Book Description

This thesis consists of three essays in commodity price dynamics. In the first essay, we embed a staggered price feature into the speculative storage model of Deaton and Laroque (1996). Intermediate goods inventory speculators are added as an additional source of intertemporal linkage which helps us to replicate the stylized facts of the observed commodity price dynamics. The staggered pricing mechanism adopted in this paper can be viewed as a parsimonious way of approximating various types of frictions that increase the degree of persistence in the first two conditional moments of commodity prices. The structural parameters of our model are estimated by simulated method of moments using actual prices for four agricultural commodities. Simulated data are then employed to assess the effects of our staggered price approach on the time series properties of commodity prices. Our results lend empirical support to the possibility of staggered prices. The second essay investigates the determinants of the percentage change in commodity prices. We apply the dynamic Gordon growth model technique and conduct the variance decomposition for the percentage change in spot commodity prices to 6 agricultural commodities. The model explains the percentage change in spot commodity prices in terms of the expected present discounted values of interest rate, yield spread, open interest and convenience yield. Empirical results indicate that the model is successful in capturing a large proportion of the variability in the 6 agricultural commodity prices. Moreover, we show that yield spread and open interest help predicting changes in commodity prices. Finally, the third essay evaluates different hedging strategies for eleven commodities. In addition to the traditional regression hedge ratio model (OLS) and the vector error correction model (VECM), we estimate dynamic hedge ratios using the conventional dynamic conditional correlation model (DCC) of Engle (2002) and the diagonal BEKK model (DBEKK) of Engle and Kroner (1995). Moreover, we propose two more advanced models, the DCC model and the DBEKK model that will account for the impact of the growth rate of open interest on market’s volatility and co-movements of commodity spot and futures returns. The empirical analysis shows that adding the growth rate of open interest improves the in-sample hedging effectiveness of the DCC model. Furthermore, the out-of-sample hedging exercise empirical results show that static models present the best out-of-sample hedging performance for 5 of the commodities. The DCC model presents the smallest basis variance for 4 of the commodities. The DBEKK model with the growth rate of open interest performs the best in terms of the basis variance reduction for corn and wheat. Our out-of-sample empirical findings provide important implications for futures hedging and highlight the fact that the use of static models to determine the optimal hedge ratio could be more effective than the use of dynamic hedge ratio models.