Public choice and the development of modern laboratory experimental methods in economics and political science

Soon after I developed conditions for Pareto optimality in a world of public goods, I discovered that similar conditions had been developed years earlier by Frisch (1959).

Downs (1957).

Black (1958), Black and Newig (1998).

Plott (1967a). This paper also contains the results regarding unanimity that first attracted my attention to the problems.

Plott (1967b).

While elements of the “fundamental equation” are evident in early writings of Public Choice, its importance as a foundation element in the development of experimental methods was only becoming recognized as laboratory experimental methods developed. See, Plott (1979). Several papers related to the development of this period are reprinted in Plott (2001).

The demonstration by Grether and Plott (1979) that preference theory could be rejected by “preference reversal” experiments performed in psychology expanded the study to include a variety of preference forms together with the possibility that preferences might be endogenous. The Grether and Plott study was an important step in expanding the study because it clearly demonstrated that preference theory as found in economics is a rejectable theory, as opposed to tautological and thus, placed the body of theory on solid scientific footing.

Fiorina and Plott (1978).

Induced value refers to the use of money to induce preferences for an abstract set of options. The resulting preferences over abstract options become parameters for models applied to the choice from the options.

Williams (1973) attempted to expand the method to multiple units but could do so only through the use of a special trading process. Similarly, a 1973 Purdue dissertation by Harvey Reed attempted to study the two unit case but inadvertently found it necessary to change the trading process. The induced preference method had not been generalized to deal with multiple markets and certainly not with complements and substitutes among variables. The issue can be made clear through a comparison of preference inducement in the market experiments of Smith (1962) and Smith (1964) with the generalization placing value on marginal changes associated with multiple units in a single market introduced at Plott and Smith (1978), or the generalization to preference interdependence in multiple markets (Forsythe et al. (1982).

The approach is influenced by Bayesian methods in the sense that it is not meaningful to reject a theory without having an alternative. However, the absence of a theory of error structure presented a special challenge.

The Fiorina and Plott experiment tested sixteen competing behavioral models within the single experimental setting. Several of the theories were found in the social psychology and sociology literature. Others were found in the political science literature and still others were found in the Public Choice and Economics literatures.

The initial experiments were all funded by an earlier NSF grant to C. Plott. By the spring of 1973, many of the Fiorina and Plott experiments were completed and the research was focused on new directions.

The case of non-equilibrium was not studied until after September 1973 when the first agenda experiments were conducted. Both Mo and I wanted to do the non-equilibrium environment but could not find a justification in terms of an understanding for what would be learned. In frustration, Mo asserted “If we move the equilibrium and if the data just follow the maximum of the interior person, it would be very embarrassing.” That comment together with the agenda theory, which had just been exposed by the flying club exercise of Plott and Levine (1978), supplied a theory. If the agent in the center proposed his/her maximum at some point, a plausible agenda step exits that could lead to the point. Thus, the agenda experiments provided a theory about what might be expected if the equilibrium did not exist. It was the justification we were seeking and the experiments were conducted. Interestingly, exactly why we were excited about the research was not obvious to everyone. Vernon Smith arrived at Caltech in the fall of 1973 and, after observing what Mo and I we were doing, asked me in a somewhat rhetorical tone, why we were doing such research, which was obviously dramatically different from what had taken place in economics and made little sense to him at the time. Mo and I knew that we were going to have a difficult time explaining what we were doing to a very skeptical audience. Vernon’s comment suggested that it would be harder than we anticipated. Except for those very close to public economics and public choice the source of excitement and curiosity was not obvious.

Berl et al. (1976).

McKelvey and Ordeshook (1978). See also, Laing and Olmsted (1978).

Early experiments pitted a Von Neumann–Morgenstern solution against the core. The issue was whether the VM solution captured data that the core would not. In particular, the experiments are asking if coalition theories had predictive power over equilibrium theories. If the committee operates by rules similar to Robert’s Rules, the answer is no.

See Levine and Plott (1977) and Plott and Levine (1978).

The closed rule was first studied experimentally in Isaac and Plott (1978).

See Levine and Plott (1977) and Plott and Levine (1978).

Plott (1983). The paper was circulated as: Social Science Working Paper 180, California Institute of Technology, 1978. As was typical of experimental papers in those days it took years to get research published.

Bohm (1972).

Dawes (1975), Dawes et al. (1977).

Marwell and Ames (1979), Marwell and Ames (1980), Marwell and Ames (1981).

Ferejohn et al. (1979).

Plott and Smith (1978); Smith (1979); Smith (1980).

In essence, the needed conditions were already known from the behavior of markets with externalities. A reduction in consumption or production of an external diseconomy is a contribution to the public good of external diseconomy reduction. Such contributions were not forthcoming in the externality experiments of Plott (1983) so the issue just turned on understanding the relationship between a positive contribution to the public good and the restraint of making a contribution to a public bad.

Isaac et al. (1985), Kim and Walker (1984).

An earlier policy application in economics was a paper by Hong and Plott (1982) who conducted a study on rate posting for the U.S. Department of Transportation. That study had an effect on the policy but the DOT did not publish the study due to the fear that Senator Proxmire would grant the DOT his “Golden Fleece” award for wasteful spending. Interestingly, the long delay in the publication of this first application is not an isolated example of the difficulties getting early experimental work published.

Grether et al. (1979). The report was subsequently published as a book: Grether et al. (1989).