Friday, March 27, 2009

Opacity in a Time of Transparency: Assessing Financial Conduits

Posted on Option ARMaggedon by Keith Allman:

Transparency is key to helping us climb out of the credit crisis. Transparency of investments, risk assessment methodologies, parties, counterparties, and so on. However, financial vehicles allowing billions of dollars to pass through them—but that do not release information regarding their underlying assets—still exist. I’m speaking of multi-seller conduits, the commercial paper variety of which have about $730 billion outstanding in the U.S. Their accounting statements offer us generalities, such as asset allocation percentages in mortgages, auto loans, leases, credit cards, student loans, and other consumer-based asset classes, but none report data relating to their exact underlying exposures. If you thought CDOs were toxic, read on….

By way of background, financial conduits are entities that issue short-term debt, typically commercial paper, and use the funds from that debt to invest in long-term assets. These have included all types of structured finance products: from standard auto loans and leases, credit cards, student loans, trade receivables, small business loans to exotic insurance products and esoteric equipment financings. And of course they also invested in residential and commercial mortgage products, which included some of the most toxic assets: adjustable rate mortgages, option ARMs, and even timeshare loans. The profit dynamic at play is well-known: borrow short at a low interest rate, lend long at a higher one.

Strong credit risk management was (or, rather, should have been) the crux of the conduit business. Large banks—with their multitude of credit analysts, transaction specialists, portfolio managers, and quantitative analysts—thought they were well-equipped to manage in-house conduits. Indeed, prior to the credit crisis in mid-2007, funds invested in just U.S. based commercial conduit transactions totaled $1.4 trillion. Since then this number has been in sharp decline due to poorly performing assets and commercial paper liquidity. And so many conduits have either been wound down or have been taken back on balance sheet by issuer banks. But, as noted above, $730 billion remains outstanding in the U.S.

Its conduits’ opacity that should have us worried: The asset-backed securities in which conduits are invested have never been disclosed publicly; Moody’s, S&P, and Fitch don’t even rate their underlying assets. Of course this opacity was part of their appeal. Originators can essentially fund themselves at a low cost without disclosing to the public information about their assets. Selective disclosure of information has its obvious benefits of course.

The rating agencies exercise a little oversight with regard to conduits, rating them as a whole. But their knowledge of conduits’ underlying assets is not required to be disclosed to the marketplace. The whole arrangement is similar to what happens with structured investment vehicles or CDOs, with one crucial difference. Investors in SIVs or CDOs buy into rated tranches: the underlying assets are typically rated and publicly known. Investing in the commercial paper of a conduit is like going into a CDO or SIV blind.
Readers may wonder: if the individual asset-backed securities in which conduits are invested are not publicly-rated, is there any way for investors to independently measure underlying credit risk? The simple answer is no. Here it helps to understand how conduits are constructed.

Transaction specialists package into conduits assets brought to them by originators; quantitative analysts then model asset performance data to create loss expectations. The transaction specialist and the quantitative analyst then work together following rating agency guidelines to determine how much of the asset pool the conduit can invest in while adhering to the lowest rating at which its commercial paper can still be marketed to potential credit investors. If the bank’s conduit credit committee agreed, then the deal was done.

The conflicts of interest are even more pronounced than for CDOs, due to the greater opacity of the deal. Those structuring conduits are bank employees, and their bonuses are based on getting profitable deals done, not on accurately measuring credit risk. Since conduits’ revenue is dictated by their net interest margin, those structuring them are incentivized to invest funding proceeds in high-risk, higher yielding asset-backed securities.

Incidentally, there is one conduit that has done well: DZ Bank’s. They have their transaction specialists put up part of their bonus as first loss on the deals. If the deals go bad, they lose bonus.

In general credit card transactions are a good example. Both Moody’s and S&P issued criteria relating to ABS backed by credit card receivables, but the criteria were very general. S&P, for instance, suggested that for a credit card ABS to maintain an “A” rating, the deal had to withstand losses 2x-3x historical averages. Naturally it was in a transaction specialists’ best interest to work on deals that were closer to the 2 times figure since the historic average loss could be higher. For instance, if a transaction specialist had to choose between two deals: one with historical loss of 5% and the other with historical loss at 7.5% they would choose the deal with 7.5%. This is because they interpreted the guidelines to the minimum published. Both the 5.0% and the 7.5% historical loss deals could be stressed to 15% loss expectation to meet an “A” rating, but a 7.5% historical loss deal is riskier, which means its yield is most likely higher. In other words, in order to increase yield, conduits travel up the credit risk curve. Maximizing net interest margin becomes a matter of flying as close to the sun as possible: trying to pack in credit risk without sacrificing a marketable rating—typically “A”—for the commercial paper needed to fund the deal. Even slight deteriorations in underlying asset performance puts conduits’ “highly-rated” commercial paper at risk of loss.

Conflicts of interest are compounded by transaction specialists’ relationship with the lenders who originate the underlying assets. Originators want the cheapest possible funding, which meant the more the conduit purchased, the less equity they would have to invest in their own assets. Basically, each deal has two components, a portion purchased by the conduit (the debt part) and a portion that the originator of the assets retains (the equity). Say an originator has $300 million of assets, the conduit might purchase 90% of them with the remaining 10% retained as equity. That means that $270 million is owned by the conduit with the remaining $30 million owned by the originator. Originators constantly pushed transaction specialists for the highest possible funding amount for each deal. The problem is obvious: leverage. The less equity contributed at the beginning of the deal, the more debt needed to fund it and the more vulnerable the conduit becomes. Conduit advance rates fluctuate depending on the asset type and loss history, but are often above 90%.

This dynamic further incentivizes transaction specialists to structure transactions to the minimum standards of a single-A rating where possible. Take the credit card example again. Assume that the historical loss was 7.5% and that the rating guidelines suggested stressing the transaction 2x-3x to achieve a single-A rating. If the transaction specialist stressed the deal at 3x instead of 2x there would be an additional 7.5% of loss to cover. This loss has to be made up somewhere in the proposed transaction and is covered by forms of credit enhancement. One method of covering that is to increase the equity, which takes the first loss. This provides more buffer to the single-A rated conduit piece. However, the originator may not want to increase their equity stake and the transaction specialist wants a higher base for the premiums he is booking towards his bonus. The easy solution is that the guidelines suggested that 2x is acceptable, so just run expected loss at 2x, lower the equity, and increase the conduit’s investment. This way the originator gets as much funding as possible, while the transaction specialist gets his or her deal approved by the credit committee and books the deal’s revenue to justify more bonus.

Like so many other structured finance transactions, conduits are a ticking time bomb of credit risk. We now know of course that securities backed by consumer related asset-backed securities are highly vulnerable when the credit cycle turns over. Take Citigroup as an example. They still have $59.6 billion of commercial paper conduit assets off balance sheet. Of that 24% are backed by student loans, 17% by commercial credit and loans, 10% by credit cards, and 8% by auto loans—all of them highly vulnerable to credit losses in recession. And this is just one bank.

To make matters worse, what happens if conduits collapse? Publicly-rated RMBS and CDOs have massive amounts of data organized in databases. There are CUSIPs that allow third parties to search for transactions; some originators make public filings; and often prospectuses can be obtained to understand deal structures. It may be very difficult to work with so much complex data, but at least the data exists. It enables investors to value distressed assets. The data for assets sitting in conduits is extremely difficult to obtain. For clogged credit markets to clear, investors need to have some ability to measure valuation. For CDOs that can be very difficult. For conduits, it may be impossible.

Regulatory overhaul is the order of the day and conduits should not be overlooked. Lax review and disclosure requirements need to be reviewed and modified; the interests of those who structure them must be properly aligned as well. If financial transparency is necessary to climb out of the economic crisis, then conduits need to be opened up and reviewed in detail.

No comments: