What happened to Argentine banks in 2001? And why?

I suspect most people who read this blog are interested in the dollar, US interest rates, the trade balance, the "hard v. soft landing debate" (outsourced today to Kash at the Angry Bear), oil, even systemic risk in US financial markets.

Retrospective analysis of what went wrong in Argentina back in 2000 and 2001 is probably not so high on most reader’s list of interests.

But Argentina’s crisis was a searing experience for me.

In my no doubt biased view, the case studies are the real strengths of this just released IMF paper (fully disclosure: I contributed to the Argentine case study). Each case study shows how balance sheet analysis can be applied to better understand crisis dynamics -- or the absence of crisis dynamics. All were written by some of the best (young) economists at the IMF (but again, I may be biased).

The Argentine cases study tries to show, concretely, why delay was costly, and specifically, how the balance sheet of Argentina’s banking system evolved during the course of 2001. The core thesis of the Argentine case study is simple: Argnetine’s banks were in far better positioned to survive a devaluation and a government debt restructuring at the end of 2000 than they were at the end of 2001. Consequently, waiting a year had real costs.

Interestingly, the deterioration of the banking system’s resilience was NOT primarily the result of new lending to the government. The banks extended far more credit to the government in 1999 and 2000 than they did in 2001. Indeed, that, in a sense, was a core cause of the government’s trouble: depositors were pulling funds out of the banking system, and a shrinking banking system could no longer extend credit to the government to help cover its ongoing deficits.

But even though the banking system’s absolute exposure to the government did not go up, its relative exposure did. Its credit to the government rose a bit, while its deposit base shrank massively. By the end of 2001, lending to the government made up a far larger share of the banking system’s assets than at the end of 2000.

Ironically, during the course of 2001, Argentine banks got rid of precisely those assets that would (potentially) have performed in the event of a devaluation and government debt restructuring. They ran down their best assets -- their liquid offshore reserves -- to pay off depositors (and to pay off maturing cross border credits). They also reduced their peso lending to Argentine firms dramatically. Peso deposits fell more rapidly than dollar deposits (that, incidentally, does not mean dollar depositors did not run: some peso depositors shifted into dollars, and some dollar depositors ran). To stay matched, currency wise, the banks had to reduce their peso lending commensurately.

That left the banks with dollar-denominated lending to the government, and dollar-denominated lending to Argentine firms. Both types of lending were almost sure to go bad in the event of a restructuring and a devaluation. Most of the banks’ dollar lending was not to firms in Argentina’s (small) export sector. Remember, so long as the peg lasted and Argentina’s exchange rate remained overvalued, the export sector was, generally speaking, a bad bet. Rather, most of the banks’ dollar lending was to firms with peso revenues. Those firms may have been foreign-owned utilities, but they were still bad bets, with mismatched revenues and liabilities. Any political risk assesment would have indicated that the utilities were not going to be able to continue to index their prices to the dollar after a devaluation.

There are more details and twists in the case study -- and similar in depth treatment of Uruguay, Brazil, Lebanon, Turkey and Peru.