A New Approach to Asset Integration: Methodology and Mystery

Abstract This paper develops a simple new methodology to test financial market integration. Our technique is tightly based on a general intertemporal asset-pricing model, and relies on estimating and comparing expected discount rates across asset markets. Expected discount rates are allowed to vary freely over time, constrained only by the fact that they are equal across (riskadjusted) assets. Assets are allowed to have very general risk characteristics, and are constrained only by a linear factor-model of covariances with the discount rate over short time periods. The technique is undemanding in terms of both data and estimation, and includes CAPM. consumption-based models, various ICAPM and other models as special cases. We provide a variety of domestic and international empirical illustrations of our technique, and find surprisingly little evidence of integration. While the S&P 500 market seems typically to be integrated, others are not, including: the NASDAQ, the Toronto Stock Exchange, and three different classes of American bonds. Further, there is little evidence of integration between these apparently deep frictionless financial markets.