Testing for Structural Breaks : Trade-Off between Power and Spurious Effects

I argue that in order to test for the effect of the foundation of the Federal Reserve on the behavior of short-term interest rates, extending the sample size on either side of the interval 190%1918 will generate spurious results. Mankiw, Miron, and Weil (1987, 1994) do not address data problems pointed out by other authors, nor do they take into account some historical events and institutional changes in the New York money market that had an impact on the behavior of short-term interest rates. Given these reasons for restricting the sample size, I argue that their Monte Carlo evidence is largely irrelevant to the issue, and that lack of power confirms my claim that no regime change can be detected astride the foundation of the Fed