Collaborative Research: Social versus individual learning in life cycle consumption and saving

点击翻译按钮获取中文摘要

This project experimentally examines whether people can learn how to perform these tasks better through observation of, and/or advice from, other people. The hypotheses under scrutiny are motivated by a large literature in sociological and psychological disciplines concerning observational, social and cooperative learning on the one hand, and group processes and performance on the other. The tasks studied are drawn from the contemporary economic literature on consumption and saving. The purpose of the experiments is to determine: (a) Whether social learning can occur in these important economic settings; (b) what factors govern the occurrence and/or strength of this learning; and (c) how such learning compares to individual learning from individual repetition.Intellectual Merit: The life-cycle consumption and saving task (hereafter LCCS task) plays a central role in macroeconomic theory, and is also an intense object of study in several microeconomic areas. Experimental examinations suggest that people do not perform LCCS tasks optimally. And while it is possible that some individual learning and adaptation could take place with individual repetition of LCCS tasks in an experimental setting, this is of questionable relevance to LCCS tasks in the external world (the world outside the lab) since, by definition, they are only encountered once in their entirety over any agent's lifetime in the external world. However, people may learn from observation of (or advice from) other people who have performed (or are performing) similar tasks. That is, there is a possibility of social learning about LCCS tasks, even though individual learning through repetition is impossible. Similar remarks apply to many other important individual decision making tasks that are, by their nature, only encountered once in a lifetime. But few are as important, ubiquitous or complex as the LCCS task.Broader Impact: Experimentalists commonly build repetition of tasks into experimental designs. This is a questionable procedure whenever: (a) The decision problem under analysis is a unique, once-over-a-lifetime task in the external world, such as the LCCS task; and (b) the experimenter wishes her results to be externally valid. However, if social learning occurs for such tasks, and social learning compares well with individual learning, then the procedure gains in respectability since it may be regarded as a "laboratory substitute" for the social learning that would take place in naturally occurring settings. So this research will add useful knowledge to the methodological base of experimental economics.