The paper attempts to determine whether persistent undershooting of the inflation target - particularly during a disinflation process - is a result of inflation avoidance preferences (IAP), namely asymmetric monetary policy with respect to inflation, or a hidden target that is lower than the announced one.

This analysis introduces a Linex-form central bank behavior function into a New Keynesian (NK) framework, which makes it possible to explicitly distinguish between the two hypotheses. Using data for Israel for the period 1994-2007, the model is empirically tested using three different regression techniques - OLS, Non-Linear OLS (NLS) and instrumental variables (IV) using the GMM method. It is shown that a hidden target was indeed adopted during the 1990s - the main period of the disinflation process - and that it equaled about 4 percent on average in comparison to the average announced inflation target of 8 percent. At the same time, an asymmetric policy was also conducted, thus suggesting that the policies - hidden target and IAP - had different motives, with the former perhaps being aimed at manipulating the public's expectations of inflation. During the 2000s, on the other hand, the evidence is inconclusive as to whether a hidden target or asymmetric policy was adopted, thus suggesting that one or the other was used to implement IAP policy. It emerges that when the CB plays a repeated game in determining monetary policy and the public goes through a learning process to understand central bank behavior, a hidden target that is lower than the announced one may be a practical way of manipulating expectations even under discretion.

The full article in PDF file