The Stability of an Experimental Market with a Supply-Response Lag

The words 'stability' and 'instability' frequently crop up in theoretical discussions of markets. In context the meanings are usually clear enough, but what would happen in an actual market if the preconditions for instability were met? Would prices, for example, go through increasingly large fluctuations, or might the early manifestations of instability lead to a change in expectations and, hence, alter or negate the preconditions themselves? If subjects were put into a simulated market, could existing theoretical models be used to explain the prices and quantities observed over a sequence of periods? The foregoing questions suggest the motivation behind this paper. Our interest here will focus on theoretical models that have assumed a supply-response lag. In other words, these models characterize markets in which an interval of time elapses between the date at which suppliers decide how much to produce or make available for sale and the date at which the supply actually becomes available. The models to be considered also assume that once the supply comes onto the market the price adjusts immediately to the existing demand so as to clear the market. The key difference between the different theoretical formulations arises in the assumption about the current supply response to what has occurred in the past. The first part of this paper deals with a number of models in the literature. The second part presents and discusses some pilot experiments in which subjects were asked to decide on a quantity to supply before the selling price was known. In the third part the experimental data are used to test hypotheses taken from the theoretical literature.