A Study toward a Dynamic Theory of Seasonality for Economic Time Series

Abstract Several economists, notably Plosser (1978), Sargent (1978), and Wallis (1978) refuted the assumption that the seasonal component of endogenous variables in a dynamic economic model has almost all of its power restricted to what is termed seasonal frequency and its harmonics. In this article, a model with a closed-form solution is formulated to provide more insight into the arguments put forward by economists. It is concluded that univariate seasonal adjustment cannot be considered a harmless simplification of data without loss of information, neither for the interpretation of economic time series nor for regression analysis.