House Financial Services Committee Hearing
Federal Information & News Dispatch, Inc. |
Introduction
I recently completed a comprehensive study that shows the FHA is engaging in practices resulting in a high proportion of low- and moderate-income families losing their homes. Based on an analysis of the FHA's FY 2009 and 2010 books of business, the FHA's lending practices are inconsistent with its mission and represent a disservice to American working-class families and communities.
The findings of this study indicate:
* An estimated 40 percent of the FHA's business consists of loans with either one or two subprime attributes--a FICO score below 660 or a debt ratio greater than or equal to 50 percent (based on loans insured during FY 2012). The FHA's underwriting policies encourage low- and moderate-income families with low credit scores or high debt burdens to make risky financing decisions--combining a low credit score and/or a high debt ratio with a 30-year loan term and a low down payment. A substantial portion of these loans have an expected failure rate exceeding 10 percent.
* Across the country, 9,000 zip codes with a median family income below the metro area median have projected foreclosure rates equal to or greater than 10 percent. These zips have an average projected foreclosure rate of 15 percent and account for 44 percent of all FHA loans in the low- and moderate-income zips.
The study found the direct and indirect costs associated with a foreclosure rate greater than 10 percent, particularly in working-class communities, are unacceptably high. Risk layering, combined with high FHA loan volumes, has a substantial impact on these communities. The resulting reduced or declining home values impact FHA and non-FHA low- and moderate-income families diligently making their payments. These families may be denied the opportunity to build equity, provide security for their family, and have the down payment for their next home as their family grows. Foreclosures also result in increased blight and crime and the larger community suffers from a reduced tax base and higher costs for providing municipal services.
The study identified specific reforms to focus the FHA on responsible lending and return it to its traditional mission:
Step 1: Do not knowingly insure a loan with a projected claim termination rate greater than 10 percent, assuming no house price appreciation or depreciation.
Step 2: Target an average 5 percent projected claim termination rate, assuming no house price appreciation or depreciation.
Step 3: Stop guaranteeing lower-risk loans and high-dollar-balance borrowers, as this allows for cross-subsidization of those loans with excessive risk. This will also let the FHA step back from markets that can be served by the private sector and allow it to concentrate on home buyers who truly need help.
Step 4: Price for risk, since not doing so deprives the borrower of the price information needed to understand the true risk of the loan. Until this is done, the FHA should disclose to the borrower his or her expected claim rate, assuming no house price appreciation or depreciation.
Step 5: Implement underwriting that results in the extension of responsible mortgage credit, by balancing down payment, loan term, FICO score, and debt-to-income ratio to achieve meaningful equity.
The Problem:
Given FHA's mission, allowing the continuation of practices that result in . . . a high proportion offamilies losing their homes represents a disservice to American families and communities. n1
Based on an extensive analysis of the
Most can agree with the FHA's mission to help creditworthy low- and moderate-income Americans and first-time home buyers by providing responsible mortgage credit that leads to homeownership success and neighborhood stability. That said, the real question is: what is the tolerance for failure? All lending entails risk. At what level of failure does lending become abusive because the direct and indirect costs associated with a high foreclosure rate are unacceptably high for FHA borrowers and the affected community?
The findings of this study indicate:
* An estimated 40 percent of the FHA's business consists of loans with either one or two subprime attributes--a FICO score below 660 or a debt ratio greater than or equal to 50 percent (based on loans insured during FY 2012).
* The FHA's underwriting policies encourage low- and moderate-income families with low credit scores n3 or high debt burdens to make risky financing decisions-- one combining one or both a low credit score or a high debt ratio with a 30-year loan term and a low down payment.
o A substantial portion of these loans have an expected failure rate exceeding 10 percent.
o Once the expected failure rate exceeds 10 percent, the resulting direct and indirect costs to low- and moderate-income families and communities are a disservice to the very families and communities it is the FHA's mission to help.
Background
By the mid-1950s FHA had demonstrated the feasibility of [relatively high LTV] lending, given the sound underwriting and appraisal standards it pioneered. n4
Many look at today's FHA and nostalgically recall their great-grandmother's Depression-era FHA. Set up in 1934, it insured fully amortizing 20-year term loans combined with a minimum 20 percent down payment. As late as 1954, down payments and terms on FHA loans still averaged about 20 percent and 20 years, respectively. n5 As a result, home buyers accumulated nearly 30 percent in earned equity after four years. Additionally, housing debt and total housing expense ratios n7 averaged 15 percent and 19.5 percent, respectively, helping to cushion any adverse income shocks. n8 This helps explain why, over its first 20 years, the FHA paid only 5,712 claims out of 2.9 million insured mortgages for a cumulative claims rate of 0.2 percent. At the same time, claim loss severity was only 9 percent of the original insured mortgage balance, or a total of
By the mid-1950s, the FHA had demonstrated the benefits of sound underwriting principles built upon 30 percent earned equity build-up over four years and a modest mortgage debt burden of 15 percent. The FHA began to abandon those principles in the late 1950s. Since then, the FHA's role has been that of the leverage leader, spurring the housing finance industry and borrowers to multiple forms of increasing leverage.
Borrower leverage takes six forms, and the FHA has promoted the simultaneous and excessive use of each, particularly with respect to low- and moderate-income families and communities.
Two types of asset leverage:
1. As the down payment percentage decreases, the asset price of the home it can leverage increases.
2. As the loan amortization term increases, asset leverage remains high because of slower earned equity buildup from amortization during a loan's early years.
Three types of income leverage:
3. As the debt-to-income ratio increases, so does the loan that may be serviced with the same amount of income.
4. As the loan amortization term increases, so does the loan that may be serviced with the same amount of income.
5. As the rate of interest declines, the size of the loan that may be serviced with the same amount of income rises. While the Federal Reserve is responsible for this increase in leverage, the FHA's underwriting policies turn virtually all of this additional buying power into increased buyer leverage.
One type of credit leverage:
6. The lower the acceptable credit score, the larger the pool of buyers. Data are not available to track this expansion back to the 1950s.
Lulled by its early success, encouraged by a housing lobby grown dependent on increasing leverage, and faced with private-sector competition for the first time, n10 the FHA, at the behest of
While the FHA utilized all forms of leverage over the period of 1954 to 2012, data are only available to track the progress of growing asset and income leverage for 1954 to 1966, picking up again in the "00" years.
* Asset leverage increased by a factor of 5 as the average down payment declined from 20 percent to 4 percent.
* Household income leverage increased by a factor of nearly 3.5 as:
o The average loan term increased from 21 years to 30 years, expanding buying power by about 12 percent.
o The average housing debt ratio increased from 15 percent to 35 percent, expanding buying power by 133 percent.
o Interest rates declined. The above comparisons are all based on a constant interest rate of 6 percent. Today, interest rates are about 3 percent, or half this rate, allowing buying power to increase by a further 30 percent.
With asset and household income leverage compounding each other, as figure 1 demonstrates, from 1954 to 2012 the average leverage on an FHA loan increased by a factor of 17 while the FHA's foreclosure start rate increased by a factor of 19. Foreclosure starts were elevated even during the boom years of the 1990s and 2000s. Today, 1 in 20 FHA borrowers enters foreclosure every year.
Figure 1. The FHA's Growing Leverage Spurs a Burgeoning Foreclosure Rate
* In calculating leverage factors for 1954-67, 2006, and 2012, interest rates were assumed to be a constant 6 percent for each year. Leverage was calculated a second time for 2012 using the current rate of 3 percent.
Sources:
The FHA's excessive use of compound leverage led to borrower and the FHA's increased dependence on unearned equity (equity accruing from house price appreciation as a result of leverage), rather than earned equity and responsible lending. Leverage is a double-edged sword. It creates a windfall of unearned equity for home buyers and reduced losses for the FHA when home prices are increasing rapidly. When prices rise more slowly or decline or when income drops, it exposes home buyers to foreclosure. Add in refinances to lower the rate and taking cash out, and burning one' s mortgage by age 60 or 65 became a distant memory.
As shown in figure 1, the FHA began expanding compound leverage in 1957, followed four years later by a burgeoning foreclosure start rate. By 1962, mounting foreclosures had caught the attention of
Homeowners of a new and unattractive breed are plaguing the
The risks that the FHA's policies present to borrowers, lenders, and the market should come as no surprise. As far back as 1970, a multidecade study on postwar lending trends and their impact on mortgage delinquency and foreclosure noted the risks posed by reduced down payments, longer loan terms, high debt ratios, and cash-out refinances. The study made numerous prescient observations, including
There have been numerous warnings that continued liberalization of mortgage terms was creating riskier loans. . . . A second effect of liberalized terms is to magnify the borrower's resources. . . . If lenders were to throw all caution to the winds and require little or no equity buildup on a property which is declining in value, defaults would almost certainly ensue. n13
The point about magnifying a borrower's resources is particularly pertinent to today's artificially low interest rate environment provided by the Federal Reserve. The Fed's purpose is precisely that--"to magnify the borrower's resources." This poses potential risk to families, particularly low- and moderate-income ones; their communities; and the FHA itself. A family with an income of
The FHA Today
Today, the FHA has 7.7 million loans outstanding and pays 12,000 claims per month, as opposed to the 5,000 it paid over its first 20 years of existence. Two of its more recent books are experiencing serious delinquency rates that are 8 (FY 2009 book) and 11 (FY 2010 book) times those of the
Figure 2 shows continuing growth in the serious delinquency rate for the FHA's FY 2010 book compared to
Sources: FHA Single-Family Mutual Mortgage Insurance Fund Programs Quarterly Report to Congress FY 2012 Q3, miscellaneous Fannie Mae Credit Supplements, and miscellaneous MGIC Portfolio Supplements.
As a result, the following description of the FHA of the 1960s and 1970s applies to the FHA today: "Frequently working with local Realtors, [FHA] lenders would solicit home purchases from families who could not, in fact, afford the acquisition." n15
The FHA's flawed policies in the 1960s and 1970s n16 were the impetus for much legislation over the years: the Home Mortgage Disclosure Act of 1975, the Community Reinvestment Act of 1977, and the Federal Housing Enterprises Financial Safety and Soundness Act 1992 (the GSE Act), to name but a few. These were largely aimed at the conventional market, not at reforming the FHA.
The problems caused by the FHA did not abate in the 1990s.
We have been fighting abuse, fraud, and neglect of the FHA program that has destroyed too many neighborhoods and too many families' dreams of homeownership for more than 25 years. . . . The FHA program has a national default rate 3 to 4 times the conventional market, and in many urban neighborhoods it routinely exceeds 10 times. In addition, the FHA program is hemorrhaging money. n17
As a result of the Community Reinvestment Act, the GSE Act, HUD, FHA, and other policy and program initiatives, trillions of dollars of private-sector capital were directed into financing low- and moderate-income housing. The resulting housing boom and bust repeated the FHA's earlier failure: once again harming the very families and communities these policies were intended to help.
Much legislation has also been enacted in an attempt to address deceptive, abusive, and predatory lending: the Home Ownership and Equity Protection Act of 1994 and the Dodd-Frank Act of 2009, to name two. Predatory lending was viewed as "exerting the same adverse impact on urban communities as the FHA scandals did in the 1960s and 1970s." n18 Once again, this legislation was aimed at the conventional market and ignored the FHA. So now we have come full circle, with the FHA acting as the primary source of lending in many communities and again setting up for failure the very families and communities it is tasked with helping.
Some would argue that the FHA saved the housing market because without it, the market would have collapsed. n19 Beyond the fact of the FHA's decades-long role as the home lending leverage leader, this statement also sets up a false choice. Many could agree that the FHA has played a countercyclical role, but the real choice is between responsible and irresponsible underwriting polices targeted at low- and moderate-income families.
Consider the last four years:
* The lowest home prices since 2002 in nominal prices and 1997 in real prices.
* The lowest interest rates in generations.
* Affordability at an historic high.
Had the FHA made two changes to its underwriting and program policies, the result would have been materially improved outcomes for low- and moderate-income families and communities, and the FHA itself:
* On rate and term refinances, use the benefit of a lower rate to shorten loan term and speed equity buildup.
* On home purchase loans, offer buyers either a lower down payment or a 30-year loan term, but not both--this reduction in risk layering would greatly enhance equity buildup.
Instead, the FHA's underwriting policies and practices have knowingly placed a high percentage of low- and moderate-income families and communities at risk of excessively high foreclosure rates. This study identified 9,000 zip codes with a median family income below the applicable metro area median, where the zips also have a projected foreclosure rate n20 equal to or greater than 10 percent. These high foreclosure rate zips account for 44 percent of all FHA loans that are in the low- and moderate-income zips, and they have an average projected foreclosure rate of 15 percent.
The FHA's 2011 Actuarial Study projects that 9.6 percent, or 330,000, of the 3.45 million loans it insured during FY 2009 and 2010 will ultimately be foreclosed upon or otherwise result in a claim against FHA's insurance fund. n21 With an annual foreclosure start rate 19 times the rate in 1954 and a loss severity rate 7 times the level over its first 20 years, this is not your great-grandmother's FHA.
The Government's Subprime Lender
As I have described and shown in figure 1, starting in the late-1950s the FHA began to encourage higher-risk lending. Today, the FHA operates as the government' s subprime lender. To be clear, the FHA's loans are subprime because of their credit attributes--for example, borrowers with impaired credit (a FICO score below 660) or a total debt ratio of 50 percent or greater. These high-risk attributes are then generally layered with the additional risks related to a low down payment and a slowly amortizing 30-year loan term.
HUD has historically defined subprime by loan rate (self-serving, since government subsidies allow FHA mortgage rates to be at or below prime loan rates) or by the use of an identified abusive product like a 2/28 adjustable rate mortgage (ARM). However, federal banking regulators define subprime by credit characteristics--typically, borrowers with weakened or incomplete credit histories or reduced repayment capacity as measured by credit scores, debt-to-income ratios, or other criteria. Two specific attributes relevant to the FHA are a FICO score below 660 and a total debt ratio equal to or greater than 50 percent. n22
Today, 40 percent of the FHA's business consists of loans with either one or two subprime attributes--a FICO score below 660 or a debt ratio greater than or equal to 50 percent (loans insured during FY 2012): n23
* 24 percent had a FICO score <660 and a debt ratio <50 percent
* 4 percent had a FICO score <660 and a debt ratio >50 percent
* 12 percent had a debt ratio >50 percent and a FICO score >660
These attributes are generally combined with a high LTV ratio (81 percent of the
For the FHA to knowingly place a high percentage of low- and moderate-income families and communities at risk of excessively high foreclosure and delinquency rates is unfair and deceptive and constitutes an abusive practice.
The
Historically, the FHA's mission was to be "a targeted provider of mortgage credit for low- and moderate-income Americans and first-time home buyers." n24 As noted in their 2011 report to
Democrats and Republicans alike can agree on this mission going forward. They should also agree with the benchmark HUD itself set for evaluating the FHA's performance: "Given FHA's mission, allowing the continuation of practices that result in . . . a high proportion of families losing their homes represents a disservice to American families and communities." n26
Does the FHA meet this standard with respect to its mission of being a targeted provider of responsible mortgage credit that leads to homeownership success for low- and moderate-income Americans and first-time home buyers? Based an analysis of the FHA's FY 2009 and 2010 books of business, the answer is no.
The FHA's underwriting policies encourage low- and moderate-income families with low 27 credit scores n27 to make a risky financing decision--one combining a low score with a 30-year loan term and a low down payment. This sets up for failure the very families and communities it is the FHA's mission to help. As a result, too many low- and moderate-income borrowers see their hope for the American dream turned into a nightmare.
Figure 3 shows that as a zip's median income declines in relation to the area median, the projected foreclosure rate increases (blue bars) and increases even more for the 50 percent of zips that are the most risky (red bars). For the riskiest zips with median income of less than 60 percent of the area median (red bar), the average projected foreclosure rate is over 17 percent. Yet this average masks shocking levels of failure in individual zip codes. In
Source: Author's data.
The FHA's low- and moderate-income families in zips where the median income is below the applicable metro area median also suffered substantially larger home price declines than in zips above the area median income. For example, zips with less than 60 percent of the median income had a 13.3 percent home price decline; the average zip had a 9.4 percent decline (FHA loans), while the average decline across
As figure 4 demonstrates, many of these families are substantially underwater on their mortgages. Even if these families do not lose their homes to foreclosure, they are unlikely to have the opportunity to accumulate equity for many years.
Figure 4. Home Price Decline, June 2009-August 2012 by Zip Income as a Percent of Area Median (FY 2009-2010)
Source: Author's data.
Some might argue that an expected foreclosure rate of 10, 20, or even 30 percent is a small price to pay, as it means that 90, 80, or 70 percent of the loans are performing.
From my point of view, if 80% of the sub-prime borrowers are managing to make ends meet and make the mortgage payments on time, then, shouldn't we as a Nation, be justifiably proud that we are dramatically increasing homeownership opportunities for those who have been traditionally left behind. n28
But strong reasons exist to dismiss the view that these levels of foreclosure are acceptable.
The collateral impact on neighbors and the surrounding community these foreclosure levels cause is substantial. These foreclosures are concentrated in low- and moderate-income neighborhoods. The resulting reduced or declining home values impact FHA and non-FHA low- and moderate-income families diligently making their payments. These families may be denied the opportunity to build equity, provide security for their family, and have the down payment for their next home as their family grows.
Foreclosures result in increased blight and crime, and the larger community suffers from a reduced tax base and higher costs for providing municipal services.
The impact of abusive lending on borrowers extends beyond excessive foreclosures. For every FHA borrower in foreclosure, another four are at an earlier stage of delinquency. Table 1a shows the impact of financial duress on FICO scores. A consumer with a 680 FICO score who experiences a new 30-day mortgage delinquency will see his or her score drop by an average of 60-80 points. n29
Table 1a. Impact of Financial Distress on FICO Scores
This impact on families is troubling because the FHA's abusive lending practices lead to excessive levels of delinquency. Today, one in six FHA loans is delinquent 30 days or more.
* At a projected foreclosure rate averaging 10 percent, nearly 1 in 5 FHA loans is delinquent 30 days or more at least once in a two year period.
* In a working class neighborhood with a projected foreclosure rate averaging 17 percent, nearly 1 in 3 FHA loans is delinquent 30 days or more at least once in a two year period.
Table 1b shows the estimated time for FICO Score to fully recover from a delinquency or other adverse credit event. It takes 9 months for a consumer with a 680 FICO score who experiences a new 30-day mortgage delinquency to get recover back to a 680 score (assuming no other changes). n30
Table 1b. Estimated time for FICO Score to fully recover
The FHA financing path keeps families in a cycle of delinquency, foreclosure, and dependency.
This abusive practice is all the more unfair and deceptive since, as I will explain more extensively below, the FHA does not price for risk. As a result, unsuspecting borrowers are offered government financing where a high-risk FHA loan is as much as 24 times more likely to experience a serious delinquency than a low-risk one.
Even if one were to take the extreme position that a failure rate in excess of 10 percent and the adverse neighborhood impacts are acceptable, there would appear to be no justification for denying prospective borrowers the facts they need to make an informed decision about the deleterious impacts of a higher-risk FHA loan.
A one-in-five chance of losing one's home and entire investment, along with being denied access to most forms of credit for three to five years and, likely, restrictions on the ability to rent a new place to live.
Even if a borrower avoids the 20 percent chance of foreclosure, he or she is substantially likely to experience a mortgage delinquency that will negatively impact his or her FICO score for years, severely limit access to credit, and greatly increase the cost of whatever credit might be available.
Equally troubling is that the FHA projects that by 2015 its median FICO score will decline from 700 to 685. Unless the FHA takes steps to reduce risk layering as it moves out the credit risk curve, its lending standards represent an even greater disservice to low- and moderate-income families and communities.
These results collectively demonstrate that the FHA is financing failure in working-class communities on a level that is a national scandal.
Credit Scores as an Objective Means of Evaluating Credit Risk
In 2007, the Federal Reserve submitted a comprehensive report to
(1) The credit history scores evaluated here are predictive of credit risk for the population as a whole and for all major demographic groups. That is, over any credit-score range, the higher (better) the credit score, the lower the observed incidence of default. These conclusions are limited to credit history scores, that is, scores calculated exclusively on the basis of individuals' credit records as assembled by the three national credit-reporting agencies (
(2) Results obtained with the model estimated especially for this study suggest that the credit characteristics included in credit history scoring models do not serve as substitutes, or proxies, for race, ethnicity, or sex.
(3) Different demographic groups have substantially different credit scores, on average. For example, on average, blacks and Hispanics have lower credit scores than non-Hispanic whites and Asians, and individuals younger than age 30 have lower credit scores than older individuals.
Table 2 is derived from the Federal Reserve study and quantifies the Fed's observation regarding different credit scores for different demographic groups.
Table 2. Credit Score Distribution by Demographic Groups
...Fed.study.interest.rate..Fed.study.credit.score.distribution.Fed.study.default.rate..FHA.serious.delinquency.rate..Non-Hispanic.white..Black..Hispanic..Credit.score.band.......
<580......9.56%....11%.....30%....30%.....8%....33%..17%
580-619...8.94%....8.50%...18%....20%.....7%....20%..13%
620-659...7.30%....10.50%..14%....11%.....9%....15%..16%
660-719...6.40%....19%.....5%......2%.....20%...17%..24%
>719......6.10%....51%.....1%......55%....16%...30%
Some results interpolated to standardize across credit score bands
Source: Federal Reserve, Report to the
The study also found:
Credit scoring likely increases the consistency and objectivity of credit evaluation and thus may help diminish the possibility that credit decisions will be influenced by personal characteristics or other factors prohibited by law, including race or ethnicity.
The analysis conducted for this study finds that credit scores consistently predict relative loan performance within all population groups; that is, for all populations, the percentage of individuals experiencing a serious delinquency on one or more of their credit accounts consistently declines as credit scores increase. The analysis also finds that some groups perform worse (experience higher rates of serious delinquency) on their credit accounts, on average, than would be predicted by the performance of individuals in the broader population with similar credit scores. For example, on average, blacks perform worse than other racial and ethnic groups with similar credit scores. Similarly, single individuals and those residing in predominantly black or low-income census tracts perform worse on their loans than do their complementary demographic groups with similar credit scores. n32
Financing Failure: Homing in on the FHA's Problems The FHA Does Not Price for Risk
Two examples demonstrate this:
* The FHA charges virtually the same mortgage insurance premium for a borrower with a 3.5 percent down payment, a 580 FICO score, and a 50 percent total-debt-to-income ratio (loan A in table 2 below) as for one with a 20 percent down payment and a 720+
FICO score and a 25 percent total-debt-to-income ratio (loan G below). n33 Yet the first loan will experience a serious delinquency rate 24 times higher than the second.
* The FHA charges the same mortgage insurance premium for a borrower with a 3.5 percent down payment, a 580 FICO score, and a 50 percent total-debt-to-income ratio (loan A in table 2 below) as the same down payment and total-debt-to-income ratio, but with a 720+ FICO score (loan F below). Yet the first loan will experience a serious delinquency rate eight times higher than the second.
Table 3. Regardless of the Risk, the Mortgage Insurance Premium Is Essentially the Same
........Loan.A....Loan.B...Loan.C...Loan.D...Loan.E...Loan.F...Loan.G
Term....30-yr......30-yr....30-yr....30-yr....30-yr....30-yr....30-yr
LTV.....96.50%...96.50%.....96.5%...96.5%....96.5%....96.5%.....80%
FICO....580-599.600-619...620-659.660-679...680-719....720+.....720+
Total.debt-income-ratio..>50%....>50%.>50%.>50%.>50%.>50%.<=25%
Serious.delinquency.rate.24%....21%.14%.8%.6%.3%.1%
Mortgage.insurance.premium.1.75%.upfront.1.35%.annually.Same.Same.Same.Same.Same.1.75%.upfront.1.30%.annually
Source: Author.
When pricing is risk based, a higher rate signals to the borrower that his or her risk profile is higher. Flat pricing by the government encourages adverse selection, promotes moral hazard, distorts market competition, and leads to higher levels of foreclosure that impose personal as well as social costs.
Adverse selection. The Federal Reserve, in its Report to
When the interest rate charged by a lender is appropriate for the average risk pool of prospective borrowers but is either to low or too high for some of the individual borrowers, the pool can suffer adverse selection, that is, a rise in the relative number of high risk borrowers. n34
High-risk borrowers have incentives to take on more risk. The Federal Reserve also observed:
High risk borrowers--those for whom the correct individual interest rate would be higher than the average rate--will perceive the single-rate offer as a good deal and accept the terms, perhaps borrowing more than they would if charged a rate more consistent with their risk profile. n35
Stronger borrowers have incentives to take on more risk. If borrowers with good credit can make a smaller down payment or take out a larger loan relative to their income without having to pay higher mortgage insurance premiums or a higher rate, they will. This is demonstrated on home purchase loans where the FHA's minimum down payment is 3.5 percent and the average is 4 percent. n36 Those borrowers now have a higher likelihood of default than less-leveraged borrowers. They benefit from the private gains benefits accruing from rising home prices and are able to socialize their losses arising from defaults, with the risk being borne by the federal government. This results in moral hazard and attendant unintended consequences.
High-risk behavior is promoted with subsidies. Borrowers have little incentive to exercise discipline to save for a down payment, improve their credit profile, or keep debt levels moderate, since these actions do not result in a lower mortgage insurance premium or interest rate. Instead, borrowers are motivated to take on more risk through subsidies. The combined FHA/Ginnie Mae subsidy on an FHA loan may be estimated by comparing pricing for a similar
* High-risk loan: A
About one-third of this subsidy results from the pricing advantage Ginnie Mae provides when an FHA loan is securitized.
The rest results from the various subsidies and cross-subsidies that the FHA itself receives or creates.
The FHA's ability to cross-subsidize high-risk loans like this one with premium income from lower-risk loans (generally those with FICO scores above 700) enables it to mask the concentrated pain inflicted on low- and moderate-income families and communities. It also enables the FHA to absorb catastrophic losses in low- and moderate-income communities.
At the same time, the FHA's so-called lower-risk loans seem that way only when compared to its higher-risk loans. While these are the loans providing the previously mentioned cross-subsidy, they themselves are high-risk loans when compared to low-down-payment, privately insured conventional loans. The serious delinquency rate on the portion of the FHA's FY 2010 book with an LTV ratio greater than 85 percent and less than or equal to 95 percent, combined with a FICO score greater than 720, was nearly four times the rate on MGIC's entire 2010 book with similar seasoning. For example, FHA loans with a balance of
Unless you are a real estate agent, it makes no sense from a policy perspective for the government to subsidize this much debt on
* Medium-risk loan: A
Housing finance participants that price for risk cannot compete with government programs that do not. Today, the FHA and the other agencies n38 using the Ginnie Mae guarantee account for about half of all home purchase loans. The failure to price for risk, along with other advantages the government bestows on the agencies, crowds out competitors since they do price for risk. This can be seen by comparing
Figure 5 shows there is little FICO credit score overlap, given 65 percent versus 13 percent of the FHA and Freddie's volumes, respectively, have credit scores below 720.
Figure 5. FHA's Subsidy and Underwriting Advantages Crowd Out Freddie Mac
Source: Inside Mortgage Finance and this study.
Figure 6 demonstrates there is also little LTV ratio overlap given that 81 percent versus 19 percent of the FHA and Freddie's volumes, respectively, have an LTV ratio of 95 percent or greater.
Figure 6. FHA's Subsidy Advantages Crowd Out Freddie Mac
Source: Inside Mortgage Finance and this study.
Higher levels of foreclosure have an adverse impact on neighborhoods. Abundant research documents the adverse impacts on neighborhoods, including reducing neighboring property values, n40 higher levels of crime, n41 impact on children, n42 older citizens, n43 and an adverse impact on health, n44 and a disproportionate impact on minorities. n45
With respect to the direct adverse impact of foreclosures on nearby families, the
Conclusions:
* The FHA not pricing for risk leads to a variant of Gresham's Law: high risk lending drives out low risk lending.
* The FHA's high risk lending policies sell hope and deliver harm, making them a disservice to low- and-moderate-income families and communities.
The FHA Does Not Underwrite for Risk
Based on an analysis of the characteristics of FHA loans, it is apparent that the FHA does not adequately evaluate the various risk factors present in a loan in a holistic manner. See appendix A for additional detail on these risk relationships.
Consider the FHA's underwriting of home purchase loans:
Down Payment: While the FHA's minimum down payment on a home purchase loan is 3.5 percent, the average is 4 percent. This indicates that down payment is not being used as a factor to offset other risks present on an individual loan.
* As indicated in appendix A, LTV ratio is generally a moderately effective indicator of default for FHA loans with FICO scores below 660.
* Loan term: 92 percent of all fully underwritten fixed rate loans (purchase and refinance) have a 30-year term. This indicates that loan term is not used as a factor to offset other risks. As indicated in appendix A, loan term is generally a moderately effective indicator of default for FHA loans, although less so for FICO scores below 620.
Total debt ratio: A slightly higher percentage of borrowers with a 620-659 scores had a total debt ratio greater than 50 percent than did borrowers with a 720+ FICO score. This indicates that total debt ratio was not used as a factor to offset other risks present on an individual loan.
* As indicated in appendix A, total debt ratio is generally a moderately effective indicator of default for FHA loans, although its effectiveness declines as FICO score declines.
Having eliminated underwriting variation based on down payment, loan term, and total debt ratio, only FICO score remains.
* As the Federal Reserve found, "Credit scores consistently predict relative loan performance within all population groups." n47
* The performance of FHA loans indicates that that FICO score is a strong indicator of default for FHA loans.
The FHA's underwriting policies largely ignore the fact that a 30-year term loan with a 95 percent LTV, a FICO score of 580-599, and a debt ratio of >50 percent has a serious delinquency rate of 24 percent, which is 24 times the 1 percent serious delinquency rate on a 30-year term loan with 95 percent LTV, a FICO score of >720, and a debt ratio of <25 percent.
Likewise, it ignores that a 30-year term loan with a 95 percent LTV, a FICO score of 580-599, and a debt ratio of >50 percent has a serious delinquency rate of 24 percent, 8 times the serious delinquency level of 3 percent on a 30-year term loan with 95 percent LTV, a FICO score of >720, and a debt ratio of >50 percent.
However, a 20-year loan term can be used effectively to mitigate risk across all FICO bands.
Figure 7 compares delinquency rates for 15-, 20-, and 30-year term mortgages. Mortgages with 15- and 20-year terms perform similarly and have far lower delinquency rates than 30-year mortgages.
Figure 7. FHA Home Purchase Mortgage Performance by Term
Source: Author's data.
Figure 8 shows that shorter-term mortgages effectively to mitigate risk across all FICO bands. Figure 8 is limited to 15- and 30-year term mortgages as FHA's 20-year is insufficient for this level of detail.
Figure 8. FHA Home Purchase and Refinance Mortgage Performance by Loan Term across FICO Bands (60-day-plus delinquency)
Conclusions:
* The FHA's failure to underwrite in a responsible manner leads it to largely ignore risk layering on individual loans.
* This leads to unnecessarily high foreclosure and delinquency rates that are a disservice to low- and moderate-income families and communities.
* Reducing loan term is highly effective as a risk mitigant.
The FHA's Irresponsible Underwriting Policies and Abusive Lending Practices
As shown in figure 9, a high percentage of the home purchase loans made to low- and moderate-income and minority borrowers are backed by the government. n48
Figure 9. Share of Home Purchase Loans with Federal Backing in 2010 (Percent)Virtually all of these loans combine a 30-year loan term with a low or no down payment, and many are to borrowers with low credit scores and/or high debt ratios. This intentional risk layering places a high percentage of low- and moderate-income families and communities at risk of excessively high foreclosure and delinquency rates. Combined with the FHA's failure to price or underwrite for risk and the fact that low-price-tier neighborhoods have historically had more volatile price movement, this policy results in disproportionate impact on low- and moderate-income families and communities. Since failing to price and underwrite for risk is done knowingly, the FHA's underwriting policies and practices are unfair and deceptive and constitute an abusive practice.
More volatile home price movement for lower-price-tier homes was generally the case during the Great Boom/Bust of the 1990s/2000s. The lowest price tier had the largest percentage price increases followed by the steepest declines. Figure 10 provides a representative example:
Figure 10. S&P/Case-Shiller Chicago Tiered Home Price Indices
Source: S&P Indices and
A study of 16 large metro areas found that in every case, the prices of homes in the lower price tier have fallen more from the peak than homes in the median and high price tiers. n49
This trend has continued since mid-2009 as lower-price-tier homes insured by the FHA in FY 2009-10 have also seen greater price declines or smaller price increases than higher-priced homes.
The combination of the FHA's underwriting policies and practices and low-tier home price volatility results in a disproportionately high level of low- and moderate-income families losing their homes.
As previously noted, this study found 9,000 zip codes where the median income is below the applicable metro area median and where the projected foreclosure rate is equal to or greater than 10 percent. These zips account for 44 percent of all loans in the low- and moderate income zips. The average projected foreclosure rate for these zips is 15 percent.
However the impact extends well beyond the 10 percent or more that may lose their homes.
* Another 20 to 30 percent will have their credit scores impacted by one or more delinquent payments.
* All homeowners in the neighborhood will be impacted as these FHA loans become delinquent and go into in foreclosure.
How the Study of the FHA's FY 2009 and 2010 Books Supports This Analysis
In this study, I undertook a detailed examination of the FY 2009 and 2010 books to identify which loans are most likely to suffer a foreclosure. I also sought to identify which zip codes will bear the brunt of the 330,000 foreclosures forecast to result from the FY 2009 and 2010 books of insurance. n50
Data included loan risk factors such as LTV, FICO score, loan term, and debt-to-income ratio along with delinquency status. Based on delinquency data along with projections from the FHA's Actuarial Study, I calculated a projected cumulative foreclosure and other insurance claim rate ("projected cumulative foreclosure rate") for each zip code. n51 Additional demographic data was appended at the zip code level, including median family income, house price change, and percent of mortgages underwater.
An analysis of the entire database revealed that while LTV, FICO score, loan term, and debt-to-income ratio are captured as part of the insurance process, the FHA did not appear to be use these risk factors in any significant way in deciding whether to insure a given loan or under what terms.
I then analyzed the results at a zip code level within individual large metropolitan (metro) areas n52 , with a particular focus on zip codes where the median family income was below the median family income for the metro area. The results were striking in that most metro areas showed an extremely wide dispersion of projected cumulative foreclosure rates, typically from a low of 3-5 percent to a high of 20-30 percent or higher across a broad range of incomes below the median for the metro area.
Further analysis was undertaken to determine what might explain this result. The study found:
1. A wide disparity in projected cumulative foreclosure rates among metro areas;
2. An even wider disparity among zip codes within metro areas.
3. At the metropolitan statistical area (metro area) level, FICO scores largely accounted for the wide dispersion of projected cumulative foreclosure rates among zip codes that had a median family income below the metro area median.
4. Many areas with heavy concentrations of FHA loans suffered declining home prices since mid-2009. Home price declines had a disproportionate impact on low-and moderate-income families and communities, the groups traditionally served under the FHA's mission.
Implications
These results raise serious policy issues:
1. The lending policies that FHA has in place today promote the financing of failure for too many families and too many communities: The study found 9,000 low- and moderate-income zip codes with projected cumulative foreclosure rates of 10 percent or greater, with an average 15 percent projected cumulate foreclosure rate. These zips accounted for 38 percent of the study's 23,800 low- and moderate-income zip codes and for 44 percent of all FHA loans studied in these same zip codes.
2. The FHA's lending policies are putting borrowers in low- and moderate-income communities in a negative equity hole that will be difficult to escape from. As noted in Figure 12 below families in zips where the median family income is below the applicable metro area median suffered substantially larger home price declines than those in zips above the area median.
3. Although many agree the FHA's activity needs to be refocused on its traditional mission of serving low-income borrowers, if this occurs and the FHA continues its irresponsible lending policies, it will be financing failure on an even larger scale. The FHA plans to continue its underwriting policies with one notable change: reducing its median FICO score from 700 to a projected score of 660. This higher level of risk-layered loans will result in a substantial increase in expected foreclosure rate. Unless reduced, we face the prospect of the FHA performing a further disservice to American families living in low- and moderate-income communities.
4. The guarantee of the taxpayer is used to put low- and moderate-income families and communities at risk. The FHA's pricing advantages and lending policies entice many low- and moderate-income families into loans that finance failure. The enablers of the FHA rely on this backstop to encourage, originate, securitize, and guarantee these risky loans.
What Is the Impact of the FHA's Lending Policies at the Individual Zip Code and Metropolitan Area Levels?
A review at the zip code and metropolitan area levels confirmed that risk layering, combined with high FHA loan volumes, leads to a concentration of high foreclosure and delinquency rates in low- and moderate-income communities. n53
This section contains an overview of the case study for
The zip boundaries are color coded based on a scale of a projected cumulative foreclosure rate of percent for the FHA's combined FY 2009 and 2010 books.
Figure 11.
The summary of the
Figure 12.
The x-axis represents a zip code's median home price as a percent of the area's median home price.
The y-axis represents a zip code's median home price change since
* The US median home price change was 7 percent since
*
* Median home prices declined 26.7 percent in those
The color of the dot represents the zip's foreclosure rate.
The size of the dot represents a zip's loan count of FHA loans in 2009-2010.
What this scatter plot tells us:
* Price declines were greatest in lower-income areas, which would also have had lower-priced homes. In this instance, the preponderance of zips with a median family income of less than 100 percent of the area median had price declines greater than 22 percent for the area (average was -26.7 percent for zips less than 100 percent of area median).
o These same zips tend to have very high expected foreclosure rates and higher-than-average loan counts.
* The lower-right portion gridded area of the scatter plot highlights the problem. Rather than helping low- and moderate-income families build meaningful equity along with stable communities, the FHA's underwriting policies are leading to equity destruction.
* Beyond the direct impact on families in terms of being likely to experience excessive foreclosure and delinquency rates, other FHA borrowers and the overall community suffer from equity destruction as a result of larger-than-average price declines.
FHA reform principles supported:
* Principle 1: Step back from markets that can be served by the private sector.
* Principle 2: Stop knowingly lending to people who cannot afford to repay their loans.
* Principle 3: Set loan terms that help homeowners establish meaningful equity in their homes with the goal of ending their dependence on FHA lending.
* Principle 4: Concentrate on those home buyers who truly need help purchasing their first home.
Figure 13.
The x-axis represents a zip code's median home price as a percent of the area's median home price.
The y-axis represents a zip code's median family income as a percent of the area's median family income.
The color of the dot represents a zip's foreclosure rate.
The size of the dot represents a zip's percentage of loans with a low FICO score (less than 660).
What this scatter plot tells us:
* The lower-left quadrant (representing zips with both a median family income and a median home price below
o The zip codes in the lower-left quadrant account for 53 percent of the loans in all the zip codes shown.
o This quadrant showed a higher percentage of loans with a low FICO score and a high foreclosure rate.
o Rather than providing responsible mortgage credit that leads to homeownership success for low- and moderate-income families and communities, the FHA's loans perform a disservice to the very families and communities it is tasked to serve.
FHA reform principles supported:
o Principle 1: Step back from markets that can be served by the private sector.
o Principle 2: Stop knowingly lending to people who cannot afford to repay their loans.
o Principle 3: Set loan terms that help homeowners establish meaningful equity in their homes with the goal of ending their dependence on FHA lending.
o Principle 4: Concentrate on those home buyers who truly need help purchasing their first home.
Table 4 sets forth the foreclosure rate, percent of purchase loans that were government loans n55 in 2009, and other data for the 10 Chicago zip codes with the lowest and 10 with the highest projected foreclosure rates. Each zip had at least 50 FHA loans. As I have noted, FHA and other government-backed loans are much more predominant among low- and moderate-income home buyers.
The chart demonstrates the strong correlation among foreclosure rate, FICO score, income, and home price. It also demonstrates the harmful effects of the FHA's policies that promote layered risk. This negative impact is magnified because zip codes with the highest foreclosure rates also have the highest incidence of home purchase lending guaranteed by a government agency (FHA,
Table 4.
Property Zip GNW Loan Count GNW Average CTR GNW Percent of FICOs <660 GNW HMDA Data: FHA, FSA/RHS& VA Loans as Percent of Total Home Purchase Loans Median Family Income for CBSA division or CBSA (FFIEC) Median family income ($) (census) Median family income (zip)/media n family income (CBSA/divis ion) -derived
Lowest
60661 121 1.6% 13.2% 13.3%
60070 60 3.8% 40.0% 28.0%
60614 144 4.2% 14.6% 9.8%
60657 207 5.1% 12.6% 13.8%
60642 78 5.4% 21.8% N/A
60463 64 5.4% 28.1% 26.5%
60610 148 5.4% 15.5% 13.2%
60607 227 5.5% 19.8% 19.2%
60613 223 5.6% 17.9% 16.9%
60187 175 5.9% 28.0% 32.6%
Highest |
60621 142 72.6% 54.2% 41.49%
60636 209 60.1% 56.5% 49.72%
60609 212 45.5% 43.9% 24.70%
60624 134 41.4% 59.0% 36.22%
60637 174 35.3% 56.3% 37.60%
60428 110 33.2% 59.1% N/A
60456 66 31.7% 57.6% 70.2%
60827 177 31.6% 54.2% 65.4%
60422 137 31.6% 47.4% 60.6%
60426 140 30.1% 60.7% 69.6%
Sources: Author's data and City-Data.com.
Figure 12 above shows the strong correlation among a zip's relative home price, its home price decline since
Conclusions:
* The FHA's risk-layered approach is irresponsible.
* The approach is a disservice to the very families and communities it is its mission to serve.
* FHA's underwriting policies are unfair and deceptive and constitute an abusive lending practice.
The Adverse Financial Impact of These Foreclosures on the
The FHA projects that its FY 2009 and 2010 books will result in 330,000 foreclosures. n56 This study found that an estimated 85 percent of these foreclosures will occur in zip codes with a median income below the metropolitan area median. There are two distinct, direct adverse financial impacts of these foreclosures:
First, the FHA's losses on the 280,000 foreclosures (85 percent of 330,000) equals
Second is the impact on the price of nearby homes owned by families located in these working class zip codes with high expected foreclosure rates. The methodology developed by Harding, Rosenblatt, and Yao in 2008 n57 is used to estimate the loss associated with each foreclosure. They found each foreclosure results in an average estimated 1 percent home price reduction on every home within a 600 foot radius (0.04 square miles) of a foreclosure. Assuming a housing density of 2,500 homes per square mile, each foreclosure results in a 1percent home reduction on each of 100 neighboring homes. At an average home price of
The FHA's 280,000 foreclosures results in 28 million homes suffering total price depreciation of
The combined total of these direct impacts is
Who Are the FHA's Enablers?
How is it that the FHA is able to continue its irresponsible and abusive lending practices targeted at low- and moderate-income families and communities?
Why is there little concern over the fact that the FHA's overall delinquency rate is habitually around one in six loans?
The FHA, as part of the
Investors in Ginnie Mae mortgage-backed securities (Ginnie MBS) are indifferent to the quality of the underlying mortgages since Ginnie MBS are backed by the full faith and credit of the US government.
Ginnie Mae is largely indifferent because it is insured against loss by the FHA, an agency backed by the full faith and credit of the US government.
Real estate agents and home builders (and the associations that represent them) are largely indifferent since they take their profits at the loan closing.
Originators continue to originate these risky loans because they are insured against loss by the FHA. n59
Borrowers are indifferent since they are charged the same insurance premium regardless of whether it is a 30-year term loan with 96.5 percent LTV ratio, a 580 FICO credit score, and a 50 percent debt ratio or a 96.5 percent LTV ratio, a 740 FICO credit score, and a 50 percent debt ratio. This policy leaves borrowers in the dark with respect to risk. A substantial percentage of FHA loans can be fairly characterized as irresponsible, even toxic loan products. But the FHA does not disclose these extraordinarily high levels of risk to borrowers and FHA is able to cover the outsized losses on these risky loans through the use of cross-subsidies.
The combination of not pricing or underwriting for risk and government guarantees allows all the above participants to ignore or be ignorant of the risks that result from the FHA's policies. In short, the FHA is built around moral hazard n60 where no one has real skin in the game.
Take real estate agents and home builders. They continually call for looser lending to qualify more marginal buyers. n61
Consider the Community Reinvestment Act where layered risks are termed "flexible" underwriting standards.
Finally, this is a result that could not continue without the tacit support of
The NAR as Enabler: Real Estate Agents Gain Even as Low- and Moderate-Income Borrowers Lose
Real estate agents get their profits at the time a home is sold, leaving them with no "skin in the game" of housing finance. Thus, it comes as no surprise that the
In a NAR news release, the organization's chief economist,
The NAR's definition of "sensible lending standards" is driven by what serves its self-interest--more agent commissions--rather than the adverse impact on low- and moderate-income families and communities. If the FHA were to follow the NAR's advice and lower its FICO average but continue using its current risk layering, this would have disastrous consequences for additional low- and moderate-income families that would now be receiving FHA-insured loans.
The NAR's push for looser underwriting is neither news nor a policy change. Near the height of the housing and mortgage bubble in 2006, then-NAR president
There is no need to recount how disastrous this was for homeowners, particularly low- and moderate-income communities in the 2000s. But what does the NAR's perennial push for risk layering mean for FHA families living in low- and moderate-income communities today? The FHA insured an estimated 1.8 million home purchase loans during FY 2009 and 2010 that were in zip codes where the median family income was below the median for the metro area. n65
* The estimated average sales price was
* The average real estate commission in America is approximately 5.1 percent, n66 for total commissions of
* These FHA home buyers have accumulated estimated earned equity of 8.5 percent (4 percent down payment plus 4.5 percent from scheduled loan amortization over three years), resulting in total earned equity of
* These FHA home buyers have had an estimated 10.5 percent home price decline since
The result is a current gross equity position of these FHA home buyers of -
The winners are real estate agents at +$14 billion and the losers are FHA home buyers in below-median-income zips at -
The estimated FHA foreclosure rate for all home buyers in below-median-income zips is 10 percent. For those home buyers in zips with highest foreclosure rates representing 50 percent of the loan volume in these below median income zips is 14 percent.
Conclusions:
* Realtors' advantage in the housing market is not surprising given that real estate agents do not have 'skin-in-the-game' and the FHA entices higher-risk home buyers with its irresponsible lending practices.
* Yet again, the NAR is more concerned about financial gains for its members than the harmful impact the underwriting changes that they espouse will have on low-and moderate-income families and communities.
The Consequences of the FHA's Excessive Leverage
The FHA's failure to apply risk-based pricing and underwriting standards guarantees failure for many low- and moderate-income families. n67 The FHA's experience demonstrates that insuring highly leveraged loans to lower-income borrowers with very low down payments and a 30-year loan term combined with impaired credit and high debt ratios results in excessive foreclosure and delinquency rates that are a disservice to these families and their communities.
Conclusions:
* It is time for the FHA to stop knowingly lending to people who cannot afford to repay their loans.
* It is time to return to what has worked in the past for FHA: insuring reduced-leverage loans for families who have impaired credit.
Reforming the FHA
Social policies that help low- and moderate-income families become homeowners have an important place, but they must balance the interest in low- and moderate-income lending against the risks to the borrowers and the interests of the taxpayers. In the past, "affordable housing" and similar policies have sought to produce certain outcomes--such as an increase in homeownership--which turned out to escalate the risks for both borrowers and taxpayers. The mortgages made in pursuance of social policies can be lower than prime quality--taxpayers may be willing to take risks to attain some social goods--but quality and budgetary limits must be placed on riskier lending to keep taxpayer losses within known and reasonable bounds.
Consistent with these policies, the FHA needs to return to its traditional mission of being a targeted provider of mortgage credit for low- and moderate-income Americans and first-time home buyers. The FHA performs a disservice to American families and communities by undertaking practices that result in a high proportion of families losing their homes.
Families looking for help from the FHA in buying a home want responsible lending solutions that do not finance failure and foster continued dependency. These include:
o A well-underwritten loan they can afford.
o Terms that will not expose them and their community to a high risk of default.
o The opportunity to build equity, provide security for their family, and have the down payment for their next home as their family grows.
o A financing path that does not force them into a cycle of delinquency, default, and foreclosure.
It is obvious that far too many families who took out FHA and other high-risk loans have seen their wealth destroyed, their credit severely damaged for many years into the future, and their neighborhoods sent into severe decline. This is particularly so for lower-income and minority families. As this study demonstrates, the FHA is now the prime mover in perpetuating such effects. Although one could focus on treating the symptoms (foreclosures, negative equity, n68 or insufficient income to support the mortgage payment), that will perpetuate the FHA's irresponsible lending practices. Focusing the FHA on responsible lending addresses the root causes: the FHA's ongoing underwriting practices that set borrowers up for failure and ensure continuing decline of neighborhoods while allowing the market and the judicial system to more rapidly correct existing problems in the housing market.
Four Principles for Responsible FHA Lending Reform
Principle 1: Step back from markets that can be served by the private sector. The FHA has significant advantages that allow it to offer much lower rates than the private sector. These include a free explicit federal guarantee and no need to earn a return on capital, pay taxes, or cover administrative costs. Unless these substantial advantages are narrowly targeted, they lead to unfair and dangerous competition with the prime and subprime private sector, political interference, and the muting of pricing signals. Over a period of three to five years, the FHA should return to a purchase market share of 10-15 percent rather than today's 30 percent.
Principle 2: Stop knowingly lending to people who cannot afford to repay their loans. Although the loans the FHA insured in 2009-11 are called its good books of business, its 2011 Actuarial Study projects that even under a rosy scenario, these guaranty books will experience an average cumulative foreclosure rate of 8.5 per 100 loans, or about 1 in 12 loans. n69 But averages can be deceiving. As this study has demonstrated, these foreclosures disproportionally impact low- and moderate-income families and communities. The FHA should target an average foreclosure rate of 5 percent, assuming no house price appreciation or depreciation, and limit the worst credit risk categories to a maximum claim rate of 10. It should also price for risk, since not doing so deprives the borrower of price information needed to understand the true risk. Until it does so, it should disclose to the borrower his or her expected claim rate, assuming no house price appreciation or depreciation.
Principle 3: Set loan terms that help homeowners establish meaningful equity in their homes with the goal of ending their dependence on FHA lending. FHA should balance the layering of risk factors such as high LTV ratio, low FICO score, 30-year loan term, and high debt-to-income (DTI) ratio so as to allow borrowers to achieve meaningful equity and build wealth. It should limit its insurance to refinance loans where the lower rate is used enable a reduction in loan term so as to speed amortization and builds meaningful equity. FHA insurance should not be used to enable cash-out refinances, since those work against meaningful equity buildup and reduce the overall equity cushion vital to reducing price volatility in low- and moderate-income communities.
Principle 4: Concentrate on those home buyers who truly need help purchasing their first home. In FY 2011, 54 percent of FHA's dollar volume went to finance homes that were greater than 125 percent of an area's median house price, up from 22 percent in 2009. Given the FHA's mission is to help low- and moderate-income home buyers, the homes it finances should cost less than the median home price for an area. Additionally, first-time home buyers should be limited to an income of less than 100 percent of area median income and repeat home buyers to an income of less than 80 percent of area median income.
Adopting these four principles would return the FHA to its traditional mission of being a targeted provider of responsible mortgage credit to low- and moderate-income Americans and first-time home buyers.
Specific Reforms to Implement These Four Principles
The following sets forth a series of steps designed to implement the four principles. Implementing all of these steps will enable the FHA to return to its traditional mission of being a targeted provider of responsible mortgage credit to low- and moderate-income Americans and first-time home buyers.
Step 1: FHA will not knowingly insure a loan with a projected claim termination rate greater than 10 percent, assuming no house price appreciation or depreciation. This should be a central part of the FHA's definition of "qualified mortgage" and will demonstrate the FHA's "reasonable and good faith determination" that "the consumer has a reasonable ability to repay the loan." n70 To allow loans where the rate exceeds 10 percent would be a disservice to low- and moderate-income communities since these loans would be concentrated in working class neighborhoods.
Step 2: FHA will target an average 5 percent projected claim termination rate, assuming no house price appreciation or depreciation. This is well over five times the historic default level for prime loans and two times the historic default rate for 90 percent LTV ratio loans with private mortgage insurance. In addition to the underwriting changes noted in Step 5, the FHA should limit seller concessions to 3 percent of a home's value.
Step 3: Stop guaranteeing lower-risk loans and high-dollar-balance borrowers, as this allows for cross-subsidization of those loans with excessive risk. This will also let the FHA step back from markets that can be served by the private sector and allow it to concentrate on home buyers who truly need help. Start by limiting loans to 100 percent of county median house price. Additionally limit first-time home buyers to less than 100 percent of area median income and repeat buyers to less than 80 percent of area median income. Returning the FHA to its mission as a targeted provider of mortgage credit for low- and moderate-income Americans and first-time home buyers has obvious benefits. It will return higher-income borrowers to the conventional lenders and private insurers market, resulting in a substantial reduction in foreclosure rate for higher income borrowers.
Step 4: Price for risk since not doing so deprives the borrower of the price information needed to understand the true risk of the loan. Until this is done, the FHA should disclose to the borrower his or her expected claim rate, assuming no house price appreciation or depreciation.
Step 5: Implement underwriting that results in the extension of responsible mortgage credit. Accomplish by balancing down payment, loan term, FICO score, and DTI ratio in a manner to achieve meaningful equity. Table 5 sets forth an approach for accomplishing this goal.
The goal of the above underwriting policies is to reduce risk layering, particularly for borrowers with FICO scores below 660, by providing a trade-off between a lower down payment and a 30-year loan term. Home buyers would be offered one, not both. In each case, the earned equity from combination of down payment and scheduled loan amortization during the first four years is the same.
This approach places a greater reliance on reduced loan term than reduced debt ratio, since the reduced term has a greater positive impact. For the 620-659 FICO borrower, the additional payment (forced savings) on a 20-year term loan would account for about 6 percent of the debt ratio and would likely reduce the foreclosure rate by about 40 percent. n71 As a result earned equity from scheduled loan amortization doubles from from 8 percent (30-year term) to 16 percent (20-year term) over the first four years.
Contrast this with the same 620-659 FICO borrower where a 6 percent reduction in debt ratio would reduce the foreclosure rate by only about 10 percent and aarned equity from scheduled loan amortization would be unchanged at 8 percent (30-year term) over the first four years.
The Positive Policy Effects of More Responsible Underwriting
Responsible underwriting for the 580 to 675 FICO score band has many positive policy impacts.
It reinforces the FHA's traditional mission as a targeted provider of responsible mortgage credit for low- and moderate-income Americans and first-time home buyers. The 580-675 FICO score band contains about 23 percent of all individuals with a scoreable credit record, but as table 6 demonstrates, it is heavily weighted with low- and moderate-income Americans, minorities, and first-time home buyers. (These weights are underlined in table 6.)
Because of lender FICO score overlays n72 and to bolster its deteriorating financial condition, the FHA reduced the percentage of its insured loans with a FICO score below 660. Increased loan limits and pricing advantages compared to a private mortgage insurance execution through the GSEs allowed the FHA to expand into the prime category (primarily in the FICO score band of 675-740). As a result, the FHA's average FICO score for FY 2012 (through August) was 695, up from 660 in 2007.
Responsible underwriting policies designed around the 580-675 FICO score band would have positive policy effects:
* 23 percent of all individuals with a scoreable credit record are in this band. n73
* 38 percent of blacks have a FICO score in this band. n74
* 30 percent of Hispanics have a FICO score in this band.
* 32 percent of the residents of low- and moderate-income zips (less than 80 percent of median) have a FICO score in this band.
* 30 percent of those 39 years of age or younger have a FICO score in this band.
The FHA's minority purchase loan volume was 32 percent of the FHA's total purchase volume in FY 2012 (through August), the same as for FY 2011.
FHA's estimated FY 2012 purchase loan volume (based on FY 2011 results) is as follows:
* Black: 9.5 percent or 69,000 out of a total purchase loan volume of 727,000.
* Hispanic: 16.1 percent or 117,000 out of a total purchase loan volume of 727,000.
FY 2007 purchase loan volume:
* Black: 14.1 percent or 40,113 out of total purchase loan volume of 283,639.
* Hispanic: 12.7 percent or 36,089 out of total purchase loan volume of 283,634.
The FHA's home purchase loan market share has been 30-35 percent since 2009. n75 A 2011 white paper by the Treasury and HUD recommended:
The FHA should return to its pre-crisis role as a targeted provider of mortgage credit access to low- and moderate-income Americans and first-time home buyers. (Today's FHA market share is nearly 30 percent, compared to its historic role of 10-15 percent.) As Fannie Mae and
Assuming the FHA's purchase market loan share returned to, for example, 15 percent by 2015 and overall purchase loan volume was 25 percent higher than in 2012. The FHA's total purchase volume would be about 450,000 loans:
* Estimated purchase loan volume for blacks would be at least 14 percent (the percentage in FY 2007), or 63,000 loans.
* Estimated purchase loan volume for Hispanics would be at least 13 percent (the percentage in FY 2007), or 59,000 loans.
n1
n2 Tabulations based on an analysis of the FHA's FY 2009 and 2010 books, which data were provided upon request by
n3 The median FICO score is about 720. Twenty-one percent of individuals have a score below 620. The remaining 79 percent have a score between 620 and 800+, the effective score range for prospective home ownership today since the FHA insures few loans where the borrower has a score below 620. The 620 to 659 score group is served almost entirely by the FHA and accounts for about 15 percent of prospective homeowners with scores above 620. Individuals with a score between 620 and 659 have a much higher risk of default than borrowers with higher scores. Source for FICO score distribution information:
n4
n5
n6 The combination of down payment and scheduled amortization.
n7 Herzog and Earley, Home Mortgage Delinquency and Foreclosure. Housing debt is the total of principal, interest, taxes, and insurance (PITI). Total housing expense (a concept no longer in use) includes PITI plus utilities, maintenance, and repair costs. It generally added about 5 percent to the housing debt ratio.
n8 Ibid.
n9
n10 In 1957, the
n11
n12 "Credit: Beware of the Walkaways," Time,
n13 Herzog and Earley, Home Mortgage Delinquency and Foreclosure, 67-68, 50.
n14 Since the FHA's and MGIC's books of business are reported on a fiscal- and calendar-year basis respectively, FHA delinquency data was lagged by one quarter so that cumulative seasoning is the same.
n15
n16 Ibid. Also see
n17
n18 Squires, Organizing Access to Capital.
n19
n20 The FHA uses the term "cumulative claim rate" (CCR) to describe the cumulative rate of foreclosures, short sales, deeds-in-lieu of foreclosure, or other actions resulting in the payment of a claim on its insurance over the life of a given book of insured loans. In this paper, the terms "projected foreclosure rate" or "foreclosure rate" will be used.
n21 Projected cumulative claim rates (CCRs) for the FY 2009 and 2010 book years were estimated at 11.45 and 7.80 percent, respectively, resulting in a weighted average of 9.6 percent. The CCRs for fixed-rate 30-year mortgages were used so as to exclude streamline refinances (FHA-to-FHA refinances) pertaining to loans made in earlier years. See Appendix G-7 in
n22
n23 HUD confirmed in a private email that 39 percent of its loans in FY 2012 had a FICO score of less than 660, a debt ratio greater than 50 percent, or both.
n24 US Department of the Treasury and US Department of
n25 Ibid.
n26
n27 See footnote 3.
n28
n29 http://www,philadelphiafed,org/communitv-development/events/2011/impact-of-workout-options-on-credit-reports-and-scores/frbp-fico-impacts.pdf
n30 Ibid.
n31 Federal Reserve, Report to the
n32 Ibid.
n33 The upfront premium is the same for both loans (1.75 percent). The annual premium is 1.35 percent and 1.30 percent for loans A and G, respectively. On a present-value basis, this is an inconsequential difference.
n34 Federal Reserve, Report to the
n35 Ibid.
n36
n37 A Fannie loan with private mortgage insurance incorporates a substantial degree of risk-based pricing. However, it is widely recognized that it is still below market-based pricing because of inadequate capital levels and return thereon.
n38 The FHA, the
n39 Freddie loan-level detail file from Inside Mortgage Finance. There are minimal distribution differences between
n40
n41
n42
n43
n44
n45 Bocian et al., "Collateral Damage."
n46 Ibid.
n47 Federal Reserve, Report to the
n48
n49 Ibid.
n50 The study used a database consisting of loan performance data for 2.4 million FHA loans from the FY 2009 and 2010 books, estimated to represent 70 percent of the 3.45 million loans insured by the FHA for these two book years. To be conservative, the CCR for fixed-rate 30-year mortgages was used to calculate the estimated number of foreclosures expected to result from these 3.45 million loans. FY 2009 and FY 2010 have CCRs of 11.45 percent and 7.80 percent respectively. The all mortgages rates are higher (12.98 and 8.75 percent, respectively), primarily because of the impact of streamlined refinances. CCRs are from
n51 FY 2009 and FY 2010 had CCRs of 11.45 percent and 7.80 percent, respectively. A FY 2009 and FY 2010 serious delinquency rate was calculated for each zip. Each of these rates was multiplied by the applicable CCR for FY 2009 or 2010. This yielded a zip-level CCR for each fiscal year. A weighted average CCR was then calculated for each zip.
n52 Metropolitan areas are officially called Core Based Statistical Areas (CBSAs). Portions of the CBSAs of
n53 All study results are based on the performance of loans from the FHA's FY 2009 and FY 2010 books. While the books were analyzed separately, the results are reported for the combined books on a weighted average basis. National results are based on all relevant zip codes with at least one FHA loan. Metro area map results are based on all relevant zip codes with at least one FHA loan. Metro area scatter plot results are based on all relevant zip codes with at least 50 FHA loans.
n54 Consisting of
n55 Government loans are those insured by the FHA,
n56 This analysis focuses solely on the FHA's FY 2009 and 2010 books using a database consisting of loan performance data for 2.4 million FHA loans from these years. This is estimated to represent 70 percent of the 3.45 million loans insured by the FHA for these two book years. To be conservative, the CCR for fixed-rate 30-year mortgages was used to calculate the estimated number of foreclosures expected to result from these 3.45 million loans. FY 2009 and FY 2010 have CCRs of 11.45 percent and 7.80 percent, respectively (weighted average of 9.6 percent). The all mortgages rates are higher (12.98 and 8.75 percent, respectively), primarily because of the impact of streamlined refinances. In terms of national impact, the MBA's delinquency data indicates that the FHA's 2003-10 books are projected to result in over 1.7 million foreclosure starts, with the preponderance almost certainly in zip codes with a median income below the metropolitan area median.
n57
n58 This is considered to be a low dollar estimate, as the FHA's foreclosures are concentrated geographically, and the Harding et al. research indicates that multiple foreclosures in the same vicinity magnify the price decline impact in surrounding homes.
n59 Originators have attempted to protect themselves from excessive liability related to loan repurchase requests by imposing higher credit requirements (credit overlays) than required by the FHA. The FHA is considering a complaint charging that such credit overlays are illegal.
n60 Where one or more participating parties have a tendency to take risk because some other party is bearing the cost of that risk.
n61
n62
n63 Ibid.
n64
n65 Purchase loans totaled 2.1 million.
n66 "What Is an
n67 For purposes of this study, low and moderate income was defined at the zip code level. Within a metropolitan area: those zip codes with a median family income below the median family income for area. For non-metro areas: those zip codes below the statewide non-metro median family income.
n68 Principal forgiveness would amount to resetting borrower equity from negative to zero or near zero. This equity reset (from a severe negative equity position to one that is merely bad) would be done mostly for a population of borrowers whose credit quality (FICO score) says the odds of failure are even greater now than when the loan was made. Thus, principal forgiveness perpetuates and extends the cycle of neighborhood decline. Also, any principal forgiveness program conditioned on delinquency will unquestionably lead to a skyrocketing delinquency rate as borrowers stand to improve their net worth by tens or hundreds of thousands of dollars if they stop making mortgage payments. The alternative is an unconditional principal forgiveness program that essentially reduces principal balances for all underwater FHA loans. The magnitude of negative equity in the FHA portfolio and the associated cost of forgiving principal are enormous.
n69
n70 Dodd-Frank Wall Street Reform and Consumer Protection Act, H.R. 4173, 111th Cong. (
n71 This estimate is based on the FHA's actuarial estimates for the period 1997-2010. The FHA estimates its 15-year loan term claim termination rate is 29 percent of that for 30-year term loans (see various FHA Actuarial Studies). This rate conservatively estimated at 60 percent for a 20-year rate. This results in a 20-year loan having a 40 percent lower foreclosure rate than a 30-year loan.
n72 A lender FICO score overlay is when the originating lender has a higher minimum FICO score for borrowers than the FHA's minimum. This was largely a defensive measure to protect customers from extremely high expected foreclosure levels if they have low FICO scores and other risks, and to reduce the potential for indemnification requests.
n73 Another 33 percent have a FICO score below 580.
n74 Another 17 percent have a FICO score below 580.
n75
n76 US Department of the Treasury and US Department of
Read this original document at: http://financialservices.house.gov/UploadedFiles/HHRG-113-BA00-WState-EPinto-20130206.pdf
Copyright: | (c) 2010 Federal Information & News Dispatch, Inc. |
Wordcount: | 15996 |
Advisor News
Annuity News
Health/Employee Benefits News
Life Insurance News