Proponents of Quantitative Easing frequently cite—inappropriately in my view—the Taylor Rule as support, saying that the rule calls for a federal funds rate as low as minus 6 percent, well below the zero bound. But in various pieces over the past year, such as Taylor Rule Does Not Say Minus 6 Percent, I have argued the contrary. If you simply plug in current inflation and output (gap) you will find that the interest rate is above zero with the policy rule coefficients I originally derived. But QE II proponents change the coefficients. Frequently they use a higher coefficient on output (around 1.0) rather than the lower coefficient (0.5) which I originally recommended. The higher coefficient on output gives a much lower interest rate now and is thus used by proponents of quantitative easing.
A new paper by Alex Nikolsko-Rzhevskyy and David Papell provides important evidence relevant to this debate. They show that, if history is any guide, the higher coefficient would lead to inferior economic performance compared with the original coefficient I recommended.
First, they look at the 1970s when monetary policy was too easy: in much of this period the interest rate was too low creating high inflation and eventually high unemployment. They show that the higher coefficient on output would have perpetuated the bad policy while the lower coefficient would have prevented it.
Second, they look at years in the 1990s when monetary policy is widely viewed as good, helping to create a long expansion. Here they show that the higher coefficient would have prevented this good policy. Actual policy was more consistent with the lower coefficient which I had proposed.
In sum, they find no reason to use a higher coefficient, and that the lower coefficient works better. David Papell’s guest blog yesterday on Econbrowser nicely puts these new results into the context of today’s policy debate and provides more details. He emphasizes that his paper with Nikolsko-Rzhevskyy does not endeavor to estimate the impact of QEII, but rather shows that the typical policy rule rationale for this discretionary action is flawed. In my view it is an example of “discretion in policy rule’s clothing.”