[proof needed] - and in Europe.

There is nothing sufficiently unique about Europe to exempts us from simple accounting relations. Cycle-averaged balanced budgets in a nominally growing economy impose an upwards unemployment bias by withdrawing fiat-money purchasing power from the economy, relative to the demand for fiat-money purchasing power.

Was French unemployment worse in 93-08 than it was in 78-93? Growth? General state of the economy?

Both of those periods are dominated by intra-European fixed-rate ForEx regimes (the ERM between '79 and '93, the EMU between '98 and '08), so you can hardly use a comparison between the two to test a hypothesis on the influence of floating versus fixed exchange rates.

I am not contesting that repeated devaluations in a fixed exchange rate system is a poor policy. This is, I believe, a theoretically well founded and clearly empirically validated conclusion. My contentions are that

1. Continuous depreciation of the currency in a floating-rate system in order to defend the real exchange rate from appreciation is not painful, as long as your financial regulator prevents people from carry trading in your currency.

2. When you are Soros attacked, you need to make a clear-headed judgement as to whether you will be able to defend against it or not. If you do not believe that you will be able to defend against it, you should not mount a defence - instead, you should use your currency reserves to minimise industrial disruption by providing bridging currency swaps to strategically important firms that require essential imports. Which firms are considered "strategic" and which imports are considered "essential" should be planned out ahead of time and continually revised, as should your estimate of your ability to defend against Soros attacks. Regular drills would be a good idea.

3. The crises in Greece, Spain and Portugal are fundamentally currency crises, not debt crises.

Now, if my third contention is correct, then the relevant comparison is between the 1993 currency crisis and the 2008 currency crisis, not between the ERM and EMU periods.

Simple inspection suggests that the 1993 devaluation was much less painful than the 2008 austerity, but in order to make a compelling case one would have to compare the magnitude of the required exchange rate adjustments to achieve real exchange rate parity with the one's choice of indicators of economic performance. I don't have 1993 price level, unemployment and exchange rate data on hand for these countries, but if you're really interested I can probably get them and run the numbers.

- Jake

Austerity can only be implemented in the shadow of a concentration camp.