Few would quibble with this table of impacts, but despite the not-yet successful efforts of exchanges to trade macroeconomic market variables, there are no real ways to trade any of them directly; stocks and bonds are not GDP futures. Moreover, the relationship between currencies and other financial markets tends to be far weaker and unstable than commonly believed.

We can illustrate this with the dollar index (DXY) relative to various financial markets. The long-dated comparisons against the S&P 500 and gold use the weekly average of the DXY against the weekly averages of both the S&P 500 index and cash gold; this eliminates some of the noise associated with using a single weekly closing price.

How many times have you heard a market commentator note how stocks moved higher “despite the weak dollar”? More than 35 years of data are conclusive in this regard: Weekly returns on the DXY lead those on stocks by one week on average, but their statistical effect is so weak as to be irrelevant. The R2, or percentage of variance explained, is 0.002 percent. Even more surprising is the sign of the relationship. It is negative. We should expect dollar strength to not depress, as evidenced by the negative slope, subsequent returns on the S&P 500.

Gold is regarded by many as a financial asset, and the dollar’s presumed relationship with gold has a strong logical footing: If each dollar is worth less, it should take more of them to claim a given quantity of gold.

But the relationship is a bit cloudier than that. The expected rate of inflation and the short-term interest rate cost of holding gold (as well as any actual supply-demand shifts in gold) have to be included in the equation.

As proven in 2006, the price of gold can disconnect from these presumed relationships quickly and dramatically. Even so, the R2 of gold relative to the DXY is several orders of magnitude higher than that for the DXY-S&P 500. Also, the sign of the relationship is strongly negative, as we should expect.

Subscribe to:
Post Comments (Atom)

0 commentsPost a Comment