A revealed preference test of rationing a Monte Carlo analysis

This paper develops and evaluates a rationing linear programming procedure that uses an “efficiency index” to allow for violations of revealed preference to be attributed to “almost” optimal choices. The procedure detects rationing using U.K. data. Various Monte Carlo simulations are performed to evaluate the ability of the procedure to differentiate between violations of revealed preferences caused by random error and those caused by rationing. •New rationing test uses Varian's efficiency measure to test data that violate GARP.•Binding rationing is detected for food and other services in the U.K.•Monte Carlo simulations show that LP rationing test has power against measurement error.