| Sumario: | A simple theoretical model is developed to illustrate that three aggregation procedures used in the estimation of long-run money demand functions from time-series data--deflating the aggregate data by population and prices, deflating by prices only, or using nominal data undeflated by population or prices--are mathematically equivalent when the data are dominated by time trend. Differences in regressions based on the three aggregation procedures have nothing to do with the degree of homogeneity in population and prices, as is often claimed, but merely reflect common time trends in the data. The model is seen to provide good agreement with data from three countries. In his restatement of the quantity theory, Friedman (1956) develops the two central themes of the "new monetarists." First, the quantity theory is a theory of the demand for real money balances rather than a theory of the price level. Second, the quantity theory rests on the empirical generalization that the demand for money, defined over some narrow group of financial assets such as currency, demand deposits, and time deposits, is stable over time. The stable demand function implies that any change in the narrow group of financial assets (the money supply) will have a direct impact on expenditure decisions.
|
|---|