Information in the Term Structure of Yield Curve Volatility

The research paper authored by Anna Cieslak and Pavol Povala back in 2016 outlines the first yield curve model that consistently prices US Treasury bonds and options on Treasury futures, and fits the realised covariance matrix extracted from high frequency yield data. To the best of our knowledge, more than a decade after this paper was written, it remains at the leading edge of yield curve modelling.

The no-arbitrage condition imposed by the model allows for a decomposition of the conditional volatility of U.S. Treasury yields into volatilities of short-rate expectations and term premia. Short-rate expectations become more volatile than premia before recessions and during asset market distress. The correlation between shocks to premia and shocks to short-rate expectations is close to zero on average, and varies with the stance of monetary policy. While Treasuries are nearly unexposed to variance shocks, investors pay a premium for hedging variance risk with derivatives.

No-arbitrage yield curve models such as the one outlined in the research paper are useful especially for managing interest rate risk. This publication demonstrates how integrating multiple large datasets from various segments of the fixed income market, along with advanced modelling techniques, can reveal new, investment-relevant insights hidden within market data.