CHEATING EXPERIENCE: WHAT EXPERIMENTAL SAVINGS GAMES TEACH ABOUT SOCIAL LEARNING AND DECISION SKILLS

Experience is a teacher. Economists sometimes defend their assumptions about rational human behaviour by resorting to just this folk wisdom. But some decision situations occur infrequently over the course of our lifetimes, or are simply very difficult to learn through personal experience. How can we learn in such cases?

By cheating, say Professors Parker Ballinger, Michael Palumbo and Nathaniel Wilcox, writing in the October 2003 Economic Journal: we can look at what others have done. Using experimental saving games, they show that people can acquire decision skills by observing other decision-makers.
Their important findings and potential implications:
" Decision skills can accumulate over sequences of people. Thus, social learning of decision skills can take place.
" Social learning has long been the bedrock of theories of cultural evolution. Yet empirical investigation of its effectiveness in the accumulation of decision skills is new.
" Social learning of decision skills is highly uneven. This may contribute to poverty and inequality. But decision skills may be teachable, providing a new weapon for battling poverty and inequality.

Ballinger, Palumbo and Wilcox's savings games are simple. A subject receives an uncertain income of experimental funny money called 'francs' at the start of each of 60 consecutive decision periods. This is added to any francs she may have saved in past periods, and she chooses how much of this to spend now and how much to save for future periods (no borrowing against unknown future income is allowed). Spending francs now buys 'points,' but each extra franc spent buys fewer extra points in each period. The total points purchased over all 60 periods determine her real currency earnings from game play.

Playing this saving game well is hard. Finding the rational saving scheme for the game is almost impossible without using a computer. No simple algebraic expression describes it. But a subject in the experiment can observe another subject who is currently playing this saving game - which she will soon begin herself. In this manner, collective experience may accumulate through sequences of subjects.

In fact, 'first generation' subjects (who begin without the benefit of watching others) make quite poor decisions, as in many previous experiments on saving decisions. But more than half of third generation subjects (the last subjects in a session, who benefit from observation of the second subject and, indirectly, from whatever the second subject learned by watching the first subject) perform better than their first generation counterparts, while few perform worse.

Aside from the fact that social learning about saving takes place, Ballinger, Palumbo and Wilcox emphasise its highly uneven nature. First generation subjects are all highly myopic, planning ahead three periods at most. But in the third generation, half of subjects plan ahead three or more periods, and a few subjects plan ahead far enough to perform nearly as well as the rational saving scheme would. But nearly half of third generation subjects learn nothing from their antecedents, and perform every bit as poorly as first generation subjects do.

Could the unevenness of social learning about saving have anything to do with poverty and inequality? The experiment cannot answer this question, but it does suggest the possibility. Usually scholars think of these things as the result of the normal functioning of labour and insurance markets, supplemented by unequal endowments of valued job skills. This work raises the possibility that unequal endowments of decision skills may also be at work.

These skills can be learned, and so perhaps they can be taught - and this suggests that educational tools might be available and helpful in the fight against poverty and inequality.

ENDS

Notes for Editors: 'Precautionary Savings and Learning Across Generations: An Experiment' by T. Parker Ballinger, Michael G. Palumbo and Nathaniel T. Wilcox is published in the October 2003 issue of the Economic Journal.

Ballinger is in the Department of Economics and Finance, Rusche College of Business

Stephen F. Austin State University, Nacogdoches, Texas.
Palumbo is in the Division of Research and Statistics, Board of Governors of the Federal Reserve System, Washington, DC.
Wilcox is in the Department of Economics, University of Houston, Houston, Texas.

For Further Information: contact RES Media Consultant Romesh Vaitilingam on 0117-983-9770 or 07768-661095 (email: romesh@compuserve.com); Nathaniel Wilcox on +1-713-743-3840 (email: nwilcox@mail.uh.edu); Parker Ballinger on +1-936-468-1878 (email: pballinger@sfasu.edu); Michael Palumbo on +1-202-452-2296 (email: mpalumbo@frb.gov).