Some Monte Carlo Results on Nonparametric Changepoint Tests

For long periods since 1982, core inflation has behaved as if it were generated by a process with a fixed mean and serially independent error term. Nonparametric changepoint tests proposed by Pettitt (1979) and Lombard (1987) suggest that since 1982, changes in core inflation have been infrequent and rather abrupt. However, little is known about the small-sample properties, the power of the tests, or the robustness of changepoint tests when a series is not i.i.d. This paper uses Monte Carlo analysis to investigate the probabilities of false positive tests under alternative assumptions about the time-series properties of the underlying process.