ECONOMIC LEARNING AND SOCIAL EVOLUTION : SUMMARY OF FINDINGS




This multi-disciplinary project, involving economists, psychologists and social anthropologists, seeks to understand the detail of the relationship between individual rationality and cultural norms, paying particular attention to the possibility that people, and societies, may learn from their mistakes and by observing others. It developed into the ESRC-funded Centre for Economic Learning and Social Evolution.

For more details contact:

Professor Kenneth Binmore, Centre for Economic learning and Social Evolution, University College, London, WC1E 6BT, UK. Phone: 0+171 380 7864; Fax: +171 916 2775.

Key Points from the Research

1. For the first time, experimental evidence was found to support the existence of Hamilton's rule in humans.

2. The conventional wisdom on evolutionary equilibrium selection in game theory was shown to require modification. With evolutionary models that incorporate noise in a more realistic fashion, the equilibrium selected need not satisfy any of the requirements of the orthodox refinement literature. Nor need the riskdominant equilibrium be selected as in the work of Young or Kandori/Mailath/Rob.

3. Pursuant to the previous point, the theories studied generated results that are closer to those observed experimentally than conventional theories. This is true, in particular, of the widely discussed Ultimatum Game..

4. Useful data on expected waiting times in evolutionary models was gathered.

Summary of Research Results

The multidisciplinary programme is described under four headings. Most of these research projects are ongoing but documentary evidence is available on all of them on request.

1. Socioeconomic models of interactive learning

a. Musical chairs: modelling noisy evolution.
In collaboration with Larry Samuelson, Binmore and Vaughan examine a model in which noise appears not only as random mistakes or mutations, but the risk-dominant equilibrium will be selected in the ultralong ruins found not to hold.

b. Conventions and social mobility.
In collaboration with Binmore's student, Giovanni Ponti, Seymour generalizes a model of Rosenthal and Landau in which the equilibrium played in a simple two-player bargaining game favours the player with higher social status. The latter increases or decreases as a result of randomizing when two players of equal status happen to play each other.

c. Cultural evolution in a structured population.
In collaboration with Chris Lowen, Dunbar studies the rate of meme replication in structured populations. The type of structure turns out to be a major determinant of the rate of spread, with extinction an ever-present risk. A survey supplemented the work by gathering data on the frequency with which subjects were in contact with relatives of various degrees and their best friends.

d. Learning to be imperfect: the Ultimatum Game.
Binmore, Samuelson and Gale use computer simulation to argue that interaction learning will not necessarily lead subjects to be subgame-perfect equilibrium of the Ultimatum Game. The results provide a possible explanation of anomalous experimental results.

e. Replicator dynamics in the Ultimatum Game: a formal analysis.
Binmore and Seymour supplement the previous paper by studying the problem mathematically.

f. Evolutive equilibrium selection.
Vaughan employs the Fokker-Planck equation to analyze the models used by Young and Kandori/Mailath/Rob in studying evolutive equilibrium selection. Traditional techniques sometimes yield more general results more easily.

g. Evolutionary drift.
Binmore and Samuelson commence a study of the effects of evolutionary drift on long-run equilibrium selection in a variety of canonical games.

2. Experiment and analysis of data

a. Hamilton' rule in humans
Dunbar and Plotkin carried out a successful psychological experiment designed to test the existence of Hamilton's rule in humans. As predicted by the rule, subjects were willing to work harder the closer their degree of relationship to the person to be rewarded for the work.

b. Hard bargains and lost opportunities.
In collaboration with Joe Swierzbinski, Binmore and Samuelson analysed the results of a bargaining experiment with outside options. The results call into question the applicability of the Coase theorem.

c. Two-stage bargaining games and the Centipede.
Under the direction of Binmore and Samuelson, Giovanni Ponti conducted extensive simulations of interactive learning processes in two-stage bargaining games and in the Centipede Game. In both cases, results were much more in line with the experimental data than the theoretical predictions.

3. Learning and rationality

a. The logic of backward induction.
Binmore explains why common knowledge of rationality does not necessarily imply that the backward induction path will be followed.

b. Imperfect recall.
Binmore comments on the Paradox of the Absent-Minded Driver revived by Piccione and Rubinstein.

4. Applications

a. Price dispersion: an evolutionary approach.
In collaboration with Ed Hopkins, Seymour studies the evolutionary requirements needed to negate the law of one price.

b. A utilitarian theory of political legitimacy.
Binmore completes Chapter 2 of the second volume of Game Theory and the Social Contract. This chapter incorporates an evolutionary theory about the origins of inter-personal comparison.

c. Altruism towards relatives.
In collaboration with Catherine Keynetta, Dunbar conducted a survey following up the work on Hamilton's rule by studying the extent to which subjects were willing to undertake risky investments on behalf of relatives.

d. Social networks.
Dunbar continues his work comparing the size of the social networks maintained by primates and humans.

e. Intergenerational altruism and Ricardian equivalence.
Vaughan uses the distributional techniques of his other work to deflate the extravagant claims that Bernheim makes about the significance of his work on intergenerational transfers for macroeconomics.

f. Mobility, inequality and welfare.
Vaughan studies the influence of dynamic considerations on standard welfare measures.

Publications

Binmore, K.G., "Playing Fair", Game Theory and the Social Contract I, MIT Press, Cambridge, Mass., 1994.

Binmore, K.G., "Rationality in the Centipede", in R. Fagin (ed.), Theoretical Aspects of Reasoning about Knowledge: Proceedings of the fifth TARK conference, Morgan Kaufmann, San Mateo, California, 1994.

Binmore, K.G. and Samuelson, L., "An Economist's Perspective on the Evolution of Norms", Journal of Institutional and Theoretical Economics, 150, 45-71, 1994.

Binmore, K.G. and Samuelson, L., "Drift", European Economic Review, 38, 851-867, 1994.

Binmore, K.G., Samuelson, L. and Vaughan, R., "Musical Chairs: Modelling Noisy Evolution", Games and Economic Behaviour, 11, 1-35, 1995.

Binmore, K.G., Gale, J. and Samuelson, L., "Learning to be Imperfect: the Ultimatum Game", Games and Economic Behaviour, 8, 56-90, 1995.

Dunbar, R. and Spoors, M., "Social Networking, Support Cliques and Kinship", Human Nature, 6(3), 273-290, 1995.

Dunbar, R., Clark, A. and Hurst, N., "Conflict and Co-operation among the Vikings: Contingent Behavioural Decisions", Ethology and Sociology, 16, 233-246, 1995.

Dunbar, R., "The Mating System of Callitrichid Primates: I. Conditions for the Co-evolution of Pair Bonding and Twinning", Animal Behaviour, 50, 1057-1070, 1995.

Dunbar, R., "The Mating System of Callitrichid Primates: II. The Impact of Helpers", Animal Behaviour, 50, 1071-1089, 1995.

Dunbar, R., "Neocortex Size and Group Size in Primates: A Test of the Hypothesis", Journal of Human Evolution, 28, 287-296, 1995.


Link to Full List of Project Publications

Link to List of Programme Discussion Papers