Predicting Interest Rates

Abstract

We examine whether asset managers contribute to price discovery in the Treasury market by exploiting regulatory bookkeeping data maintained by the Financial Conduct Authority (FCA) in the UK. Our sample covers all secondary-market trades in gilts and contains detailed information on each individual transaction, including the identities of both counterparties. A simple calendar-time portfolio that goes long the quintile of gilts heavily bought by asset managers and short the quintile heavily sold generates a monthly return of 24bps, with an annualized Sharpe Ratio of 1.4. Controlling for the level, slope, and curvature factors have little impact on our finding. This return predictive pattern does not revert in the subsequent twelve months and is stronger for the subset of asset managers with better performance in the past year. Additional analyses reveal that most of this superior performance is due to asset managers’ ability to forecast changes in shortterm interest rates, over and beyond public information.