Advances in Financial Economics, (1995): 123-154
Campbell R. Harvey
Duke University and NBER
Tests of a nonlinear asset pricing model are presented using an economic activity proxy for the marginal rate of substitution. Recent research has been restricted to quarterly data because the Department of Commerce does not report unfiltered monthly consumption. The proxy introduced here is constructed from retail sales data. A model is tested that allows for seasonal, holiday and trading-day taste shifters in the utility function. The model's restrictions are strongly rejected using one month real returns on a Treasury bill and a value-weighted stock index. However, there is less evidence against the model when longer horizon returns are examined.