"It's for teaching, not believing": Comments on teaching, learning, and problem solving through economic experiments

Journal of Agricultural and Applied Economics, Aug 2003 by Nelson, Robert G, Wilson, Norbert L W

This series of papers is an excellent opportunity to reacquaint agricultural economists in the Southern region with the exciting field of experimental economics and is indeed opportune in light of the recent awarding of the Nobel Prize in Economics to Vernon Smith, considered by many to be the father of experimental economics. Rather than try to share the limelight with the authors on their far-reaching and comprehensive topics, we plan to take this opportunity to share some of our views of the role of experimental economics in the research laboratory and classroom.

Problem Solving and Hypothesis Testing

In this session, Hudson argues that experimental economics offers a means to counter criticisms of economics as a science by addressing both the rigor of empirical tests and the quality and timeliness of data needed for policy prescriptions. He describes what experimental economics is and why we do experiments and suggests that experiments can be classified dichotomously according to whether subjects are paid or not. This, we believe, is not a particularly helpful distinction, because payment, and motivation in general, are better integrated under Smith's (1982b) criteria for a "valid, controlled microeconomic experiment," which we describe in a subsequent section. As an alternative, we suggest Davis and Holt's framework for classifying experiments in the dimensions of institutional and environmental complexity.

Institutional Complexity

Confusion often arises with the term institution, because of its association with the field of institutional economics, which is concerned with the ways that institutions evolve in response to individual incentives, strategies, and choices and how institutions affect the performance of political and economic systems. For experimental economists, the term institution simply refers to the rules governing economic interactions in an experiment. In a recent interview (Lynch and Gillespie), Vernon Smith emphasized: "The thing that's not very explicit in much of economics is what the rules of trading are and how they affect outcomes. Experimental economics asks how the performance of a market is influenced by its rules."

The rules that characterize institutions include:

(1) The nature and timing of messages allowed between agents (e.g., Is face-to-face communication allowed? Who can make offers? Are offers sequential or simultaneous?);

(2) the kinds of decisions or actions that are observable (e.g., bids, offers, contracts, forecasts, draws from an urn, side payments);

(3) the mapping of these decisions into the payoff or incentive structure (i.e., the reward a subject gets for making a certain decision); and

(4) stopping rules for ending the session or trading period (e.g., random, fixed, known, unknown).

Game theorists also have an interest in precisely specifying institutional complexity, but the domain available for exploration by experimentalists is much more extensive and often more expedient.

Environmental Complexity

The term environment may also cause some confusion, because of its association with natural resources, ecology, and environmental concerns. For experimental economists, the concept is again far more simple and refers to the structural characteristics of the economic setting, including:

(1) the number of agents (e.g., monopoly, duopoly, oligopoly, "many" buyers and sellers);

(2) their initial endowments (e.g., money, information, experience, market power, property rights);

(3) production technologies (e.g., supply schedules derived from production costs; single vs. multiple units; constant vs. varying costs; storable vs. nonstorable units);

(4) demand conditions and structures (e.g., demand schedules derived from redemption values vs. home grown preferences; stationary demand vs. cyclical or aperiodic shifts); and

(5) the number of periods for continued interaction (e.g., "one-shot" games vs. repeated contact; 10 vs. 200 rounds).

An example of an institution with a high degree of complexity is a voluntary assessment mechanism for funding a public good like a church. An institution of modest complexity is the Dutch auction, in which the price for an item falls sequentially until it is sold to the first buyer who makes an offer. An institution of very low complexity is the mandatory assessment mechanism of a commodity check-off program, which has the authority of the police powers of the state and is functionally equivalent to a simple excise tax. Recently, an increasing number of commodity boards have had their check-off rescinded and are showing a renewed interest in and appreciation for the complexities of voluntary assessment mechanisms.

Market environments of graduated complexity could range from a simple monopoly setting with a single seller facing a computer-simulated buyer population with a fixed demand curve (Nelson and Beil 1994), to an oligopoly setting with two to four sellers (Nelson and Beil 1995), to a market with four buyers and four sellers, each with multiple units, trading in either forward or spot markets with random supply and demand shocks (Menkhaus et al.).