Differential Interpretation of Information in Inflation Forecasts

We test the hypothesis associated with a standard assumption in the theoretical literature on learning: that economic agents interpret information identically. We use a data set based on a survey of Israeli business executives forecasting future inflation. One of the main advantages of using this data is that a major change in the inflation regime in 1985 can be treated as a natural experiment in new beliefs formation. We develop a methodology for testing this hypothesis and find evidence that is inconsistent with the identical-interpretation hypothesis, but is consistent with the proposed alternative.