"Concentrations of risk, plagued with deadly correlations"

The FT’s Gillian Tett makes a simple but important point: AIG’s role in the credit default swap market meant that a lot of risk that the bank regulators thought had been dispersed into many strong hands ended up in a single weak hand.

Tett:

What is equally striking, however, is the all-encompassing list of names which purchased insurance on mortgage instruments from AIG, via credit derivatives. After all, during the past decade, the theory behind modern financial innovation was that it was spreading credit risk round the system instead of just leaving it concentrated on the balance sheets of banks.

But the AIG list shows what the fatal flaw in that rhetoric was. On paper, banks ranging from Deutsche Bank to Société Générale to Merrill Lynch have been shedding credit risks on mortgage loans, and much else. Unfortunately, most of those banks have been shedding risks in almost the same way – namely by dumping large chunks on to AIG. Or, to put it another way, what AIG has essentially been doing in the past decade is writing the same type of insurance contract, over and over again, for almost every other player on the street.

Far from promoting “dispersion” or “diversification”, innovation has ended up producing concentrations of risk, plagued with deadly correlations, too. Hence AIG’s inability to honour its insurance deals to the rest of the financial system, until it was bailed out by US taxpayers.

If the US creates a "systemic risk" regulator, it should be on the lookout for similar concentrations of risk.

One other point. The fact that several of AIG’s largest counterparties are European financial firms is by now well known. What is I think less well known is that the expansion of the dollar balance sheets of "European" financial firms -- the BIS reports that the dollar-denominated balance sheets of major European financial institutions (UK, Swiss and Eurozone) increased from a little over $2 trillion in 2000 to something like $8 trillion (see the first graph in this report) -- played a large role in the US credit boom.

As the BIS (Baba, McCauley and Ramaswamy) reports, many European banks were growing their dollar balance sheets so quickly that many started to rely heavily on US money market funds for financing. And if an institution is borrowing from US money market funds to buy securitized US mortgage credit, in a lot of ways it is a US bank, or at least a shadow US bank.

Consequently I think it is possible to think of AIG as the insurer-of-last resort to the United States’ own shadow financial system. That shadow financial system just operated offshore. There was a reason why investors in the UK were buying so many US asset backed securities during the peak years of the credit boom.

UPDATE: I want to second those in the comments who recommended this post on AIG by a credit trader. Those familiar with some of the technicalities of risk management should particularly enjoy it.

A simple historical example of "wrong way risk" is buying insurance against a big fall in the value of the ruble from Russia’s government (or Russian banks). The problem with the insurance? A big fall in the ruble is likely to be correlated with bad things happening in Russia, which would tend to make it less likely that Russia would be able to honor its promises. This was exactly the case in 1998.

The credit trader notes that writing insurance against a default on a lot of different high quality CDOs backed by mortgages is ultimately different than writing a lot of insurance against say fires. The odds of a large nation-wide fall in home prices is a lot higher than the odds of a fire that sweeps across a lot of geographically dispersed cities. I continue to be struck by the damage done by the historical data that seemed to suggest that a nation-wide fall in home prices was rare. Surely a large rise in home prices correlated across several regions would increase the odds? Moreover, it wasn’t hard to realize that the structure of the market had changed over time. When I was growing up -- not so long ago -- interstate banking was a novelty. Kansas and Missouri had different banks. That meant Kansas banks were very exposed to a downturn in the Kansas economy, but it also tended to reduce the correlation among different housing markets.