Learning rational expectations in a policy game (Q805487)

The author is concerned with the question whether the process of learning rational expectations will have important effects on the time profile of optimal policy. The process of learning rational expectations starts at the beginning of a government's term of office when the public is discovering the government's propensity to create inflationary surprises. Comparing two different types of learning - rational and nonrational learning - the author concludes that if rational learning is assumed there will be low levels of initial inflationary bias in government policy and the government slows the public's learning. If learning is nonrational by applying weighted least squares, there will be a permanently reduced inflationary bias and thus, there will be a reduction in the credible sustainable level of inflation.