The Nonadjustment of Nominal Interest Rates: A Study of the Fisher Effect
NBER Working Paper No. 836 (Also Reprint No. r0521)
This paper critically re-examines theory and evidence on the relation- ship between interest rates and inflation. It concludes that there is no evidence that interest rates respond to inflation in the way that classical or Keynesian theories suggest, For the period 1860-1940, it does not appear that inflationary expectations had any significant impact on rates of inflation in the short or long run. During the post-war period interest rates do appear to be affected by inflation. However, the effect is much smaller than any theory which recognizes tax effects would predict. Further- more, all the power in the inflation interest rate relationship comes from the 1965-1971 period. Within the 1950's or 1970's, the relationship is both statistically and substantively insignificant. Various explanations for the failure of the theoretically predicted relationship to hold are considered. The relationship between inflation and interest rates remains weak at the even low frequencies. This is taken as evidence that cyclical factors or errors in measuring inflation expectations cannot account for the failure of the results to bear out Fisher's theoretical prediction. Rather, comparison of real interest rates and stock market yields suggests that Fisher was correct in pointing to money illusion as the cause of the imperfect adjustment of interest rates to expected inflation.
Published: Summers, Lawrence H. "The Nonadjustment of Nominal Interest Rates: A Study of the Fisher Effect." A Synbosium in Honor of Arthur Okun, edited by James Tobin, pp. 201-241. Washington: Brookings Institution, 1983.