Inflation targeting has become a popular monetary policy strategy during the last two decades. This has given rise to a lively debate about the empirical effects of the adoption of inflation targeting. Some influential empirical studies have argued that the apparent improved performance of inflation targeters is merely regression to the mean, and controlling for the initial condition they conclude that inflation targeting does not matter.

This paper challenges these findings that the apparent benefits of inflation targeting have basically been a mean-reverting mirage. It finds that these tests of the effect of inflation targeting have low power. It is shown analytically how they could fail to find any effect even if inflation targeting has in fact been highly effective. The low power of the tests is further illustrated using simulation results. As a result, prominent empirical findings that inflation targeting does not matter due to regression to the mean are misleading as the tests lack power to distinguish an oasis from a mirage.