Session Report: Market Dynamics
This session, chaired by Malcolm Brady, contained three papers dealing with dynamic modeling in microeconomics and institutional economics. The growth (and collapse) of a firm and an industry was a common theme of the three contributions, but the technical approaches selected by the authors were substantially different.
The first presentation, “Advertising effectiveness and spillover: simulating strategic interaction using advertising,” by Malcolm Brady, presented a dynamic model of two firms competing using different advertising policies. The model is a dynamic extension of an analytical, Cournot duopoly model. The latter model is known to be static, producing equilibrium values of quantities sold and unit price of a product. After advertising is introduced in the model and the impact of advertising on demand is endogenously modeled, the model becomes dynamic, but also analytically intractable. Simulation is therefore used in order to gather insight in the model behavior. In his research, Brady analyzed the model behavior for a number of competitive scenarios, with some interesting observations. Small changes in model parameters can alter significantly the firm and industry behavior. High levels of advertising effectiveness lead to reinforcing loops dominating while low levels make balancing loops dominate. The reinforcing loop dominating one firm leads to explosive growth for that firm and decay for the rival firm, ultimately driving the rival out of business. Positive spillover (cooperation) reduces the impact of advertising, negative spillover (predation) increases the impact of advertising.
The second presentation, “Self-organizing markets,” by Fernando Buendía, presented a dynamic model of self-organizing markets characterized by a highly skewed distribution of firms’ size. One way to explain the behavior of such markets is by considering increasing returns to the growth of a firm. Multiple sources of increasing returns to the growth of the firm can be identified, including economies of scale, scope, integration, and expansion. The other increasing returns mechanisms, capable to offset scale diseconomies resulting from difficulties of managing large organizations, are related to technological progress. They include Schumpeterian learning, learning-by-doing, learning-by-using and demand-side increasing returns. In his model, Buendía considers all of these sources and their combined effect and suggests the use of urn theory or Pólya processes for formalization of the general concept of market self-organization via increasing returns. He comes up with a generalized urn model that can explain combinations of negative and positive feedback, jumps and other types of perturbations. Specifically, he extends the classical models to the cases where multiple dependent or independent urns with several additions or withdrawals and several colors are considered. Generalized urn models appear powerful enough to model complex dynamic behaviors such as the formation of industrial clusters.
The final presentation in this session, “Non-equilibrium industry dynamics with knowledge-based competition: an agent-based computational model,” by Myong-Hun Chang presented a model of the evolutionary dynamics of an industry subject to knowledge-based competition with entry and exit. Empirically, many industries exhibit a common pattern of growth characterized by the initial rise in the number of producers, followed by a sharp decline and convergence to a stable level. The pattern is accompanied with the increasing but slowing output and declining price. In order to give an endogenous explanation to the pattern, Myong-Hun Chang models the production process as a system of inter-dependent activities, the firm as an adaptive entity whose survival depends on its ability to perform various activities with greater efficiency than its rivals, and the industry as a population of myopic agents decide on adoption of new technologies by following simple decision rules. These rules are combined with the rules governing the entry and exit behaviors of the potential entrants and the incumbent firms, respectively. The resulting model is not only capable of reproducing the empirically observed regularities, but it yields a very flexible platform to perform comparative dynamics analyses, in which the impacts of various parameters on the resulting industry dynamics can be examined.