Need for Reform
Since the end of 2007, FHA’s reserves have declined from $22 billion to currently around $3.5 billion. While some decline is to be expected, given the bursting of the housing bubble and continued weakness in the labor market, further declines could easily erode the remaining reserves and require a direct appropriations to cover future claims.
The potential for a bailout of FHA is not a remote possibility. According to the FY2010 Actuarial Review, the net present value of future cash flows from FHA’s current book of business is a negative $25.4 billion. The FY10 Actuarial Review projects a positive economic value for FHA solely on the basis of assuming that future business will generate revenues sufficient to cover imbedded losses. In order for that assumption to turn out correct, the credit quality of FHA’s lending must be improved considerably. It should be noted that a critical assumption driving the positive expected value of future business is the continued prohibition of seller‐financed down‐payments.
Although FHA’s market‐share was relatively small during the height of the housing boom, that did not protect FHA from guaranteeing loans that currently have a negative net present value. Values for the FY06 book are a negative $1.6 billion. Of course, this becomes relatively small when compared to the FY08 (-$7.8 billion) and FY09 (-$6.6 billion) books of business. These values also depend heavily on what I believe are relatively optimistic projections for the housing market. Further price declines will dig these holes even deeper.
As FHA guarantees the credit risk on mortgages that underlie GNMA securities, FHA bears the majority of the risk. Interest rate risk is transferred to the investor. Accordingly, most of my testimony will focus on FHA’s Single Family 203(b) program.
Programs Costs Should Be Accurately Priced
If there is any lesson we should take away from the recent financial crisis, it is that when borrowers, lenders, investors and governments do not face the actual costs of their decisions, such decisions are likely to have negative consequences. FHA, and its Congressional oversight, have long suffered from poor decision‐making due to gross underestimates of cost.
First among those is that FHA premiums are not structured to cover the administrative costs (including salaries) of running FHA. No private business would last long if it did not price to cover the costs of its employees. Such costs for FHA are covered by appropriations that directly come at the expense of the taxpayer. In recent years, these costs have averaged about $350 million. Given that FY10 insurance‐related cash flows were approximately a negative $510 million, excluding administrative costs underestimates current negative cash flows by at least 40 percent.
Subsidy rates for FHA are calculated under procedures specified by the Federal Credit Reform Act of 1990 (2 USC 661). In addition to excluding administrative program costs, FCRA excludes any adjustment for market risk. Under insurance programs, such as FHA, where the private sector pays to transfer risk‐bearing to the government, the private sector is also protected from market risk. A clear benefit is being provided that is not included under FCRA. CBO has estimated that calculating FHA’s subsidy costs under a fair value method — which CBO believes “provides a more comprehensive measure of the cost” — would shift an expected budgetary savings of $4.4 billion in FY12 to a budgetary cost of $3.5 billion.1 It should be noted that fair value accounting has been used in other federal contexts; for instance Section 123 of the Emergency Economic Stabilization Act of 2008 requires the Treasury Secretary to take into account market risk in the context of the Troubled Asset Relief Program.
When one ignores administrative expenses and fair value, FHA has long been presented as “making money”. Yet these assumed “negative subsidies” were based upon erroneous estimates on the part of FHA. A comparison of original estimates and subsequent re‐estimates of FHA subsidy rates for the 203(b) program show that from 1999 to 2011 actual subsidy costs were revised upward by a net total of $44 billion. These re‐estimates have been large enough, in the years from 2002 and 2009, to change “negative subsidies” into actual positive subsidies. As the following chart clearly illustrates, the errors in FHA’s subsidy estimates have been quite large. For instance, the FY06 book was initially projected to create cash equal to 2 percent of book. Upon re‐estimate, FY06 actually cost FHA over 4 percent of its book. An error that has costs billions. The chart also illustrates that the bias of estimates has consistently been in one direction: the underestimation of costs.
Given the gross under‐pricing of actual risk by FHA, I urge the following changes to made to FHA’s premium pricing:
- Require charged premiums to cover projected administrative costs, including employee compensation.
- Require charged premiums to be estimated on a Fair Value basis.
Towards Sustainable Homeownership
The performance of FHA single‐family mortgages during the last decade at times made sub‐prime lending look safe. For 2002 to 2007 the delinquency rate of FHA mortgages actually exceeded that of sub‐prime. This should of course come as no surprise given that in the 2005 book of business about 60 percent of FHA borrowers had FICO scores under 640. Until 2004, FHA did not regularly collect credit scores for its borrowers. Once it began the collection, it readily became apparently that FHA was one of the largest sources of credit for sub‐prime borrowers. In 2009, the credit profile of FHA borrowers improved considerably, raising the expectation that future books of business may see a reduced incidence of loss.
Losses from sub‐prime borrower credit are usually manageable when there is significant equity on the part of the borrower. It is the combination of poor credit history and low/no down‐payment that have resulted in tremendous losses, both for FHA and private sub‐prime mortgage lending. As the following table (Anderson, Capozza and Van Order) illustrates, as low equity is combined with weak credit defaults sky‐rocket. Note that the table is normalized so that a loan with a credit score between 680 and 720 and a LTV between 71 and 80% equals “1”. Other figures are either fractions or multiples of this number. The magnitudes are nothing short of shocking. Loans with a FICO below 620 and down‐payments of less than 10 percent display default rates 20 times that of the base group.
Such high levels of default are not healthy for the borrower, the lender or the taxpayer (not to mention the economy). We know, with near certainty, that borrower credit quality and equity are the drivers of default, both in FHA and in the mortgage market generally. If we wish to protect the taxpayer and avoid a future bailout of FHA, these are the policy margins along which we must make substantive changes. Given the relatively “safe” features of an FHA loan, we do not have to guess about loan characteristics driving the borrower into default. We know it is equity and credit history that drives losses.
To insure that FHA guarantees loans that are both sustainable on the part of the borrower and also represent a minimum risk to the taxpayer, I urge that the following:
- Immediately require a 5 percent cash down‐payment on the part of the borrower.
- Require FHA to allow only reasonable debt‐to‐income ratios.
- Restrict borrower eligibility to a credit history that is equivalent to no worse than a 600 FICO score.
- Require pre‐purchase counseling for borrowers with a credit history that is equivalent to a FICO score between 600 and 680.
- Require a 10 percent down‐payment, immediately, for borrowers with a credit history equivalent to below a 680 FICO score.
- Borrower eligibility should also be limited to borrowers whose incomes do not exceed 115 percent of median area income, so as to mirror the requirements of section 502(h)(2), as amended, of the Housing Act of 1949.
Towards a Fairer Sharing of Risk
It is not solely the behavior of the borrower that matters for default. Incentives facing the lender also greatly contribute to default. Where the lender bears the full cost of default, we can expect prudent and careful underwriting to prevail in the long run (as the imprudent eventually fail, unless we rescue them). Where the lender, with little penalty, can pass along the cost of default to another party, for instance the taxpayer, then poor or negligent underwriting is to be expected. Accordingly, we must change lender incentives under the FHA program. As has been repeatedly detailed by HUD’s Inspector General, FHA has long shown a lax attitude toward lender fraud and misbehavior. Given the legitimate due process concerns that arise when any party receives a government benefit or participates in a government program, FHA’s ability to effectively eliminate fraud ex post will always be somewhat limited. Of course this does not eliminate the necessity of doing so. It does imply, however, that alternative means must be found for improving the incentives facing lenders.
In order to provide the appropriate incentives for lenders to conduct sufficient due diligence and quality underwriting, I urge the following:
- Immediately reduce maximum claim coverage from 100% of loan to 80%.
- Require lenders to “take back” any loan that defaults within six months of origination.
- FHA should also end the process of letting the lender choose the appraiser and return to the safeguard of an appraisal board.
The history of FHA has been one of an almost constant reduction in standards, usually as an excuse to “re‐start” the housing market. Indeed the first substantial legislation changes were made just four years after its creation, when in 1938 Congress lowered down‐payment requirements from 20 to 10 percent and extended the maximum loan duration from 20 years to 25. This did little for the housing market, which did not begin to recover until after World War II.
The recent housing boom and bust has witnessed a similar reaction. Attempts to re‐start the bubble by transferring massive amounts of risk to the taxpayer. Again these efforts have accomplished little at great cost. We should not repeat the same mistake that has followed almost every housing bust in the last 100 years. Instead of leaving these additional stimulants in place, we should begin moving federal mortgage policy towards a sounder footing. Only then can we hope to avoid having the taxpayer left holding the bag when the next bubble inevitably bursts.
Future projections of FHA’s financial health depend critically upon a significant increase in credit quality. In order to protect the taxpayer, Congress should begin making efforts to guarantee that increase in credit quality today.
1 Congressional Budget Office. Accounting for FHA’s Single‐Family Mortgage Insurance Program on a Fair‐Value Basis. May 18, 2011.