Why The Financial System Collapsed

Banks and financial traders always try to outrun the regulators to cop as much cash as they can. But this time, the game swelled to epic proportions.

Saturday, April 11, 2009
by Corine Hegland

Interactive Graphicso low as to create serious safety and soundness concerns," which Freddie said would culminate in a "net reduction in the amount of capital in the banking system to protect against credit risk."

"Freddie Mac would be pleased to assist the agencies by making available our expertise in mortgages and mortgage-backed securities," the mortgage giant said.

Nobody called, of course. Almost none of the regulators involved in the important rule change that took effect on January 1, 2002, which opened the door wide to these innovative securities, even remember the two cogent letters of comment from Fannie and Freddie in 2000. The letters contain pages and pages of prescient analysis of the unusual risk characteristics of rated, structured, asset-backed securities -- the exact financial instruments that would, some years later, explode through every regulatory system under the sun.

Fannie and Freddie, of course, went on to have remarkable adventures in accounting and subprime loans themselves. But the question remains: If they saw this train wreck coming, why didn't anyone else?

Magical Marvelous Money Machine

Fannie and Freddie were lousy oracles: Not only did they fail to predict the collapse of the Western financial system, they didn't even foresee their own swoon into government conservatorship. They did, however, understand the securitization industry that they invented in the 1970s and still dominated in 2000, when they wrote their warning letters.

Securitization is a marvelous, magical, surefire money machine. To use it, banks make loans and sell those loans to create a larger pool of loans; then that pool sells shares of itself, the infamous asset-backed securities, to investors. Banks, in turn, use the cash proceeds to make more loans.

Anything with a cash flow can be turned into an asset-backed security -- credit cards, auto leases, business loans -- but mortgage loans are the granddaddy of them all because of the government-sponsored enterprises involved: Fannie, Freddie, and Ginnie Mae. Private banks started securitizing loans in the 1980s, originally to duck their capital requirements through regulatory arbitrage. But privately issued asset-backed securities didn't surpass the GSEs' activity until 2004.

The original mortgage-backed securities gave all investors equal shares of the loan pool. In the early 1980s, however, Wall Street fine-tuned the machine to build structured securities instead. These complex instruments result from slicing a pool of loans (or securities from a pool of loans) into tranches with varying degrees of blended risk. If the pool loses money, the bottom tranches are likely to be hit first, because they carry most of the pooled loans' risks. The middle, or mezzanine, tranches take the next hit, followed by the senior or top ones.

The junior tranches, therefore -- both bottom and mezzanine -- are far more exposed to losses than the senior ones, although they compensate somewhat for that risk by paying higher yields to investors. The senior tranches, in turn, pay lower yields because they rely on the support -- or credit enhancement -- provided by the junior tranches.

Confused? Try picturing a house: The full weight of the house rests upon the foundation; much of the weight also falls on the load-bearing walls, while the roof handles just wind and rain. Measured in square feet, the roof is the largest part of the house; measured by the weight they support, the foundation and the load-bearing walls are the most important components.

You can dramatically reduce your home-building costs by putting specialists in charge of each section. But if you hire a construction crew that undercuts legitimate contractor bids by building faulty foundations and walls that can handle only wind and rain, you have a problem. If you like the fraudulent crew's low price so much that you give it a package deal to build all of your houses, you have a bigger problem. "And if you can no longer sell your houses, because the word on their [slipshod] quality is out, so you rent them instead, you have a multi-trillion-dollar toxic-assets problem," said Ann Rutledge, who runs a structured-finance firm, R&R Consulting, with her husband, Sylvain Raynes.

Rutledge and Raynes met in the structured-finance department of Moody's Investors Service, one of the three big credit-rating agencies, in the mid-1990s. They opened their own shop in 2000 in the hope of replacing Wall Street's reliance on pixie dust with a more fact-based, analytical approach to the value and risk of structured securities. In addition to evaluating structured-finance deals, they teach securitization classes and have written a textbook on the subject; they have also watched the events of the past few years with horror.

Until the early part of this decade, Rutledge says, most of the blueprints for building structured securities were imperfect but honest. Sometime around 2002, however, fraudulent crews invaded the market by using the magical securitization machine to apparently erase some of the risk in the underlying loan pools. But that's impossible, as it turns out. Securitization redistributes risk, it never eliminates it. The fraudulent crews simply hid the excess risk within the various tranches.

And in an unfortunate coincidence, on January 1, 2002, the Federal Reserve, the FDIC, and the regulators at Treasury outsourced the supervision of securitization to the credit-rating agencies.

"It may not be entirely coincidental that, in the subsequent years, financial service companies swung into high gear creating new classes of rated securities," FDIC Chairwoman Sheila Bair said last year, looking back on the 2002 rule. "These products were very attractive to banks that wanted to boost returns on equity, and to economize on regulatory capital," meaning reduce the amount of capital they had to hold against unexpected losses. "In retrospect, regulators may have unintentionally encouraged banks to bet heavily on a new class of non-transparent securities."

Oops

Putting the credit-rating agencies in charge seemed like a good idea at the time. Several banks had collapsed after representing riskier junior tranches as safer senior ones, actions that cost the FDIC $1.5 billion. Regulators reasoned that the credit-rating agencies, which were already working with investment banks and investors on securitization formulas, could handle the securitization machine better than the regulators' own outdated and restrictive capital requirements.

Under the 2002 rules for securitization, the grades that credit-rating agencies bestowed on pieces of a structured security determined how much capital a bank needed to hold in reserve against those pieces. Because the risky bottom tranche isn't rated, it requires a bank to look through the tranche to determine how much risk it entails from the underlying loan pool, and then set aside a hefty dose of capital to back it up. For the safer, senior tranches, a bank doesn't need to evaluate the underlying loans (unlike foundations, which should be measured by the weight they support, roofs can be measured in square feet, which is called "face value" in this case); a bank can measure a senior tranche by its face value and set aside a smaller amount of capital to secure it because it carries a AAA rating, the highest. (General Electric, by way of reference, had a AAA rating until it was downgraded last month by Standard & Poor's to AA+. Of course, so did American International Group before its collapse, which tells you something about relying on ratings.)

"The intent was to create a more risk-sensitive regime that would give credit for taking less risk and require more capital for holding more risk," said an official at Treasury's Office of the Comptroller of the Currency, which oversees all nationally chartered banks. "There was nothing particularly unique to this ratings-based approach to securitization that appeared problematic at the time."

But regulators didn't know much about load-bearing walls then. Both mezzanine and bottom tranches, remember, support the senior tranches, and so both carry more risk than indicated by their size, or face value, alone. Purchasing a mezzanine tranche on the basis of its face value is like pretending that a wall inside your house is no different than a sheet of plywood at the lumber yard -- you can do it, but you'd better check your insurance policy first.

The 2002 rule treated most mezzanine tranches like senior tranches, or sheets of plywood -- measured by size (face value) instead of weight (relative importance to the structure) -- and required banks to set aside an amount of reserve capital to back them that was pegged to ratings between a speculative BB and a near-perfect AA. As a result, the risk within mezzanine tranches disappeared: A bank that wanted to buy a $100 million portfolio of residential loans could set aside either $4 million in capital to buy that portfolio whole, or $2.276 million to buy the securitized version of that same portfolio: bottom, mezzanine, and senior tranches. Better yet, by selling the $94 million AAA "roof," or senior tranche, and keeping all the junior tranches -- the riskiest ones -- a bank could reduce its required reserve capital down to $772,000 while retaining nearly all of the risk of the $100 million loan pool. In other words, risky pieces required less capital than safe ones. That basically turned banking rules upside down.

Both Fannie and Freddie had watched some of their own structured securities collapse in volatile housing markets, and they knew how easily mezzanine tranches get wiped out if losses rise even slightly above expectations. Even before predatory subprime loans and cheating securitization machines, handling a junior tranche without looking through to the underlying loan pool was just asking for trouble.

What eventually became the 2002 rule, Freddie wrote in its warning letter, could encourage banks "to structure securitizations that reduce their capital requirements to a fraction of what they would otherwise be required to hold, even though the risk exposure remains the same. The result could be a net reduction in the amount of capital in the banking system to protect against credit risk."

Chimed in Fannie Mae, in its letter: "There should be equal capital for equal credit risk, regardless of the form in which the risk is held." (Treasury Secretary Timothy Geithner came to the same conclusion nine years later, announcing that the U.S. regulatory system should govern financial products and institutions by "the economic function they provide and the risks they present, not the legal form they take.")

Fannie provided several examples of how the proposed rule would violate the equal-capital-for-equal-risk principle by handling mezzanine tranches by size, or face value, instead of looking through to the risks they carried from the underlying loan pool, and it challenged the regulators to publicly defend their approach: "If the agencies choose to maintain the use of the face-value treatment, we believe the agencies should provide for public comment an explanation of the economic rationale for using the face-value approach to set capital requirements as opposed to an approach that is based on the economic risk of the underlying collateral."

Furthermore, Fannie warned regulators that if rating agencies were in charge of banks' capital requirements, banks would shop around for the best grades. "Such practices could undermine the reliability of ratings over time," Fannie said, quoting a comment from Moody's. Then Fannie added: "The proposal requires additional controls and oversight if ratings are to be used to establish capital standards, especially in combination with face-value treatment."

Some of Freddie's and Fannie's technical concerns were addressed in the final rule, but their stark and repeated warnings about the implicit risk hiding within pieces of structured securities were not. Interviews for this story with numerous regulators on background indicate that many did not understand the concerns at the time; those who did were unable to find a better compromise for dealing with banks' holdings of bottom tranches, which by then had been a problem for 10 years.

It probably didn't help that some people felt that the 2002 rule was aimed directly at Fannie and Freddie's dominance in the mortgage field. Critics had long argued that the GSEs hold an unfair advantage in securitizing mortgages, partly because their perceived backing by the U.S. government lets them operate cheaper than a bank, and partly because many of their mortgage-backed securities are purchased by banks and thus carry the same light capital requirements that the 2002 rule extended to any AAA or AA asset-backed security -- i.e., to any senior tranche, or "roof," and a few walls, too. "It was very self-serving for Fannie and Freddie to talk about how ratings and structure are no substitute for a guarantee. They had a free guarantee from Uncle Sam," said Guy Cecala, the CEO of Inside Mortgage Finance Publications.

Uncle Sam insists that Fannie and Freddie do not have a guarantee from Treasury, but he did bail them out, and no one really believes his protestations. On the other hand, would anybody but a direct competitor like Fannie or Freddie have enough incentive to study bank capital rules and enough experience to tell regulators they are making a mistake?

Hindsight

The 2002 rule change did not cause the financial crisis. Undeniably, some large banks made reckless bets under it, but many of the bad bets were also done under earlier and later rules. Furthermore, European banks, American securities firms, insurance companies, and pension funds all made the same bad bets as well, each under the watchful eyes of their own regulators. The late, unlamented Lehman Brothers, for example, was not subject to the 2002 rule, and it still collapsed.

But if anybody had checked on Fannie and Freddie's warnings about the hidden risks in mezzanine tranches and the pressure on rating agencies to give them good grades, they would have seen the following:

A six-year mortgage boom began in 2001 when homeowners flocked to refinance under low interest rates, and a small structured securities boom started in 2002. By 2004, "everybody in the universe who could refinance had," Cecala said, and the expanding mortgage-backed securities industry stumbled on the bright idea of looking for new loans in all the wrong places -- and that's how the subprime market ballooned. These loans covered the gamut from undocumented "liar loans" to predatory subprimes and adjustable-rate negative amortizations. People were encouraged to buy houses with no money down or turn their home into an ATM. The mortgage industry used just about anything to create new mortgages that it could feed into the maws of banks hungry for structured securities and the profits that those securities produced. Everyone seemingly forgot about, or didn't care about, the quality of the underlying loans.

Along the way, the big three credit rating agencies -- Moody's, S&P, and Fitch -- forgot about the credibility that had carried them through nearly 100 years of grading corporate bonds. Lured by skyrocketing profits in structured securities, they began bestowing ratings based on models so flimsy that, as one analyst noted, half the risk had disappeared. A security "could be structured by cows and we would rate it," the analyst wrote.

Few people thought about the effects that unpayable mortgages would have on home values and structured securities. Too many people forgot, or never understood, or didn't care about the hidden risks that gave mezzanine tranches a better yield than identically rated corporate bonds. And for some reason, banks and rating agencies started pretending that the bonds springing from the mezzanine tranches of structured securities were akin to high-quality corporate bonds. So they used them to build still-more-profitable restructured securities, called collateralized debt obligations. These CDOs are manufactured much like structured securities, but they start with a collection of bonds instead of an underlying pool of loans. The financial wizards then tossed the bonds derived from mezzanine tranches of structured securities, which had a lot of subprime loans in their pools, into their collection of CDO bonds. These wizards, however, did not look back at the first structured security or its underlying loan pool -- in effect, they plopped a house on the roof of another house, "a house of cards," as another rating analyst wrote. In short, the CDO bonds were about three steps removed from the true value and risk of the underlying mortgage loans.

In fact, the high yields and hidden risks of mezzanine tranches based upon subprime loans made CDOs so profitable that there weren't enough of them to go around; banks in 2005 or so started using derivatives, such as credit default swaps, to create fake mezzanine tranches. The banks would build one house, take the load-bearing wall from that house, and jam it into a second house, the CDO. They would then use credit default swaps to copy that same load-bearing wall and stick it into still more CDOs. This, in effect, meant that the banks were plopping four or more houses on the roof of the first house and constructing dizzying profits from dizzying leverage and risk. "That's how you took a small portion of the residential mortgage market and went around the world to lose 15 times as much," said Phoebe Moreo, a partner in Deloitte & Touche's securitization transactions team.

Between 2004 and 2007, the wizards created more than $1.4 trillion worth of these unstable houses of cards. They did not build all of them from pieces of risky subprime loans, but the blessings of the rating agencies made the CDOs attractive to everyone. Commercial banks, securities firms, pension plans, money-market funds, insurance companies, governments -- nearly every sector of the financial markets somehow bought into the mass delusion of a bottomless pot of gold.

Cheating Markets

Fannie and Freddie worried that the sloppy handling of mezzanine tranches would encourage banks to invest in the risky products; they did not anticipate that the entire financial industry's sloppy handling of mezzanine tranches might, instead, facilitate a $1.4 trillion CDO market that looked like the Wild West.

"I think the regulators -- and most policy makers -- truly thought the market would properly regulate itself, and they continued to think so even as the excesses built to extraordinary levels," said Timothy Howard, who was involved with Fannie Mae's comment letter as its chief financial officer at the time. He was forced out in 2004 after being accused of manipulating Fannie's earnings.

So, the global house of cards was built on top of the inherently weak foundation of asset-backed securities, which were structured in a way that made it easy and profitable to hide the underlying risks. There was another problem, though: Even though many people were fooled, there were not enough foolish players to purchase all of the structured securities being produced. So banks concealed the leftover pieces of risk from everybody -- regulators, themselves, and the markets -- by either pretending they had voluntarily purchased the leftovers to resell later, or by thrusting them into the short-term debt markets in the guise of top-grade, AAA-rated, asset-backed commercial paper. Without those outlets for hiding the skyrocketing risk, the whole debacle likely would have stopped a lot earlier and forced the financial markets to find another trough at which to feed.

Instead, by the time the commercial paper markets realized that they were financing a house of cards and pulled up roots in the middle of 2007 (an action that sparked the initial market correction that policy makers called a "liquidity crisis"), all those leftover chunks of risk, combined with the pieces that banks had deliberately retained, had built up to an estimated $1 trillion in toxic securities. These securities are now on the books of American banks, in addition to another $1 trillion in toxic whole loans. Worldwide, the crisis might cause $4 trillion in damages, the IMF estimates.

"Anybody that knew anything about this market would have understood that leverage, and leverage, and leverage was an accident waiting to happen," said Gustavo Dolfino, a former investment banker who now runs an executive placement company called the Whiterock Group. "The capital ratios were insufficient, and people in the business knew they were insufficient, but there was a lot of pressure to create wealth, and if you didn't do that, you were fired."

One lesson for policy makers, then, is that when the markets stumble onto a bottomless pot of gold, somebody needs to play the grown-up and figure out what's going on. "Ultimately, the risk has to go somewhere. If you can't find it, you've lost it and you probably better figure out where it is," said Joseph Mason, a banking professor at Louisiana State University and a former employee of the Office of the Comptroller of the Currency. "At some point, somebody got fooled into thinking the risk went away and they didn't need to track it."

American regulators, Mason points out, knew that banks were hiding risk and debt from the rules governing such things by sticking it in both their traders' books and in the commercial paper markets; the regulators simply failed to put 2 and 2 together and realize that this regulatory arbitrage, or regulatory avoidance, was also hiding risk and debt from the markets.

If regulators had taken a closer look at banks' use of the commercial paper markets to finance their complex structured securities, they should have realized they had a problem.

In 2004, regulators had an opportunity to look: The Financial Accounting Standards Board issued a rule closing the accounting loopholes that Enron had used -- with the help of major banks -- to hide shenanigans from its balance sheet before it collapsed. The securitization industry -- commercial banks and securities firms alike -- protested vehemently, saying that the accounting change would force them to take about $300 billion of assets onto their own balance sheets, which would increase banks' capital requirements.

Why? When banks finance assets in the cheap, short-term commercial paper markets, the assets technically belong to an off-balance sheet "conduit"; these conduits resemble a holding area where the bank doesn't need to hold capital in reserve against the assets.

The rating agencies are not complete fools, however. Before they give commercial paper issued by the conduits the coveted top grades required by the short-term debt markets, the agencies make banks promise to support their conduits. This is so investors don't lose money if the conduit can't "roll over" its paper or find new investors after the short-term loan expires. Even conduits without such promises usually get bailed out by their bank sponsors, which know that if investors ever lost money on a conduit's investment-grade securities, the banks would have trouble securitizing future loans.

Now, a lay person might have wondered why the banking industry that helped Enron manipulate its balance sheet by engaging in complex structured finance activities with Enron's conduits was so concerned about a rule intended to prevent another Enron. If that lay person had a hunch that something was rotten, he or she might have taken a closer look at banks' commercial paper conduits. There, the lay person would have discovered conduits holding collateralized debt obligations built from the mezzanine tranches of some marginal subprime loan pools. If that person were a regulator with a wide streak of skepticism, he or she should have detected the rot with a simple whiff. Such skepticism, says Richard Herring, co-director of the Wharton Financial Institutions Center, "would have been a damn good thing."

Both regulators and academics, however, tended to believe that the regulatory arbitrage behind the commercial paper conduits "was serving a fairly useful function," said Charles Calomiris, a professor of financial institutions at Columbia University, because it allowed banks to escape a fairly useless set of global banking rules known as Basel I.

In 1988, the world's bank supervisors had negotiated Basel I. These rules threw assets into ludicrously crude categories of risk: Loans to General Electric and General Motors, for example, require 8 percent capital reserves because they are corporate loans; mortgages given to Warren Buffett and Bernard Madoff, meanwhile, carry only 4 percent capital charges by virtue of being residential mortgages. In response to Basel I, banks immediately began arbitraging -- avoiding -- the rules by securitizing loans and financing them through the commercial paper markets, which both banks and regulators thought was a reasonable compromise with Basel I and its overly restrictive risk categories.

Basel II was supposed to fix Basel I by dispensing with the crude risk categories and instead allowing banks to use the credit rating agencies' letter grades in determining how much capital they had to hold in reserve for various kinds of debt instruments. Big banks under Basel II could dispense with the ratings altogether and rely on their own internal assessments of risk instead. Bits and pieces of Basel II have been implemented over the past decade, but the basic accord wasn't finalized until 2004. U.S. securities firms, such as Lehman Brothers, began operating under a slightly modified version of Basel II in 2004; European banks fell under Basel II in 2007; and U.S. commercial banks were supposed to do the same in 2008.

Although global regulators are today frantically revising Basel II in the wake of the financial crisis, the task uncomfortably resembles redesigning the Titanic. "Every phase of Basel II has proven unreliable in this past year," Herring said. "The ratings haven't worked, the internal models haven't worked, and the guards against regulatory arbitrage haven't worked."

So in 2004, U.S. regulators, who were stuck with Basel I and anticipating the arrival of Basel II, decided not to study banks' conduits and did not think about the market implications of financing complex structured securities, such as CDOs, with commercial paper. Instead, they issued a rule saying that if the post-Enron accounting changes forced conduit assets onto banks' balance sheets, banks could ignore the assets when calculating their capital reserve requirements under Basel I. This turned out to be a superfluous move, as banks swiftly found loopholes in the new accounting rules anyway, and regulators agreed that banks could continue holding almost no capital against conduits that remained off their balance sheets. At the industry's request, regulators explicitly included in the 2004 rule a newly popular type of risky conduit, a Structured Investment Vehicle, which financed itself with both commercial paper and medium-term debt. Regulators acknowledge today that nobody was watching the SIV sector; it subsequently swelled into a $400 billion nightmare and collapsed in 2007.

The paper conduits "were the black hole of the system," Rutledge said. By allowing banks to conceal the true scope of their risk and leverage from regulators and from the markets, and sometimes even from themselves, the conduits encouraged the collapse of the house of cards.

Commerce And Thieves

The Federal Reserve, the FDIC, and officials at Treasury regulate only American banks. They were not in charge of banks' competitors such as insurance companies, hedge funds, foreign banks, and some securities firms. The regulators, as a result, were caught "between a rock and a hard place," Calomiris said. "That doesn't justify every mistake they made," he said. "There were things that were crazy, and there were regulators who said, 'I'm not so comfortable with this.' But do you want to stand in front of a several trillion-dollar train and tell it that it has to be derailed?"

Even if U.S. regulators had faced that train, they likely would have been hauled in front of Congress to explain why they wanted to knock banks out of the subprime CDOs powering the profits of every other major financial institution in the world.

Capital markets are fast and fluid; as President Obama said, they "move around the globe in a second," seeking whatever investment is underpriced, even if it is built on a house of cards. Bank regulators tried to set consistent standards around the world with Basel I, but politics are politics, and they needed nearly 20 years to negotiate Basel II. Banks, on the other hand, immediately began arbitraging Basel I from its inception, and then banks collapsed within moments of Basel II being implemented.

So how do you go about establishing rules for all the money movements on Earth? Because that's really the challenge underlying the G-20's resolution to overhaul failed regulatory systems, and nobody has a good answer yet. "So long as the standards remain inconsistent across the credit supply chain, or don't reflect what they purport to address, cheating will remain a more profitable business for financial institutions in the short term than serving clients," Rutledge said.

One option is to rethink globalization and prevent U.S. institutions from becoming too big to fail; if that means conceding a long-held U.S. competitive advantage to countries less concerned about protecting their taxpayer dollars, then so be it. (Seriously -- as a taxpayer -- so be it.)

A complementary if harder option is to rethink everything. "We have to go back to the very beginning of regulation here," Mason said. "This has gotten so far out of balance that it is not at all clear to me how to put the rules back into balance by picking through them piecemeal."

Since World War II, American banks attracted the world's investments because they were thought to be under tight, transparent supervision. The U.S. global edge in financial services, in fact, is one of the key things keeping U.S. trade deficits from being even worse. Our recent piecemeal regulatory approach to financial institutions and products, however, has undermined the world's trust. Perhaps Washington can regain it by figuring out what powers it has wrongly ceded to hyperfast markets searching for the next house of cards. U.S. laws, for example, exempt some financial transactions from bankruptcy proceedings, which are governed by rules that are supposed to help stabilize financial markets. But such exemptions also appear to be undermining players' incentives to check one another before markets do collapse.

"We're putting too much emphasis on supervisory discipline rather than market discipline," Herring said. "The supervisors are always going to be behind: They don't have the resources, the pay scale, or the ability to keep up with innovative industry. The people who know these banks best are their counterparties, and we've developed a body of private law that spares them from any loss if things go wrong."

So it may be best to start over. Assume that thieves and cheats will inevitably slip into financial systems. Establish -- and enforce -- transparent rules giving U.S. regulators and U.S. banks a competitive advantage in catching them faster than anybody else. And then stick a trip wire in the new rules, triggered by vanishing risk. Because when things get out of whack -- when the markets, under the eyes of regulators and financial players, stumble onto a seemingly bottomless pot of gold -- there must be an outside Cassandra whose warnings are heeded.