Shifts in the Interest-Rate Response to Money Announcements: What Can We Say About When They Occur?

Numerous studies find shifts in the interest-rate response to money-announcement surprises following the Federal Reserve policy changes of October 1979, October 1982, and February 1984. In recent issues of this journal, LeSage used a multiprocess mixture model to detect shifts in the interest-rate response and reported that shifts occur at times other than these dates. We provide simulations that show that use of the mixture model produces inaccurate estimates of when response shifts occur and often identifies shifts when none occur. These results are consistent with existing work that shows that detecting structural change is difficult when the independent variable has zero mean.