Thre's not much left in FHA's Mutual Mortgage Insurance Fund to cushion against future claims.
Mortgage Banking
June 2012
By Robert Stowe England
The Federal Housing Administration (FHA), after coming to the rescue of a busted housing mortgage market after 2007 with very-low-down-payment mortgages, now finds itself on thin ice with excess capital reserves near the vanishing point.
Rising delinquencies in single-family mortgages that FHA insures are depleting the reserves for its Mutual Mortgage Insurance (MMI) Fund. Some now fear that continued high delinquencies could deplete not just the $2.6 billion excess capital reserves but also all $33.6 billion in available capital resources. Should that happen, it would require a taxpayer bailout for the first time in the agency’s 78-year history.
Last November, the Department of Housing and Urban Development (HUD) reported that the much-watched capital reserve ratio for FHA’s Mutual Mortgage Insurance Fund fell to a tiny 0.24 percent--far below the 2 percent buffer Congress mandates.
It marked the third year the ratio has been below 2 percent, and the latest number is half again the level of the prior year of 2010, when the capital reserve ratio was calculated to be 0.50 percent.
The capital ratio sank below 2 percent in 2009 to 0.53 from 3 percent in 2008. In 2007, it was a relatively robust 6.3 percent.
In spite of the vanishing capital reserves, HUD has continued to be upbeat about the long-term outlook for the MMI Fund, based on projections for gains from new business over the next eight years. The projections were made in the annual independent actuarial review conducted by Integrated Financial Engineering Inc. (IFE Group), Rockville, Maryland, and based on economic scenarios provided by Moody’s Analytics, West Chester, Pennsylvania.
The actuarial review also found that FHA is on a path to restoring its excess capital fund to the minimum 2 percent by 2014, according to Raphael Bostic, assistant secretary for policy development and research at the HUD.
“The [actuarial] report demonstrates the long-term strength of the fund, while not shying away from the challenges it faces in the near term due to the ongoing stresses in the housing market,” he says.
Hiking premiums
Even so, Acting FHA Commissioner Carol Galante announced on Feb. 27 that FHA would increase its annual mortgage insurance (MI) premium by 0.10 percent for loans of less than $625,5000 and 0.35 percent for loans greater than that amount. Upfront premiums were also increased by 0.75 percent.
Congress authorized the higher annual premiums last year. FHA increased the upfront premium under statutory authority that gives it the option to change the upfront premium without congressional approval.
The premium increases appear to indicate that both Congress and the FHA are worried about the MMI Fund’s financial health and want to shore things up sooner rather than later.
Under last year’s actuarial review by IFE Group--which does not include the new increases in premiums--the MMI Fund was expected by 2014 to regain its 2 percent excess capital level under the baseline scenario of a continued recovery and no second recession.
There were also other more pessimistic scenarios in the review. Under one alternative scenario, the United States experiences a second mild recession scenario. In this case, the capital reserve ratio would reach 2 percent in 2016. Under a deeper second recession, the 2 percent goal would be reached in 2017; and under a protracted slump, the goal would be reached in 2018.
No doubt part of the reason for premium hikes comes from the fact that last year’s projections, which were completed in July 2011 but not released until November, assumed that house prices would rise 1.2 percent in 2012 after a 5.6 percent decline in 2011. So far this year, however, house prices have continued to decline.
Skeptics
Not surprisingly, there are skeptics about the official positive outlook for the MMI Fund. Chief among them is Joseph Gyourko, professor of real estate, finance and business and public policy at the Wharton School of the University of Pennsylvania, Philadelphia.
“FHA’s present state is precarious,” Gyourko charged in a paper titled (ital) Is FHA the Next Housing Bailout? (end) and published by the American Enterprise Institute (AEI), Washington, D.C., in November, ahead of the FHA’s annual report on its insurance fund.
Based on his analysis of the independent actuary’s 2010 review of the MMI Fund, Gyourko contends that the agency would have needed a $12 billion capital infusion in 2010 to be solvent.
The capital hole is bigger now, claims Gyourko, because FHA has been underestimating its future losses and overestimating the value of new business. To bring the insurance fund up to the 2 percent mandated capital cushion would require between $50 billion and $100 billion--“even if housing markets do not deteriorate unexpectedly,” Gyourko wrote.
“If FHA was a state-regulated private insurance company, it would be shut down by now,” says Mark A. Calabria, director of financial regulation studies at the Cato Institute, Washington, D.C.
Things are worse than they appear from the actuarial report, he says, because “FHA has the ability to move and hide losses” when it conducts its annual reviews. Calabria expects FHA will have to seek taxpayer funds in the next five to six years.
“Even though the Federal Housing Administration was representing only about 2 percent of mortgages at the height of the housing bubble, it now has over a fourth of all mortgages with negative equity,” Calabria explains.
A rising share of underwater mortgages
Because the agency rapidly expanded its market share after the financial crisis erupted in 2008 and offered mortgages with down payments as low as 3.5 percent, the number of underwater FHA loans has been rising steadily because house prices have continued to decline. The higher share of underwater mortgages is contributing to higher delinquencies, Calabria explains.
A Jan. 4 white paper from the Federal Reserve, titled (ital) The U.S. Housing Market: Current Conditions and Policy Considerations, (end) estimated that as of September 2011, 3 million of the 12 million homes with underwater mortgages are FHA-insured. According to the Fed, those estimates are based on data from Santa Ana, California-based CoreLogic Inc. and Irvine, California-based LPS Applied Analytics Inc.
The Fed also points out that the estimate does not include second liens and, thus, underestimates the number of underwater mortgaged properties. According to CoreLogic, in the fourth quarter of 2011 4.4 million households with mortgages also had a second lien and were, on average, $84,000 underwater with mortgages at a 138 percent loan-to-value (LTV) ratio.
The Fed recently provided Mortgage Banking an updated estimate that 3.56 million in FHA-insured mortgages were underwater as of December 2011, while the overall number of homes under water rose to 12. 48 million. The FHA share of underwater mortgages, thus, rose from 25 percent to 28.5 percent. Underwater mortgages, in turn, represented 24 percent of the 51.96 million homes with mortgages, according to the Fed.
Out of 48.7 million mortgages tracked by CoreLogic, 11.1 million had negative equity in the fourth quarter of 2011, while another 2.5 million had “near” negative equity, meaning they are within 5 percent of being in negative equity.
“Of course this was all very predictable,” says Calabria, who did, in fact, predict FHA would end up with lots of underwater loans. “If you decide, as did our federal government, to get lots of borrowers into loans with very little equity at a time when prices are falling, you will create a whole lot of loans with negative equity,” he adds.
New scrutiny of monthly delinquency data
Worries about FHA’s financial condition have drawn new attention to the agency’s monthly reports on its delinquencies. The loans that were delinquent by 90 days or more stood at 5.79 percent in June 2011, the delinquency levels on which last year’s actuarial report is based. The number rose steadily until December 2012, when it hit 5.72 percent and then fell to 4.57 percent by March 2012.
Given FHA’s high delinquency rates, if you used generally accepted accounting standards (GAAP) with the FHA’s insurance fund, “they would have no capital; indeed, they would have negative equity of $12.05 billion and a capital shortfall of $13 billion to $50 billion,” says Ed Pinto, resident fellow at the American Enterprise Institute and former chief credit officer at Fannie Mae.
These estimates are based on applying to the MMI Fund the private-sector reserving practices of Genworth Mortgage Insurance Corporation, a mortgage insurer based in Raleigh, North Carolina.
Pinto made his calculation of FHA’s capital position based on the fact there were 847,573 loans in March 2012 that were more than 60 days delinquent--representing 11.2 percent of FHA’s insured loan portfolio of 7,471,930.
Genworth reserves for losses using an assumption that 58 percent of known 60-plus-days-delinquent loans would ultimately lead to a claim.
FHA’s most recent claims loss rate for 2009 is 64 percent with an average gross paid claim of $127,821 and an average loss per claim of $83,384. That would represent, Pinto calculates, an overall reserve requirement of $42.66 billion if FHA were a private mortgage insurer.
Pinto contends that delinquencies are continuing to rise because the 2009 and 2010 books “are now becoming more seasoned and, as a result, are beginning to enter their peak years of high delinquency.”
The higher delinquencies during the last half of 2011 for FHA-insured loans occurred against a backdrop where delinquencies were generally declining for all types of mortgage loans. The seriously delinquent rate for all mortgages fell to 7.76 percent in the fourth quarter of 2011, according to the Mortgage Bankers Association’s (MBA’s) National Delinquency Survey (NDS). That was down from 8.6 percent a year earlier.
Higher credit scores helping newer books
One factor that is mitigating against losses is the significant increase in credit scores for borrowers with FHA-insured mortgages in recent years, according to Mike Fratantoni, vice president of research and economics at the Mortgage Bankers Association in Washington, D.C. “The average FICO® [score] is 700, up from 660 a few years ago,” he notes.
Fratantoni points out that if one looks at the vintage of seriously delinquent loans in the entire mortgage market, one sees the problem loans are overwhelmingly from the years 2005 to 2007. In the fourth quarter of 2011, for example, 65 percent of the seriously delinquent loans were originated between 2005 and 2007, 19 percent were originated prior to 2005, while 14 percent were originated in 2008 and 2009, and 2 percent originated in 2010 or later, according to MBA’s fourth-quarter 2011 NDS.
One sees a different distribution of serious delinquencies for FHA-insured loans, which stood at 9 percent in the fourth quarter, according to Fratantoni.
Because FHA originated far fewer loans in the 2005-2007 period, those loans represent only 24 percent of FHA’s overall serious delinquencies, according to MBA. The agency’s biggest share of delinquencies--53 percent of the total in the fourth quarter of 2011--comes from loans originated in 2008 and 2009, when FHA was rapidly expanding its role in the marketplace. Another 6 percent comes from loans originated in 2010 and later, while 17 percent were originated prior to 2005, according to MBA data.
By the latter part of 2008 and up through today, FHA, the Department of Veterans Affairs (VA) and the U.S. Department of Agriculture (USDA) together account for one-third of residential originations--and about half for purchase mortgages (as opposed to refinancings), particularly owner-occupied properties.
“That’s pretty much a dramatic increase,” says Fratantoni. The one-third share is somewhat higher than FHA’s historic average of 20 percent, he adds.
“The 2008 cohort still hasn’t peaked [in terms of delinquency],” Fratantoni says. “The typical rule of thumb is that mortgages will hit peak delinquency three or four years after they are made.”
FHA could, however, see better ultimate loan performance results from mortgages originated after 2008. “The 2009 cohort is performing better than the 2008, at the same point from the time of origination,” Fratantoni says. “And the 2010 is performing better than 2009 and 2011 is a little better than the 2010 book.”
After the 2008 delinquencies peak, “we should see some improvement in FHA performance, given the better credit quality of the later books,” Fratantoni says.
Fratantoni acknowledges that “FHA-insured loans made to first-time homebuyers are at heart risker loans.” The key question, he adds, is this: “Did FHA charge enough for the risk?”
Fratantoni is less worried about underwater borrowers than other observers, who see the rising share of underwater mortgages as a harbinger of higher delinquency rates.
“That really comes into play once a borrower falls behind in their mortgage. If they lost a job, then they are really forced to take a look at their equity position in the property. If they have positive equity, then they can sell. If they are forced to bring cash to the table, that’s when you see more delinquencies and defaults,” he explains.
Growth via high LTV
For some, however, the growing share of underwater mortgages is viewed with concern. In his paper, Gyourko sees significant risks in FHA’s growth since 2008, driven by insuring mostly very-high-LTV mortgages.
In 2007, FHA’s market share of the overall mortgage market was 4 percent, according to the actuarial report. Just over 60 percent of the insured loans had LTVs of greater than 95 percent in 2007.
In fiscal year 2010, FHA’s mortgage origination market share was 19 percent and close to 69 percent of its insured mortgages had LTVs higher than 95 percent, according to the actuarial review. For 2011, with partial data for nine months of the fiscal year, FHA’s market share of new mortgages at the end of June was lower at 15 percent. Yet, 67 percent of the guaranteed mortgages had LTVs above 95 percent.
The most troubling trend, however, is FHA’s vast expansion of its 3.5 percent down-payment program beginning in 2008, creating year after year of ever-larger exposures to mortgages on properties leveraged at ratios of 30 to 1, according to Gyourko. With leverage so high, the FHA becomes more vulnerable to small declines in home prices.
Using data from the actuarial report on LTVs and the price trends reflected in the Federal Housing Finance Agency (FHFA) House Price Index, Gyourko plots the change in net equity in homes financed with a 3 percent or 3.5 percent down payment mortgage, which is the most popular FHA program. (The minimum was raised to 3.5 percent in October 2008.) He finds that loans with 3 percent down payments that were originated in 2007 would have seen a 19.5 percent decline in their net equity by 2011. This would push the home value significantly below the value of the mortgage.
Homes financed in 2008 with 3 percent and 3.5 percent down would lose 13.4 percent in net equity, while homes with mortgages from 2009 would be down 7.5 percent. Even those loans insured in 2010 would have equity down 5.8 percent. Thus, all 3 percent and 3.5 percent down FHA-insured mortgages from 2007 to 2010 would be underwater, according to Gyourko’s analysis of FHFA’s UPI data.
The actuarial study also divides the loans guaranteed by FHA into LTV categories, based on their LTV at origination. Using that data, Gyourko estimates that “over half of [the] existing FHA-insured portfolio suffers from negative equity.” That would put the number of underwater loans at 3.7 million as of March 2011, very close to the latest Fed estimate of 3.56 million.
The number of underwater loans is high because FHA lending grew so rapidly in 2008, 2009 and 2010 as home prices were declining, according to Gyourko. Those three years represent 76.5 percent of all outstanding FHA insurance exposure at the end of fiscal year 2010, he writes.
Gyourko argues the FHA should have been increasing its capital cushion at the same time it was rapidly expanding its guarantee business. But that’s not what happened. In 2005, FHA had $14.41 of insurance-in-force for every dollar of capital resources, Gyourko notes in his paper. By 2010, this number had doubled to $28.83.
Unobserved high credit risk
In its examination of IFE Group’s actuarial reviews, Gyourko looks at whether or not the review dealt thoroughly with unobserved default risk; that is, risk that is not reflected in the variables contained in the computer model developed by the reviewer.
Gyourko reports that IFE Group has since 2006 added a variable to cover unobserved higher credit risk of more recent loan pools, made at a time of declining home prices. This variable assumes the loans will default at about double the rate of loans of similar borrowers before the housing crisis.
Gyourko commends IFE Group for including such a variable, although he finds it somewhat wanting, in that it does not separately measure loan, borrower or market traits that might account for higher defaults.
The Wharton professor, however, takes issue with FHA’s decision to gradually assume away the unobserved credit risk variable in 2012 and 2013, and does away with it entirely in 2014.
The consequences of getting it wrong by taking away the unobserved credit risk variable are huge--losses could be double those projected beginning in 2014, according to Gyourko. So, it is important to try to gauge the extent to which the unobserved high credit risk in mortgage pools since 2006 is truly “a thing of the past,” Gyourko notes.
He insists unobserved credit risks will continue for the MMI Fund because those risks include the potential impact of continuing high unemployment. This is especially important because FHA does not separately track unemployment for its borrowers and, thus, cannot provide that information for any analysis of how it affects loan performance.
Given that unemployment is expected to remain high, it leads to “the inescapable conclusion . . . that IFE’s forecasts of future losses are biased downward,” Gyourko writes. He estimates that a proper measurement that included unemployment would increase projected losses by $30 billion to $50 billion or more--a figure in line with Pinto’s projections.
Homebuyer tax credits
One additional area where Gyourko faults the HUD and IFE Group is for potentially underestimating the future risk of FHA-insured loans where the borrower may have obtained a tax credit to cover the down payment.
Borrowers who are unable to save enough for a down payment are higher credit risks and FHA recognizes three different levels of risk in the model used by IFE Group, depending on the source of the funds, Gyourko explains. If the funds are from a relative of the borrower, the probability of default is 35 percent above the baseline, according to FHA. If the funds are from an unrelated third party--so-called seller-assisted down payments--then the risks are 72 percent higher. Finally, if the funds are from the government, the risks are 44 percent higher.
Losses to FHA from seller-assisted financing have been three times higher than losses for loans without such assistance, leading Congress in 2008 to ban FHA from insuring such loans. The termination of those programs is one reason FHA is more optimistic about insurance claims against loans issued since 2009. The 2010 actuarial review, for example, identified only 0.12 percent of loans as those with seller-assisted down payments.
Gyourko, however, finds that FHA is missing the fact the $8,000 first-time homebuyer credit constitutes a third-party risk and that it could be quite large. The tax credit was adopted by Congress in 2009 and expired in the spring of 2010. The credit was not provided at closing on a mortgage but was later reimbursed to the taxpayer. FHA allowed state and local agencies to provide bridge loans for the tax credit, according to Gyourko.
According to the IRS, there were 2.6 million first-time homebuyer credits issued between January 2009 and October 2010. FHA insured loans on 1.66 million first-time homebuyers in the same two years.
“Theoretically, all of those [FHA-insured loans] could be tax credit-financed deals, and it seems likely that FHA got a disproportionately high share of the total,” Gyourko contends. He estimates FHA’s share to be 832,000 to 1.25 million loans.
FHA’s response
HUD’s Bostic publicly criticized Gyourko and some of his claims in a HUD blog posted Nov. 18, 2011.
Bostic charged that Gyourko failed to take into consideration improved borrower quality, noting that while half of FHA borrowers in 2007 had credit scores below 620, only 3 percent in 2010 and 2011 had scores that low.
While the 2009, 2010 and 2011 books of business have become 75 percent of FHA’s portfolio, Bostic notes, those loans are performing substantially better than earlier books of business. Seriously delinquent loans after 12 months in the 2010 book are just over 1 percent, he writes. By comparison, it was 6 percent in 2007.
Bostic took particular issue with claims the homebuyer tax credit adds risk to FHA’s portfolio, labeling Gyourko’s claims “completely false and irresponsible.” Bostic cites FHA data that show that only 4 percent of first-time homebuyers used a taxpayer credit during 2009 and 2010.
Bostic disagrees that borrowers who used taxpayer credits would perform similarly to those who relied on seller-financed down payments. “The tax credit helps makes these purchases more affordable. They don’t necessarily change the terms of the mortgages themselves or how they are underwritten,” he explains.
“We still require the same amount of down payment, we still require the credit standards. And the benefit to the borrower happens on the back side after the fact,” he contends. “So, I’m not sure we would expect there to be very large differences in performance, given how the borrowers have to get into the mortgage to begin with.”
Bostic also takes exception to Gyourko’s claims that FHA was systematically underestimating future defaults.
Most of the risk is near term
In an interview with Mortgage Banking Bostic elaborated on his objections to the Gyourko report.
“The actuarial report is a very specific exercise, which is designed to assess the health of the fund but also assign some probabilities and run through some scenarios to give perspective on what could happen,” he says.
“I think part of our disagreement is really about the probability that some of these scenarios are more or less likely,” Bostic says.
“Some of the more vocal critics are really focusing and anchoring outcomes that we think are less likely, and so we have a fundamental difference of opinion about the likelihood for the need for an additional influx of funds,” he says.
Bostic underscores the strengths of FHA’s new books of business. “Those [newer books of business] are going to be much safer, much more secure and will add to the stability of the FHA in ways that I don’t think were fully emphasized by some of the critics,” he says.
While the 2011 book adds $9.3 billion, as noted earlier, the 2013 book is expected to raise excess reserves to $15.6 billion, according to the actuarial report.
Most of the risk for losses is in the short term, for 2012 and 2013, Bostic notes. “I haven’t seen anyone suggest that what we’re doing doesn’t make sense from a long-run perspective,” he says.
Noting that even under an extended slump, the actuarial review anticipates FHA will return to 2 percent capital ratio, Bostic notes, “Everyone would have to acknowledge we have taken steps in all scenarios to get us to where we need to go.”
He adds, “The real issue is how the broader housing market functions. That’ll affect how quickly we get there. I don’t think there’s any question in any of these whether we’ll get there. It’s just when.”
Is the potential for higher unemployment incorporated in the alternative scenarios and, thus, addressed by the actuarial review?
“The strength of the economy and creating jobs is going to be a key element in how quickly we would get to the 2 percent. And so it will be important for us to continue to work hard to get the economy back on its feet and [growing] at a more robust pace than we are now,” Bostic says. “So, that’s important.”
“You’ll note in the actuarial report they talk about the scenarios in terms of recessions or broader macroeconomic performance. I think that’s an acknowledgement that unemployment, job creation and job growth are central to the performance of the fund,” Bostic says. “Look. If people don’t have income they can’t buy houses. And they’re not going to be able to continue to pay their mortgage. So I think you to pay attention to that.”
Who’s right?
Gyourko penned a response to HUD’s criticism of his paper on FHA that was published Nov. 21 on AEI’s website. He reaffirmed all the arguments he made in his paper and challenged the claim that he ignored the fact that capital resources are expected to grow from new books of business. “This is true, but it is only half the story,” Gyourko contends.
“Any conclusion about how risk changed necessarily involves a comparison of how FHA’s potential liabilities grew relative to its capital.” In fiscal year 2011, while its capital was increasing $400 million, FHA issued $213 billion in new guarantees. “Thus, it grew its potential liabilities by well over $500 for every dollar increase in its total liquid assets,” Gyourko states.
Gyourko cautioned that one should always look at both sides of the balance sheet--assets and liabilities--and not a single component when assessing the financial condition of a program. In fiscal year 2011, FHA had $1 trillion in outstanding insurance-in-force with total capital resources of $28.2 billion. That’s a leverage ratio of 35.8 to 1. By comparison, FHA was leveraged 16.6 to 1 in 2005, Gyourko notes.
Gyourko also questions the calculation that the 2012 book of business will add $9 billion in value to the MMI Fund.
“That number comes from the same outside reviewer who has estimated that every future book of business since 2006 would generate positive value to FHA,” Gyourko writes in his response. “Given that forecasting track record and the fact that FHA’s excess reserves have dwindled from 6.5 percent to 0.12 percent of its single-family insurance portfolio over that time period, FHA would be well advised to let others ‘stress test’ those predictions.”
Finally, Gyourko points out that FHA’s accounting might camouflage the true condition of the MMI Fund. Because the 1 percent upfront premium is often financed, “FHA’s accounting permits it to book the fees on loans that its actuarial model expects to default,” Gyourko writes. He asks how much of the $2.13 billion booked from $213 billion in new business in 2011 is actually net cash.
Gyourko, when contacted at the Wharton School, declined to comment further beyond his published comments.
Some market analysts have their own questions about the adequacy of the insurance fund. “I think you don’t have to be a brain surgeon to figure out that the [FHA insurance] fund is most likely broke,” maintains Paul Miller, managing director, financial services, FRB Capital Markets, Arlington, Virginia.
“Their audit that is supposed to be independent had, I thought, some pretty rosy outlooks to keep the fund relatively solvent. I think the fund will have to either tap the Treasury or do some steps at some point in the next couple of years,” Miller says. “It’s very difficult for me to look at that thing and think that 24 basis points is enough to overcome the level of defaults in that portfolio.”
Does Miller expect FHA will be able to restore its excess reserves to 2 percent? “Oh, I think eventually they will. It’s a program that is not going to go away. They don’t go bankrupt,” he says. “But they’ve got to tighten up their underwriting standards. Part of the problem with the housing boom is the loose standards that the FHA brought to town.”
He adds, “Right now there are about $100 billion to $140 billion in loans in various points of delinquencies” at FHA. “If that number doesn’t grow, you can lose tens of billions of dollars on that alone,” he says.
“They do collect money. There are fees coming in. They collect 60 bits on $1 trillion. That’s $10 [billion] to $11 billion a year on all fees. It’s not like the thing doesn’t generate some income and cash flow. [Even so,] they’re going to lose tens of billions of dollars.”
Miller believes it is very difficult for HUD and FHA to fully recognize the losses that are building up. “It makes everyone’s life on Capitol Hill a lot easier to pretend that those things are money good,” he says. “If we have a report where there’s money bad [in the next annual review], you know it’s not going to come out until after the election. No one wants that embarrassment on their books.”
All of which means, it would seem, that the public remains somewhat skeptical of claims by government officials that everything is going to be OK.
MB
bio: Robert Stowe England is a freelance writer based in Arlington, Virginia, and author of Black Box Casino: How Wall Street’s Risky Shadow Banking Crashed Global Finance, published by Praeger and available at Amazon.com. He can be reached at
Copyright © 2012 Mortgage Banking Magazine
Reprinted with Permission
