Measuring the Inflation of Parallel Currencies: An Empirical Reevaluation Of the Second Hungarian Hyperinflation

During the hyperinflation of 1945-46, the Hungarian government issued two types of money-like liabilities with differing rates of return: a traditional non-indexed currency (called the Peng õ), and an indexed central bank liability (called the Tax Peng õ). In this paper I reexamine the empirical characteristics of the Hungarian hyperinflation using a definition of the money supply that includes both of these components. I argue that regulations limited the degree of substitutability between these two classes of money. This suggests a specific way of aggregating the indexed and non-indexed components of the money supply, and implies that the appropriate measure of currency depreciation is the decline in the purchasing power of this aggregate. When the inflationary process is represented using the proposed aggregate measures, some of the empirical anomalies traditionally associated with the second Hungarian hyperinflation are readily resolved. In particular, (1) the magnitude of the inflation appears much less extreme than previously thought, (2) seigniorage income remains positive throughout the episode, and (3) unilateral Granger-causality from inflation to money creation, observed during most other hyperinflations, is verified for the entire episode. I would like to thank seminar participants at Stanford, Texas A & M, Purdue, Haverford, UCSD, IUPUI, the Federal Reserve Banks of New York, Cleveland and San Francisco and the Board of Governors for helpful discussions. I am especially grateful for the insightful comments of Valerie Bencivenga, William Bomberger, Gail Makinen, Tom Sargent, Pierre Siklos and Bruce Smith.