Yona Rubinstein and Nathan Sussman
The Hebrew University of Jerusalem

One of the most persistent puzzles in monetary economics is the positive correlation between interest rates and the price level. this empirical regularity Ð defined by Keynes as the ÒGibson ParadoxÓ Ð seems to defy the Fisher hypothesis. Numerous empirical studies and theoretical suggestion have emerged to reconcile this contradiction between economic theory and the data. In this paper we specify an Augmented Fisher Equation that allows shocks and nominal rigidities, reflected in the market for real balances, to affect the real interest rate and for the long run anchors of the price level to affect the nominal component of the interest rate. In a classical fixed exchange rate regime the money supply is endogenous and the anchors of the long-run price level are the exchange rate and the foreign price level. In a classical free float, the money supply is exogenous and the long-run price level is determined by the quantity theory. Real and nominal shocks will have a different effect on real balances in the two monetary regimes. Therefore, we expect them to have a qualitatively different effect on the real interest rate. We test our Augmented Fisher Equation by selecting Britain in the years 1797-1850. During that period, a change in monetary regimes and policy provides us with a unique Ônatural experimentÕ to study the relationship between two classical monetary regimes and the variables that make up our Augmented Fisher Equation. Britain was forced, due to the Napoleonic wars, to suspend gold convertibility in 1797. It maintained a free float until 1821. We show using the Augmented Fisher Equation that we can find empirical evidence in support of FisherÕs hypothesis. We also showed that the Gibson effect is part and parcel of the Augmented Fisher Equation and in no way constitutes a Paradox. We find that the Gibson relationship is not limited to war time or war finance; on the contrary, the Gibson relationship is stronger in the post war era. We also find that when we allow for the effect of supply shocks a positive correlation between the price level and interest rates existed also in the Suspension era. We find that the war raised the real interest rate by roughly one percentage point. However, we also find that the Bank of England financed part of the war with inflation tax. According to our results the increase in the real money supply of 5.5% per year contributed to a decline in the average real interest rate throughout the war of about 0.5%. Our results provide evidence in support of the operation of the fundamental long run relationships of the PPP and the quantity equation in their respective regimes during the nineteenth century. Results obtained from estimating the Augmented Fisher Equation allow us to claim that British investors behaved in accordance with the monetary regime; we cannot reject the hypothesis that rational agents living at the time understood the difference in the monetary regimes and the economic policy pursued and formed their expectations accordingly.