Volatility and Risk Premia: Lessons from the Eurodollar Markets

We estimate affine models using the Eurodollar futures and options data. The rationale for this exercise comes from a combination of recent theoretical and empirical work, which documents a trade-off in models abilities to match the expectations hypothesis and generate conditional volatility, and suggests to break this tight connection by explicitly removing volatility from the term structure dynamics via the unspanned stochastic volatility (USV). However, the USV property is surprising in the light of the interest rate proxies studies, which find strong evidence for stochastic volatility. We believe that these results could be explained in part by estimation methodology typically used, and in part by a limited span of swaps data. We use a more general methodology in this paper, and expect options data to bring in information about volatility that is absent in the underlying market.