an unsubsidized market-rate mortgage. For variable-payment loans with MC&S, the MCC could be reissued at a decreasedrate (say, a lo-percent credit rather than a 20-percent credit), as warranted by periodic review of income. A MCC reissued at a lower rate would decreasethe subsidy amount and maintain housing pay- ments as a percent of income. The variable payment approach can lead to better targeting of benefits and is similar to the HUD section 236 and MA section 602 home owner- ship programs, except that QMBloans and MCCS provide shallower subsi- dies. (Seeapp. III for descriptions of the HUD and F~HAprograms.) As with the roll-over loan, the income-basedvariable payment loan is con- sistent with the ability-to-pay principle. Also, it tests for income growth periodically, rather than when the home is sold, and it doesnot rely on a voluntary move to halt the subsidy for recipients who no longer need it to afford housing payments. The income-basedvariable payment mort- gagefulfills this objective indirectly by increasing mortgage payments to a level above that which the recipient with rapidly increasing income would pay in the private mortgage market.’ ‘The variable payment approach could be combined with a recapture mechanism as the HUD section 236 and FmHA section 602 programs do. Regardless of the possibly desirable equity outcomes, such a mechanism would entail administrative complexities and costs that are not part of the current recap- ture approach. Page42 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubeidy Chapter4 AlternativesThat WouldBetter Identify and DenySubsidyto ThoseWith Rapid IncomeGrowth Increased Administrative As with the roll-over loan, income-basedvariable payment loans will Costs and Complexity increase administrative cost and program complexity over the current recapture approach. As with the roll-over loan, a reexamination at 3- or S-year intervals would be less costly than annual reexaminations. Overall, however, periodic income reexamination and adjustment of mortgage payments for variable payment loans would be more costly than for roll-over loans. These activities must be performed at each des- ignated time interval (say, every 3 years) for each outstanding loan under the variable payment approach. However, under the roll-over approach, loans that had “rolled-over” would not be subject to further examinations and adjustments. It is possible that, over the life of the loans, the increasedcost and effort to reexamine income and adjust vari- able payment loan payment schedulescould make this approach unworkable. The final consideration of how the increased payments are treated by the loan servicer and the housing agency is similar to that of the roll-over loan. Conclusions The decision to substitute a roll-over loan for the existing recapture mechanismis not clear-cut. The roll-over mechanism presented in this chapter better achievesthe objective, as stated in the conferencereport, of not providing subsidiesto those who experience rapid income growth. On the other hand, the roll-over approach adds someadministrative costs and complexity. However, the additional cost and effort that would be required could be lessenedif the period for reevaluating need were set at 3- or S-year intervals, rather than annually. A variable payment loan, which is an income-basedalternative to a roll- over loan, would also be more equitable than the present recapture mechanism becauseit would tie the amount of the subsidy directly to income growth. However, becauseit requires constant monitoring of changesin recipient income, it probably is not administratively feasible within the operations of the QMBand MCCprograms. If the Congressdecidesto replace the existing recapture approach with Matter for one that will terminate assistanceto recipients who experience rapid Consideration by the income growth and, as a result, can afford housing payments on a Congress market-rate loan of comparable terms, it may wish to enact a roll-over approach that doesnot require external refinancing. I Page43 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy chapter 4 Alwnativea That WouldBetter Identify and DenySubsidyto Thaw With Rapid IncomeGrowth Both the Association of Local Housing Finance Agencies and the Agency Comments National Council of State Housing Agencies disagreedwith adopting a roll-over type approach as an alternative to the existing recapture mech- anism. Both cited our discussion of the added administrative burden and complexities in reexamining owners’ incomesin this type of approach. Treasury did not state a position on whether a roll-over approach should be substituted for the recapture mechanism. However, it did sug- gest that our report should discussthe administrative and compliance costs associatedwith a roll-over approach. Both our draft and this report recognizethe additional burden and com- plexity of the roll-over approach and contain an estimate of the cost for periodic recertifications. The added burden and complexity is the trade- off for better identifying those whose income is rising to the point where further assistanceis not warranted. The existing recapture mechanism requires that the home be sold before that examination of income increasestakes place. Under the roll-over approach, that examination would be done periodically. Page44 GAO/RCED-!bO-117 Recapturingthe QualifiedMortgageBondSubsidy Page46 GAO/RCED-IH)-117 Recapturingthe Qualified MortgageBondSubsidy Appendix I Example of How to Compute the Recapture Amount J, a single individual, purchases a home with a $66,000 QMBloan. At the time of purchase, the applicable income limit for a single person is $20,000. He marries S, and they have two children, E and M. They sell their QMB-assisted home 6 years and 2 months later and realize a gain of $12,000 on the sale of the home. In the year of sale, J and S’s household income (adjusted gross income plus tax-exempt interest) is $32,000. Before adjustments, the maximum recapture amount would be $3,438 (0.0626 x $66,000). However.,becausethe home was held more than 6 years, the holding period adjustment reducesthe recapture amount to $2,636 ((120-74)/60 x $3,438), where the adjustment is based on the number of months (74) out of the lo-year period (120 months). The recapture amount is further reduced becauseJ and S’s household income ($32,000) is $2,900 less than the estimated income eligibility limit of $34,900.’ Sincethe recap- ture amount is reduced by 2 percent for every $100 that modified adjusted income exceedsJ and S’ household income, the recapture amount owed is $1,108. Since $1,108 is less than the amount determined by the holding period adjustment, the maximum recapture is $1,108.2 SinceJ and S realized a $12,000 gain on the sale of their home, the 60- percent gain-on-salelimitation did not further reduce the $1,108 recap- ture amount due. If the gain on sale had been $2,000 instead of $12,000, then the recapture amount would have been $1,000, which is the lesser of the computed amount of $1,108 or 60 percent of the gain on sale (0.6 x $2,000 = $1,000).3 ‘If $20,000 was the applicable income limit for a one-person household, $23,000 would be the appli- cable 3-or-more-person household limit ($20,000 x 116%). The $23,000 amount is increased by 6 per- cent per year for each of the 6 full years to $29,900 ($23,000 x 0.06 x 6 = $6,900 and $23,000 + $6,900 = $29,900) and $6,000 is added to this amount ($29,900 + $6,000 = $34,900). 2Multiply 2 percent for every $100 that household income ($32,000) is less than the estimated income eligibility limit of $34,900. Since ($34,900 - $32,000)/100 = 29, the reduction to the recapture amount is 0.02 x 29 x $2,636 = $1,628. The recapture amount due after this step is $2,636 - $1,628 = $1,108. 3We were also asked to determine the effect on the marginal tax rate resulting from the recapture provisions. Since the provisions do not require any modifications to the taxpayer’s adjusted gross income, no tax bracket changes will occur. Page46 GAO/RCED-90-117 Recapturingthe Qualified MortgageBondSubsidy Appendix II Income IncreasesNeededto Pay the Recapture Amount for an Area With a $26,000 Median Income Chapter 2 provides examples of the recapture amount due for hypothet- ical householdsin an area in which the median income was $34,000. As discussed,the income test is sensitive to the median income level. In this appendix, we show how much income would have to increaseto pay the recapture amount for three householdsin an area where median income is $ZS,OOO.l In this example, we use three households,each with three or more mem- bers. The householdsdiffer only in their income at the time they purchase their assistedhomes. Household C has the maximum income to be eligible to participate, $29,900 (116 percent of the area median income); Household D has an income at the area median, or $26,000; and Household E has an income of $20,800 (80 percent of the area median income). Figures 11.1,11.2,and II.3 show the average annual income increases neededto pay the recapture amount for HouseholdsC, D, and E, respec- tively. The increasesneededfor HouseholdsA and B in chapter 2, where our calculations were basedon an area median income of $34,000, are smaller than the increasesneededfor HouseholdsC and D, respectively. Average annual income increasesneededto pay the maximum recapture are higher the lower the area median income becauseof the $6,000 con- stant in the income test formula. The increasesare higher becausethe $6,000 amount has a larger influence when the area median income is lower: a $6,000 increase from $26,000 to $31,000 is a 19-percent increase, and a $6,000 increase from $34,000 to $39,000 is a 16-percent increase. 1For example, $26,000 is about equal to the estimated 1989 statewide median incomes for Alabama and Kentucky, according to HUD. Page47 GAO/RCED-90-117 Recapturingthe Qualifkd MortgageBondSubsidy AppendixII IncomeIncreasesNeededto Paythe RecaptureAmountfor au AreaWith a 826,000 MedianIncome Figure ll.1: Average Annual Incnsre In Income Needed to Pay the Recepture Amount for Houeehold C 28 Avorago Annual lncnase In Income 20 15 0 1 2 3 4 s 6 7 6 9 10 YOUIn - Full recapture amount for Household C -I - - Any recapture amount for Household C Note: Household C has an income at time of purchase of 115 percent of the area median and a house- hold size of three or more. Area median income is $26,000. Page48 GAO/RCED-BO-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixII Income Incream Neededto Paythe RecaptureAmountfor an AreaWith a $26,999 MedianIncome Figure 11.2:Average Annual Increase in Income Needed to Pay the Recapture Avemgo Annual lncmarc In Income Amount for Hourehold D 40 30 2 3 4 6 6 7 8 9 10 Yoam - Full recapture amount for Household D - - - - Any recapture amount for Household D Note: Household D has an income at time of purchase of 100 percent of the area median and a house- hold size of three or more. Area median income is $26,000. Page49 GAO/RCED-99-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixLI . Income Increaeea Paythe Needed to RecaptureAmountfor an AreaWith a 826,000 MedianIncome Figure 11.3:Average Annual incresee in income Needed to Pay the Recapture 80 Avomgo Annual Incnaaa In Income Amount for Houeehoid E 0 1 2 3 4 I 6 7 8 9 10 VOW3 - Full recapture for Household E ---- Any recapture amount for Household E Note: Household E has an income at time of purchase of 80 percent of the area median and a household size of three or more. Area median income is $26,000. Page60 GAO/RCEDBO-117 Recapturingthe Qualified MortgageBondSubsidy Recapture Mechanisms in Federal, State, and Local Housing Programs Certain federal, state, and local housing programs recapture someor all of the assistancethey provide to home buyers or have recapture-like mechanisms.The recapture may be levied programwide or limited to a specific housing project. This appendix describesseveral housing pro- grams with recapture or recapture-like mechanisms. The HUDsection 236 home ownership assistanceprogram provides for a HUD Section 235 mortgage interest reduction for lower-income families. Loans are made Program by commercial lenders at market interest rates. Eligible families pay at least 20 percent of their adjusted income toward monthly mortgage pay- ments (or 28 percent, depending on which portion of the program the assistedhousehold is participating in). The difference between the market-rate mortgage payment and the amount that subsidized owners are paying is made up by federal assistancepayments (the “interest sub- sidy” or interest credit). The interest rate may be subsidizedto a level as low as 1 percent. The commercial lender servicesthe loan and reexam- ines the home owner’s income annually, and the recmired home owner payment is increasedor decreasedaccordingly. , Section 206(b)(l)(C) of the Housing and Community Development Act of 1980 (P.L. 96-399, Oct. 8, 1080) added a recapture provision for those home owners who disposeof their property or rent it for more than 1 year. Under the act, HUD recaptures the lesser of (1) the amount of interest subsidy actually received or (2) 60 percent of the net apprecia- tion of the property.’ The legislative history behind the act doesnot explain why the recapture was added or the problems it was intended to address.On disposition, the recapture due constitutes a hen on the Property. When the assistedowner proposesto disposeof the home, the mortgage lender notifies the local HUD field office of the total amount of assistance payments that HUDprovided. The HUD field office then calculates the recapture amount due and the recapture amount must be paid to convey title. ‘Net appredation is defined as the increase in value over the original purchase price, less (1) reason- able coets of sale, (2) reasonable costs of home improvements, and (3) certain increasersin the mort- gage amount over the original balance. Page 61 GAOjRCEDW117 lkaplm’hg the Qualified Mortgage Bond Subsidy ___ ___ - .- -.- &pendix III RecaptlueM- in Federal, St&e, and Local Iiontling FrograuM F’mHA Section 502 FmHA’s section 602 rural home ownership loan program provides low- Program and moderate-incomehouseholdswith home purchase loans. F~IIA makes fixed-rate loans to eligible buyers at an interest rate approxi- mately equal to the cost of long-term government borrowing at the time of loan origination. However, most borrowers receive a subsidy, called an interest credit, to reduce the loan’s effective interest rate to as low as 1 percent, depending on borrowers’ incomeswhen they receive the loans. Section 602 participants initially pay at least 26 percent of their adjusted income for principal, interest, real estate taxes, and insurance. FMU annually recalculates the borrower’s required payment using the borrower’s current income. Section 606(a)(3) of the Housing and Community Development Amend- ments of 1978 (P.L. 96-667, Oct. 31, 1978) added a recapture provision to the section 602 program. The legislative history behind the act does not explain why the recapture was added or the problems it intended to address.The assistedowner must repay the lesser of (1) the amount of the subsidy received or (2) a percentageof the property’s net apprecia- tion, ranging from 9 to 78 percent (depending on the averageinterest rate paid by the owner over the life of the loan and the number of months that the mortgage was outstanding).2The mortgages or deedsof trust signed by those receiving the interest credit contain a provision making the amount of the subsidy a lien against the property. In 1989, the National Council of State Housing Agenciessurveyed its State Housing membership to determine the extent to which recapture mechanismsare Programs used and how they are structured. The Council reported that eight of its members had someform of recapture mechanism as part of one or more of their home ownership programs. None of the eight applies recapture to its QMBprograms. Rather, the Council reported, the recapture mecha- nisms appear to be used only with special programs of limited duration or availability. The Council reported the predominant objective of the state recaptures was to recapture funds for future program use. Four states had a sec- ondary objective of preventing speculation. Two states had a third 2Net appreciation is defined ss market value of the home, less the amount of non-FmHA prior loans, unpaid balance of FInHA liens, sales expenses, principal paid at note rate, principal reduction attrib- uted to subsidy, and original equity. Page 62 GAO/WED-fMb117 Recapturing the Qualified Mortgage Bond Subsidy AppendixIll RecaptureMechanisms in Federal,State,and LocalHousingFrograms objective of removing the subsidy when borrowers were no longer eli- gible for it. The Council’s summary included a short description of the recapture provisions in sevenstates:3 . The amount recaptured is the amount of down payment and/or interest rate subsidy or a portion of the net appreciation of the property (Massa- chusetts, Minnesota, North Carolina). . During a specified control period, 95 percent of the appreciation of the home salesprice (net of capital improvements) is recaptured (New Jersey). . If the assistedowner sells or transfers the house within the first 16 years, any proceedsin excessof the state’s home ownership affordability index must be paid at a rate of 2 percent of the salesprice or appraised value (Rhode Island). . At time of sale, an owner receiving state-funded down payment assis- tance in conjunction with QMB loans must pay 20 percent of gained equity during years 1 to 6, 10 percent in years 6 to 10, 6 percent from years 11 to 20, and 2.6 percent from year 21 to loan maturity (New Hampshire). l If a home is sold within 7 years, the assistedowner must pay (1) the interest rate differential between the (assisted) note rate and 8 percent and (2) on a sliding scale,a percentageof the original salesprice. The recapture cannot exceedthe gain realized on sale (Connecticut). We surveyed 128 local housing agenciesthat are membersof the Associ- Local Housing ation of Local Housing Finance Agencies.We (1) asked the agenciesto Programs tell us if they recaptured someor all of a subsidy they provided and (2) allowed the agenciesto define what recapture might entail. Because there are other local housing agenciesthat do not belong to the Associa- tion, our survey was not meant to be statistically representative of local housing agency practices. Rather, the survey was meant to provide some insight into how somelocal agenciesstructure recapture mechanisms. 3Alaska, one of the eight states with recapture mechanisms, did not provide a program description. Page63 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixIII RecaptureMechauiemain Federal,State,and lQc8.lHondngPrograms Twenty of the agenciesthat returned our survey indicated that they employed a recapture mechanism.4Thesemechanismstook five forms, and three agenciesuse more than one technique. In a number of cases, the documents sent to us showed that the recapture was limited to a single development or small set of properties rather than being applied to the agencies’housing programs in general. Also, several of the recap- ture mechanismswere applied to non-oMBand -MCC programs. The fol- lowing mechanismsare used: l Limitations on resale price (eight agencies).The assistedowner agrees to resell the house at no more than the original purchase price plus the value of improvements, plus someor all of the percentageincrease in the area’s median income (or increase in the salesprice for existing homes in the area) for the period that the home was owned. l Shared appreciation loans (seven agencies).Most frequently, these loans were deferred payment secondtrusts. For example, when the assistedhome is resold, the owner pays the agency the principal amount and the lesser of (1) 10 percent simple interest on the loan or (2) 60 percent of the net appreciation on the home. Several programs have pro- visions that reduce the proportion of appreciation taken the longer the assistedowner holds the assistedunit. . Equity appreciation loans (five agencies).The loans are somewhat similar to shared appreciation loans, but the amount recaptured is based on the contribution made by the housing agency.For example, if the agency made a secondtrust loan that represented 16 percent of the total loan amount received by the assistedowner, then the amount to be recaptured at time of sale would be the outstanding principal balance and 16 percent of the appreciation of the property. l Roll-over loans (two agencies).If periodic examination of the bor- rower’s income shows that the subsidy is no longer needed,then the owner is required to refinance the loan or pay it off. Otherwise, the sub- sidy is extended. . An interest rate surcharge (one agency). Assisted owners receive a subsidy that is repaid as a S-percent(simple interest) loan. If the home is sold within the first 6 years to other than a low- or moderate-income buyer, a S-percentsurcharge is added (total of 10 percent). The surcharge declines to 4,3, 2, 1 and 0 percent in years 6 to 10, respectively. 4About 91 percent of those surveyed responded, although not all operated home ownership programs. We did not consider the simple repayment of principal and interest to be a recapture mechanism, as indicated by several agencies. Also, one of the respondents was not an Association member, but returned our survey when another agency sent the questionnaire to it. Page54 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixIll RecaptureMecWm6 In Federal,State,and LocalHousingprorpame In addition, several agenciesrestricted resale of the assistedunits to low- and moderate-incomebuyers. While this is not a recapture, per- se -’ it may tend to keep resale prices somewhat lower than if the restriction were not in place, effectively limiting the gain on sale. The local jurisdictions using the limitation on resale price were Fremont, San Mateo County, San Francisco,Santa Barbara, and Palo Alto, Cali- fornia; Boulder, Colorado; Montgomery County, Maryland; and Fairfax County, Virginia. The jurisdictions using shared appreciation loans were Hayward, San Francisco, and two agencieswithin the city of Los Angeles, California; Baltimore and Montgomery County, Maryland; and Seattle, Washington. Thosejurisdictions using equity appreciation recaptures were Sacra- mento, California; Minneapolis and St. Paul, Minnesota; Raleigh, North Carolina; and Portland, Oregon. The localities using a roll-over provision were Anaheim, California, and Baltimore, Maryland. The city of Santa Ana, California, used the interest-rate surcharge mechanism. PageI56 GAO/RCED90-117 Recapturingthe QualifiedMortgageBondSubsidy Appendix IV Ekpected Recapture Effects on Program . Participation and Tax Administration Matters This appendix discusseswhether the (1) imposition of a recapture mech- anism will be likely to affect a household’s decision to apply for the assistance,(2) assistedowner will have sufficient liquid assetsto pay the recapture amount when due, and (3) IRS will be able to identify those who sold their assistedhome and are subject to recapture. Our earlier report and the example in table IV. 1 indicate that the QMB No Expected Adverse program generatesrelatively small subsidy levels.’ In the former, the Effect on Program participating population was found to be fairly typical of first-time Participation home buyers. In the latter, the present value of the subsidy that is con- tinued for 10 years is only $6,989. Even with relatively small subsidy levels, home buyers who do not have to changetheir housing choicesto receive a subsidy would be expected to participate in the program becauseit would be in their financial interest. If a larger subsidy were provided through an MCCor other form, potential recipients who would line up to participate would be likely to include this group plus others who may be willing to make someadjustments in their housing choices to receive the (larger) subsidy. Table IV.l: Example of Present Value of QMB Assistance Wlth and Without Present value of subsidy Recapture Provisions at time of purchase Nominal Without With Time In housing unit subsidy recapture recapture0 5 years $5,542 $4,347 $i,04ab IO vears i I ,083 6.989 6.989C Note: The example is based on an $85,000, 30.year, fixed-rate mortgage. The market and QMB mort- gage interest rates are 10 percent and 8.5 percent, respectively. Present values are calculated by dis- counting with the lo-percent rate. aAssumes no reductions due to income or gain-on-sale provisions. bAt time of sale, the recapture amount equals the amount collected at that time, or $5,313. CAfter 10 years, no recapture is due. We would expect that potential recipients who could afford and would choosehome ownership will participate as long as the expected present ‘Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GAO/ , al-.2% 198% PP. x-38. Page56 GAO/RCED-90-117 Recapturingthe Qualified MortgageBondSubsidy AppendixIv ExpectedRecaptureEffectson Program Participation.andTax AdministrationMatters value of participation, net of recapture amounts and other program-spe- cific costs, is positive.2Economic theory suggeststhat even the hypo- thetical recipient characterized in table IV. 1 who would receive a subsidy stream net of recapture with a present value of about $1,000 would still probably participate in the program, as long as the recipient’s housing choicesdid not have to be altered. While someassistedowners’ recapture liability could be several hun- Ability to Pay dred dollars or less, for others it might be several thousand dollars. The Recapture With recapture amount would probably be paid anywhere from several Income Tax Filing months to over a year after the assistedhome is sold. For those with substantial liability, the recapture amount might present payment problems unless owners set aside funds from the sale of the assisted home or have other financial assetsto enable them to pay the recapture amount when due. Alternatives that do not present cash flow problems are discussedin chapter 4. IRS’ Ability to Identify The Internal RevenueService needsto be able to identify assisted owners subject to recapture to ensure that they will pay the appropriate Owners Subject to recapture amount. To accomplish this, the recapture provisions in the Recapture 1988 act require that the real estate broker must send an information return to the IRS stating whether or not the seller’s financing was from a &MBloan or a loan with a MCC.The Code defines “broker” to include any person who acts as a middleman with respect to property and ser- vices. The American Land Title Association expressedconcern that the broker, in many instances,may find it difficult to comply with the reporting requirements3 We contacted officials at three state housing agencieswho tended to mirror the Association’s concerns.These are as follows: l The information submitted to brokers is often not substantial. They may not have the original loan commitment to determine whether the seller haa a QMBloan or a MCCprovided along with the conventional loan, 2This suggests that people make decisions using the best financial information available to them. However, the (1) smaller the subsidy, and (2) greater the uncertainty of the expected recapture amount, the more likely it will be for households not to participate. 3The Association is a trade association whose goal is to promote the safe and efficient transfer of ownership of real property. It represents about 2,300 abstracters and agents, title insurance compa- nies, and associated members nationwide. Page57 GAO/RCED-90.117 Recapturingthe QualifiedMortgageBondSubsidy AppendixN Exgmctd RecaptureEffectaon Program ParticipationandTax AdministrationMatters . Even if the seller’s loan documents are available to the broker, the loan documents will not always contain the necessaryinformation to deter- mine whether the seller Is subject to recapture. . To the extent that a principal in the real estate transaction choosesto withhold information not necessaryto the closing of the transaction, the broker will be unable to comply with the reporting requirement. One approach to foster identification of assistedloans would be for mortgage lenders making loans for housing agenciesto “flag” loan files, so that a notice that the owner is subject to recapture is provided to the broker. A secondapproach would be for housing agenciesto add wording to their loan documents (or in amendments)that the loan is a QMBloan and the owner is subject to recapture. However, the second approach would not be effective for those settlements in which loan doc- uments are not submitted. Also, both approachesrequire the voluntary cooperation of all state and local governments making these loans, since the act’s reporting requirement doesnot require these actions. A third approach would be for brokers to request that lenders identify whether the loan was a &MB loan or made with a MCCwhen it makes its request for information relevant to the transaction. However, if lenders (1) do not flag their files to identify loans as &MBloans or loans with MCCsand/or (2) choosenot to comply with the broker’s request, then the approach will not work. Page68 GAO/RCJD90.117 Recapturingthe QualifiedMortgageBondSubsidy Appendix V Approach Used in Applying a Total-Debt- Expense-to-IncomeTest to Assisted Buyer Data Base In our March 1988 report, we found that over half of the assistedbuyers covered in our review probably could have purchased the samepriced home with a conventional fixed-rate mortgage loan.1This result was basedon a housing-expense-to-incomequalifying test. For this report, we performed a secondtest, the total-debt-expense-to-incometest, to determine the extent to which performing a secondqualifying test to those who passedthe first test affected our earlier finding. This stan- dard determines whether total debt payments (housing expenseand other debt, such as automobile loan payments) exceed36 percent of income. For the “other debt” amount, we assumedthat each of the assistedhouseholdsthat passedthe housing-expense-to-incometest had monthly “other debt” payments of $160. This amount is equal to the median dollar amount of installment payments for home owners in a survey conducted by the Survey ResearchCenter of the University of Michigan.2The Center’s survey data were basedon statistical samplesof householdsacrossthe country. The median $160 installment payment is for all home owners, rather than for first-time buyers alone. However, the Center’s results for three reported measures(age of head of household, race, and income) suggest that the $160 median installment payment amount for all home owners is likely to be a fair approximation of the payment amount for &MB loan recipients. In this regard, our March 1988 report showed that QMB- assistedbuyers were typically younger, white, and had a median income of $26,000 (seetable V.l below). In comparing the Center’s reported monthly debt payments for those with similar age,race, and income range to QMB-assisted buyers, the monthly installment credit payment amounts cluster around the Center’s $160 figure for all home owners. As such, we believe the Center’s estimate is a reasonableapproximation for use in our total-debt-to-income test. ‘See pp. 30-36 of Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to ThoseinNeed(Gwll,M - - al,. 28t 19881. 2SeeRobert B. Avery, et. al., “Changes in Consumer Installment Debt: Evidence from the 1983 and lQS6 Surveys of Consumer Finances,” Federal Reserve Bulletin, vol. 73, no. 10, Oct. 1987, pp. 761-78. Page69 Recapturing GAO/RCED-O-117 the QualifiedMortgageBondSubsidy Appendix V ApproachUsedin Applyinga Total-Debt- Expenseto&womeTestto AssistedBuyer DataBnse Table V.l: Compariron of QMB-Aa8irted Buyers to Hourehold Attributer in the Median installment payment in Survey Survey Research Center Survey QAO’a March 1988 Rerearch Center resultr (monthly Household attribute report payment,) Median installment (Not reported) $160, for all home owners debt Income Median income was $150, for all households with incomes of $26.000 $20.000-$34.999a Age of head of 61 percent of $173, for all households of age 25~34~ household assisted buyers were age 25-34 Race 82 percent of $150, for all white households8 assisted buyers were white aHome owner and renter households. Note: All dollar amounts are in 1986 dollars. Sources: See text Use of the median value for debt payments implies that half of those in the Survey’s results had consumer payments in excessof $160 per month. As such, a closer estimate of the extent to which consumer debt payments would affect our test results would be to apply the distribu- tion of higher debt payment amounts to assistedowners in our data base.However, we did not have the Survey data from which to make such a distribution. Nor did we have a way to allocate higher debt pay- ments to those in our data base. Instead, we performed our debt test again to determine how the results would change if all assistedbuyers had debt payments in excessof the $160 median value using three different amounts. We found that, if all buyers in our data basehad consumer debt payments of $180 per - month, the percentageof assistedowners passing this test decreased only slightly from 61 percent (if all assistedbuyers had installment pay- ments of $160 monthly) to 49 percent. Increasing the debt payments to $200 per month and $220 per month reduced the number of households passing the debt test to 46 percent and 41 percent, respectively. Since it is unlikely that all buyers had debt payments of this magnitude, we believe that our results are reasonable. To determine the effect of this secondtest on the assistedowners, we applied it only to those owners who had a housing-expense-to-income ratio of 28 percent or less, using the samegeneral procedure outlined in our March 1988 report3 If the assistedowner’s (1) housing-expense-to- 3Seepp. 79-80 of (GAO/RCED-88-111, Mar. 28, 1988). Page00 GAO/RCED-90417 Recapturingthe QualifiedMortgageBondSubeidy AppendixV ApproachUeedin ApplySnga Total-Debb Expens&o-IncomeTestto AssistedBuyer DataBase income ratio was 28 percent or less and (2) total-debt-expense-to-income ratio was 36 percent or less, we then concluded that the assistedbuyer may have been able to afford the samehome at the sametime with a conventional fixed-rate loan. Housing agency files that we obtained for our March 1988 report sizable amounts of missing data and often required substantial effort to make them usable. As a result, we elected to use the Center’s data rather than trying to obtain a complete data set for the total debt expensevariable from housing agency files. Page01 GAO/RCED-ftO-117 Recapturingthe QualifiedMortgageBondSubsidy Appendix VI Comments From the Association of Local Housing F’inanceAgencies Note: GAO comments supplementing those in the report text appear at the end of this appendix. alhfa ASSOCIATION OF LOCAL HOUSING FINANCE AGENCIES 1101 Connecticut Ave.. N.W.. Suite 700. Washington, D.C. 20036 * 202/857-1197 May 30, 1990 omurs Mr. John M. Ols, Jr. P188ldBnl Director, Housing and Community Walter D Webdalo Development Issues Fa~rlax County, Varginta U.S. General Accounting Office Vice President Washington, DC 20548 Charles Brass New York, New York Dear Mr. 01s: Secrelery Jay Jensen This letter responds to your invitation to Muvuapolis, Mnnesola the Association of Local Housing Finance TlClilSW3~ Agencies (ALHFA) to provide comments on GAO's Barbara T Smllh draft report "Limiting Mortgage Assistance to San Frsrwsco. Cal~lorn~e Owners with Rapid Income Growth." lmmedlale Past President Kennelh Johnson The report makes several findings: St Paul. Mlnnasola DIrectora o The current recapture mechanism does not treat equitably all owners with income George Arendas Allegheny County, Pennsylvania increases; o Some owners with large increases in income WIlllam R Bruce will continue to receive assistance after Memphis, Tennessee their incomes could support housing Darlene K&da payments without it; and P~nellas County, FlorIda o Bond-assisted loans are made to those who Angus 1 Olson could afford market-rate conventional Alexand,,a. V,rgin,s loans with comparable terms, the mortgage bond program imposes high costs on the David Perel Los Angeles. Caklorn~a Treasury, and the program benefits are relatively small. Robert L (Robb) PII& Atlanta. Georgia Macheel F Schubert “Promam Not Needed” ChIcago. lll,no,s Charles Taylor This last finding is a reiteration of that See comment 1. Los Angeles County. CaMorrva which GAO made in its 1988 report, "Home Ownership: Mortgage Bonds Are Costly and ALHFA Slall Provide Little Assistance to Those in Need." John C Murphy ALHFA challenged that finding then and does Execuwe D~reclor so again. Mary Kale Unbe DlreClOr Conventional Underwriting Criteria Program Development Kalher~ne E Llboro Apart from mentioning a 28-percent mortgage Admux%wve Aswanl payment-to-income ratio, the GAO staff in 1988 did not enumerate any other criteria. Wdkam F N,emeyer Prqecl Assislanl It appears that on this basis alone, GAO Y SPRING EDUCATIONAL CONFERENCE. MAY 9-l 1, 1990. WASHINGTON, D.C. Page62 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVI Comment-gFkomthe Associationof Local HousingFinanceAgencies 2 concluded that fully 67 percent of assisted homebuyers could have purchased homes with unassisted conventional financing (56 percent with 30-year fixed-rate loans, and 11 percent with adjustable-rate loans). How did GAOqualify these assisted buyers for conventional mortgages? Did GAOuse the different underwriting criteria applicable to adjustable-rate loans? Did GAO consider credit histories, employment histories, and other financial obligations? Did GAO calculate the standard second ratio that compares the mortgage payment and other regular monthly expenses with income? Did GAO actually determine that mortgage bankers would have extended the conventional mortgages to assisted buyers simply because they had a front ratio of 20 percent or below? The conclusion reached by this dubious methodology is even more lacking in validity today because of the income limits imposed by the 1986 and 1988 Tax Acts. ALHFA contends that this simplified qualifying procedure reflects the GAO's failure to perform reality testing on its conclusions. Based on our conversations with association members contacted by GAO for this report, we learned that the study amounted to an exercise in data gathering and manipulation, in other words a preoccupation with data files. Regretfully, it appears that the study group made little or no effort to learn the program content or context in the individual localities. ALHFA maintains that a methodology reliant entirely on numbers is inadequate and inappropriate for such program analysis. Revenue Loss Estimates The GAO staff reported in 1988 that they referred to work done by the Joint Tax Committee staff and "others" to estimate revenue losses from tax exemptions on the bonds. Who are the other sources? What are their assumptions and methodologies? What are their predispositions regarding bond-financed housing programs? Interested parties often use these revenue loss estimates to comment on the relative efficiency of bond-financed programs in promoting homeownership. ALHFA contends that GAO's revenue loss discussion missed an important efficiency consideration. Housing finance agencies make carefully informed decisions to issue housing bonds at very specific times when they perceive a demand for affordable mortgage financing. In the absence of this demand, agencies do not issue bonds; they take action that results in federal revenue loss only in an economic environment that otherwise discourages homeownership opportunities for low- and moderate-income, first-time homebuyers. In some instances, the economic environment in which agencies issued mortgage revenue bonds changes so that the Page83 GAO/RCEDSO-117 Recapturingthe QualifiedMortgageBondSubsidy CkmunentvFromthe Association of Local HousingFinanceAgencies 3 affordability of assisted mortgages resemble that of conventional mortgages. This results from unexpected falling interest rates, over which housing finance agencies have no control. In this environment, first-time buyers tend to opt for the administratively-simpler conventional loans; the assisted loan funds do not move, the agency calls the bonds, and the federal revenue loss related to that issue ceases once the bonds are retired. Federal revenue loss discussions should not ignore or underestimate the ability of housing finance,agencies to make responsible economic decisions about issuing bonds; they do not compulsively or carelessly issue housing bonds. Likewise, such discussions should not ignore the fact that Congress and the President accepted revenue losses generated from this program as a tradeoff for the public gains also generated. Benefit Estimates ALHFA believes that the analysis in 1988 emphasized "statistical significance" at the expense of what is really "important." GAO must not lose sight of the fact that Congress created the MRB program for housing finance agencies to achieve public policy objectives and to create public benefits. Over the years, Congress has modified the program to further ensure this public purpose. ALHFA contends that GAO largely ignored this public character in its analysis, as revealed most vividly in the program benefit estimates. It does not appear that GAO considered the intangible benefits of homeownership - improved self-esteem, self-worth, pride, and sense of responsibility and of the community; the benefits from achieving public policy objectives such as increasing homeownership rates, improving affordability, and expanding the housing stock; the community development impacts of the MRB program; or the value of rejuvenating residential neighborhoods in distressed areas. GAO did not count the employment, income, and tax revenue benefits generated by new housing construction. GAO’s data indicate that over 100,OO units of new construction occurred, creating approximately 176,000 jobs as estimated by the National Association of Homebuilders. It did not estimate the value of protecting low- and moderate-income first-time buyers from interest-rate risk inherent in the conventional market. GAO did not address the counter-cyclical benefits yielded by bond- financed homeownership programs. It even underestimated the buyers' direct economic benefit by focusing on monthly payment savings alone and ignoring the up-front benefits associated with the program. In fact, lenders are the first to observe that the downpayment and points are more critical factors than interest Page64 GAO/RCED-90-117 Recapturingthe QoalifiedMortgageBondSubsidy Comment4Fromthe Aawciation of Local HousingFinanceAgencies 4 rate8 in determining affordability. The MRB program further enhances affordability by permitting these things to be financed in the loan. Assertingthat the MRB program yields only lower monthly mortgage - as implied in GAO’s coats benefit estimate - grossly underestimates the economic, public policy, and personal benefits which our agencies have observed in their communities over the life of their bond-financed mortgage programs. It should also be noted as ALHFA did in 1988 that the 1988 GAO study included findings which reflected favorably on MRB programs: o Most assisted buyers were in the "25 to 29" age category and lived in a household comprised of two people; this suggests that single-family bond programs bring low- and moderate-income individuals into the housing market early in their adult years permitting them to enjoy homeownership benefits sooner than the conventional market permits; o Forty percent of assisted buyers purchased new homes and 80 percent of bond issuers eet aside some portion of bond proceeds for developers; this demonstrates that single- family bond programs add significantly to new housing construction and the nation's overall affordable housing stock, providing increased homeownership opportunitiest and o Assisted homebuyers purchased homes that cost 70 percent of the average purchase price and 80 percent of assisted buyers had incomes at or below 11.5 percent of the area median income; this suggests that housing finance agencies administered single-family programs before the 1986 Tax Act in a manner that generally complied with the Act's new income and price restrictions. These findings suggest that housing finance agencies have largely succeeded in achieving the very fundamental objectives of the Mortgage Revenue Bond Program enunciated by Congress: to encourage homeownership among low- and moderate-income households by providing an incentive to purchase in the form of an affordable mortgage, and to expand homeownership opportunities for such households by expanding the affordable housing stock. It is not surprising then that the MRB program enjoys widespread, bipartisan support within Congress. H.R. 1200, the bill to extend the sunset on authority to issue MRBs/MCCs, currently has 364 cosponsors while its companion, 5.355, has 85 cosponsors. The GAO should recognize this fact and lay to rest its notion that the program is somehow not needed. Y Page85 GAO/RCED-80417 Recapturingthe QualifiedMorrgageBondSuMdy APPsndix M C!ammente Prom the Auodation of Imal HolmingFlnrnee AgencIee 5 “RWaRtwe Does Not Treat Euuitsblv Those with Income Increaseegg GAO states that, while the recapture provisions will recover some or all of the subsidy received by assisted homeowners, it does not treat equitably all owners with income increases. Some owners will receive benefit larger than the uniform 1.25-percent interest rate differential, while others will receive less benefit. In such cases the owner could pay more in recapture than the owner received in subsidy. GAO suggests that this situation could be overcome by basing the recapture amount on the monthly payment reduction. Of all the elements of the recapture mechanism, the 1.25- percent differential is perhaps the simplest, easy to understand by borrower and lender alike. The situation which GAO See comment 2. ;~~thseisee of a borrower receiving a subsidy of less than 1.25 . ., 1.00 percent) is not likely to occur. Housing finance agencie8 simply would not irsue bonds when the spread between tax-exempt and conventional rates is this small. In instances where the subsidy was greater than 1.25 percent, the public purpose of assisting homebuyers through providing a meaningful oubeidy would be achieved. Therefore, ALHFA does not believe any change in the 1.2%percent assumed subsidy differential is warranted. Requiring lenders to calculate the actual benefit against some measure of conventional interest rates would needlessly enaumber the program. GAO also aassrte that the limitation of the recapture amount to no more than 50 percent of the gain-on-sale is inequitable "because those with rapid income increases may escape paying it if they live in areas with low or no housing price appreciation." The Report goes on to say "by allowing owners to retain some equity for a rsplacement house purchase, the gain-on-sale provision fulfills a purpose: it does not force the owners to return to renter status because of a large recapture amount." ALHFA could not agree more that denial of some share of the gain- on-sale to the owner could very well return such owner to renter status. This event would be completely contrary to the fundamental purpose of the MRB program! to stimulate homeownership. Permitting the owner to retain 50 percent of the home's appreciation does enable the owner to move up, thereby freeing up his/her more modestly priced home for potentially another first-time homebuyer. That the gain-on-sale provisions treats those in high appreciation versus low/no appreciation housing markets differsntly is no reason to eliminate the gain-on-sale provision. See comment 3. GAO then asserts that the phaseout of the recapture from years 6 to 10 'I... is not consistent with the benefits principle since those who receive greater total benefits over time are subject to a lesser recapture amount than those who receive the Page66 GAO/RCED-9lH17RecapturingtbeQuaMedMortgageBondSul~idy AppendixvI CommentsFkomtheAssodationofLocxd HonsingFinanceAgendea 6 same annual benefit, but over a shorter period of time." GAO suggests the this situation could be remedied by reversing the recapture provision to allow the recapture amount after 5 years to increase at the full recapture rate until the home was sold, allow it to increase at a reduced rate, or establish a plateau from which the amount would neither increase nor decrease. GAO’s discussion of the phaseout treats it as a negative effect: that persons assisted should be discouraged from remaining in their home. Recapture should have the opposite public policy objective as the phaseout has -- encouraging See comment 4. neighborhood stabilization through long-term homeownership. "Subsidy Allowed to Continue Despite Obvious Significant Income Increases" GAO states that, because recapture is imposed after the owner sells the house, it is easier and less costly to administer, since the amount is computed once, as opposed to periodic determination(s) of whether there is a continuing need for assistance. ALHFA agrees that it is easier and less costly to administer. GAO goes on to say that some owners' income will increase to the point where they do not need continued assistance to remain homeowners. Again, ALHFA agrees with this statement. However, ALHFA sees one of two things happening in such a situation: one, the owner sells the home and pays off the MRB-financed loan. These proceeds are then used to retire the bonds, thereby saving the federal government money, or the proceeds are relent to a qualifying homebuyer; or, two, the owner remains in the home lending the stability of homeownership to the neighborhood. In either case a proper public purpose is served. “A Prospective Affordability Test is Needed" The Report again notes that GAO's 1988 study concluded that a majority of those assisted between 1983 and 1986 would have obtained fixed-rate conventional loans at the time they received MRB assistance. Even more could have qualified for adjustable- rate mortgages. GAO then asserts that to remedy this situation lenders would have to determine whether those applying for MRB loans could qualify for conventional loans. If they could, they would be denied the MRB assistance. GAO admits in a footnote, as it did in 1988, that it reached its conclusion that most of those receiving MRB assistance could have qualified for a conventional loan, based solely on the industry housing-expense-to-income test. It readily concedes that it @I... did not use other information used in the loan- Page67 GAO/RCEIHO-lI7RecapturingtheQualifiedMortgageRondSuhsidy AppendixVI Cvmmenta From the Atmoeiationof Local HousingFinanceAgencies 7 making process, such as housing-expense-total-debt ratios, credit histories, and employment records, " dismissing their importance by saying that the housing finance agencies computerized records did not contain it and that regardless it would not have changed See comment 5. the outcome. This statement is so incredible that it calls into question the validity of the 1988 report as well as this draft report. These additional factors are critical and are often the deciding factor in helping low- and moderate-income households who cannot qualify for conventional loans. Housing finance agencies use private lenders to originate loans and to work with borrowers to make loans. MRB-financed loans have more qualifying and administrative requirements than conventional loans with which both lenders and borrowers must comply. Given a choice between a conventional loan and a MRB- financed loan, we believe that the borrower who anticipates a near-future rise in income will choose the conventional loan to avoid the administrative and programmatic complexities associated with an MRB-financed loan. ALHFA contends that they have made this choice all along considering that, as GAO pointed out in the 1988 study, the average life of an MRB-assisted loan was 7 years. This relatively long tenure, coupled with the targeting restrictions enacted in 1986 and 1988, makes the MRB program one not for the 'tyuppie*V as GAO would have us believe, but for owners who cannot qualify for a conventional loan. ALHFA concludes therefore that such an affordability test is unnecessary and See comment 6. would further complicate rather than simplify the program. %lternatives to the Rxisting Recapture MechanisW GAO asserts that "the current recapture mechanism is preferable to none at all because some or all of the subsidy will be recaptured from at least some owners who probably could have waited a short time and bought a home without assistance due to rapid income growth." GAO suggests two alternatives: a roll- over loan where assistance is discontinued, after periodic reviews, based on preestablished affordability criteria, and income-based variable payment mortgage loans whereby housing payments are increased periodically as long as income increases. In the former alternative, "if the owner's income rises to a point where the owner could afford a comparable market-rate loan, the owner is required to refinance the assisted loan with a market-rate loan. If income growth is insufficient for the owner to afford a market-rate loan, then assistance is continued." GAO correctly states that a roll-over loan introduces costs and other complexities to the program for both borrowers and lenders. Borrowers must bear refinancing costs which might be substantial and more than the subsidy received. It will also increase administrative burdens for the lender by (rejcertifying Page08 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBond Subsidy AppendixVI CommentsFromtheAsseciationofLecsl HoueingPinanceAgeneies a income and determining what the market rate should be. ALHFA finds no justification for Congress to choose this as a See comment 7. preferred alternative to the existing recapture mechanism. In the second alternative, the owner's payment for principal, interest, taxes, and insurance are a fixed percentage of household income. As income increases, so do monthly payments even to the point where the interest rate exceeds the conventional rate. At this point, according to GAO, there would be an incentive for the household to leave the program. GAO again correctly asserts that this alternative 'I... will increase administrative cost and program complexity over the current recapture approach." It would even be more costly than the roll-over approach due to the continuing need to adjust loan payment schedules leading to increased costs. Here again ALHFA finds no justification for Congress to choee this alternative over the current recapture. Moreover, such a requirement would conflict with normal mortgage underwriting practices whereby lenders assume some income growth over time on the part of the borrowers, thereby increasing the loan's security. If the payment-to-income ratio See comment 8. were held constant through an adjustment to the interest rate, lenders would compensate by imposing lower credit-to-income ratios thereby targeting away from lower-income to higher-income borrowers, which would not be a desirable public policy outcome. In wnmary, ALHFA finds no recommendation in the report which would make the recapture mechanism, or alternatives thereto, simpler or easier to administer. We find too that GAO's continued assertions that the program is not needed based on faulty and aged data are inaccurate, and a disservice to the program's congressional supporters, the local and state agencies which administer the program, and to the first-time homebuyers whom Congresa has chosen to assist. Sincerely, Page69 GAO/RCED-~117Recapturingtl1eQualWiedMortgageBondSubsidy AppendixVI Canme~~t.e Fromthe Aeeoclationof Lintal HoudngPlnanceAgencies The following are GAO'S comments on the Association of Local Housing Finance Agencies’ letter dated May 30, 1990. 1. The first four pagesof the Association’s comments reiterate, mostly GAO Comments word-for-word, its 1988 comments disagreeing with most aspectsof our March 1988 report. The comments largely relate to topics in our earlier report and not to those in this report. Seepp. 94-106 of Home Owner- ship: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GAO/RCED-88-111, Mar. 28,1988) for the Association’s comments and our responses. 2. The Association states that home owners are not likely to receive an (after-tax) subsidy of less than 1.26 percent becausehousing agencies will not issue bonds when the interest rate-spread between tax-exempt bonds and conventional mortgage rates is this small. The Association’s comments ignore the fact that interest rates sometimesdecline after the bonds are issued and before loans are made, thus reducing the interest rate differential the home owner can achieve. The 1982-87period was one of generally declining interest rates. Our 1988 report estimated that one-quarter of the buyers in our 1988 study received interest rate reduc- tions of three-quarters of a percentagepoint or less from the conven- tional interest rate. Finally, the Association believes that tailoring the recapture amount more closely to the actual interest-rate reduction would needlessly encumber the program. We believe that the procedure we propose in chapter 3 would only minimally increase administrative requirements. The alternative is to decreasethe attractiveness of the program to some prospective buyers who learn that the recapture mechanism may, in someinstances,take away more than the subsidy they are likely to receive. 3. Neither GAO'S draft report nor this final report recommendedelimi- nating or changing the gain-on-saleprovision. 4. The Association n&characterizes our discussion.The enacted phase- out mechanism reducesthe recapture amount for those who benefit the most (as measured by length of time in the assistedhome) and who exceedthe income test thresholds. 6. We added an analysis of housing agency loan files using a total debt- expense-to-incometest in this final report. Mortgage lenders qualifying Page70 Becspturjng GAO/RCED-90-117 the QuaUf’ied MortgageBondSubsidy Chnment.8 From the Association of Local HouelngFinanceAgencies prospective QMBloan and MCCrecipients review credit histories and employment histories to ensure that the applicants are good credit risks. We did not need to do so again. 6. As discussedin appendix IV, a typical prospective buyer may find it financially worthwhile to obtain a QMBloan, no matter what the buyer’s income expectations, even with the imposition of a potential recapture. 7. Our draft report outlined an approach that would not have required loans to be refinanced. The approach we suggestedfollows that of FII-IHA’Ssection 602 program and avoids interest-rate risk on the part of the assistedbuyer and the imposition of refinancing costs.The Associa- tion is incorrect in suggestingthat our approach would require refi- nancing and the attendant costs to do so. However, we have reworded our discussionto further clarify this point. 8. Adjustable rate mortgages,which can result in increasedmonthly payments-even if income does not increase-are used widely in the conventional marketplace. Thus, the assertedconflict does not exist. Roll-over approachesare also used elsewhere.Seeappendix IV of this report for a description of HUD'Ssection 236 and F’mHA’Ssection 602 home ownership programs and two local housing finance agenciesthat use a roll-over loan mechanism. Page71 GAO/RCED-O-117 Recapturingthe Qualified MortgageBondSubsidy Appendix VII I Comments From the National Council of State Housing Agencies Note: GAO comments supplementing those in the report text appear at the end of this appendix. National June 5, 1990 Mr. John M. Ols, Jr. Director. Houslng and Community Development Issues Unit;;eStates General Accounting Washington, D.C. 20548 Dear Mr. 01s: Thank you for providing the National Council of State Housing Agencies (NCSHA) the opportunity to comment on the General Accounting Office’s draft report, “Homeownership: Limiting Mortgage Assistance Provided to Owners With Rapid Income Growth.” The enclosed comments are submitted to you on behalf of NCSHA’s member State Housing Finance Agencies. If you have any questions on our comments, please do not hesitate to call me. We look forward to your final report. Sincerely. :r John T. McEvoy Executive Director cc: James Ratzenberger Dennis Fricke enclosure 444 North Gpitol Street, N. W., Suite 118, Washington. D.C. 20001 (202) 624.7710 fix (202) 624.7719 Page 72 GAO/RCED-!30-117 Recapturing the Qualified Mortgage Bond Subsidy AppendixVII Comments Fromthe NationalCouncilof StateHousingAgencies Comments on the General Accounting Of&x’s Draft Report, “Homeownership: Limiting Mortgage Assistance Provided to Owners With Rapid Income Growth” Submitted by the National Council of State Housing Agencies June 1.1990 The National Council of State Housing Agencies (NCSHA) is pleased to be able to submit these comments on behalf of our member Housing Finance Agencies (HFAs) on the General Accounting Office’s (GAO’s) report, “Homeownership: Limiting Mortgage Assistance Provided to Owners With Rapid Income Growth.” That report comments upon the MRB recapture provision of the current Internal Revenue Code. NCSHA considers that the MRB recapture provision represents a highly complex overreaction which burdens the entire MRS program to penalize a few theoretical homebuyers whose existence is documented only anecdotally at best. Because the recapture provision is expensive to administer, raises difficult issuer and lender liability questions, and may frighten eligible homebuyers away from buying their first home, we believe it should be repealed. If it is not repealed, we believe it urgently needs substantial simplification. These NCSHA comments are limited to the simplification issue, since they are a response to the GAO report on improving the existing recapture provision. NCSHA is a national, not-for-profit organization created in 1970 to represent the interests of State HFAs in low and moderate income housing. HFAs in every state, the District of Columbia, Puerto Rico, and the Virgin Islands respond to low and moderate income housing needs through the financing, development, and preservation of affordable rental and ownership housing. HFAs collectively operate more than 350 affordable housing programs, which range from homeownership to homeless initiatives. NCSHA members have helped more than l.OOO,OOO low and moderate income Americans buy their first home through Mortgage Revenue Bond (MRBl programs in every state. These HFAs issued approximately $6 billion in MFtE3s in 1989 under the private activity bond cap which strictly limits the overall volume of such bonds. In 19 states, these HFAs also administer Mortgage Credit Certificate (MCC) programs. Local government agencies also administer MRS and MCC programs in several parts of the country. NCSHA’s member HFAs have also financed over 500,000 units of rental housing for low and moderate income tenants through reduced interest rate mortgages financed by tax-exempt bonds. Since 1986, NCSHA members have financed 235.000 low income rental units with the Low-Income Houslng Tax Credit for households with Incomes at 60 percent or less of area median. Y Page73 GAO/WED-99-117 Recapturingthe Qualifkd MortgageBondSubsidy Append& VU CommentaFromthe NationalCouncilof StateHousingAgencies The MRB/MCC (MRBl program, which is scheduled to expire this year, is the only generally available federal mortgage payment assistance program in today’s disheartening environment of climbing home prices and declining homeownership rates. An overwhelming 85 percent majority of the Congress -- 454 of the 535 House and Senate members, including 70 percent of the members of both tax-writing committees -- have co- sponsored legislation in this Congress to extend this time-honored, well- tested, and Congressionally limited program to make the national goal of homeownership attainable by lower income Americans. The NCSHA and its member agencies hope that Congress will this year extend MRBs permanently to eliminate what has become a pointless and time-consuming review for both Congress and MRB issuers of a program which has been tightly targeted by Congress and has proved its value and popularity in the states. At the same time, we hope Congress will substantially simplify the program’s recapture requirement. found in Section 143(m) of the Internal Revenue Code, which becomes effective for MRB loans made and MCCs issued after December 31. 1989. Without such a simplification, the cost of administering the MRB program will be significantly increased without any benefit to the federal government. Moreover, deserving, but unsophisticated, borrowers may be frightened away by the complexity of the recapture provision, and many MRB-assisted homebuyers will face the need to obtain professional assistance to file their tax returns when they sell their homes. This response to the GAO report is divided into three parts: l First. analysis of the goals and complexity of Section 143(m) of the Internal Revenue Code, the “MRB recapture” provision enacted in 1988. which is the subject of the GAO report. l Second, a discussion of the unfounded bias against the MRB program displayed in the GAO report which undermines the credibility of its recapture recommendations. l Third, an analysis of the GAO’s recapture recommendations. Section XWfml: The Recwture Provision Existing provisions of the tax code limit MRB assistance to lower income purchasers who buy less than average priced homes. The price of an MRB-financed home cannot exceed 90 percent of the average area home purchase price. MRB borrower income may not’ exceed 115 percent (100 percent for families of fewer than three persons) of the higher of the area or statewide median gross incomes for the area in which the residence is located. 2 Page74 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Appendix VII Comments Fromthe NationalCkmncilof StateHousingAgencies In practice, HFAs frequently set their MRB-assisted home prices and urchaser incomes well below the federal limits. In many areas, however, Rousing costs are high relative to incomes. In such areas, incomes even hlgher than the federal limits are frequently necessary to purchase even lower priced homes, so the MRB program limits inhibit use of MRRs in these markets. Last year, conventionally financed home prices averaged 37 percent more than State HFA MRB-financed homes. Conventional borrowers’ incomes averaged 44 percent above MRB borrowers’ average incomes. Congress enacted Section 143(m) of the Internal Revenue Code in 1988 to discourage buyers who expect their income to increase faster than a simple five percent per year above the MRB program limits from using MRB mortgages. When an MRB-assisted home is resold, Section 143(m) requires MRB-assisted owners to repay to the federal government some or all of any MRB benefit they obtained while they owned the home. The amount of recapture penalty MRB purchasers might owe under Section 143(m) depends upon how much their incomes have risen, their family size at the time of sale, the size of their mortgages, and the length of time they owned their home. NCSHA believes that the recapture proposal represents a highly complex solution to a problem of minor proportions, if a problem exists at all. Thus, NCSHA believes that the recapture provision should be simply repealed. NCSIiA agrees with the objective of discouraging those from getting an MRB mortgage for whom conventional financing is a reasonable alternative. However, the Congress has made no finding beyond anecdotal evidence of significant and rapid income growth among MRB beneficiaries. Nor has NCSHA found evidence to support the perception that MRB beneficiaries experience significant and rapid income growth. In fact, occupational surveys conducted by several HFAs suggest that MRB borrowers are not upwardly mobile families whose salaries climb rapidly, but instead are teachers, firemen, factory workers, and service industry people whose salaries rise, if at all, only at approximately the rate of inflation. We agree that recapture may be appropriate for those whose incomes rise outside the MRB eligibility range while occupying an MRB-assisted residence. The current provision, however, is like a cannon trained upon a gnat, in terms of the complexity imposed upon the entire MRR program, to “catch” a few buyers whose incomes may rise rapidly and who may exist more in anecdote than in reality. For these reasons, we are urging the Congress to consider simplifying current law without diminishing the intended effect of recapture and to avoid adding additional complexities to the law which meet no demonstrated need. 3 Page75 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII 0xnmenta Fromthe NationalCouncilof StateHoueingAgencies Section 143(m) involves a process of computation which contains up to 13 sometimes complicated mathematical computations to determine the recapture amount. A summary example of how Section 143(m) would work in an ordinary home sale consumes a single-space page of the GAO report, complete with two footnotes. When the Congress enacted the recapture provision in 1988. it delayed the effective date in order to permit consideration of alternatives to attain its policy goal. As part of that review, Congress asked GAO to study the recapture provision’s effectiveness and administrability. GAO has now rendered its report. We commend GAO for acknowledging the complexity of the recapture mechanism in that report. However, we are disappointed that GAO offers no solution to this complexity, which will confound borrowers and lenders alike and discourage their participation in the program. The GAO suggests no simplification of the current recapture provision. Instead, demonstrating the truth of the axiom “the perfect is the enemy of the good,” the GAO report poses new reca ture alternatives, the cost, inconvenience, and complexity of which dwar P those of the existing version of Section 143(m) without any assurance of improving its results. In fact, GAO’s proposed revisions and alternatives to the current provision would further complicate its application and are completely unworkable. Each of the GAO suggestions would add additional layers of complexity to a provision whose intended purpose can be well served by a limited simplification of its existing provisions, without the introduction of still new complications. possible RcCaDttW SimDlification The current recapture provision contains up to 13 mathematical calculations to determine the recapture amount. At least one simplification (described below) of Section 143(m) reaches virtually the same mathematical result as the current law, but reduces the required mathematical computations to a maximum of eight. Under current law, recapture is assumed to be potentially payable by all MRB purchasers, but the amount is reduced potentially to zero by application of a set of complex calculations involving up to 13 separate computations. Under a possible simplification, no recapture would be owed unless income exceeded a certain threshold level in the year of, sale. That Threshold Income would be determined by multiplying .05 times the number of years the home was owned and further multiplying that result by the maximum eligible income which would have been available to a family the size of the MRB seller at the time he or she purchased the home (maximum eligible income) and adding that product to the maximum eligible income. 4 Page70 GAO/RCED-90-117 Recapturingthe Qualifkd MortgageBondSubsidy AppendixVII CommentsPromthe NationalCouncilof StateHousingAgencies The current law provision calculates the maximum recapture amount by multiplying 6.25 percent times the amount of the MRB mortgage and multiplying that result by a “holding period percentage” which is derived through a two or three step calculation, depending on whether the home is sold in years one through five or In years six through ten. Under the simplification alternative, the maximum recapture amount would be determined by multiplying the origlnal mortgage amount by ,001 and multiplying that result by the number of months the home is owned up to 60. (The ,001 is the result of dividing 1.25 by 12 months. Sixty months is the period during which current law recapture rises to its maximum possible amount.) The simplification proposal differs significantly from current law in its application after year five. Current law increases the recapture amount by 1.25 percent per year to a maximum of 6.25 in year five, reduces the maximum amount by 1.25 percent per year in years six through ten and eliminates recapture after year ten. Like current law, the simplification proposal assesses the maximum recapture amount in year five, but that amount remains constant in years six and seven and recapture is eliminated in year eight. The reason for this change is that the targeted “yuppie” population will not remain in their homes beyond seven years, if they stay that long. In fact, MRB mortgage prepayment statistics indicate that the vast majority of MRB homeowners sell their homes between years five and seven. The Congress recognized this in the softening of the current recapture penalty in years five through ten. The simplification proposal just eliminates these additional and unnecessary computations. Under current law, the actual amount of recapture is determined by adding $5000 to an income level determined in exactly the same fashion as in the simplified alternative and then reducing the total recapture amount otherwise payable by two percent for every $100 the owner’s income is below this amount. Additionally, the recapture actually owed could not exceed one-half the MRB seller’s net gain on the sale. Under the simplification proposal, no recapture would be owed if income did not rise over the Threshold Income. If the owner’s income exceeded the threshold, recapture would equal the lowest of the following: (1) the maximum recapture amount, (2) 50 percent of the net gain on sale, or (31 .0002 times the amount of income over the threshold times the maximum recapture amount. The effect of both current law and the simplification is that for every $100 the seller’s income Is above the Threshold Income, the seller would pay 2 percent more of the maximum recapture amount, until it is $5,000 above the Threshold Income, at which point the seller would pay full recapture. The actual operation of current law and the possible simplification are demonstrated in the example set out below. The following description illustrates how current law and the proposed simplification would apply to the “Example of How to Compute the 5 Page77 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII CvnuuentePromthe NationalCouncilof StateHousingAgencim Recapture Amount” contained in’ Appendix I to the GAO Re ort (page 52). In the example, the MRB Mortgage is $55,000; the applicab Pe income level for the seller’s household is $23.000; the home is held 6 years and 2 months: the seller’s income at time of sale is $32.000; and the gain on the sale is $12,000. w - Thirteen Steps to Determine Recapture Maxtmum Recapture Equals: $55.000 x .0625 = $3.438 (one step) Holding Pertod AdJustment: ((120-741/60)x $3.438 = $2,636 (three steps) Income Eltgibtllty Llmlt: ($23.000x .05 x 6) + $23,COO+ &WXIO= &34SXl (four steps) Recapture Reductton: (($34,ooo- $32.000)/ 100)x .02 x $2,636 = $1.528 (Toursteps) Recapture Computatlon: $2,636 - $1.528= $1,108 (one step) Recapture equals lesser of the Recapture Computation ($1.108). the Hotding Period Adjustment ($3.6361, or one-half the gain on the house ($6.0001. - Three to Eight Steps (No Recapture below Threshold Threshold Income: ($23.W x .05 x 6) + $23.000 = $29.900 (three steps) M&mum Recapture: $55.OKJx ,001 x 60 = $3,300 (two steps) AdJusted Recapture: 0002 x @32.000 - $29.900) x $3.300 = $1.386 (three steps) Recapture equals lesser of Adjusted Recapture ($1.3861, Maximum Recapture ($3,300). or one-half the gain on the house ($6.0001. In this example, the recapture under the simplification proposal ($1.3861 actually exceeds the recapture under current law, because current law ($1.108) reduces the recapture amount for each year over iIve the house is held, whereas the simplification proposal does not. 6 Page78 GAO/RCED-SO-117 Recapturingthe QualUledMortgage BondSubsidy AppendixVU CommentsFkomthe NationalCouncilof StateHoueiugAgencies The Unfortunate Bias in the Q,AO Resort. The GAO approach to the recapture issue not only overlooks simplification alternatives to existing law. but also reflects an unfounded and undocumented assertion that most MRB-assisted buyers need no assistance and, hence, should have to prove they cannot get credit elsewhere before they get an MRB loan. The GAO has traditionally opposed the MRB program, which is the only federal homeownership mortgage assistance program generally available to lower income individuals and families. Today’s MRB program is the product of a decade of Congressional refinement designed to tighten the MRl3 program as far as practicable to assure that MRB benefits are targeted to those who need them in order to qualify for homeownership at all. Because the MRB program has been caught up with a number of other tax programs - popularly known as “the extenders” -- which have been continued from year to year during the 1980’s for revenue accounting reasons, it has been reviewed about as often - five times - and amended to assure its cost-effectiveness - four times - as any tax provision ever considered. And a nearly unprecedented majority of the Congress - more than 80 percent of all Members in both Houses and 70 percent of each House’s tax-writing committees - have co-sponsored legislation to continue the MRB program in its present form. As a matter of fact, the MRB program is one of the least expensive of the “extenders” or many other provisions of the Internal Revenue Code. The Joint Tax Committee staff estimates that the MRB extension represents only about three percent of the aggregate cost of all 12 extenders. In light of these refinements and overwhelming Congressional support. the Senate Finance Committee last year reported a permanent extension of the MRB program as part of its version of the 1989 Tax Act, notwithstanding the Gramm/Rudman Act, which has led Congress to provide only short-term extensions for other “extenders,” in order to meet deficit reduction targets. That permanent MRB extension, however, along with most of the rest of the Finance Committee bill, was dropped in the subsequent maneuvering over the capital gains issue. But just as Congress has consistently found the highly targeted, recapture-limited MRB program to be overwhelmingly worthwhile. some academic critics have persisted in comparing it to goals Congress does not share and ignoring the bulk of all the evidence which impeaches their criticism. Regrettably, the GAO study of Section 143(m) partakes richly in this prejudice. It ignores the Congressional determination embodied in the MRB program to provide a limited chance for homeownership to lower income families. The GAO report is undermined by repeated reference to a previous 7 Page79 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII CommentsProm the National Council of State Housing Agencies GAO report, “Homeownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those In Need,” in which GAO criticized the MRB program. But this 1988 report was fatally flawed, because it was based on data the GAO See comment 1 gathered selectively and admitted was at best inconclusive and manipulated with invalidly hypothesized assumptions about the actual real world practice of making MRB mortgages. Proceeding on this unfounded prejudgement that the MRF3 program is full of undeserving participants, the GAO unsurprisingly finds fault with the recapture mechanism Congress enacted in 1988 to discourage buyers from the program who expect an extraordinary income increase or to recover the benefit from others who receive such an increase unexpectedly while MRB owners. Surprising, however, is the additional complexity and further burden GAO asks the Congress to consider imposing on the MRB program for all buyers, without any convincing statement of need for such steps based on any serious examination of current or statistically valid MRB program data. We strongly object to GAO’s repeated assertion throughout its report that a majority of MRB beneficiaries could have purchased the same house without help. using a conventional mortgage. We are dlsmayed that GAO continues to rely on a view first expressed in its dated 1988 report, which was based on incomplete and obsolete information, contained highly controversial conclusions. and became completely irrelevant after Congress further restricted MRB eligibility and enacted the recapture provision later that same year. NCSHA successfully rebutted the conclusions of GAO’s 1988 study in a 1988 Occasional Paper of Georgetown University’s Center for Public Policy prepared by Dr. Margaret Wrightson and funded by NCSHA and in a 1988 report -- A Referendum on the American Dream -- prepared in cooperation with the Association of Local Housing Finance Agencies, the National Association of Realtors, and the National Association of Home Builders. The flndings of these reports have been supported by data collected by NCSHA in the years since their release. GAO has never updated its 1988 report to reflect the MRB program tightening Congress enacted that year, acknowledged the thorough rebuttal provided by the Georgetown and NCSHA responses, or seriously addressed the inaccuracies those studies demonstrated in GAO’s conclusions regarding MRBs. We are aware of no credible evidence that MRB beneficiaries could, as GAO claims, buy a home without help. The 1988 GAO report containing See comment 2. these assertions was not based on a random or impartial selection of state program data, but instead, by GAO’s own admission, was conducted with a skewed methodology. As the 1988 report stated on page 16 . . .“Because we selected housing agencies judgmentally, we cannot assert that our findings See comment 3. are representative of qualified mortgage bond activity nationwide.” 8 Page80 GAO/RCED-90-117 Recapturingthe QualifiedMortgage BondSubsidy ‘, AppendixVII CommentsFromthe NationalCouucUof StateHousiugAgencies GAO’s conclusions are contradicted by any impartial, statistically valid review of the same and comparable data. The Georgetown study used data from the same years as the GAO study, collected from an unbiased sampling of data from 17 HFAs. The Georgetown study included eight of the states GAO studied whose data was sufficient to permit a statistically valid review and an additional nine states selected to replace the states in the GAO study in which the data was inadequate to support a statistically valid review. These replacement states, if anything, tended to be “high cost areas,” data from which would tend to support GAO’s conclusions more strongly, if, indeed, those conclusions were correct. The Georgetown study found that MRB household income and purchase price limitations established by Congress in 1980 (and gradually made more stringent since) have been very effective in targeting MRB See comment 4 assistance to lower income homebuyers. It revealed that many states are serving even lower income people than mandated by the federal guidelines. It found that the incomes of the population served and the prtces of the homes financed with MRBs are well below those represented by conventional, or even FHA or VA, sales. According to the Georgetown study, in 1987 (the last year studied by either GAO or Georgetown). the median income of an MRB recipient was $27,000, compared to the 47 percent higher $39.600 income of the conventional buyer. The average price of a MRB-assisted home was $62.000, compared to the 34 percent higher average conventional buyer purchase price of $82.900. The gap between MRB and conventional borrower home purchase prices and income continued to widen in 1988 and 1989. According to data compiled by the NCSHA in its Annual Survey of the 58 State HFAs, the average purchase price of an MRB-assisted home was $52,597 in 1988 and $59,377 in 1989, compared to conventionally financed average first-time buyer prices of $94,400 in 1988 and $95.000 in 1989. The average income of an MRB borrower was $23.071 in 1988 and $25.019 in 1989. compared to conventional borrower incomes of $44,200 in 1988 and $45,000 in 1989. In addition, GAO’s conclusions were based on underwriting assumptions which it invented and did not reflect actual market practice, such as the FHA underwriting standards used at that time which were commonplace in the MRB program. The underwriting ratio of permitted See comment 5. expenses-to-income GAO used in its study helped predetermine its conclusion that MRB recipients could have qualified for conventional loans. Standard underwriting ratios, on both conventional loans and adjustable rate mortgages, are stricter than those used by GAO. These standard ratios include a formula that penalizes the amount of the borrower’s household debt in determining mortgage credit worthiness. This debt factor is often the toughest hurdle for first-time homebuyers to jump. Yet, GAO did not include it in its underwriting assumptions. Y Page81 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII CommentsFromthe NationalCouncilof StateHousingAgencies As an consequence of this bias, the GAO recommends requiring that potential MRB borrowers first obtain proof that they could not qualify for a conventional loan before receiving MRB assistance. NCSHA believes that See comment 6. such a requirement is unwarranted, as the existing law recapture provision is effectively designed to discourage participation in the MFLB program by any potential buyers who are within the MRB income eligibility range and who may also qualify for a conventional loan. The threat of recapture will outweigh any possible MRB interest rate advantage for such buyers. To require lenders to essentially deny conventional credit to borrowers as a precondition for making an MRB loan, could potentially cause them to issue adverse actions against borrowers which may become part of the borrowers’ credit histories and may leave lenders vulnerable to legal See comment 7. action. Even a lesser requirement of prequalifying borrowers for both MRE3 and conventional loans would impose additional administrative requirements on lenders, who already have little incentive to participate in the MRB program. and Alternatives Are Unworkable GAO suggests that the Congress either retain the current recapture mechanism with some modifications or replace it with one of two alternative approaches. The proposed modifications to the existing recapture provision would further complicate it and make its administration more difficult without furthering recapture’s public policy objective. The alternatives GAO suggests are even more complex than the current law recapture provision and would impose insurmountable administrative and financial burdens on MRB borrowers and housing agencies alike, wfthout measurably improving recapture efficiency. GAO suggests that the current law recapture provision be modified to base the recapture amount on the actual interest rate benefit received by MRB borrowers compared to conventional rates. The 1.25 percent per year provided in current law is an approximation of that benefit. Under the GAO proposal, lenders would have to document the difference in monthly principal and interest payments between the MRE3 loan and a comparable conventional loan at the time the MRB loan is closed. When the MRJ3- assisted home was later sold, the amount of recapture would be the amount of that monthly savings times the number of months the loan was outstanding. We strongly object to this approach. First of all, it is unworkable since there is no single conventional rate to which MRB rates can be compared in most real estate markets. Rates vary from bank to bank on a daily basis. For example, the May 26 Washfngton Post survey of 19 lenders in the D.C. See comment 8 metropolitan area showed that these institutions offered as many as eight different effective interest rates with differences as great as 5/8 percent on IO Page82 GAO/RCED-SO-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII CommentaFrom the National Council of StateHousingAgencies a 30-year conformlng conventional loan on a single day. Thus, two Identically situated borrowers could be subject to significantly different recapture amounts If they borrowed from different banks on the same day or even the same bank on different days in the same week. This is an absurd result Congress should not allow. It is important to note that the basis for this GAO suggestion is that the present recapture provlsion might take loo rnti recapture from some MRB-assisted buyers, not too little. The GAO points out that the arbitrary 1.25 percent per year recapture formula of current law overstates the benefit actually received by many MFtB borrowers. GAO’s attempt to achieve mathematical perfection would, however, only further complicate an already exceedingly complex recapture formula. See comment 9. It would replace current law’s standard calculation applicable to all MFW- assisted purchasers with documentation which would have to be individualized for each and every borrower and which would require further administrative and record keeping requirements of lenders. Moreover, GAO would fall to obtain the parity it seeks, for some MREi owners would be required to return every dollar of MRB benefit even if they legitimately qualified for the benefit for most of the time. For example, a borrower’s income could remain well within the MRB eligibility range for the first three years of owning the home and then jump considerably in the See comment 10. fourth year due to a change in employment or the addition of a second household Income. In the year of sale, the total interest rate benefit recefved over the life of the loan would be recaptured, even though the beneflt was deserved in at least years one through three. GAO’s solution to this problem -- suspending or adjusting the benefit as income rises over the life of the mortgage-- introduces a whole new set of complexities and admlnistratlve hurdles which are discussed further on. GAO is also concerned that MRB owners who experience large income increases might remain in their homes for long periods in order to continue receiving a benefit which they do not deserve. In addition, some MRB owners may pay little or no recapture, because the recapture penalty is phased out under existing law during the second five years of ownership and eliminated after ten. GAO also argues that this phase-out will prompt owners to stay in their homes for longer periods than they would otherwise, They suggest that the recapture amount should either increase through year ten, increase through year ten at a reduced rate, or plateau in year five and remain constant through year ten. NCSHA believes that the so-called “yuppies.” the upwardly mobile families with rapidly increasing incomes that recapture was designed to See comment 11 “catch,” are unlikely to remain In their homes after they can afford to move. They will not stay long periods Just to avoid paying recapture and, in fact, are better off selling in the early years to take advantage of the one-time 11 Page83 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII CommentsFromthe NationalQmncil of StateHousingAgencies $5,000 factor added to income under current law to compute the actual recapture total. Finally, GAO suggests that the 50 percent of “gain-on-sale” recapture limitation will cause households in low-appreciation areas to pay less recapture than comparable households in high-appreciation areas. GAO at least concedes that this result is not easily remedied for the purpose of the limitation 1s to leave a portion of the home equity to be applied to the urchase of a replacement home. It does not recognize that the 50 percent Pimitation is essentially a protection for those living in the poorest neighborhoods that they may have something left to move on with after See comment 12. recapture is paid. We strongly advise the Congress to leave this protection in place. Recapture was never intended to prevent households from eventually moving up in the housing market and certainly was not designed to return owners in the poorest, least likely to appreciate neighborhoods, to renter status. Lower-income households’ ability to move up is already strained under the current recapture requirement. The GAO suggests two possible alternatives to the current recapture rovision: a “roll-over” loan approach and an income-based variable payment Poan approach. The “roll-over loan” would dtscontinue MRB assistance when the owner’s income rose by raising the interest rate on the loan to conventional rates in subsequent years. Under the income-based variable payment loan approach, housing payments would increase as a fixed percentage of income, if income rose. Both approaches would require housing agencies or their lenders or servicers to establish a system for monitoring the income of every MRB borrower. Borrowers would be required to disclose their federal tax returns periodically. This process would be an extremely intensive, costly, and administratively complex task for servicers who must service hundreds of thousands of loans. Those servicers would have to make sure borrowers’ tax returns are actually collected and reviewed, determine if the borrowers’ incomes could support higher monthly payments, and when necessary. dispatch a new payment schedules to the borrowers. Another serious problem with these approaches is that increases in income are incorrectly presumed to mean increased “ability to pay.” See comment 13. Increased tax withholdings, utility costs, property taxes and other household costs must be considered to avoid financially overburdening homeowners and potentially triggering a loan default. To take such increased costs properly into account, a lender would have to completely re-underwrite a homeowner’s loan each and every time his or her income was reviewed. This would require substantial financial information from the homeowner 12 Page84 GAO/RCED-90-117 Recap- the QuaUf’iedMortgage BondSubsidy AppendixVU CkmuuentuFromthe NationalCouncilof StateHousingAgencies and significant time and financial analysis from the lender. This is a task unprecedented in mortgage servicing. Under the Section 8 Certificate program, income monitoring and verification is required, but the administering agency is paid a fee by the federal government to cover such costs. Moreover, unlike Section 8 where only the amount of tenant income is in question, the servicer under the GAO approach would be required to analyze the owner’s income against the income which would be required to support a conventional housing payment. The substantial cost of administering such an MRB monitoring program would be passed on to all borrowers in the form of increased loan points or servicing fees. Neither approach is practically workable, and neither improves on the falrness or effectiveness of the current law provision. The GAO suggests testing income under either approach at three and five year intervals. Since exlstlng law recapture peaks after five years and ends after ten, these approaches would provide less disincentive for purchasers who expect extraordinary income increases than current law does, since current, law See comment 14. Imposes a definite. unavoidable penalty for any sale within ten years. In contrast, the periodic income testing approach would not impose any penalty until the year income is tested, which might be years after the assisted buyer’s income rose substantially, with no recapture of the benefit received in those prior years. Under the roll-over loan approach, as GAO acknowledges, borrowers could be required to refinance thelr loans, at potentially higher conventional rates than prevailed when they first took their MRB loan and, in any case, with the payment of points which might well exceed the value of any MRB subsidy received. Their inability to pay could force them out of their homes. Thla risk would be particularly acute when the income increase was only temporary. Establishing a market-rate at the time the borrower enters the MRB program and adjusting the original mortgage to that rate without refinancing when the borrower’s income rises could overcome some of See comment 15. these risks. However, the difficulties involved in determining what the conventional rate is at any particular time which has already been discussed and the periodic need to examine income and all its associated costs and complexities would remain. Under the variable payment loan approach, housing payments for principal, interest, taxes and insurance would be set at a flxed percentage of household income. As income increased, so would monthly payments. Unlike the roll-over loan which caps the interest rate at some predetermined rate, the variable payment loan allows the owner’s effective Interest rate to rise above the market as income continues to increase. Of course, the owner would have the option of refinancing when his or her payments exceeded payments on a conventional loan, but then would incur all of the associated costs of refinancing which might exceed the MRB 13 Page86 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy .. AppendixVII CommentsPromthe NationalCouncilof StateHousingAgencies benefit received and would face the risk of not being able to service the refinanced loan. Both of these alternatlves would Impose recapture on all Ml3B borrowers whose incomes rise steadlly and conservatively over the life of their mortgage and not just the “yuppie” population Congress intended to See comment 16. target with recapture. In conclusion, the current recapture provision must be substantially revised. so that it is simple, administrable, dlscloseable and does not adversely affect lower income borrowers. An alternative along the lines of the simplification proposal contalned In these comments would achieve these goals. The rev&ions and alternatlves GAO poses in Its report would See comment 17 not accomplish these objectives, would In no way increase the effectiveness of recapture and in fact, would alter Congressional intent. These recommendations should be soundly rejected. The NCSHA Is committed to working with the Congress to achieve a slmpllficatlon of the current recapture provision which presences Congresstonal Intent. 14 Page86 GAO/RCED-80-117&capturing the QualifiedMortgageBondSubsidy AppendixVII CommentaFromthe NationalCouncilof StateHousingAgencies The following are GAO'Scomments on the National Council of State Housing Agencies’ letter dated June 6, 1990. 1. We disagree with the Council’s characterization of our report1 The GAOComments Council’s characterization of our report ignores other supporting anal- yses contained in the report that reached similar conclusionsusing dif- ferent data sourcesand analytical techniques. 2. The studies cited by the Council do not offer an assessmentof the extent to which owners could have purchased homeswithout &MB assis- tance. To our knowledge, this type of analysis has not been undertaken on a widespread basis elsewhere. We have conducted an additional anal- ysis that further supports our earlier results. We continue to believe that many assistedowners may have been able to purchase their first homes without QMB loans. As a piece of anecdotal evidence, a Rhode Island mortgage lender stated that its analyses of its loan files showed that at least 26 percent of approved QMBloan applicants would have been able to obtain an alterna- tive mortgage from that lender.2Additionally, it believed that 20 percent of all Rhode Island participants could have purchased a home without a QMBloan. Since Rhode Island was not included in our review, we cannot specifically comment on the numbers derived. However, while their results are less striking than ours, they do show a significant portion of buyers served were not in need of the assistancethey received. 3. We object to the Council’s characterization that by “GAO'S own admis- sion, [its 1988 study] was conducted with a skewed methodology.” We believe that the approach we used was appropriate to provide an esti- mate of conventional affordability. In addition, the home ownership test was one of several analyses contained in the report using different approachesand data sources.All reached the sameconclusion. ‘For the Council’s prior comments in this vein, and our discussion, see pp. 107-l 11 of Home Owner- ly and Provide Little Assistance to Those in Need (GAO7 2”Expiring Tax Provisions,” Hearings Before the Subcommittee on Taxation and Debt Management of the Committee on Finance, United States Senate. S. Hrg. 100-1002, Mar. 28, 1988. Pp. 86-91. Page87 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVII CommentsFromthe NationalCouncilof StateHoueIngAgencies The Council objects to our use of a judgmental sample of about 178,000 loan files.3The Council essentially argues that decisionsabout the effec- tiveness of the QMBprogram can only be answered by statistically reli- able studies. If that is so, the studies cited by the Council do not meet that test. For example, in claiming statistical validity of the studies it supported, the Council substantially misrepresented the statistical nature of these studies. First, it claims that the study conducted by Dr. Wrightson was a “statistically valid review.” However, disclaimers in that study specifically state that it was not and should not be taken as such.4Second,notably omitted from the Council’s statement is any refer- enceto local housing finance agency performance. Neither of the studies cited by the Council analyze local housing agency &MBloan performance. As such, these studies can hardly be consideredto be statistically valid reviews of QMBloan activity. Finally, many policy questions do not lend themselvesto statistical anal- yses for a number of reasons,including the onesthat hampered us from obtaining a statistical sample of buyers. In addition, even those analyses conducted under statistical procedures are often open to different inter- pretations. In the absenceof iron-clad proof, which the Council suggests is the only basis from which to judge the merits of QMB loan activity, policy makers make their decisionsby weighing the merits of available evidence. 4. Our 1988 report recognizedthe efforts of several housing agenciesto target assistancemore strictly than required by the Code.However, the question central to the affordability test is the extent to which QMBloans increase home ownership opportunities to first-time buyers. A finding that someagenciesprovide set stringent eligibility requirements does not addressthe question of whether the assistedowners could not have purchased their home without QMBloan assistance. 6. We added an analysis of housing agency loan files using a total debt- expense-to-incometest to chapter 3 of this final report. Our results did not change markedly. “These files represented home purchase loans made by 29 state and local housing agencies and in their automated files for the January 1983-June 1987 period. These were the latest data available at the time of our review. We did not perform a statistical sample of state and local housing agency loan activity because of the lack of a central data base from which to do so. Nonetheless, we believe that these loans represented about one-third of the loans made during that period. 4The “Referendum” paper also cited by the Council relies heavily on Dr. Wrightson’s study. Page88 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy Appendix VU Comment From the Natsonal Council of StateHoudngAgencies 6. We disagree.The recapture mechanism,either as enacted in law or proposed by the Council, tests for certain income increasesafter an assistedhome is sold. As we discussin appendix IV, the recapture formula would not be expected to induce buyers to decline to participate. 7. We seeno basis for the Council’s contention that affordability testing would subject lenders to lawsuits. If the applicant could afford a con- ventional loan of comparable terms, then the lender would be agreeable to making a loan to the applicant. If, as a result of the affordability test, the applicant could not afford the conventional loan, then the applicant would be offered a QMBloan (assuming the applicant met eligibility requirements and was a good credit risk). Therefore, applicants meeting either conventional or &MB loan-making requirements would not be denied credit becauseof the affordability test. If they are not denied credit, then a lawsuit would be groundless. Finally, if the applicant sued on the basis that he or she was denied the lower-interest rate QMBloan, the suit would likely be dismissed as long as the lender acted in good faith in that the lender was following affordability test requirements set out in law. 8. The terms that lenders offer are those at which they are willing to make loans. For that lender at that time, the stated rate for the prospec- tive buyer is “the rate.” That interest rates changeover time or differ between lenders is the basis of our proposal. The QMBloan rate is fixed, while conventional interest rates change.In this environment, the enacted recapture formula will almost always recapture more or less than the assumedinterest rate reduction. 9. We believe that the Council overstates the administrative burden involved. Seechapter 3 for a discussionof the steps involved. 10. We would not support the Council’s concept becausewe believe that it would be too difficult to administer. Rather, the approachesthat we suggestform a middle ground between “real time” testing and the enacted version. Chapters 3 and 4 discusswhat we believe are the bene- fits from the approacheswe suggest,as well as the increased adminis- trative effort, relative to the enacted version. In deciding whether to changethe recapture formula, the Congressmust weigh these expected benefits against the increased administrative effort. 11. The Council ignores the fact that substantial transaction costs in selling a home can substantially eliminate gains on salesin early years Page89 GAO/RCED-SO-117 Recapturingthe QualifiedMortgageBondSubsidy Y AppendixVII CommentsFromthe NationalCouncilof StateHousingAgencies creating a disincentive to sell within the first several years of ownership. 12. Neither our draft report nor this final report advocate eliminating the gain-on-saleprovision. However, the gain-on-saleprovision benefits all assistedowners, not just “those living in the poorest neighborhoods.” 13. We disagreethat additional underwriting would have to occur. See our discussion in chapter 4 which describeshow assistedowners’ loan payments could be increased administratively without underwriting or refinancing the loan. It is interesting to note that conventional fixed-rate loans are not underwritten again when home owner costs increase or if an owner’s income remains flat or decreases.Similarly, conventional adjustable-rate loans that can result in higher housing payments when interest rates rise are not underwritten again, even if other costs increase and income remains stable or decreases. 14. The 3- or S-year testing period is presented as a trade-off between examining income changesat the time of sale or doing so more fre- quently, say, annually. 16. The Council has misstated our suggestion.The roll-over approach that we discussedin our draft report doesnot involve refinancing of loans, including interest-rate risk, or the payment of refinancing costs. In fact, our suggestionspecifically stops this from occurring. However, we have added further clarifying language.Seealso comment 8, above, for our view on determining a market interest rate when the loan is made. 16. The Council has misstated our approach. The roll-over approach tests for whether an assistedowner no longer needsthe subsidy. Whether income increased “slowly and conservatively” or more quickly, we seeno public purpose being served in continuing the assistancewhen those receiving QMBloan or MCCassistancecould remain home owners without it. A roll-over approach satisfies that goal. 17. We disagree with the Council’s assertion that our proposals alter congressionalintent. The intent is quoted in chapter one of this report, and our analyses clearly fall within the language contained in the con- gressional conferencestatement. Page90 GAO/RCED-W-117 Recapturingthe Qualified MortgageBondSubsidy Appendix VIII Comments From the Department of the Treasury Note: GAO comments supplementing those in the report text appear at the end of this appendix. DEPARTMENT OF THE TREASURY WA*HINOTON July 9, 1990 A88ISTANT 8CCRCTARY John M. 016, Jr. Director, Housing and Community Development Issues U.S. General Accounting Office Washington, D.C. 20540 Dear Hr. 01s: Thank you for providing the Department of the Treasury with the op ortunit to comment on the draft GAO report, “Home Owners It ipr Lim r: ting Mortgage Assistance Provided to Owners With Rapid Income Growth.” The report evaluates the recapture provision of the Qualified Mortgage Bond (QMB) and Mortgage Credit Certificate (NC) programs in terms of how effectively it limits the homeownership subsidy to households with rapid income The recapture provision was enacted and the report %%d by the Technical and Miscellaneous Revenue Act of 1988 (P.L. 100-647). Background The QMS program assists first-time homebuyers with below market interest rate loans financed with the proceeds of tax-exempt mortgage bonds issued by state and local housing a encies. The WCC program also assists first-time homebuyers by a f lowing them to claim a credit against their Federal income tax equal to between 10 and 50 percent of the mortgage interest paid during the year on a market rate loan. Those who claim a credit, howeve f , must reduce their home mortgage interest deduction by the amount of the credit. Under both programs, homebuyers must meet income, purchase price, and other eligibility requirements. In the 1988 Act, Congress added a provision designed to recapture tha sirtsidy provided by these progrsme from households that experienced ra id increases in income subsequent to their purchase of a f: ome. This provision becomes effective on January 1, 1991. An assisted homebuyer may become subject to recapture only in the year in which he disposes of his home, and only if his income in that year exceeds the maximum amount of income that would have allowed him to qualify for the assistance in that year. The recapture amount increases from zero to the full recapture amount on a pro rata basis as the homebuyer’s income increases from the maximum qualifying income to the maximum qualifying income plus $5,000. During the first five years of the mortgage, the full recapture amount is equal to the product of 1.25 percent of the original mortgage amount and the number of years since the home Y was purchased. The full recapture amount decreases by 1.25 percent of the original mortgage amount for each year thereafter so that after year ten there is no recapture. In addition, the Page91 GAO/RCED-90-117 Recapturingthe QualifiedMortgageBondSubsidy AppendixVIII CbmmenteFromthe Departmentof the Treasury -2- racapture amount is capped at 50 percent of the gain on disposition of the home. The recapture amount is added to federal income tax liability for the year in which the homebuyer disposes of his home, The report concludes that the proposed recapture mechanism is referable to no recapture mechanism, because it would recover at Peast some of the subsidy from those who would have delayed home purchases for a short time until they could have afforded an unassisted loan. The report recommends that Congress consider tailoring the recapture amount to the actual amount of assistance received and eliminating the phaseout of the recapture after year five. It also recommends that Congress consider limiting QMB and HCC program assistance to individuals who could not otherwise obtain a market rate loan with comparable terms. Finally, the report discusses two alternatives to recapture: a “roll-over’ loan and an income-based variable payment loan. Comments on GAO Recommendations GAO Recommendation: With respect to the QMB program, Congress should consider basing the recapture rate on the actual difference between the market and below market interest rates at the time the home was purchased. Comment: It is inappropriate as a matter of fairness to mre an amount which exceeds the actual subsidy received by a homebuye r . The report cites a 1988 GAO report*/ which estimated that the actual median difference between the market and below-market interest rates received was 1.44 percent. This translates into an actual median after-tax benefit of 1.22 ercent assuming the marginal income tax rate of the homebuyer is L35 percent. This is less than the after-tax benefit of 1.25 percent assumed under the current law recapture provision. While this difference would not be important if there were little variability in the actual subsidies, it appears that there is a substantial amount of variability. The report cites the 1988 GAO report which estimated that 25 percent of the assisted buyers received reductions of about three-quarters of a percentage point or less. The implication is that the recapture amount could exceed the actual subsidy by a significant amount in a significant number of cases. I/ V S General Accounting Office, Home Ownership: Mortgage Bonds Xre Coitly and Provide Little Assistance to Those in Need (GAO/ - - ) March 28, 19SS # hereinafter referred t o as the 1988 GAO Report. ’ Page 92 GAO/RCED-90-117 Recapturing the Qualified MortgageBondSubsidy . AppendixVIII Comments From the Department of the Treasury -3- Equating the recapture amount with the actual benefit would generally be appropriate as a matter of fairness, but as a matter of tax policy, this objective would have to be balanced against the associated administrative and compliance costs. The report would be improved by the addition of a discussion of the See comment 1 magnitude of these costs in comparison with the potential recap- ture amount. The report might also indicate that the recapture amount may understate the benefit of below-market rate financing See comment 2. because of the assumption implicit in its calculation that the current value of the benefit received in any given year is constant. While the benefit of below-market rate financing in any given year may be 1.25 percent of the mortgage amount, it must be compounded at the prevailing interest rate over the rubsequent years of the mortgage to obtain the true economic See comment 3. benefit. In addition, it would be helpful to clarify whether interest rates are before- or after-tax rates. GAO Recommendation: Congress should consider the following alternatives to the phaseout of recapture after year five: eliminate the phaseout8 reduce the rate at which the recapture phases out; and cap the recapture amount at year five. Comment: The report argues that because the recapture amount decreases after year five, the homebuyer has the opportunity to reduce or eliminate the recapture amount by choosing when he sells his home. A related concern, not addressed in the report, la that homebuyers whose income remains below the maximum qualifying income in the earlier years of the mortgage but whose income increases in the later years may be subject to the full recapture amount. In theory, the appropriate policy would be to recapture the benefit only for the years in which a homebuyer’s income exceeded the maximum qualifying income. This would require a See comment 4. continuous monitoring of a homebuyer’s income which would increase administrative and compliance costs. The report might acknowledge this problem and discuss the administrative and compliance costs associated with recapturing the benefit only for the years in which the homebuyer’s income exceeds the maximum qualifying income. GAO Recommendation: Congress should consider requiring that QMB loans and MCCsonly be provided to those who could not otherwise obtain a market rate loan with comparable terms. Comment: Providing assistance to those who could otherwise affora market rate loan with comparable terms is not consistent with the goals of the QRB and MCC programs. The report cites the 1988 GAO report which estimated that over half of the assisted homebuyers could have obtained a market rate mortgage with comparable terms to purchase the same home. The report, however, does not explain how this recommendation would be implemented. The report ml ht, for example, address how much tighter the See comment 5 income eligib 9 lity requirements would have to be to target the assistance to those Congress intended to benefit. Page93 GAO/RCED-99-117 Recapturingthe QualifiedMortgageBondSubsidy Appendix VlIl Commenta Prom the Department of theTressury -4- GAO Recommendation: Congress should consider “roll-over” loans as an alternative to the recapture mechanism. Comment : A roll-over loan would require a homebuyer to refinance mrket rate if his income increased above the maximum qualifying income. This would require continuous monitoring of a homebuyer’s income and would impose substantial administrative and compliance coats. In addition, if a homebuyer were forced to refinance, he would bear substantial refinancing costs. Moreover, if interest rates were to rise during the period in which he owned his home, then requiring a homebuyer to refinance at a higher rate could make him worse off than he would have been if he had originally financed his home with a market rate loan. There considerations suggest that the roll-over loan be pre-approved, so there would be minimal refinancing costs, and that the rate on the roll-over loan be capped at the market rate in existence when the homebuyer purchased his home. The report should discuss the administrative and compliance costs associated with continuously monitoring income , pre-approved refinancing See comment 6. costs for roll-over loans, and the costs associated with capping the refinancing rate at the market rate in existence when the individual originally purchased his home. GAO Recommendation: Congress should consider an income-based variable payment loan as an alternative to the recapture See comment 7. mechanism. Comment: An income-based variable payment loan would establish a -or’s mortgage payments at a fixed percent of his income. This a f ternative could allow the implicit rate to rise above the market rate that existed when the home was originally purchased. It also could allow the implicit rate to drop below the subsidized rate if the homebuyer were to experience a drop in income. The appropriate policy might be to allow the implicit interest rate to vary between the subsidized rate and the market rate that existed when the home was originally purchased. This alternative would impose additional administrative and compliance coats on the homebuyer and the administrative agency. General Comments The report might consider the alternative of having no re- capture mechanism (if only for the purpose of exposition). In- deed, it is oasible that the administrative and compliance costs See comment 8 associated wP th recapturing an amount which is close to the actu- al benefit may exceed the actual recapture amount. The report cites the 1988 GAO report which estimates that the after-tax value of the subsidy to the typical homebuyer is $40 a month. Assuming this is the potential recapture rate, the maximum oeaible recapture amount will increase on a pro rata basis to $ 2,400 at the end of year five and then will decrease to zero at the end of year ten. Unless the recapture phaseout is Page 94 GAO/RCED-fJO-117 Recapturing the Qualified Mortgage Bond Subsidy . Appendix VIII Comments From the Department of the Treasury -5- eliminated, the amount by which the recapture amount exceeds the administrative and compliance costs may not be substantial, especially since the recapture amount is capped at 50 percent of the gain on dieposition of the home. The QMB and MCC programs may also be inherently self-policing. This is in part due to the fact that there is a ceiling on the amount of QMBs and MCCs that a state or local houain agent may issue, which suggests that states have an incent 9 ve to x nsure that the subsidy goes to those for whom it was intended. It is also in part due to the fact that the eligibility requirements limit the pool of potential beneficiaries and the value of purchased housing. In general, See comment 9. beneficiariee must be first-time homebuyers with incomes that do not exceed 115 percent of the area median income, and assisted mortgages can be used only to purchase a home that costs less than 90 percent of the area average. The report should clarify whether income eligibility refers to the maximum amount of income that a homebuyer could have and See comment 10. (Page still qualify for assistance or the amount of income that would reference in final report subject a homebuyer to full recapture (see, for example, the firet full paragraph on p. 19). is p.16.) Finally, the analysis of the percent by which income must increase in order to trigger recapture under different scenarios could also be clarified. Because recapture is triggered if income exceeds the maximum qualifying income in the year of dis osition, the percent by which income must increase is obv ousl dependent on the amount of starting income. The See comment 11. h othet ‘I II cal scenarios included in the report may only confuse t KP s straightforward issue. If you have any questions concerning these comments, please contact William Trautman of my staff at 566-2203. Kanneth W. Gideon Assistant Secretary (Tax Policy) - Page 96 GAO/RCED-90-117 Recapturing the Qualified Mortgage Bond Subsidy . AppendixVIU CommentaPromthe Departmentof the Treasury The following are GAO'Scomments on the Department of the Treasury’s letter dated July 9, 1990. GAO Comments 1. While we did not attempt to quantify the added costs involved in tai- loring the recapture formula more to the benefit received, we expect that the additional cost per home owner would be small. The additional effort would entail computing the difference in monthly housing costs between the reduced interest-rate QMBloan and a market-rate loan of comparable terms. The results of this computation would be provided to the assistedowner and retained by the lender or the housing agency. The cost of this approach should be weighed against the benefits derived from it. Regarding compliance, the seller’s tax return could, at the Internal Rev- enue Service’s discretion, identify the dollar amount of the monthly reduction, the number of months the home was held, and the product of these two amounts as the unreduced recapture amount. Should the Ser- vice decide to audit a taxpayer’s return, it would request all pertinent documentation, as it does for other taxpayer audits. 2. Treasury’s observation is correct if the only alternative to the QMB loan is an adjustable-rate mortgage that provides for annual adjust- ments. However, the fixed-rate, market-rate loan provides for a constant differential over the life of the loan. 3. This final report clarifies this point. 4. We added Treasury’s comment to this final report. 6. We believe that the draft that we provided to Treasury and this final report outline how this approach could be implemented. Changesto income eligibility limits are not necessaryto carry out the conventional affordability test described in this report. 6. Both the draft report provided to Treasury and this final report dis- cuss that refinancing is not necessaryunder the approach we present. We have added languagethat further clarifies this point. Also discussed is the administrative cost trade-off between annual income reexamina- tions and less frequent reexaminations. Finally, both the draft report and this final report provide an estimate, based on similar activities, of the administrative cost of the roll-over approach. Page96 GAO/RCED-O-117 Recapturingthe QualifledMortgageBondSubsidy AppendixVIII CommentaFromthe Departmentof the Treasury 7. Neither our draft report nor this final report recommendthat the Congresssubstitute an income-basedvariable payment loan as an alter- native to the recapture mechanism. 8. We did not consider the alternative of having no recapture require- ment becausethe statutory study requirement was clear in its charge that we study recapture and other alternatives that would achieve the congressionalobjectives set out for it. 9. We agreethat state and local agencieshave an incentive to make loans to those who meet the first-time buyer, income, and home purchase price requirements. If these requirements were not met (and if the Service enforced these provisions), then the underlying bonds would lose their tax-exempt status. If a bond losesits tax-exempt status, it would becometaxable. On future issues,investors would require higher interest rates becauseof the uncertainty over the tax-exempt status of the bonds. This discussion was included in our March 1988 report1 How- ever, we do not believe that it is germane to this report. 10. We believe that the term “income eligibility limits” clearly identifies eligibility for QMBloans or MCCS. 11. We do not believe that the analysis is as straightforward as Treasury suggests.The starting income, the recapture formula’s $6,000 constant, and the number of years before the home is sold all contribute to the different shapesand levels of the curves. This very point is illustrated by the charts in chapter 2 and appendix II. ‘We reported in 1988 that the Service had no ongoing compliance review program for QMBs. See pp. 92-93 of Home Ownership: Mortgage Bonds Are Costly and Provide Little Assistance to Those in Need (GAO/R CEDS8- - Page97 GAO/RCED-9@117 Recapturingthe QualifiedMortgageBondSubsidy Appendix IX . Major Contributors to This Report Dennis W. Fricke, Assistant Director Resources, JamesC. Ratzenberger,Evaluator-in-Charge Community, and Patrick B. Doerning, Operations ResearchAnalyst Economic Alice G. Feldesman,Social ScienceAnalyst Carol Herrnstadt Shulman, Reports Analyst Development Division Washington, DC. Page98 GAO/RCED-BO-117 Recapturingthe QualifiedMortgageBondSubsidy . Page BB GAO/RCED-90417 Recapturingthe QualUledMortmge BondSubaId~ Related GAO Products Home Ownership: Targeting Assistance to Buyers Through Qualified Mortgage Bonds (GAO/RCED8%~~OBR, June 27,lSSS). “Role of Qualified Mortgage Bonds in Providing Home Ownership Opportunities,” Testimony Before the Subcommitteeon Taxation and Debt Management,SenateCommittee on Finance (GAO/T-~~~~88-31, Mar. 28,1988). Home Ownership: Mortgage Bonds Are Costly and Provide Little Assis- tance to Those in Need (GAO/RCED-%iii, Mar. 28,lSSS). “S-1698, The First Time Homebuyer Assistance Act of 1983,” Testimony Before the SenateCommittee on Finance, Sept. 13,1983. “The Costs and Benefits of, Single-Family Mortgage RevenueBonds,” Testimony Before the House Committee on Ways and Means,June 16, 1983. “The Costs and Benefits of Single-Family Mortgage RevenueBonds,” Testimony Before the Subcommittee on Taxation and Debt Management, SenateCommittee on Finance, May 13,1983. The Costs and Benefits of Single-Family Mortgage RevenueBonds: Pre- liminary Report (GAO/RCEP~~-146,Apr. 18,1983). (385177) Page100 GAO/WED-ml17 Recapturingthe QuellfledMortgageBondSubsidy Ordt~riug IuforruaCiou ‘I’ht~ first five copies of each GAO report art’ frt*tk. Atiditiomtl t+opiths iire $2 each. Orders should be st*ut t,o the followiug atltirt*ss, ilt*t~)I~ I~arrieti by a check or mont*y order made out, to t.he Superint.t~ntit~tll of hcurn~nt.s, when uecessitry. Ortkrs for 100 or mm’ copit to t,t* uritilt~cl t,o a single atIdrt*ss m-e tIiscouut.rtI 25 percerrt. I1.S. Gt~nt~ral Acmuuthg Office I’.(). Hex 60 I5 Gaithershurg, MI) 20877 Ortkrs may also btb plwctd by calling (202) 2756241. .-_-I.--- -..--- - .-.^_ ------ ---. - .._-._. ..___ - .-.” ._“I. .., I_.I_ .., _I” ,.“l,l ll.“,“*l”lllll,l*,l-*-“-,- 1--1 ~- 1-
Home Ownership: Limiting Mortgage Assistance Provided to Owners With High-Income Growth
Published by the Government Accountability Office on 1990-09-26.
Below is a raw (and likely hideous) rendition of the original report. (PDF)