This article was published at various websites in April 2008, the response and feedback have been overwhelming. I decided to re-post it here at my blog site.
We have all heard about CDOs (collateralized debt obligations) and probably about the insurance of CDOs through CDSs (credit default swaps), which transfer the credit risks of CDOs between two parties (financial institutions).
The CDS is a bet between two parties on whether or not a company or a financial product will default. It is third party speculation on the outcome of the CDO. The CDS provides insurance to cover a fixed income product in case of its default. If a company declares bankruptcy or a debt is downgraded, a claim is triggered.
Currently, the outstanding notional amount of all credit default swaps is about $65 trillion, more than half of the entire asset base of the global banking system. Why are financial institutions are so interested in them? Why have they created so many of them to make this market so big and out of control?
There are many incentives, some of them are larger, and some smaller. I will discuss the three major reasons.
Transference of Risk
First, credit default swaps are not normal insurance policies; each side can trade them to make a quick profit (spread) if there is a willing counterparty. Commonly, after the original CDS contract is engaged, each of the original two parties will try to engage another party to further hedge their risk and earn a small spread. Pretty soon there are layers and layers of counterparties involved, with the total notional amount increasing several fold, until no one knows who they are really dealing with anymore.
You can't monitor this risk since you don't know your counterparty down the CDS chain, and whether they are able to pay in the event of a default. If you throw the counterparty risk out the window, you can always find a sucker to do the trade with you and earn a small spread.
This kind of entanglement has never been seen before in the usually highly-regulated insurance industry. This is why CDSs are traded on OTC (over the counter) derivative markets which bypass all government regulations.
This entanglement creates a chain reaction. If something happens - even a downgrade but not a default - the claim will trigger a domino effect of many claims cascading down through various parties. It will break at the weakest joint (counterparty), probably a highly leveraged hedge fund, and will ripple through all parties involved and likely break the whole chain. It amplifies counterparty risk to the hilt, beyond the default risk of the CDO itself.
Let us use the analogy of new type of car insurance: Car insurance company A trades our car insurance policy costing us $1,000 per year (or $1,000 revenue for them) for another $1,100 similar policy from another company B, and pockets a $100 quick profit. Or we trade our $1,000 policy (company A) with $900 policy from another company C. If we find cheap auto insurance, we just simply cut A out and make a switch to C. Our relationship with the insurance company is always one on one.
However, imagine what happens if both parties trade the policy with a third party. If an accident occurs, company A will not want to pay us, since we engaged company C, and company C will not want to pay either, since it did not issue the original policy. And if company A pays us eventually, they will have to file a claim against company B, who will most likely deny such claim.
This is what happened to the insurance company AON in a story that surfaced last year from a lawsuit. In this real story, Bear Stearns loaned $10M to an entity in Philippines. To hedge this default risk, Bear purchased a protection contract from AON for $0.4M. To hedge this risk, AON purchased protection from Societe Generale for $0.3M. AON thought they were geniuses, offsetting the risk and at the same time earning an easy quick $0.1M profit. Who says there is no free lunch on Wall St.? Think again!
You guessed it - the loan went bust as expected. Bear sued AON for $10M based on the first CDS contract. AON lost the case and paid. Of course, AON sued Societe Generale. Due to some legal technicalities, a different court and judge had a different opinion.
The judgment was that the first and second CDS contracts were two separate contracts. Legally, the resolution of first CDS lawsuit did not automatically grant the similar status to the second CDS. The first judgment can't be referenced in the second lawsuit. As a result, the risk can't automatically be transferred and offset each time.
AON lost the case, and the $0.1M "profit" turned into $10M loss in principal. It was an expensive lunch. This sets an important legal precedent for future CDS lawsuits. A small $10M default in the Philippines impacted three parties. Maybe this is an unexpected downside of globalization?
The market cap of GM is about $11Bn. However, based on estimates in the CDS market, there are about $1 trillion (notional amount) in CDSs betting on GM and their bonds. Any change in GM's situation will create a ripple effect in this $1Tn CDS community of GM.
There are obviously not $1Tn of GM assets to serve as collateral, so you have to trust all parties involved in this wild casino betting that they won't go under water. As a matter of fact, you better pray, because if one goes under, which is a high probability, it will throw a monkey wrench in the whole community, as everyone is trying to unwind and get out at the same time. It becomes a "no way out situation."
Bill Gross at PIMCO did a simple calculation in January in his famous article "Pyramid Crumbling." The total amount of CDS contracts was $45 trillion at that time. The historical default rate is 1.25%, so $500Bn CDS contracts are likely to default. Assuming a recovery rate of 50%, the resulting loss is $250Bn alone.
In the case of CDOs created from sub-prime mortgages, Gross’ assumptions are too optimistic. The assumed default rate is far too low. His recovery rate assumption is probably also too high if you consider the long process of home foreclosure in a deteriorating real estate market with no buyers and legal maneuvers which can be implemented by homeowners (refer to my early articles on foreclosure). I expect to see real losses doubling his estimate, or approximately $500Bn in the sub-prime market.
Using CDS to Smooth Earnings
Second, and more importantly, besides the quick profit earned through a small spread, there is a big incentive to use credit default swaps to smooth earnings from quarter to quarter and to hide losses.
If a CDO is rated AAA by rating agency (almost as good as US Treasuries) why would Wall St. want to buy insurance for protection? At the same time, assume a AAA CDO enjoys a 50 basis point spread above US Treasury rates. It is almost as good as a free lunch. Why would Wall Street firms want to eat into the 50 bp spread (and their profit) to buy CDS insurance?
The answer lies in different accounting treatments. Wall Street firms are not stupid, and they are smart enough to know their CDOs are not US Treasuries, even if their structured product groups and sales people claim they are, with the backing of the rating agencies. It is similar to the promotion of internet stocks in late 1990s, when Wall Street put out a "strong buy" ratings on supposedly great internet companies with unbelievable growth stories, while their internal memos referred to them as "pieces of garbage".
Due to GAAP (generally accepted accounting principles) requirements, investment banks need to mark their CDO products to market if they do not carry CDS insurance. This creates a problem, since both interest rates and credit spreads fluctuate, making it harder to manage earnings. What happens if investors or regulators suddenly realize they are really a piece of garbage? They don't want earnings volatility and surprises, especially at the time when their bonuses are at stake. Banks want to defer paper losses due to write-downs until they actually sell their CDOs, which would mean real bonus reductions for executives and structured finance groups.
By purchasing CDS insurance, according to GAAP regulations, there is no need to mark-to-market, Investment banks need to declare losses only if the CDO is permanently damaged and a claim will have to be paid. No more quarter to quarter fear of marking to market.
This effectively transfers the price risk to a counterparty, who in turn dumps it to another party, and so forth. By the time it comes back full circle, no one needs to worry about marking to market and earning surprises. If a paper loss happens, they can point to the insurance and claim it is only temporary. Financial statements will not be impacted.
CDS provide a vehicle to allow participants to hide any losses to a point where they really can't hide them anymore. They act as an earnings smoother and, worse, they hide the actual risk of investment bank holdings from the public. This is a reason why so many strange things have happened on Wall Street over the last several years.
For example, for a AAA rated CDO with a 50 basis point spread, investment banks would buy insurance from a small second tier bond insurer (such as ACA) whose rating was only single A as a firm (now at junk triple C). If they believed a rating agency with AAA rating on this CDO, why would they want to cover it with a lower rated policy which eating into their profits?
In the example earlier, GM has an $11Bn market cap with $1Tn in CDS outstanding. Use home insurance as an analogy here. If your house is only worth $200k, why would there be policies on your home with a combined notional amount of $20 million?
Using CDS to Accelerate Earnings Recognition
The third and most important use of CDSs concerns the strong incentive to book the next ten years' profit today.
CDSs offer investment banks something called a “negative-basis trade,” which is another accounting loophole like the earnings smoothing discussed above. Using the same example of a CDO with a 50 basis point spread over US Treasures, banks will buy a CDS, costing 20 basis points. By doing so, even though they seem to make less profit (50 vs. 30 basis point spread), banks can book the difference in spread for the whole life of this CDO instantly, through the magic of a negative-basis trade.
If the life of the CDO is ten years, banks can book the whole ten years of phantom profits this year, even if the CDO defaults sometime in the future. This has obvious implications for the bonuses of the structured product groups at Wall Street firms.
Who cares whether this CDO defaults next year? Let’s realize the next ten years’ of bonuses today! There is a common secret at Wall Street - it doesn't matter whether a product is good or bad; the only thing that matters is how you structure it. As former Secretary of the Treasury, John Connally, said to European central bankers in the 1970s' "It might be our currency (the US dollar), but it is your problem." The same thing applies here. If a CDO defaults, investment bankers have already bumped up their stock price, cashed out their stock options and their vested shares, and collected their year end bonuses. Now it is the shareholders’ problem.
This kind of accounting manipulation can fool people for a few years, but not forever, since the well of CDOs gets sucked dry very quickly when every single firm on Wall Street has found out about this and is doing it. A firm owning a mortgage originator has a competitive advantage since it guarantees the source for the well. This is why Stanley O'Neal at Merrill Lynch wanted to buy First Franklin (a mortgage loan originator) so badly, because for every loan First Franklin originates, Merrill Lynch executives and their structured product groups could accelerate ten years of their firm's earnings and future bonuses.
Wall Street wants to package and collateralize everything from residential to commercial mortgages, from credit card to auto loan debt, resulting in a major shift and increase from traditional M&A fees to the so-called trading "profit" in recent years "earned" by investment banks.
How real were the past earnings reported by both Wall Street firms and hedge funds with large CDO profits? If managers can trade a minor reduction in profit (from 50 to 30 basis points) for an immediate bonus of 10 times (by accelerating ten years of earnings to the current year) what would they choose?
If a CDO defaults next year and take a hedge fund under their high water mark, it's no a big deal. The fund already collected a 20% fee from the "profit" the year before. The manager can close the fund and open a new one, raising money probably from the same pool of investors. If you want to see a pyramid scheme, there is none more vivid than this.
How about those unbelievable earnings reported by Wall Street investment banks over the last several years?
Frankly and openly, early this decade, investment banks had repeatedly expressed their dissatisfaction about relying mainly on the traditional banking fees from M&A and IPOs. There was very little room for manipulation since they only got paid when a banking deal or IPO was completed. By discovering the CDO and CDS markets, they suddenly found their Holy Grail, with profit becoming more and more skewed toward the asset "structuring" (or manipulation?) and the trading side now representing the majority of their "earnings."
This kind of accounting abuse is not unusual. It happens in the option ARM (adjustable-rate mortgages) market too. Homeowners (borrowers) for the first year or two pay a teaser rate of 2%. However, in their financial statements, banks (lenders) report the full amount of interest, say 6%, as "profit," while they actually only collect 2%. The net 4% shortfall is added to the borrower's balance. Banks have nothing to lose, but homeowners see their balance increasing and home prices dropping.
WaMu reported $1.4B profit from this kind of ARM last year, while Countrywide earned approximately $600M from them in 2007. How much of those earnings were real? How much real cash have they actually collected or will collect? They would be lucky to collect one third, and the rest they may never see. ARM default rate jumped from 1.5% from last summer to 5% by end of 2007. There are two big waves of ARM rate resets coming - one this summer and another in October this year. By the end of this year, I forecast ARM default in double digits. The total outstanding ARM loans are estimated to be approximately $4-5 trillion. A 10% default will put $400Bn of loans in default.
There are many other accounting manipulations and abuses in the mortgage market. For example, when MBS (mortgage-backed securities) are sold to investors, banks will record the cash flow from interest and servicing rights over the whole life of this bond up front, as something called a "gain on sale". In another example, since mortgages classified as "loans held for sale" on the balance sheet must be marked to market, banks move them to another category called "loans held for investment," which has no such requirement. Only when banks believe their losses are not temporary, but permanent, must they mark to market. Banks can always argue the current real estate plummet is only temporary, if you think long term.
If you are using the last several years' earnings as a reference point, banks don't look very expensive today (setting aside the risk associated with all the off-balance items in their long footnotes). But do you think those trading profits will return in the future? I really doubt it.
Accounting and legal loopholes will get closed by new regulations, and the CDO market will dry up for the foreseeable future. Banks will not be able to return to their previous earning power without their trading "profit" from accounting loopholes. Even if the economy comes back soon with lots of M&A and IPO deals, which is very unlikely, the good old days of trading "profit" are likely gone forever.
I don't see why all the sovereign wealth funds (SWFs) are rushing to invest in Wall Street banks. There is not much upside - only a lot of downside. It is also very interesting to note the stated investment strategy of many SWFs is to seek a 5% stable annual return in very safe financial products.
SWFs are investing in financial institutions like Blackstone and investment banks which are anything but safe and stable. This totally contradicts their investment strategy. Nothing they have invested in can be regarded as safe and stable, let alone yielding a 5% annual return. Perhaps they will see a 5% fluctuation weekly if not daily.
No SWF has factored in currency risk, which may be 5% in US dollar terms, but could result in a negative 25% return in their own currency (if the US dollar depreciates 30% in the next several years, as I believe is likely). Their investments could not be further from their stated investment strategy, and this raises a credibility issue with regard to everything else SWFs have claimed.
Of the outstanding $45 trillion CDO credit default swaps at the end of last year, JP Morgan owned about $15 trillion, or one third of the whole market. This is may be the reason for the Bear Stearns' acquisition by JPM. Even though Bear Stearns' CDS position was “only” around $2.5 trillion, the default of Bear would bring a shock wave of counterparty risk across the whole CDS market and would inevitably expose JPM, with its huge CDS risk, to the global financial market. Who can afford to pay for this? Who has a large enough capital base to absorb such a loss, especially since these contracts are concentrated in only a few CDS derivative dealers like JPM? No one.
It always comes down to the deep pockets, as in any liability litigation, where litigators will skip all the smaller players but will jump on whoever has the deepest pockets and largest exposure and position. JPM currently seems to fit the picture. When such a time comes, all the other weaker and smaller players will try to dump their risk to JPM to unwind their positions. By buying Bear Stearns, JPM can postpone the CDS debacle for another year, but not forever. I expect JPM will eventually suffer very large losses in this area - bigger than one third of their market share. Bernanke and Paulson better get their helicopter and $500Bn of cash ready for their Wall Street friends.
Early this decade, Warren Buffett publicly turned against derivatives. "When Charlie [Munger] and I finished reading the long footnotes detailing the derivatives activities of major banks, the only thing we understand is that we don't understand how much risk the institution is taking," he told investors. He said "Derivatives are financial weapons of mass destruction [WMD], carrying dangers that, while now latent, are potentially lethal."
I believe what he was referring to at that time were not CDSs, but other derivatives such as interest rate swaps. The day will come when investors become aware of the OTC CDS scheme, which has no footnote in annual reports, no market, is not regulated, leaves no trace whatsoever, has no clearing house, and where everything depends on the credit and liquidity of the weakest player in the CDS chain. If an interest rate swap can be called a WMD by Warren Buffett, I can't come up with a name for the CDS. Just as Paul Volcker said last week, the current crisis is "the mother of all crises." The CDS is the mother of the current credit crisis.
The CDS is a Wall Street vehicle used to manipulate loopholes in accounting and legal regulations in order to move and hide losses, to record future profits today, to manipulate, realize and increase reported earnings in today's financial statements to accelerate bonus payments, all done with the help of willing and eager accomplices, in both bond insurers and rating agencies.
At the same time, it is the same old game of quick profits, phantom earnings, rip-offs, manipulation, distortion and cover-up, played on Wall Street since its inception. Only this time it is greatly exacerbated by the financial deregulations enacted during the Greenspan era.
Thomas Tan, CFA, MBA