JP Morgan - Is High Correlation Between Stocks and Exchange Rates Due to Japan's High Dependence on Exports?

We have pointed out that the relationship between Japanese stocks and the yen may change. The close correlation of Japanese shares with the yen over the last several years of higher (lower) share prices equating to a weaker (stronger) yen may change to higher (lower) share prices equating to a stronger (weaker) yen. The relationship between the two is not merely correlative but causal. We think Japan’s high export dependency could be behind this. According to this logic, the cause and effect relationship of a strong (weak) yen and lower (higher) share prices will continue as long as Japan’s export dependency is high (in other words, as long as domestic demand-led economic growth cannot be expected).In this case, a weaker yen is an absolute condition for share prices to rise.

Is that really the case? Figure 1 plots the correlation coefficient every five years for the TOPIX (yearly change) and the yen-dollar rate (yearly change) juxtaposed over the trend in export dependency. It is true that, over the last several years, export dependency has risen to its highest level since 1960 at least. However, when export dependency was high in the first half of the 1980s, the correlation between the stock market and the yen-dollar rate was negative, and when the correlation between the stock market and the yen-dollar rate was strongly positive in the first half of the 1990s, export dependency fell to its lowest point since 1960. We do not see any correlation between the rise/fall in export dependency and the correlation between the stock market and the yendollar rate. Even if export dependency is high, the correlation between the stock market and the yen-dollar rate is negative. In other words, higher (lower) share prices may equate to a stronger (weaker) yen.

The performance of the Japanese stock market is a function of corporate earnings, and if corporate earnings are a function of production volume, share prices should rise when there are signs of a recovery in production. Figure 3 plots the correlation coefficients for TOPIX (yearly change) and industrial production (yearly change), and for the yen-dollar rate (yearly change) and industrial production (yearly change) (five-year estimation period). As we would expect, both of the correlations are basically positive during the period, with the exception of the bubble era, when share prices rose all out of proportion to production (and thus earnings, roughly speaking). However, it is difficult to explain movements in the yen-dollar rate based on production.

In conclusion, we can say that share prices will rise on the back of a recovery in production, but we cannot say what the yen-dollar rate will do. As we have noted in the past, market participants seem to be overly sensitive to exchange rate movements.