forum.jpg (4424 bytes)     "Inside  every small problem is a large problem struggling to get out."

Rules Forum Contributors [For contributors only]


Experimental Economics
General Equilibrium
Other Topics
Prisoners Dilemma
Zero Sum Games


Thread and Full Text View

Ask a question about: Experimental Economics
Respond to the question: What have we learned about the theory?

05/27/2011 04:53 AM by name withheld; plase help
i have a problem whit this task Trading Company should provide 5 types of goods. p - income from sale l - a loss of storage. To determine the optimum allocation of supply n=5 p1=p2=p3=p4=p5=40 l1=19 l2=18 l3=17 l4=16 l5=15. I
[View full text and thread]

01/04/2000 01:10 PM by David K. Levine; How does theory work in the laboratory?

Conventional economic theory posits a "Nash" equilibrium in which rational self-interested players have rational expectations about one anothers behavior. Does this theory hold up in the laboratory?

  1. There is overwhelming evidence [Plott, 1987] that after subjects have had a chance to learn how to play the theory works well in "competitive environments." These are settings in which no player can have much impact on another player's welfare. From a theoretical perspective, these environments satisfy Ostroy's no surplus condition [Friedman and Ostroy 1991]. From a practical point of view, these are the settings in which economists have traditionally looked for competition - many buyers or sellers competing with each other to buy or sell the same products. Because there are many buyers or sellers, no one buyer or seller matters very much.
  2. The theory works less well when players can have an impact on their opponents. Occasionally players engage in altruism, spite or other forms of non-self-interested behavior. However, it is important to emphasize the word occasionally. In practice 70% or more of the subjects do engage in rational self-interested behavior. One interpretation of equilibrium - that of an exact equilibrium may fair poorly, because the behavior of a small number of players may matter a great deal to the majority. However, more sophisticated notions of equilibrium, such as that of approximate equilibrium [Fudenberg and Levine, 1997] or quantal response equilibrium [McKelvey and Palfrey, 1995], that allow for deviant behavior by some individuals, do predict well.
  3. Theories of equilibrium that come about because of internal calculation by players do poorly. Equilibrium behavior is typically observed only after players have had some opportunity to learn and adapt. All learning and adaptive theories [Camerer and Ho, forthcoming] give a reasonable description of the basic groping towards equilibrium that is observed in the laboratory, but none seem to do especially well on the details.
  4. The size of the stakes matter. The more money is at stake, the better the theory does [Slonim and Roth, 1998].

Although there are theories that are consistent with what happens in the laboratory, the "tightness" with which these theories make predictions depends on the environment - in some environments theory is useful, in others it is not. So despite the general success of theory in the laboratory, there is a need for theories that make tighter predictions.


Plott, C.R. [1987]: "Laboratory Experiments in Economics: The Implications of Posted-Price Institutions," Science 232: 732-738.
Friedman, D. and J. M. Ostroy [1991]: "Competitivity in Auction markets: An experimentl and theoretical investigation," UCLA, Department of Economics Working Paper No 633
Slonim, R. and A. E. Roth [1998]: "Learning in High Stakes Ultimatum Games: An Experiment in the Slovak Republic," Econometrica, 66: 569-596.
[Manage messages]