Three essays on commodity price dynamics

The idea that commodity prices are inherently dynamically stable has played a major role in studies of price fluctuations over the past decades. Indeed, the dominant view of the classic theory of commodity price formation is that prices are self-stabilizing and that fluctuations result from shocks in the environment of the market, from the weather, technology, macroeconomic variables or from demand. For instance, Schultz (1945) thought that the main source of price instability was shocks from the industrial sector, and Cochrane (1958) supported the view that uncertainty is created by weather conditions, but he also set forth a structural feature of the agricultural sector that may lead to instability: the low price elasticity of the demand for food items. However, there is still much unexplained price dynamics in markets for storable commodities. That the possibility of storage is relevant for the evolution of prices in commodity markets is all too obvious. Variability of prices will depend fundamentally on the variability of production, harvest, extraction, on one side, and of demand for final consumption on the other. Whether the product can be stored or not, however, makes a world of difference in the ways in which the variability in the fundamentals gets transferred to the equilibrium price. Therefore, the possibility of storage and its consequences should always be considered in empirical analyses of commodity price formation. Among theories that tried to explain the role of storage in the dynamics of commodity markets, the “theory of competitive storage”, developed within a stochastic framework, has undoubtedly experimented the greatest success. Under this theory, uncertainty is introduced in the supply and demand equations through stochastic variables, economic agents are treated as being risk neutral, economic systems follow linear relations, and price expectations are rational (e.g., Lowry et al. (1987) for a model of intra-year storage,Williams and Wright (1991) for a model of inter-year storage). Recent work, however, on the exogenous price formation hypothesis declined in popularity due to the unsatisfactory empirical results of the theory of competitive storage and its overly simple assumptions on linearity, rationality and risk-neutrality. The empirical test of this theory has not rendered entirely satisfactory results. Deaton and Laroque (1992) have tested this theory against the behaviour of 13 commodities. The assumptions of their model are consistent with those of the theory of competitive storage with the exception that a nonlinearity is introduced through the non-negativity constraint of stocks. Their model does manage to explain the rare and sudden rises of annual prices, but fails to reproduce the high level of autocorrelation of actual price series. Hence, in the second chapter of the thesis, An Empirical Estimation of the Rational Expectations Competitive Storage Model, Fabio aims at testing the theoretical validity of the competitive storage theory by applying simulation based econometric methods, namely Simulated Method of Moments (SMM) (Duffie and Singleton, 1993; Lee and Ingram, 1991). The idea is to put this theory to a test while verifying the capacity of its model of reference to mimic actual data. By so doing, if some parameterization of the model would generate values close to the actual ones, this may be taken as evidence that storage could be among the determinants of the observed commodity price dynamics. The third chapter, Modelling the Relationship between Speculation and Gold Prices, adds to that stream of the literature that claims recent increasing investments into commodity-related assets did not actually led to the skyrocketing price behaviour witnessed in the last few years. The results of this work support the idea that a nonlinear approach is needed when modelling the relationship between speculative trading behaviour and gold price returns. Using weekly data from the June 1995 to October 2006, the analysis first focused on speculators’ and hedgers’ response to price shocks, and vice versa, by making use of vector autoregressions, and then studied the dynamics of speculators’ long positions. A relative measure of position size, which aims to capture the net position of the average trader in a CFTC classification, was derived. This study found that the relationship between speculation and nearby futures contract returns for gold follow a nonlinear path, which can be well characterized by a logistic smooth transition process, as proposed by Granger and Teräsvirta (1993) and Teräsvirta (1994). The transition from one state to another is smooth but asymmetric. It implies that, within the context of gold price returns, speculators may react differently to price changes according to the price regimes. The transition from one regime to the other occurs when price changes are close to zero, indicating a different behaviour takes place during price expansions (positive returns) and contractions (negative returns). Trading activity induced by price changes appears to be much more intense during periods of increasing price levels. In addition, former speculative activity only plays a significant role in expansions. Results also suggest herding behaviour and positive feedback trading of speculators during price booms. The last chapter, Modelling commodity price dynamics: a new paradigm?, extends and very briefly analyses the issue of whether the increasing presence of financial investors in commodity markets may have affected the price dynamics of these markets. The main purpose of the study is to build an as simple as possible price model, but also that may be able to effectively modelling spot price dynamics, while quantifying the impact of changes in market fundamentals on price. The choice of basing the analysis on a fundamental point of view is driven by the fact that such an approach may enable one to deal well with the most difficult stage of market analysis, that is, quantitatively and objectively to link actual market conditions to the cash price levels. Furthermore, the motivation lies on the conviction that conventional market analysis leads one towards only one of the many price drivers that actually act simultaneously. After selecting a few fundamental drivers, including a proxy, different from the one derived in the previous chapter, to represent the so-called speculative pressure, results support the idea that financial speculative investments may be taken into consideration in order to achieve a better fit and properly explain the recent dynamics of aluminium prices.