The value of the US dollar

There are few less understood phenomena in macroeconomics than the behaviour of exchange rates. Of all the theories we have about this most basic of variables, none do very well at explaining day-to-day fluctuations in the currency. But that doesn't stop economists from trying.

Take, for example, this recent discussion by Menzie Chin (http://www.econbrowser.com/archives/2007/04/a_new_era_for_t.html), arguing that the value of the US dollar will continue to fall. It sounds plausible, and I cannot, to be honest, fault the logic. But for every economist who agrees with this argument, there is likely to be another who disagrees. And there is close to a 50% chance that this view will turn out to be wrong, and the USD will appreciate rather than fall, especially in the short run. In general, the most robust empirical model we have of the value of currencies (in the short run at least) is that they will follow a random walk. That is, the best prediuction of the future value of the currency is its value today. It might go up; it might go down. We just don't know.

My preferred explanation for our inability to predict the movement of exchange rates is the ease with which people can speculate on currencies (again, in the short run). If we truly believed that the US dollar will depreciate, then we should take short positions in the US dollar and long positions in other currencies and profit from the depreciation. But it's very easy (and cheap) to take such positions, and the act of taking them causes the USD to depreciate today, implying that it should depreciate less in the future, all other things equal. If many traders follow such a strategy until they no longer expect to gain from a depreciation, then the actual value of the exchange rate today ends up being approximately equal to the expected value of the future exchange rate. And if their expectations are rational, the exchange rate rate will look like a random walk.

In the longer run, however, it is difficult to speculate on currency since futures markets thin out, and are virtually non-existent at horizons of more than one year. Thus at longer horizons, it should be possible to predict the direction of exchange rates. So while I'm unwilling to make a firm prediction of which way currencies will move in the short run, I think that the USD will have appreciated relative to the Euro and depreciated relative to the Yen by at least 20% within 5 years. You'll need to come back in April 2012 and see if I'm right!