A new test of the theory of storage comparing historical and contemporary data.

In this paper we investigate the relationship between commodity price volatility and market fundamentals comparing the 1920s with the present decade and focusing on cotton and tin. The theory of storage provides the theoretical reference for the analysis. Our first result is to find that the series have widely different properties which reflect the speedier diffusion of information in the markets today. This emerges both in the order of autocorrelation of the VECMs used to analyze the dynamics of the spot and futures returns and in the structure of the GARCH parameterization. Our second finding is to show that, based on full sample correlations, the theory of storage seems to capture the dynamics of data with the exception of historical tin. Rolling correlations, however, qualify this result and show that dynamic correlation for historical tin largely corroborate the theory of reference while recent inroads of financial agents in commodity markets seem to have affected the cotton market, giving prominence to financial risk factors.