Submitted by Peter Tchir of TF Market Advisors
AIG – Still More Questions and Fallacies than Answers
So, as the government is about to unload some of its AIG shares on the market, we will hear about it all day long. It is impossible to listen to a report on AIG without someone mentioning how credit derivatives were directly responsible for the collapse of AIG which is proof that credit derivatives are bad. It also seems that everyone is convinced that it wasn’t the fault of the banks and that the bailout was justified.
The reality is that AIG lost money on customized regulatory capital arbitrage pass through trades where they underestimated the risk but also gamed the system. The banks themselves were sloppy on the credit terms that they engaged AIG on, creating the margin call death spiral. Finally, it has never been made clear that AIG had to get dragged into the problem, and that AIG FP could have been ring fenced and not impacted the parent companies.
AIG Losses Came From Highly Customized Trades Designed to Minimize Regulatory Capital
The trades that caused the losses for AIG are referred to as credit derivatives, but the reality is that most of the trades that cost AIG billions are better referred to as regulatory capital arbitrage pass through trades. AIG did not lose money writing credit protection on single name CDS. They didn’t even lose much money on CDO’s linked to corporate credits. Virtually all of their losses came from selling protection against specific one off deals.
In the dark ages (early 2000’s) banks would buy a pool of mortgages and structure them into tranches. They learned that rather than selling the AAA tranche (by far the largest tranche), they could buy credit protection from the likes of AIG and earn more money than by selling the tranche outright. The regulatory arbitrage was beautiful in its simplicity. Banks bought credit protection from AIG FP using a document that was close enough to a credit derivative contract that they got full regulatory relief from their purchase of protection. That let banks hold huge inventories of AAA mortgage bonds on their books and earn the interest (net of their funding costs and protection purchase cost) with minimal regulatory capital. The document included enough insurance-like terms and provisions, that AIG could treat it as an insurance contract rather than a credit derivative. That let AIG hold negligible regulatory capital against the trades, so although the premium received was fairly small, the return on capital was incredibly high. The return on capital was much higher than if they bought the assets outright. The trades eventually evolved into purely synthetic deals, but the principle was the same – the actual risk to the financial system was unchanged, but the combined regulatory capital was a fraction of what it would have been if either the bank or AIG held the assets outright. This regulatory failure is a big part of why the problem grew and has somehow continued to escape scrutiny. To this day there seems almost no coordination to ensure that risk in the banking system is treated similarly to risk in the insurance system. Until that happens we will wind up with risk in the financial system that is not properly capitalized. But the main point here is that the trades, although called credit derivatives generically, had very little to do with what most people think of as CDS – trades referencing a single corporate or sovereign entity. To continue to call these CDS and imply they have anything to do with the bulk of CDS trading is misleading and doesn’t help the financial community come to the correct conclusions on how the market should work in the future. I do not believe the market needs trades like the one AIG did; whereas, I believe the market as a whole is better off with the standard CDS.
AIG’s own policies made the deals less fungible.
Making the deals even less transparent and harder to price and trade was the fact that AIG liked to write protection on entire deals. They could get some additional rights by writing protection/insurance against the entire tranche. These rights may have been worth something but it helped ensure that no dealer other than the underwriter knew enough details to provide good secondary valuations. In any case it’s another example of how different these trades were from true CDS.
The Banks’ Counterparty Risk Management Failed
The next phase of the problem was a result of a complete breakdown of risk management by Wall Street. While AIG FP was writing all this protection and remained AAA, AIG would argue that they should not post any collateral. In fact they argued that they should not post collateral until they lost their investment grade rating. Banks, in their rush to do these massively profitable deals with AIG, agreed to these terms. Even banks that tried to impose less lenient collateral terms caved in and provided very high initial threshold for mark to market losses before they could ask for collateral. Maybe risk managers didn’t realize the size of the exposure AIG had to these deals with other firms, or underestimated how much these deals could cost AIG, or both, but this was a recipe for disaster.
AIG’s rating had become correlated to the mortgage market. The worse the mortgage market performed, the worse AIG looked. Ultimately this correlation reached the point that AIG’s mortgage exposure was so big it dragged its credit ratings down which in turn triggered collateral calls.
The death spiral scenario had been put in place by the bankers failing to see the potential scenario:
1) Losses in AIG’s portfolio causes AIG to be downgraded
2) AIG gets margin calls once it is downgraded
i) AIG cannot liquidate assets to raise money in a falling market
ii) AIG cannot issue new debt to raise money because the rating downgrades and margin calls are scaring bond investors
3) Banks add to their hedges against AIG default as they negotiate with AIG and the underlying portfolio continues to deteriorate, causing more margin calls
Had AIG been forced to post collateral as the market deteriorated, they would likely have cut some of the position. Had AIG treated these as credit derivatives and marked them to market, the losses would have been reflected quarterly and they likely would have cut some of the position. They probably thought about cutting positions as it was, but it takes a strong management to take a 15% hit on a position where you are only earning 0.1% per annum. The mark to market losses quickly dwarfed the maximum potential profit AIG could have ever made, but they kept the position.
Once the collateral triggers started happening, AIG was in big trouble. The losses they owed collateral on were massive and of course, occurred at a time where raising money was not easy. Had Wall Street been tougher on initial margins and had they required collateral sooner in the process, this situation would not have gotten as bad as it did.
I am convinced that Wall Street is much better about that now, and no one (except probably Berkshire Hathaway) gets away with the credit terms AIG got anymore. On a side note, I’m willing to bet that GS had collateral triggers that kicked in prior to other dealers, because GS is very good at risk management. They would have wanted rating triggers higher than anyone else to get their money first. That is just smart business, and they likely had the negotiating power to get those provisions.
Asides from whatever hedging was done as the situation deteriorated, had Wall Street imposed stricter terms up front on AIG, this problem would never have gotten as bad as it did. Bailout out Wall Street for its own failure to manage counterparty exposure is an incredibly bad precedent but I believe Wall Street has learned its lesson.
The death spiral problem was a common in feature in other products that caught the market by surprise in terms of how quickly they unraveled and the size of the losses. SIV’s, conduits, Leveraged Super Senior, and CPDO’s, all of played a role in the financial crisis and all relied on the fallacy that they could raise money to meet margin requirements at the same time that the only asset they invested in was rapidly declining in value.
Did AIG Actually Owe Money or was it only an AIG FP obligation?
One issue about the AIG bailout that I have not been able to find a satisfactory answer for is what entity was really responsible for potential payments. The employees who lost the money all worked for AIG FP. Was AIG FP the entity that the banks faced, or was AIG actually the counterparty on the trades? If it was AIG FP, as I suspect, then was AIG an explicit guarantor or was it implicit? If AIG FP had an explicit guarantee from AIG what accounting or regulatory game was being played? If it wasn’t an accounting or regulatory game, then why keep AIG FP as a separate sub? If AIG didn’t initially fully guarantee AIG FP obligations, when did that occur and why? We need to know if the losses were largely ring fenced at the AIG FP entity or if AIG was always at risk for the losses, or if at some point AIG assumed the risks of AIG FP? These are clearly technical details, and someone would have to examine each ISDA master document, and each trade, but clarity on which entity actually owed money would help to evaluate the bailout.
If AIG FP was ring fenced, and AIG could have walked away from it without repercussions then the fact that we provided a bailout is unconscionable. Earlier in the year, the insurance regulators had been adamant that the insurance entities of AIG were not exposed to the risks taken on in AIG FP. If that was true, why did we bail out AIG? If it AIG FP was ring fenced, the banks could have been in deep trouble, as their protection on AIG would have been worthless if they weren’t required to make whole AIG FP’s losses. IF AIG FP wasn’t ring fenced, we need more explanations of why the entity existed in the first place. Someone needs to do a study of when and how the risk moved out of the AIG FP entity and into AIG. Until that is done, no one will really know the full story, and I suspect, if you examine the details closely, you will see a culture at AIG, that refused to acknowledge the mistakes made at FP and exposed the entire AIG family to the risk in a desperate attempt to keep the trades alive in the hope that the losses will decrease.