Alternative Explanations of the Money-Income Correlation

Standard explanations of the bivariate correlation of money and income attribute this correlation to an inability of agents to discriminate in the short run between real and nominal sources of price shocks. This paper is an empirical comparison of the standard explanation with two alternatives: 1) the"credit view", which focuses on financial market imperfections rather than real-nominal confusion; and 2) the real business cycle approach, which argues that the money-income correlation reflects a passive response of money to income. The methodology, which is a variant of the Sims VAR approach, follows Blanchard and Watson (1984) in using an estimated, explicitly structural model to orthogonalize the VAR residuals. (This variant methodology, I argue, is the more appropriate for structural hypothesis testing.) The results suggest that the standard explanations of the money-income relation are largely, but perhaps not completely, displaced by the alternatives.