from Follow the Money

Systemic risk

February 16, 2005, 10:57 am (EST)

Blog Post
Blog posts represent the views of CFR fellows and staff and not those of CFR, which takes no institutional positions.

More on:


John Plender of the Financial Times has a nice (subscription only) survey of the risk that "systemic risk" (think LTCM/ 1998) may make an unwelcome comeback.

Plender’s analysis clearly draws on New York Federal Reserve President Tim Geithner’s recent speeches. It certainly seems that Geithner is assuming the role of bad cop (Don’t forget about systemic risk, avoid "sustained fiscal and external imbalances that increase the risk of large macroeconomic shocks") while Greenspan generally is more comfortable playing the role of good cop (Advances in financial technology make higher degrees of leverage safe) -- though today, he too seems intent on warning against excessive complacency.

Plender makes an important point:

As short-term rates rise, and long-term rates fall, the "carry" shrinks. The same is true when credit spreads fall. Getting the same return, therefore, can require more leverage.

"The snag is that, since the Fed started to raise rates last Hune, the margin has become thinner. To achieve the same profit, people are taking on more leverage and making bigger bets. This makes for more volatile markets ... Yet financial institutions continue to set targets for high and increasing revenues and profits that assume the good times spawned by freakish monetary policy will continue to roll."

Sounds right to me.

Someone clearly has been making big, and I would assume quite leveraged, bets that the Treasury curve will flatten -- though perhaps right now the trend is to take profits on this trade rather than to up the basic bet. No doubt there are other big leveraged bets out there as well.

One final point: Argentina’s default did not put a dent in the US banking system (though it did put a dent in the earnings of Citi/ the Bank of Boston). But Greenspan is wrong, I think, to attribute the resilience of the US banking system to the development of markets that permit credit risk to be unbundled and transferred -- i.e. new financial technology. The more parsimonious explanation: US banks, and the US financial system writ large, did not have much exposure to Argentina.

Argentina had tons of bonds -- roughly $100 billion -- outstanding. But most of those bonds were held by Argentines, not Americans. 50% of the bonds, roughly speaking were in the hands of Argentines, 25% in the hands of retail investors, and 25% in the hands of big institutions in the US and the UK. Generally speaking, US firms were trimming their exposure through out 2000 and 2001, while Argentine banks and pension funds were forced to keep adding to their exposure (lots of arm twisting). The result: Argentina’s default devastated Argentina’s financial system, but not the world’s financial system. Some local banks (reportedly, the Spanish owned banks) were also selling protection against default to international investors. That added to their losses, but it hardly transferred risk to those most able to bear it.

Citi and other US banks with local operations did take enormous losses. But Argentina’s banking system only had $80 billion in deposits. I think Citi had no more than say $8-10 billion in deposits (ballpark, I have not gone back and checked). It took huge, and not entirely expected, losses (over $2 billion, if memory serves) on its local operations, as did Boston. But its overall exposure was limited -- pesification shifted some losses to depositors even as it imposed others on the owners of local banks. Citi was not going to lose more than its equity stake in Citi BA -- and what ever additional losses it took on its offshore loans to Argentine firms (Offshore = Lending funded not by Argentina deposits, but by Citi’s global operations).

All in all, though, the most best explanation for the resilience of the US financial system to Argentina’s default is limited exposure and plenty of time to prepare -- not gee whiz new financial technology.

More on: