Combining brain and behavioral data to improve econometric policy analysis

From brain imaging data one obtains new economic theory with economically relevant policy implications. In particular, one would like to use data from a small number of subjects in an imaging experiment to predict a larger population’s response to proposed policy changes. This paper develops a method by which one can combine behavioral and imaging experiments’ data to provide empirical content to economic theory and improve econometric policy analysis. I develop probability bounds on the behavioral effects of a policy change, and show that these bounds depend on the probability that certain brain activation patterns are present. Moreover, I show that these activation probabilities can be estimated from a combination of behavioral and imaging experiments, so long as decisions in the behavioral experiment are sufficiently dependent on the activation patterns of interest.