Regression Models with Variables of Different Frequencies: The Case of a Fixed Frequency Ratio

An increasing variety of data frequencies available in economics, finance, etc. gives rise to a question how to build and estimate a regression model with variables observed at different frequencies. In a unifying framework of ("m,d")-aggregation we consider various approaches by discussing some potential and limitations. A Monte Carlo experiment and an empirical example illustrate that the traditional fixed aggregation approach, widely used in applied economics, might be inconsistent with data and highly inferior in terms of model precision. Copyright (c) Blackwell Publishing Ltd and the Department of Economics, University of Oxford, 2010.