Collateralized debt obligations were never the main avenue for the sale of banks' trust-preferred securities. But it's on such credit-market side streets where their buyers — fellow banks, especially small ones — have gotten mugged.
During the boom years, few innovations expanded credit to smaller banks more aggressively than the pooling of so-called TruPS. For as little as 130 basis points above the London interbank offered rate, banks and insurance companies overlooked by the capital markets blended more than $40 billion in unsecured, unrated issuances and sold them as collateralized debt obligations.
Today, trust-preferred CDOs are a multibillion-dollar drag on the industry, a source of buyer's remorse. Though some larger companies have significant holdings (Citigroup Inc. and M&T Bank Corp. hold nine figures' worth, and Zions Bancorp. owns $1.6 billion of them), $2.5 billion more are held in chunks by small and midsize institutions that may not have fully understood that they were buying trust-preferred CDOs in the first place. Nor do many of them grasp how much their holdings are worth now; many banks, analysts and rating agencies are still ill-equipped, even with years of hindsight and default data, to evaluate the riskiness of such securities.
"Someone with other capital markets instruments in their portfolio, they probably have more valuable paths to find out what the pricing really is," said James Moss, a managing director at Fitch Ratings. To value banks' true trust-preferred CDO holdings, "you've got to predict how the banking industry is going to emerge. It becomes a bit circular."
The first bank trust-preferred CDO was born in 2000, an event that a 2003 Moody's Investors Service note on rating methodology called an "illustration of the flexibility of the CDO structure" that was "likely to flourish for years to come."
Trust-preferred CDOs gave small banks unprecedented access to the capital markets, enabling individual banks to issue trust-preferreds — hybrid securities that regulators treated as equity because they permitted a bank to defer payments up to five years.
Prohibitively expensive for small institutions to issue on their own, the unrated trust-preferred securities could be pooled, tranched and sold off as rated bonds expected to be redeemed within five or 10 years.
Unlike subprime mortgages, which even in good times were recognized as risky, geographically diverse banks had never defaulted at significant rates. Therefore, rating agencies required only a thin buffer of collateral, barely 2%, to protect a deal's investment-grade tranches.
In addition to providing most of the deals' inputs, small banks also loaded up on the structured securities they produced.
"This was a relationship-driven market," said Gene Phillips, a director for PF2 Securities Evaluations, a small firm that helps small banks value their holdings. Many times the broker that was putting together a trust-preferred CDO would sell slices of other deals to issuers. "Issuing banks would call up their guy, and he'd say, 'Buy some of this different CDO,' " Phillips said.
It was a tempting offer. Investment-grade slices offered solid returns and reasonably short investment horizons. The securities would perform as long as small banks like their own remained healthy.
Not much needs to be said about that assumption. When banks started defaulting in large numbers in 2008, the prices of the CDOs plunged. The problem wasn't just that the expected default rate has been exceeded by orders of magnitude — 28% of the banks that issued the securities have failed or deferred payment, according to Fitch data — but that the CDOs began to perform in unanticipated ways.
One problem was structural — some of the deal documents allotted effective control of the CDO trust to equity holders.
With their positions often worthless, those players have an incentive to turn against other classes of owners. In one case already, the hedge fund TPG persuaded the lower-ranking stakeholders of a trust-preferred CDO called Tropic CDO IV to sell the fund its performing collateral at pennies on the dollar, potentially robbing more senior noteholders of income and sending the deal into litigation.
That wasn't the only weakness. The bonds were meant to be retired within five to 10 years, but the realization of how heavily the initial risk had been mispriced and how high deferrals have been will likely discourage even healthy companies from retiring the trust-preferreds before a bullet payment of principal is due in 20 to 30 years. The onset of "maturity extension" –— akin to having a low-yielding, short-term bond extended to 30 years — means that even securities expected to perform are selling for well below their par value. Essentially, no one expects unsecured small-bank borrowing will ever be so cheap again.
Given these flaws, trust-preferred CDOs are understandably worth less than they used to be. Most tranches began their lives as investment-grade, Tier 1 capital but are now rated as junk. But the question of what type of junk can be controversial, as Zions' portfolio demonstrates.
"There's a huge disparity on ratings among some of the agencies," Zions Chief Financial Officer Doyle Arnold said at an investor conference last month. "There are many cases in which Fitch would have [a trust-preferred CDO security] as 'AA' and Moody's would have it as' B'."
One reason for the wide discrepancies is the timing of downgrades. Analysts are still struggling with matters such as the correlation between bank performance in different regions and the number of banks deferring their trust-preferred payments that will eventually pay off their securities.
This has led to significant variations in how the major credit agencies are approaching the monitoring and re-rating of the securities. According to Fitch's Derek Miller, the agency is "in the process of reviewing all our assumptions" on trust-preferred CDO defaults and deferrals. For recent rating actions, he said, Fitch did not do precise cash-flow modeling, because it felt that the nuances of the capital structure have been drowned out by the sheer volume of defaults and deferrals that determine payouts. Other rating agencies have taken different approaches.
Overall, the changes in approach do appear to be working, some observers said, as the accuracy of the agencies' marks has shown improvement in recent months. The ratings are starting to correlate with actual price, said Elton Wells, head of structured products for Second Market, a company that matches buyers and sellers of illiquid products like trust-preferred CDOs.
"The methodology they're putting out now, they spent a lot of time revisiting it, they're probably closer to the mark now that they've ever been," he said. "But it's going to take time. People have gotten burned, frankly, and don't trust them anymore."
In the absence of trusted benchmarks, the secondary market has become dominated by proprietary research and models. Along with shops like PF2 and R&R Consulting, companies like Intex sell software that allows hedge funds and others to model the cash flow of deals based on their own deferral and default assumptions.
In recent months, the pricing for the most senior tranches has slightly recovered as investors have grown more optimistic that broader economic improvements will trickle into the small-bank sector, Wells said. Trust-preferred CDO buyers are still largely hedge funds demanding superlative returns on their capital, and the market seems likely to be one of the last in the CDO sector to truly become liquid.
"Only over the last few months people are really starting to focus on this," Wells said. "The first ball to drop was residential mortgages, then credit cards, then commercial mortgage-backed securities. Then as an afterthought, people thought of the community banks."