EC Liftoff: Empirical Assessment Of The Implications For The Dollar

I have been stressing the international implications of a potential interest rate increase as a rationale for deferring monetary tightening. Export growth is slowing and economic activity in the tradables sector (manufacturing output, manufacturing employment) as the dollar has appreciated. [1] [2] How much more appreciation should we expect should the Fed tighten?

This is actually a harder question to answer than one would think. That’s because over the past seven years, the correlation between the US short term interest rate and the dollar has essentially disappeared as the zero lower bound has placed a floor on rates. This point is illustrated in Figure 1.

Figure 1: Log nominal USD broad index, 1973M01=0 (blue, left scale), and Fed funds – ECB refinance rate differential (pink, right scale). An increase in the log USD index from 0.021 to 0.178 implies a 15.7% appreciation. NBER Defined recession dates shaded gray. Dashed line at 2008M12 when ZLB effectively encountered. Source: Federal Reserve Board, Bundesbank, NBER, and author’s calculations.

A regression (in first differences) over the 199M03-2015M08 period yields a statistically insignificant negative coefficient on the interest differential, with zero adjusted R-squared. Augmenting with VIX leads to an increase of adjusted R-squared to 0.15, but no change in the coefficient on the interest differential (the explanatory power is provided by the VIX).
As I’ve pointed out before, the US dollar does seem to be correlated with the shadow Fed funds rate. Without a good measure of rest-of-world interest rates, I use the shadow rate for the euro area. Both series are calculated by Xia and Wu.

Figure 2: Log nominal USD broad index, 1973M01=0 (blue, left scale), and shadow Fed funds – shadow ECB rate differential (red, right scale). An increase in the log USD index from 0.021 to 0.178 implies a 15.7% appreciation. 2015M06-09 differential observations assume ECB shadow rate at 2015M05 level. NBER Defined recession dates shaded gray. Dashed line at 2008M12 when ZLB effectively encountered. Observations for differential for 2015M06-09 are based on no-change in ECB shadow rate from 2015M05. Source: Federal Reserve Board, Xia and Wu, NBER, and author’s calculations.