Forecasting Pre-World War I Inflation: The Fisher Effect Revisited

We consider the puzzling behavior of interest rates and inflation in the United States and the United Kingdom between 1879 and 1913. A deflationary regime prior to 1896 was followed by an inflationary one from 1896 until the beginning of World War I; the average inflation rate was 3.8 percentage points higher in the second period than in the first. Yet nominal interest rates were no higher after 1896 than they had been before. This nonadjustment of nominal interest rates would be consistent with rational expectations if inflation were not forecastable, and indeed univariate tests show little sign of serial correlation in inflation. However, inflation was forecastable on the basis of lagged gold production. Investors' expectations of inflation should have risen by at least three percentage points in the United States between 1890 and 1910. We consider in an information processing context alternative ways of accounting for this failure of interest rates to adjust, for example the possible beliefs that increases in gold production might be transitory. We conclude that the failure of investors to exhibit foresight with regard to the shift in the trend inflation rate after 1896 is not persuasive evidence that investors were negligent or naive in processing information.