Using various regression models and a vector autoregressions (VAR) framework, this paper examines causal relations and dynamic interactions among stock returns, interest rates, real activity, inflation, and money growth using the post-war United States data. While stock returns have long been presumed as an important indicator of real activity in theory, the predictive ability of the stock market as perceived by macro-economists has been questioned. The purpose of this paper is to answer following two main questions: (1) What causal relationship exists among macro-economic variables behind the apparent negative stock returns-inflation relations in the post-war U.S. data? and (2) Is the stock market a good predictor of the future real activity? Major findings of this paper are that (1) the negative inflation-stock returns relations are in part induced by the negative inflation-future output relations during most of the post-war U.S. data. (2) The persistent component of inflation predicts future output better than temporary output in the regression results. (3) In a dynamic framework of real stock returns, real activity, inflation, and money in the VAR system, the theory by Fama (1981) holds well during the period 1972 to 1995, which suggests aggregate supply shocks have been pronounced in the recent two decades. (4) Stock returns (both real and nominal) seem to Granger cause and explain a substantial fraction of the variance in real activity. (5) Real activity also explains a substantial fraction of variance in inflation. This shows that additional spending has been purely inflationary for the post-war U.S. economy. (6) The asymmetry effects of real stock returns have been captured. Positive real stock returns seem to better predict the real activity, and the 'proxy' hypothesis of Fama does not hold in both cases.