Mortgage Banking
January 2008
Estimates keep rising for total losses expected from higher defaults and markdowns on mortgage assets and related derivatives held by financial institutions worldwide. The bottom line isn't pretty, and it's not even over yet.
By Robert Stowe England
Like a rock thrown into a tranquil pond, the subprime mortgage meltdown that kicked off in early 2007 and transformed into a much broader credit market event has spread ripples in ever-widening circles over the intervening year.
Yet, unlike the proverbial pebble tossed in a pond, the intensity of the effects seems to have grown as they spread wider and wider.
The repercussions have included mounting losses, growing provisioning against losses, big markdowns of mortgage-related securities, falling capital ratios and plummeting share prices.
Since last summer, fear and panic have repeatedly roiled global markets, striking blows at private-label mortgage-backed securities (MBS) and related collateralized debt obligations (CDOs) along with related derivatives, asset-backed commercial paper and, finally, structured investment vehicles (SIVs). SIVs are off-balance-sheet entities set up by banks and other financial institutions and which issue commercial paper to buy higher-yielding, long-term assets such as mortgage-backed securities.
During July and August, as the first wave of losses was reported, the U.S. private-label MBS market basically collapsed. This key component of the Wall Street bond markets, which began stumbling and falling in late July, imploded on Aug. 9.
To gauge the magnitude of the fall, one need only ponder that in 2006, Wall Street firms packaged $1.15 trillion in nonagency mortgages (including $449 billion in subprime) into pools and issued private-label residential mortgage-backed securities (RMBS) against the loans, according to Inside Mortgage Finance. The full extent of last summer's collapse can be seen in the data for non-agency RMBS issuances. Those issuances fell 20 percent in August to $37.4 billion from $46.5 billion in July, according to data provided by FBR Investment Management Inc. (FBRIM), Arlington, Virginia. While RMBS issuance leveled off in September at $38.9 billion, it fell again by 50 percent in October to $19.5 billion. To put that decline in perspective, October 2007 volume was an incredible 81 percent below the $104.1 billion level of October 2006.
At the same time, investors began to run from asset-backed commercial paper-an enormous market in its own right. That market peaked at $1.196 trillion on Aug. 8. Asset-backed commercial paper imploded in the subsequent weeks and months, shrinking 31 percent or $370 billion by Nov. 28, as investors declined to provide new rollover funding for this type of short-term paper, which matures in 270 days or less.
Next began the unraveling of SIVs, which increasingly made it impossible to issue commercial paper (even those with super senior non-subprime mortgage assets). As a result, banks had to extend credit lines in order to keep the SIVs from dumping their $350 billion in assets and driving down their value. Because if the SIVs were forced to dump their assets, it would set in motion a process that could ultimately move the off-balance-sheet assets in the SIVs onto the bank's own balance sheet.
Following the Federal Reserve's actions in August (a surprise 50-basispoints reduction in the discount rate) and September (both a 50-basis-points reduction in the Federal Funds Rate and a 50 percent reduction in the discount rate), increasingly frozen markets began to function better, and spreads of fixed-rate assets over risk-free Treasuries narrowed. By late November, however, conditions in the nation's housing markets had deteriorated further. This was reflected in the S&P/case-Shiller® Home Price Index, which showed a 4.9 percent decline in national home prices in October (year-over-year). That trend was confirmed by a National Association of Realtors® (NAR) report that said median home prices in October had fallen 5.1 percent over the same month a year earlier.
"We haven't faced a [housing] downturn like this since the Depression," Bill Gross, chief investment officer at Pacific Investment Management Company LLC (PIMCO), Newport Beach, California, told the Associated Press on Nov. 25. While Gross indicated he was not saying current conditions are as terrible as the bleak days of the 1930s, he expects the housing downturn will definitely have significant consequences for the overall economy.
"Its effect on consumption, its effect on future lending attitudes, could bring us close to the zero line in terms of economic growth," he said. "It does keep me up at night."
While the economy will slow, most economists continue to predict it will grow. "We have not yet seen fully the impact of the credit shock to the United States and world economies, and the severity of that impact will depend on how long it takes for the markets to return to normal functioning and where credit spreads ultimately settle," according to Doug Duncan, chief economist and senior vice president of research and business development at the Mortgage Bankers Association (MBA). "In terms of housing, we expect 2008 sales to be below 2007 levels until late 2008, but given the oversupply of homes in a number of markets, any significant increase in home building is probably years off," Duncan says.
"The drag on GDP [gross domestic product] growth from the housing sector is being at least partially offset by strength from international trade. Between the fourth quarter of 2006 through the third quarter of 2007, residential investment in constant dollars fell by $97 billion, while net exports rose by $53 billion. Strength from the external sector will surely continue, given robust growth abroad, a declining dollar and the impact of slower growth at home moderating the rise of imports," Duncan says.
Delinquencies and foreclosures rise
The housing sector will continue to remain vulnerable to rising delinquencies and foreclosures. In its latest National Delinquency Survey for the third quarter of 2007, MBA found that the total delinquency rate had risen to its highest level since 1986, while the rate of foreclosure starts and the percentage of loans in the process of foreclosure were at their highest levels ever.
"This is the first quarter which registers the full combined effects of the seizure of the nonconforming securitization market, broad-based home-price declines, continued weakness in some regional economies and rate adjustments on monthly payments," says Duncan. "The predictable results are increased delinquency and foreclosure."
MBA reported in early December that the delinquency rate for mortgage loans on one-to-four-unit residential properties had risen 47 basis to 5.59 percent of all loans outstanding in the third quarter of 2007 on a seasonally adjusted basis, and was up 92 basis points from the third quarter of 2006. The delinquency rate does not include loans in the process of foreclosure.
The percentage of loans in the foreclosure process was 1.69 percent of all loans outstanding at the end of the third quarter, an increase of 29 basis points from the second quarter of 2007 and 64 basis points from the second quarter of 2006.
The share of loans entering the foreclosure process was 0.78 percent on a seasonally adjusted basis-13 basis points higher than the previous quarter and up 32 basis points from one year ago.
The increase in foreclosure starts was due to increases for all loan types, but increases were greatest for adjustable-rate mortgage (ARM) loans, both prime and subprime. From the previous quarter, prime fixed-rate mortgage (FRM) loan foreclosure starts increased 4 basis points in the third quarter of 2007 to 0.22 percent. However, prime ARM foreclosure starts rose 40 basis points to 1.02 percent. Subprime fixed-rate foreclosure starts increased 3 basis points to 1.38 percent, while subprime ARM foreclosure starts were up 88 basis points to 4.72 percent.
Florida and California are the two largest states in terms of mortgages outstanding, and they are the key drivers of the increase in the national foreclosure rates. While California and Florida together have 36.4 percent of all the prime ARM loans in the country, they had 42.4 percent of the nation's foreclosure starts for prime ARMs. Similarly, California and Florida together have 28.1 percent of the subprime ARMs and 33.7 percent of foreclosure starts for subprime ARMs.
Florida, Ohio, Michigan and Indiana have 16.4 percent of the prime fixed-rate loans in the country, but 29.3 percent of the foreclosures started on prime fixed-rate loans, 18.9 percent of the subprime fixed-rate loans and 26.3 percent of the foreclosure starts.
While the third quarter's numbers show the highest level of foreclosure starts (on a seasonally adjusted basis) for prime-fixed rate mortgages in the last 10 years, that increase is largely due to increases in Florida, Ohio, Michigan and California. "In most states, the increase in prime fixed-rate foreclosure starts is due to borrowers who will fall behind on their payments for the traditional reasonssuch as loss of job, medical expenses, divorce or separation-but who cannot sell their homes due to market conditions," says Duncan.
Gloomier credit performance outlook
The autumn swoon in the capital markets was set off by reports of growing deterioration in the credit performance of the mortgages originated since the fourth quarter of 2005.
While expectations about how low the elevator of credit performance would go were already fairly glum, the outlook deteriorated even further in early October, as new delinquency and default data came out for the month of September.
During September, credit performance deteriorated sharply from the level in August for adjustable-rate subprime mortgages that back residential mortgage-backed securities, according to data evaluated by FBR Investment Management, including data from INTEX Solutions Inc., Needham, Massachusetts.
According to FBRIM's calculations, the default rate for 2005 subprime RMBS adjustable-rate mortgages jumped by 10.2 percent to 14.85 percent. That conclusion was based on weighted-average default-rate data from 363 metropolitan statistical areas (MSAs) from First American LoanPerformance, San Francisco, and from data and analysis from FBR Investment Management.
For 2006 subprime RMBS issues, the default rate jumped by 13.8 percent to 14.02 percent. And, for first-half of 2007 subprime RMBS issues, default rates soared 30.4 percent for 2007 to 5.27 percent, even though the loans were only a few months old.
The spike in defaults for 2007 vintage mortgages was especially disappointing to the markets, given that at the time those mortgages were underwritten, it was already widely known that the 2006 vintage had been poorly underwritten. Thus, lenders should have already begun to tighten their underwriting criteria. Instead, based on early evidence, the underwriting for the first half of 2007 is actually proving worse than in 2006.
The October default rates for 2005 exceeded the weighted average of all adjustable-rate subprime RMBS at comparable seasoning by 1.3 times, according to Michael Youngblood, managing director of ABS research at FBR Investment Management. For the 2006 vintage subprime ARMs, the October default rate was 1.9 times the weighted average of all subprime RMBS at comparable seasoning, while it was 2.7 times for the 2007 vintage ARMs. These numbers provide useful indicators of the increasing liberalization of underwriting standards that apparently lasted from late 2005 to mid-2007.
The credit performance of fixed-rate subprime mortgages packaged into RMBS also declined, but less dramatically. While 2005 fixed-rate subprime RMBS were in line with expectations, the default rate for the 2006 vintage rose 13.7 percent from August to 5.63 percent, which was 1.3 times the weighted-average default rates for all subprime fixed-rate RMBS at comparable ages.
"It is striking that the default rates of adjustable-rate subprime RMBS are both relatively and absolutely higher than the fixed-rate subprime RMBS," Youngblood wrote in the Oct. 19,2007, issue of FBRIM's newsletter, Structured Finance Insights.
As default rates soared for securitized adjustable-rate subprime mortgages, prepayment rates were falling, which FBR Investment Management found troubling. It implied subprime borrowers with ARMs were having difficulty either refinancing or selling their homes if they were encountering financial trouble.
During September, the one-month conditional prepayment rate for adjustable-rate subprime RMBS issued in 2005 fell 37.3 percent to 17.17 percent. For 2006, prepayment rates fell 4.3 percent to 17.17 percent, while 2007 prepayments actually rose 27.5 percent to 10.92 percent. The prepayment rates were low relative to the weighted-average prepayment rates for all adjustable-rate subprime RMBS.
FBR Investment Management reported in its Oct. 19 Structured Finance Insights newsletter that "we are more concerned" about the decline in prepayment rates in September than the increase in default rates, "because they strongly suggest that subprime borrowers are progressively less able to refinance."
FBRIM had expected that 74.7 percent of subprime borrowers who were never delinquent through July 2007 would have refinanced into conforming mortgage loans by September, especially since the 30-year conforming mortgage rate averaged 6.71 percent in August (the refinance window for September). This rate compares with a 7.85 percent weighted-average rate for subprime mortgage loans originated in 2005, 8.35 percent for 2006, and 8.36 percent for 2007, according to FBRIM.
The research firm identified falling house prices as a likely "countervailing force" slowing down the rate of prepayments and refinancings into fixed-rate mortgages.
There was also bad news on non-agency alternative-A loans. FBR Investment Management reported that the credit performance of securitized non-agency alt-A and subprime loans "deteriorated sharply" in August, while securitized jumbos deteriorated only marginally. This marked the 27th month of consecutive decline in the performance of alt-A and subprime loans.
The default rate for alt-A loans reached "the highest level FBRIM has ever recorded" in August when it rose to 3.96 percent from 3.34 percent in July and 1.04 percent in August 2006, according to FBRIM's Structured Finance Insights newsletter of Oct. 29. The default rates for securitized non-agency subprime mortgages rose to 16.14 percent from 14.65 percent in July and 7.74 percent in August 2006, according to FBRIM.
Subprime defaults were found to be concentrated in only 49 MSAs out of 363, representing only six of the 50 states plus the District of Columbia: Arizona (six MSAs), California (26 MSAs), Florida (20 MSAs), Nevada (three MSAs), Oregon (two MSAs), the District of Columbia (one MSA), Hawaii (one MSA) and Idaho (one MSA). FBR Investment Management identified house-price declines as a key factor beyond liberal underwriting and labor market issues in Arizona and California.
FBRIM identifies 43 of the 49 MSAs as markets with house-price bubbles, whereas it counts only 69 house-price bubbles in all the 363 MSAs combined.
FBRIM found that during the span of time that the Office of Federal Housing Enterprise Oversight (OFHEO) has released its House Price Indexes (HPIs) (first quarter of 1975 to the second quarter of 2007), once a house-price bubble has burst, house prices have declined an average of 18 quarters (four and a half years). FBRIM has found that only 13 of the 49 MSAs with the sharpest increases in default rates in August are experiencing the highest default rates.
FBR Investment Management identifies a separate list of 49 MSAs that have the highest default rates for non-agency RMBS. They are distributed in 23 states as follows: California (nine MSAs), Colorado (two MSAs), Florida (five MSAs), Indiana (four MSAs), Massachusetts (three MSAs), Michigan (11 MSAs), Minnesota (two MSAs), Ohio (nine MSAs), Illinois (one MSA), Mississippi (one MSA),
Tennessee (one MSA) and Wisconsin (one MSA).
"The elevated default rates of these MSAs reflect, with a few exceptions, either the chronically weak local economic conditions of the Rust Belt, the more recent economic weakness in New England [since the third quarter of 2005] or the malingering impact of Hurricanes Katrina and Rita on the Gulf Coast," stated FBRIM's Structured Finance Insights report on Oct. 29.
The rising default rates in August 2007 led FBR Investment Management to update its econometric models. In a new forecast released on Nov. 2, FBRIM predicted higher defaults for non-agency RMBS than previously assumed for subprime, but no change in predicted increases in alt-A and jumbo prime default rates.
For subprime non-agency RMBS, defaults will rise to 19 percent in August 2008 from 14.65 percent in August 2007, the forecast notes. This represents a 7 percent higher default rate than the 17.83 percent FBRIM had earlier forecast for July 2008. This overall higher subprime default rate is based on a forecasted increase in subprime default rates in 281 of the 363 MSAs, while FBRIM expects 80 MSAs will see lower subprime default rates and two MSAs will see default rates remaining constant.
An October AAA plunge
As the financial markets began to assume higher default rates, prices on the ABX subprime index began to fall again, after staging an earlier modest recovery from the summer collapse. Importantly, the AAA-rated tranches began to fall dramatically.
The ABX Index is a series of creditdefault swaps based on 20 bonds that consist of subprime mortgages. While it is widely followed as one indicator of the value of subprime MBS and related derivatives, it is viewed by some as imperfect. It increasingly has been faulted as an indicator of true market value-yet remains an indicator of trends.
FBR Investment Management's Youngblood is one of several mortgage market observers who question the value of the index. "We and others have long argued that these [ABX] index securities are not representative of the credit performance of the subprime market as a whole," says Youngblood.
"Furthermore, if anything, the securities chosen for the indexes [represent], in most cases, the worst quartile of performance," he adds. "Yields on the index are not representative of the yields on the great majority of cash securities."
Youngblood cites as an example the second-quarter 2007 BBB-minus-class ABX index, which was trading at a yield that was 6,667 basis points over one-month London interbank rate (LIBOR) in mid-November. Meanwhile, FBRIM calculates the yield of newly originated BBB-minus subprime mortgages at 1,500 basis points over one-month LIBOR.
"So, if my thesis is correct that the collateral of the ABX index is not representative and in fact represents an adverse or negative selection of outstanding securities, what we just observed is exactly what we would expect," says Youngblood. "We would expect a much higher yield on the ABX than on comparable obligations from the third, second or first quartile [of] performance and not the fourth quartile of performance," he says.
Further, Youngblood notes, "Credit default swaps on individual name obligations are trading on [the] BBB market at a spread of 1,230 to one-month LIBOR, which is consistent with the cash price and, like the cash price, wildly out of line with the ABX indexes."
Keeping in mind the questions raised about the value of the ABX indexes, the prices on the indexes nevertheless show a stark deterioration in market values during October, especially for AAA and AA tranches.
On Oct. 1, second-quarter 2007 ABX index for AAA tranches of mortgage-backed securities stood at 95.61 (par = 100), while AA was at 88.38. Single-A was at 62.78 and BBB was at 40.17. By Oct. 30, AAAs had fallen below the 80 barrier to 79.53, with AAs sinking below 50 to 47.97. Yet the deterioration was far from over.
The slide continued in early November, with AAAs falling below 70 on Nov. 9 to 69.93, while AAs sank to 40.21. AAAs reached a monthly low on Nov. 23 when they were priced at 66.41, with AAs down to 35.04. Single-As were at 24.20 and BBBs at 19.88. At month's end, AAAs were at 71.81, AAs at 39.22, single-As at 21.06 and BBBs at 20.56.
FBRIM does not expect defaults to peak until the fourth quarter of 2008, which means financial institutions can be expected to report higher losses in the coming year. "Rising default rates will keep loans as well as securitized loans under pressure," explains Youngblood.
"Falling house prices in many markets-in fact, we think house-price declines accelerated in the third quarter quite sharply-will translate these defaults into elevated realized losses, although the losses will take a year or longer to appear after the defaults," he says.
The delay in realized losses occurs because of the time it takes to foreclose and sell a foreclosed property. These foreclosures are likely to lead to significant losses in the value of the house and the mortgages, Youngblood adds.
"It is a bleak picture for mortgage lenders, in terms of defaults for the next year and in terms of losses for the next two years," he says. The pressure on asset prices, however, could abate before then. "Obviously, at some point, the equity markets will anticipate better credit performance in the future and begin to reprice accordingly," he says.
Putting a number on the damage
During November, research analysts began to publish gloomier assessments about the full extent of expected losses. On Nov. 15, for example, New York-based Goldman Sachs & Co. Chief Economist Jan Hatzius said that "a back-of-the-envelope calculation" of the likely credit losses suggests a number around $400 billion-a number four times the top of the $50-billion-to-$100-billion range estimated for subprime credit losses made by Federal Reserve Chairman Ben Bernanke back in July.
Hatzius says his calculation is based on historical default and loss patterns in different home-price environments. He starts with the roughly $2.5 trillion in subprime and $1 trillion in alt-A mortgages originated since 2005 (recognizing that some have already prepaid). Next, he assumes that home prices will fall 15 percent from peak to trough, assuming the economy stays out of recession. In such a home-price environment, historical patterns suggest subprime mortgages will have remaining lifetime loss rates after recoveries of 15 percent, while alt-A losses would be 5 percent.
"This would imply a total loss of $275 billion for the 2005 to 2007 subprime and alt-A originations alone," Hatzius says.
Estimated losses from earlier vintages of subprime and altA plus losses from jumbo prime and prime conventional loans raise the implied loss to $400 billion, he calculates.
Hatzius recognizes that "even a $400 billion loss does not look all that large compared to the vast size of the U.S. financial markets, and one sometimes hears that it is just equivalent to one bad day in the stock market."
However, that comparison does not capture the potential impact of the losses. "There is a big difference between stock market losses, which are mostly borne by long-only [long-term] investors, and mortgage credit losses, which are mostly borne by leveraged investors such as banks, broker-dealers, hedge funds and government-sponsored enterprises [GSEs]," he says. "The key difference ... is that long-only investors passively accept a hit to their net worth, while leveraged players [such as banks] actively scale back lending to keep their capital ratios from falling," Hatzius explains.
The macroeconomic impact of such losses could far exceed what people might expect, Hatzius says. "For example, if leveraged investors see $200 billion of the $400 billion aggregate credit loss, they might need to scale back their lending by $2 trillion," he says. "Even if this occurs gradually, and even if there are some offsets from reduced credit demand and increased lending by other sectors, the drag on economic activity could be substantial."
Not surprisingly, reasoning like that has prompted some economists to alter their forecasts. Goldman Sachs economists raised their recession probability forecast for 2008 from 30 percent to 43 percent and predicted that the Federal Reserve Open Market Committee (FOMC) would cut the Federal Funds Rate to 3 percent by mid-2008. Previously, they had predicted the Fed would lower the Federal Funds Rate to 4 percent.
An independent estimate
In the Dec. 2 edition of The New York Times, columnist Ben Stein criticized Hatzius for spreading doom and gloom potentially in support of short positions taken by Goldman Sachs, sparking a debate about whether losses-however high the estimate-would lead to sharply lower lending by banks and, thus, have a more negative effect than would appear from the overall size of the potential mark-down. He also questioned Hatzius' estimate that home prices would fall 15 percent peak to trough.
Stein wrote: "This, as I see it, is a conclusion that is an estimation based upon a guess. I found especially puzzling the omission of the highly likely truth that the Fed would step in to replenish financial institutions' liquidity, if necessary. In a crisis like that outlined by the good Dr. Hatzius, the Fed-any post-war Fed except perhaps that of a fool-would pump cash into the system to keep lending on track." Indeed, one could also make the case that if banks move to raise their capital base by issuing new equity shares-as New York-based Citigroup Inc. is already doing-then the theoretical contagion effect seen by Hatzius on overall lending might be diminished.
What is clear, however, is that Goldman Sachs is not alone in its estimate of the magnitude of any potential loss. The independent research firm Lombard Street Research Ltd., London, has come up with similar back-of-the-envelope calculations of the potential losses from defaulted mortgages in the United States, and has provided Mortgage Banking details on how it arrived at its estimate. First, Lombard Street looked at historical trends in industry foreclosure and default data, and found that the ratio of defaults to foreclosures is 62 percent, according to Charles Dumas, chief economist with
Lombard Street Research.
Based on that ratio, Lombard Street estimates a total of 1.5 million defaults for 2007. Lombard Street is assuming 2008 will also generate 1.5 million defaults as a huge wave of resets hits borrowers. Adding in 800,000 or so defaults from 2006, Lombard Street comes up with a total of roughly 4 million defaults in the period of 2006 to 2008.
Next, Lombard Street looked at the loss rate from the 1991-1992 housing slump and found that the average loss rate was 20 percent-or put another way, the average recovery rate was 80 percent [of the value of the loan], according to Dumas. "At that time, the typical loans being advanced were only about 80 percent of the value of the houses," Dumas says. Today the loss rate would be greater against the value of the house because foreclosed loans are typically around 100 percent loan-to-value (LTV) and, for this reason, Lombard expects recoveries to be closer to two-thirds of the value of the loan. "On that basis, we get $250 billion, assuming the same sort of loss recovery rate, and perhaps more-so you could say it's $250 billion to $300 billion," Dumas says.
There are more losses on top of that, according to Dumas, when you take into consideration the fact that the mortgages "are sold on to CDOs and SIVs at premiums of 5 percent to 10 percent," he says. "That money has been taken as profit, but will, in effect, have to be given back or taken back as losses as these things dematerialize," he adds.
This would amount to another $40 billion to $80 billion or more in losses. Adding together the loan losses and the recovery of premiums, the cost of the defaults rise to a range of between $290 billion and $380 billion, Dumas estimates.
There are yet more losses to be taken due to the hit on the pricing of CDOs and other derivatives tied to outstanding mortgages that did not foreclose. The value of even the best assets will decline, "because no one thinks they're AAA anymore," says Dumas. These additional losses could "easily" push overall losses toward $400 billion, he adds, in line with the estimate made by Hatzius.
The next question, of course, is how much of that is going to be borne by U.S. banks and financial institutions. Dumas notes that a lot of losses have already been taken by many European institutions and a few Chinese ones.
"So, we really don't know at all how much of this comes back to the U.S. commercial banking system," Dumas says.
U.S. banks have total capital of $1 trillion, Dumas says. "If you're talking about $300 billion to $400 billion of losses spread over two to three years, it's a sizable chunk and maybe half of that going to U.S. banks," he estimates. Thus, one could say that over a period of two to three years, U.S. banks will be losing a whole year's profit. At the same time, income from the kind of activity that generated the losses would be missing going forward, thereby lowering profits even more. "It's not a catastrophe, but it's a very serious situation," he says.
Like Goldman Sachs, Lombard Street Research sees this having significant impact on the economy. First, the huge premium of earnings growth above the employment cost index is going to go away, "because that premium owes itself to a lot of activity showing up in [Wall Street] bonuses, real estate commissions, stock options, profits and so forth," Dumas says. Earnings on those accounts are likely to be falling rather than increasing, much faster than ordinary income.
"That means that the U.S. economy goes into a patch in which hourly pay is growing, at best, 3 percent, maybe less [because of] slowing jobs growth," he says. Add to this the fact that consumers will likely increase savings because of the reversewealth effect of falling house prices. The slower wage growth, when adjusted for inflation, will mean no real increase in consumer spending. Indeed, Lombard Street Research is predicting no real increase in consumer spending for the rest of 2008.
While business capital spending will grow modestly, it will be outweighed by the drop in the housing sector, sending private investment down for 2008. And, while the housing decline will ease over 2008, business capital spending will "go negative," Dumas predicts.
"All together, it's a private-demand recession, but it's not a GDP recession, because you're going to get positive numbers from government spending and positive numbers from the net export story," says Dumas.
While the U.S. economy could possibly be negative for two successive quarters, the economy will not decline on a yearover-year basis in any quarter in 2008, Lombard Street Research forecasts. The firm expects European and Japanese growth to "be significantly worse in the next few quarters."
"So you end up with quite a gory story, but by no means a crash," he says. It will not necessarily tank the U.S. stock market as more of the Asian glut in savings comes into U.S. equities (instead of fixed income assets), Dumas says.
Estimating subprime losses and exposures
Deustche Bank securities Inc., New York, has also made a set of calculations about the ultimate losses from the mortgage meltdown and bursting of the housing bubble. A Nov. 12, 2007, report by Mike Mayo, research analyst at Deutsche Bank, predicted that total subprime losses alone will reach $300 billion to $400 billion globally (see Figures 1 and 2).
Deutsche Bank expects 30 percent to 40 percent of the $1.2 trillion in subprime loans will default, and the losses will range from 40 percent to 50 percent or $120 billion to $250 billion. Derivatives tied to subprime loans, which Deutsche Bank estimates with some degree of uncertainty at $200 billion, could have loss rates as high as 80 percent-or total losses of $150 billion. Deutsche Bank further predicts that by yearend, there would be $60 billion to $70 billion of subprime markdowns by the end of 2007.
The Deutsche Bank estimate is based on known charges as of Nov. 12 ($43 billion) and estimated additional writedowns of $1 billion at Charlotte, North Carolina-based Bank of America; $4 billion at Merrill Lynch; and $5 billion or so each at Zurich, Switzerland- based UBS AG, Edinburgh, Scotland- based Royal Bank of Scotland Group PLC, London-based Barlcays PLC and London-based HSBC Holdings PLC. If these estimated writedowns are taken, it would represent 50 percent of the eventual losses of $100 billion to $130 billion (perhaps reduced by hedging) based on a "seat-of-the-pants estimate" that banks and brokers will have one-third of the total estimated subprime losses of $300 billion to $400 billion, according to Mayo.
Based on Mayo's analysis, Deutsche Bank predicted that by year-end 2007, financial institutions will likely have taken half of the estimated total outstanding exposures.
Behind the assumptions in the Deutsche Bank analysis is a grimmer outlook on how the housing bubble will unwind. The bank provided additional insight into its analysis during a Nov. 15 phone conference, when Karen Weaver, global head of securitization, advised that "much of this crisis is still ahead."
Noting that the process of moving from default to clearing a bad loan from the bank books and reselling the foreclosed home takes up to 18 months to unwind, Weaver forecast that "downward pressure [on home prices] will be with us through at least mid2010."
Weaver offered some possible scenarios on how far housing prices might fall. If one wanted to return housing prices to the affordability levels of the 2000-2003 period, it would require a 42 percent reduction in home prices in Los Angeles from the levels in effect in January 2007 (assuming current income levels), she said. Taking another approach, if one wanted to restore the balance between the cost of renting and owning that existed in 1999, housing prices in Los Angeles would have to fall 49 percent.
Despite these worrisome comparisons, Deutsche Bank foresees the home-price decline from peak to trough to come in at somewhere in the low to mid-teens, but warns that the "order of magnitude could be significantly higher." This is in line with the 15 percent price decline Goldman Sachs has said it is expecting.
Super-SIV to the rescue?
As 2007 drew to a close, some of the attention on prospective ideas for reducing some of the market fallout from the mortgage meltdown focused on the creation of a super-SIV to buy the assets of trouble SIVs. Efforts to develop a plan for a super-SIV were begun in September by three major banks- Citigroup; Bank of America; and JPMorgan Chase & Co., New York-with the blessing of Treasury secretary Henry Paulson.
The banks temporarily dubbed the potential entity the Master Liquidity Enhancement Conduit on Oct. 15, but did not release any details. A month later, on Nov. 12, The New York Times reported the three banks had agreed on a more simplified version for the super-SIV than the original design, which failed to attract the level of support originally envisioned.
Under the revised version, a consortium of banks is expected to provide the super-SIV with "at least $75 billion" in funds, according to Times reporter Eric Dash. None of the three banks would provide on-the-record details on the proposed super-SIV to Mortgage Banking.
The hope, according to sources knowledgeable about the mechanism, is that it will prevent a fire sale of assets that would send the value of the assets plummeting. That, in turn, would force big writedowns on financial institutions that hold them. The assetbacked commercial paper market continues to shrink as investors decline to roll over maturing assets, while more and more SIVs are expected to unwind and be liquidated.
Citigroup's $80 billion in assets in SIVs is almost one-fourth of the $350 billion SIV market, according to a fact sheet on the super-SIV released by investment consulting firm DiMeo Schneider & Associates LLC, Chicago. JPMorgan Chase and Bank of America do not have SIVs, although each has exposure through money-market funds that hold commercial paper issued by SIVs, according to DiMeo Schneider.
One key concern about any super-SIV is that it will create moral hazard and plant the seeds for the next asset bubble. Paul McCulley, a bond manager at PIMCO, has expressed reservations about the super-SIV or any type of bailout. In an Oct. 31 interview on CNBC, McCulley said that many of the SIVs should probably be liquidated, and that it is "better to know the value of SIVs now [rather] than later."
In an Oct. 19 interview with Emerging Markets, the newspaper of record for meetings of the International Monetary Fund (IMF), the World Bank and regional development banks, former Fed Chairman Alan Greenspan warned that the superSIV runs the risk of further undermining already brittle confidence in besieged credit markets. "It's not clear to me that the benefits exceed the risks," Greenspan said.
He drew a distinction between the bailout of a single large hedge fund to prevent the widespread sell-off of assets-such as happened in 1998 with Long-Term Capital Management (LTCM)-and efforts to prop up an entire asset class, as in the case of the proposed super-SIV. "It could conceivably make [conditions affecting investor psychology] somewhat adverse, because if you believe some form of artificial non-market force is propping up the market, [then] you don't believe the market price has exhausted itself," he told the newspaper.
Greenspan outlined the conditions he thinks are necessary for the resolution of the mortgage crisis. "What creates strong markets is a belief in the investment community that everybody has been scared out of the market, [and they have] pressed prices too low and they're wildly attractive bargain prices there," Greenspan said. Intervention could scare away the vultures, who can be "very useful" when they come in to buy depressed assets, essentially establish a floor for prices and providing a basis for recovery, he said.
Greenspan said the fundamental problem is investor behavior, which swings from euphoria to fear "virtually overnight." He added, "You cannot really defuse them until the speculative fever breaks. When it breaks, it's very abrupt and you just have to wait it out."
The thorny issue of pricing
Setting a real market value for mortgagerelated bonds is a difficult undertaking in itself, while the challenge is even thornier for CDOs and other derivatives, according to Lombard Street Research's Dumas. The ABX index does not directly track CDO prices, where most of the coming markdowns are expected. Noting that AA securities for the second quarter of 2007 had moved from an ABX index reading of a little under 90 in early October to 43 in November, Dumas found the slide very troubling.
"I must say that when you have mainstream indices of the AA tranches of mortgage-backed paper trading below 50 percent-AA is really the senior stuffthen either the structuring is being extremely ineffective or [else] the market is discounting almost [the] worst-[case scenario]. You can take your pick," he says.
"Frankly, I don't think anyone knows. It seems to me the only sensible way to do it is to start conducting some auctions, having a controlled exchange between a few players, and see who it is that is prepared to come in and buy this stuff in bulk and at what prices," Dumas says. He believes that banks and other financial institutions may have "more exposure than they've let on." He contends there's a difference between the assets held on the balance sheets and other exposures that are not as apparent, such as "loans [banks have] to hedge funds which have this stuff on their balance sheets," he says.
The banks' exposure would not be something that a superficial scan of the balance sheet would indicate, Dumas contends, because it may be sitting there in the unused credit facilities and similar places. "The only way you can know that sort of thing is to be inside the bank and be a chief credit officer," he says. Everyone on the outside is probably just guessing, he adds.
Dumas suggests that beyond having markets evaluate the losses, financial institutions can "do what Morgan Stanley has done, which is to say that 'if the whole damn thing is worth nothing, then this is our total maximum exposure,'" he says. ( Morgan Stanley provided guidance on Nov. 7. It is expected to take a $2.5 billion loss in the fourth quarter, but warned that the losses could rise to $6 billion.)
In essence, this approach assumes there is no bid for the assets and, thus, no value. From the time a financial institution can confidently say "This is the worst state we're in," it may be able to put the problem behind it, Dumas says.
"I think the super-SIV mechanism is only going to do any good if it is linked to some mechanism for trading this stuff in a realistic way that creates properly analyzed bids," says Dumas. "You fundamentally get workout groups, maybe private-equity groups, that will go in there and buy this stuff and work out loans at a profit," he says. Private equity could be put together to do this, Dumas says.
The situation is also complicated by the fact that some financial institutions will see this as an opportunity to "play their own hands competitively," Dumas says.
Thus, there's "an administrative vacuum" because the banks themselves are not responsible for the SIVs, CDOs and other derivatives, because they are not on their balance sheets. "It's really hard to see who can be called on to step up to the plate," he says. Thus, the disposition of bad assets is "a real industrywide problem of a structural kind, not of a financial kind," Dumas says.
Capital infusions
Whether or not there is a concerted private or government effort to buy up busted mortgage-related assets and derivatives, individual companies, financial institutions and SIVs will have to find ways to deal with their own pressing problems. Indeed, there are some early signs on how workouts might occur. Capital infusions, for example, have been one way to address the fallout from bad loans and severely written-down assets.
Take the case of Citigroup. According to the Deutsche Bank analysis, Citigroup's exposure is enormous. In addition to the $6.5 billion charge taken in the third quarter and a $10 billion loss guidance in the fourth quarter, Deutsche Bank estimates another mind-boggling $72 billion exposure yet facing Citigroup. This includes $43 billion in CDOs, $11.7 billion in direct RMBS and $17.3 billion in direct subprime lending.
Citigroup chose to get a capital infusion to shore up its capital and avoid a dividend cut. On Nov. 26, Citigroup announced it had agreed to sell $7.5 billion in equity units, with mandatory conversion to common shares, in a private placement to the Abu Dhabi Investment Authority (ADIA). Citigroup had to pay a premium for the investment, agreeing to pay ADIA a fixed annual payment of 11 percent, payable quarterly, until the equity units convert in blocks in 2010 and 2011.
Another way out is to take a cash infusion and, at the same time, sell off troubled assets. That was the approach of the discount brokerage firm E*TRADE Financial Corporation, New York, which was faltering in mid-November as customers withdrew assets from the company after it took a $2.2 billion write-off on MBS in its portfolio. The company was rescued on Nov. 29 when a big hedge fund, Citadel Investment Group Inc., Chicago, provided a $2.55 billion cash infusion into the company. The funds were used to bolster the company's capital with an 18 percent stake by Citadel, which also bought E*TRADE's $3 billion portfolio of asset-backed securities for $800 million, or 27 cents on the dollar.
The Citadel deal provided Wall Street some assurance that there may be a floor on valuations of troubled CDOs. It also showed that "there is indeed money out there, and certain parts of the broader credit and liquidity concerns are being addressed," according to Derrick Wulf, a portfolio manager at Dwight Asset Management, Burlington, Vermont, quoted in the Dec. 2 edition of The Wall Street Journal.
Fannie and Freddie also affected
In view of the meltdown in the nonagency market, high hopes were placed on the ability of Fannie Mae and Freddie Mac to close the gap in providing new mortgage funding to homebuyers and homeowners who are refinancing. Both Fannie and Freddie have been able to go to the secondary market for funding because they are government-sponsored enterprises and are seen as having an implicit government guarantee. Further, both of the GSEs had avoided direct subprime lending for the most part.
Nevertheless, some of the hope transferred to the GSEs was dashed in November when both reported mounting losses from their loan portfolios, as well as from markdowns on non-agency mortgage-backed securities held in their portfolios.
Freddie Mac reported a $2 billion net loss in the third quarter, and took $3.5 billion in markdowns on the value of non-agency mortgage-backed securities on its books. Freddie Mac's writedowns were part of its efforts to "move aggressively," as Chairman and Chief Executive Officer William Syron put it, to deal with the problems rather than string them out. Its losses could potentially imperil the capital cushion Freddie is required by OFHEO to maintain. To address the concern, on Nov. 27 Freddie Mac announced plans to sell $6 billion in preferred stock and to cut its dividend in half.
Fannie Mae reported a loss of $1.4 billion in the third quarter, and reported writedowns on its non-agency portfolio. On its $14 billion of subprime private-label securities (abbreviated as PLS in Fannie's reports), Fannie Mae recognized a markdown of $300 million. On its $28 billion of subprime PLS, Fannie reported an unrealized loss of $600 million. Fannie Mae also revealed that on $5 billion of alt-A PLS classified as trading losses, it took a $100 million markdown. And, on altA PLS held on the company's books, Fannie reported an unrealized loss of $300 million. Like Freddie Mac, Fannie Mae has issued new equity to boost its capital base against losses. On Sept. 28, Fannie raised $1 billion in a variable-rate stock issue, and on Nov. 16 it sold $500 million in preferred stock at a 7.625 percent dividend rate.
On Dec. 7, Fannie announced it would issue another $7 billion in preferred stock not convertible to common stock. The shares would pay at a fixedrate dividend of 8.25 percent until Dec. 31, 2010, and thereafter a variable-rate dividend. Fannie Mae can, at its option, redeem the preferred shares in the future. "This preferred stock issuance completes our previously announced capitalraising program," said David C. Benson, Fannie's senior vice president and treasurer.
Fannie Mae further muddied the water at the time of its quarterly earnings, when it also announced a new method for disclosing its losses-an approach that quickly ran into a buzzsaw of criticism. "Fannie Mae's numbers effectively make its credit look better than it is," charged Peter Eavis in the Nov. 14 issue of Fortune magazine.
Former OFHEO Chief Armando Falcon was not happy with the change, either. "This just smacks too much of the accounting games the company was playing a couple years ago," Falcon told the Associated Press on Nov. 16. "They have very little room to play with here when it comes to trust and credibility," Falcon said, adding, "it doesn't bode well for the new management." Fannie did not respond to a request for comment on the criticisms from Fortune and Falcon.
Fannie has been pressing OFHEO to further raise the cap on its mortgage investment holdings, set at $735 billion after a 2 percent increase several months ago, as a way to fund more loans in the wake of broad cutbacks in loan volume.
Parting advice
The toll from the housing bubble and the subprime mortgage meltdown is rising. And it is likely to continue rising with unforeseen potential impacts beyond the housing, mortgage, banking and securities sectors. Indeed, market worries have risen steadily as more details emerge on where the risks are to be found.
One measure of that fear can be found in the flight to quality being demonstrated by movement into 10-year Treasuries. Demand for such risk-free assets has pushed down yields on 10-year Treasuries steadily from 5.03 percent at the end of June to 4.56 percent on Oct. 1 to 3.83 percent on Nov. 26.
So perhaps the best advice one can follow going forward is a slight variation on the famous quip Bette Davis gave partygoers in the acid-tongued flick All About Eve: Fasten your seatbelts. It's going to be a bumpy ride.
Copyright Mortgage Bankers Association of America. All Rights Reserved. Repritned With Permission.
