Some tests of the consumption-based asset pricing model

Abstract

This paper tests a simple version of the consumption-based asset pricing model. Simulations show that when consumption is measured with error tests of the model's restrictions on the serial dependence of a vector of asset returns and consumption growth can be biased towards rejection. These simulations also reveal that, whether or not consumption is measured with error, estimates of a representative individual's relative risk aversion, computed under these restrictions, can be biased downwards substantially in finite samples. Tests of the model's restrictions on the marginal means of a cross-section of returns cannot reject it.