Regression and rate of change in economic time series

Abstract

This note comments on Geary's (1972) use of regression to calculate the rate of change in economic time series. His paper refers to "the rate of change", but it does not distinguish between the instantaneous rate of change, with which he is concerned, and the common discrete rate, which is a rate per period. The bias and variance of the estimators of the discrete rate of change for the two methods discussed by Geary are derived, using his assumptions.