The Great Debate: Empiricism vs. Rationalism

The usual divide in the market is between fundamentalists, who look at economics and securities values, and technicians, who forecast future prices by studying historical price patterns. With the latest technology making access to news and charts cheaper and more user-friendly, most professionals seem to use some combination of the two approaches.

Yet there is a more illuminating division in the market between the rationalists and the empiricists. At their best, rationalists use flawless logic to deduce the likely future course of events. Empiricists, on the other hand, tend to base their prognostications on an examination of the specific facts. The different approaches lead to radically different trading strategies.

The rationalists begin with the premise that foreign-exchange prices are a reflection of the relative supplies of money between two countries. The middle term in the syllogism is that Japan's money supply is growing slowly and that the Bank of Japan's balance sheet is contracting. Therefore, the rationalists conclude that the yen should be strengthening. The policy implication is that by easing monetary policy, Japan alone could stem the yen's advance.

For the empiricists, the relationship between two time series -- the BOJ's balance sheet and the yen (against the dollar) -- is a statistical question rather than a theoretical issue. At first blush, the empiricists might run a correlation between the two. Recognizing that what they are really interested in is the correlation between the change in the yen's value against the dollar and the change in the BOJ's balance sheet, empiricists would run the correlation at what statisticians call the first degree of difference.

It results in a value called r-squared, which measures to what extent the two time series move in the same direction. Running a correlation between the change in the yen/dollar relationship and the change in the Bank of Japan's balance sheet produces an r-squared of 0.079, which, on a range from plus 1 to minus 1, implies an insignificant relationship, statistically speaking.

But what if the relationship is more complicated? What if the yen/dollar reacted with a lag to changes in the Bank of Japan's balance sheet? And yet performing the same calculation on the two series, using a time lag of 3, 6, 12 and 24 months, actually produced a lower r-squared than the contemporaneous correlation.

Does this mean that there is no relationship between the yen/dollar and the BOJ's balance sheet? No, not in the least. It merely suggests that if there is a relationship, it is quite a bit more complicated than most observers think when they argue that by simply expanding its balance sheet, the BOJ could drive the yen lower. The policy prescription that follows from the empiricists is the possibility that international cooperation may be needed to slow the yen's rise.

One area of inquiry suggested by the empiricists' findings is that perhaps U.S. monetary policy also needs to be tightened to stop the dollar from falling against the yen. Although the Federal Reserve raised interest rates twice in the past four months, some observers, like Goldman Sachs, argue that U.S. monetary conditions, broadly defined, remain nearly as accommodating as they were at the end of last year.

The rationalists and the empiricists also come to quite different conclusions about the relationship between current-account deficits and foreign-exchange prices. The rationalists in essence argue that current-account deficits tend to be a negative influence for a currency. The U.S. has a current-account deficit approaching 4% of GDP. Therefore, the dollar will have to fall.

The empiricists might begin with some statistical analysis again. But there is no need to reinvent the wheel. The academic work shows little relationship between currency fluctuations, especially in the large, open economies, and a country's external trade balance.

The empiricists would want to look at the "facts" and leave the theory aside. The U.S. has been running a sustained current-account deficit for the past 20 years. During that time, the dollar strengthened in the first five years of the period, weakened for the middle 10 years and has generally trended higher since the spring of 1995. Dividing the world's major currencies by the home countries' current-account position is not very illuminating. For example, Switzerland enjoys a current-account surplus relative to its GDP that is even larger than Japan's, yet the Swiss franc has been much weaker than the Japanese yen.

Or take the opposite case. The U.K. has a growing current-account deficit. In fact, in the middle of last week, the government reported that the first-quarter deficit was really 3.56 billion sterling, not the 2.54 billion of the preliminary estimate. And the deficit in the second quarter was 3.73 billion sterling, significantly larger than the market had expected. Rather than weaken, as the rationalists would suggest, sterling shot up. It registered new highs for the year against the European currencies (which for the most part enjoy current-account surpluses) and new six-month highs against the dollar.

Mark Twain once advised not letting one's schooling interfere with one's education. To that we might want to add a corollary: Don't be so blinded by economic theory that you lose sight of the facts. If currencies moved as theory suggests, wouldn't making money trading currencies be like shooting fish in a barrel? If only.