Stock returns and economic activity; evidence from wavelet analysis

Many empirical studies have analyzed the interations between the stock market and aggregate economic activity by examining either their short run or long run relationships, as the time series methodologies employed (cointegration analysis) may separate out just two time scales in economic time series, i.e. the short run and the long run. But the stock market provides an example of a market in which the agents involved consist of heterogeneous investors making decisions over different time horizons (from minutes to years) and operating at each moment on different time scales (from speculative to investment activity). In this way, the nature of the relationship between stock returns and production growth rates may well vary across time scales according to the investment horizon of the traders, as the small time scales may be related to speculative activity and the coarsest scales to investment activity. Thus, for example, if we think that big institutional investors have long term horizons and, consequently, follow macroeconomic fundamentals, we should expect the relationship between stock returns and economic activity to be stronger at the intermediate and coarsest time scales than at the finest ones. The main aim of this paper is to reconsider the positive relationship between stock returns and real activity by using signal decomposition techniques based on wavelet analysis. In particular, we apply the maximum overlap discrete wavelet transform (MODWT) to the DJIA stock price index and the industrial production index for US over the period 1961:1-2004:7 and analyze the correlation between stock prices and industrial production at the different time scales stemming from the multiresolution decomposition of wavelet analysis. The results suggest that: i) the degree of correlation between stock returns and economic activity tends to be stronger at the intermediate and coarsest time scales than at the finest ones, ii) at the coarsest and intermediate scales the timing of the correlation pattern indicates that stock prices tend to lead economic activity. Our findings confirm previous results about the usefulness of wavelet analysis in analyzing economic and financial relationships where the time horizon of returns plays a crucial role.