Fool's Errand? Robust Identification of Risk Premia and Expectations in Asset Prices

In this paper I show that the difficulty in estimating unconditional means from time series data alone is the cause for the lack of robustness in empirical estimates of the workhorse model in macro-finance. Using US and UK yield curve data and an extensive Monte Carlo study I show that using survey forecasts is effective in producing robust estimates of term premia, while alternative estimation methods and model restrictions are ineffective. I propose a new way to measure robustness that is simple and allows sharper comparisons across models and sheds new light on the nature of hidden and unspanned factors