A simple method to account for measurement errors in revealed preference tests

Revealed preference tests are widely used in empirical applications of consumer rationality. These are static tests, and consequently, lack ability to handle measurement errors in the data. This paper extends and generalizes existing procedures that account for measurement errors in revealed preference tests. In particular, it introduces a very efficient method to implement these procedures, which make them operational for large data sets. The paper illustrates the new method for both classical and Berkson measurement errors models.