Research

Expectations, Fundamentals and Asset Returns: Evidence from the Commodity Markets (with A. Beber) 

• Frontiers of Finance  2017 
  
• World Finance Conference (Cagliari, 26-28 July, 2017)

​We study the links between expectations, fundamentals, and asset returns using the rich empirical setup offered by commodity markets. We find that survey-based expectations predict future fundamentals, but are not significant predictors of future returns. Expectations of returns are correlated with the slope of the commodity futures term structure and with trading flows. Furthermore, dispersion in analysts' forecasts helps in explaining the options implied volatility risk premium. Interestingly, time-series momentum exhibits a strong negative relation with survey-based expectations. We rationalize these findings using a simple model with heterogeneous beliefs, where professional forecasters can still have rational expectations.

---

Expected Spot Prices and the Dynamics of Commodity Risk Premia (with D. Bianchi)

• Presented at: NBER Economics of Commodity Markets Meeting 2016; European Winter Meeting of the Econometric Society 2016 
• Scheduled: Society for Economic Dynamics Meeting 2017 (Edinburgh, 22th-24th June 2017);  Financialization of Commodity Markets Workshop (Bolzano, 22th-23rd June 2017)

We analyse a novel time series of investor expectations of future commodity spot prices, and provide evidence that survey predictions are extrapolative and inconsistent with a strong form of rationality. We show that a model of adaptive expectations can replicate investor forecasts, and use this to back out the dynamics of the monthly (ex-ante) risk premia, as postulated by the theory of normal backwardation, for different commodities and maturities between 1995 and 2016. The empirical analysis demonstrates that commodity risk premia are time-varying and their dynamics is predominantly due to risk sharing channels and the changing demand for risk insurance and appetite, as proxied by open interest and hedging pressure. Time-series momentum and value factors also significantly generate time variation in commodity risk premia. In this respect, we provide evidence that the explanatory power of diverse factors is not constant over time, both across commodities and time horizons.