Can behavioral biases explain the rejections of the expectation hypothesis of the term structure of interest rates?

We test whether the rejections of the expectations hypothesis can be explained by two behavioral biases: the law of small numbers and conservatism. We use the term structure to decompose excess bond returns into components related to expectation errors and expectation revisions, enabling a direct test of behavioral models using the expectations of market participants. We find systematic patterns in expectation errors, and expectation revisions, which are consistent with these two biases. We show that a trading strategy that exploits these biases delivers significant economic profits and that our results are unlikely to be driven by a time-varying risk premium.