The so-called dollar crisis of the past month needs to be kept in perspective. The overall trade weighted dollar index - cushioned by the huge fall in the Mexican peso - has fallen by only about 5 per cent against all other currencies, and it is trading at about the same level it reached last October. But it has fallen this year by almost 10 per cent against both the mark and the yen, and has hit new all-time lows against both these currencies at a time when it was already heavily "undervalued" in competitiveness terms.
For what it is worth, the Goldman Sachs estimates of "fair value" for the dollar in real terms stand at the following rates: DM/$ 2.12, Y/$ 180 and $/£ 1.59. These are the exchange rates that would ensure that purchasing power parity (PPP) was achieved, implying that prices for similar goods would be equalised in each economy.
So far is the dollar below PPP at present that it is tempting to dismiss the concept as wholly irrelevant. However, recent academic work has largely resurrected PPP as an operational concept. For example, Jeffrey Frankel and Andrew Rose have just completed a paper for the American National Bureau of Economic Research that shows fairly decisively that PPP holds over long periods of time. These authors have reversed the results of earlier academic work by examining a much larger sample of countries and time periods - in fact, their data collection covers 150 countries over 45 years.
Their conclusion is that around 15 per cent of any deviation from PPP tends to be eroded each year. This implies that the "half life" of a currency's deviation from PPP is around four years - that is how long it takes a currency to get halfway back to PPP from any given starting point.
Admittedly, it is not clear whether the adjustment towards PPP can be expected to occur because the nominal exchange rate shifts while inflation remains constant, or because inflation shifts while the exchange rate remains constant. Possibly, a little of both might occur. Certainly, if the Frankel rule held for the dollar, and if the Federal Reserve keeps the lid on American inflation, then we would expect to see a few years of solid gains in the exchange rate before very long.
Seen in this light, it is a big puzzle why the dollar has been able to maintain such a massively under-valued level in inflation-adjusted terms for so many years. One explanation is that the quality of US products has been slowly declining compared to international standards, so that American goods need to become cheaper each year in order to compensate for lack of quality. The chronic current account deficit recorded by the US, and the persistent current account surplus recorded by Japan (although no longer by Germany) seem to offer proof of this.
The existence of these current account imbalances at exchange rates well below PPP has caused economists to search for other measures of medium term equilibrium for the dollar. The best known alternative - the fundamental equilibrium exchange rate (FEER) - was invented by John Williamson. It represents an estimate of what level the dollar would need to attain over the medium term in order to balance the balance of payments, assuming that all countries are working at normal levels of employment.
Obviously, many debatable assumptions are needed to make such calculations, but the outcome for the FEER is almost always substantially lower for the dollar than any estimate of PPP. Even so, FEER estimates of around DM/$ 1.60 and Y/$ 106 have been left behind long ago by the tumbling dollar.
We therefore need to delve a little deeper to find an explanation for the dollar's "undervaluation". Let us assume that we know that, in the medium term, the dollar will trade at close to its FEER. Accepting that this is a medium-term phenomenon, where should the dollar be trading today, and how should it approach its FEER on an equilibrium dynamic path?
This all depends on two other factors - interest rate differentials between the US and other countries, and the risk premium (if any) which the market demands to hold dollars.
Compensation for risk
Essentially, the dollar must be expected by the market to move up or down towards the FEER on a path where the change in the exchange rate exactly compensates the holder of dollar assets for any interest rate differential and for any risk premium demanded. Consequently, if interest rates in the US are below those in other countries, or if there is a positive risk premium on the dollar, the dollar must trade below its FEER so that it can be expected to appreciate back towards it over the medium term.
This explains a conundrum which often puzzles the market, which is why the consensus of economic forecasts for "weak" currencies with poor fundamentals always seems to show the currency rising over time. The reason is that the currency must be expected to rise in order to compensate holders for the risk of holding it. If it were not expected to rise at any point in time, it would need to fall further, and would continue until the market became convinced that it had overshot downwards.
At present, there is little difference between interest rates in the US and in Germany, while US rates are substantially above those in Japan. We therefore cannot blame interest differentials for the fact that the dollar is trading below its FEER. This means, by deduction, that there must be a large risk premium attached to the dollar, and the dollar's recent depreciation suggests that this risk premium has risen quite sharply this year.
The risk premium - or the expected excess return in dollar assets - is needed to attract sufficient short-term capital flows into the US to offset the chronic deficits that it has been running for some time on the current and long-term capital accounts of the balance of payments. However, with the outflow on long-term capital account having shrunk sharply in recent months, this does not explain why the risk premium has risen abruptly lately. Nor do events like the Mexican crisis seem sufficient to explain this.
As the graph shows, Goldman Sachs estimates for the risk premium on the dollar are now phenomenally high - 3.5 per cent per annum against the mark, 6 per cent against the yen. These risk premia are clearly extreme - in fact, almost as extreme as they were in the other direction when the dollar surged to unprecedented highs in the mid-1980s.
We learned then that markets are capable of driving currencies away from their fundamental levels by vast distances for several years, but that they do eventually adjust. Sooner or later (and quite possibly much later), these extreme risk premia will subside, and the dollar will recover.