Commodity cycles, a function of market complexity? Extending the cobweb experiment

We examine market behavior in a series of cobweb-like experiments. As in previous studies we find no cyclicality in the simple supply lag design with five players. Step by step we add investment lags and capacity vintages, and thus external validity to the basic design. As complexity increases, prices become autocorrelated with cyclical tendencies. Regressions and simulations suggest that players put too little weight on capacity vintages when making investment decisions. Different from the original cobweb design, adaptive price expectations do not suffice to eliminate fluctuations. Thus, our experiment indicates that the basic cobweb design is not well suited to test endogenous theories of cyclicality in commodity markets.

► Current market theory is not congruent with observed commodity cycles.
► Step by step investment lags and capacity vintages are added to the Cobweb model.
► Laboratory experiments show that increased complexity leads to cyclicality.