ON TESTING THE UTILITY HYPOTHESIS

In order to be able to conduct a test of the (core) utility hypothesis that is not confounded with tests of (subsidiary) hypotheses that eco nomic agents all have the same preferences and that their preferences are weakly separable, it is necessary to use data that are disaggregat ed and complete. In order to carry out a utility hypothesis test with complete data, revealed preference theory must first be extended. The generalised axiom of revealed preference is extended to include labour supply and portfolio choice. Then a complete, disaggregated data set is used to test the utility hypothesis and two preference separability hypotheses.