An empirical investigation of the relationship between the real economy and stock returns for the United States

Abstract

This paper tests for the relationship between excess returns and economic growth rates in the U.S., using a Seemingly Unrelated Regression (SUR) approach. The system includes monthly data for inflation, consumption, narrow money supply and personal disposable income and each equation has up to 24-lagged Autoregressive terms. After removing the four major shocks associated with Black Monday, the Asian Crisis, “9·11” and its anniversary, we cannot find any ARCH behaviour in either the excess returns or the money series. The models are reduced to their parsimonious forms and the inflation and real consumption equations are corrected for ARCH. To make the result more robust we reduce our system to four equations by conditioning on income and testing the remaining equations for stability. The SUR model suggests strong long-run feedback between the financial sector and the real economy firstly through inflation, then consumption while the influence of real money supply appears transitory. Consumption is more sensitive to the economic variables in short and long run as compared with stock market windfalls.