| Inflation persistence is one of the phenomena which modern macroeconomic theory based on the New Keynesian (NK) model has not seemed to explain. In the last decade, macroeconomists poured a large amount of effort to resolve the discrepancy between the theory and reality, and some of them accused the convention of assuming rational expectations (RE), not the NK model itself, of the discrepancy.;In fact, economic agents' expectations play a crucial role in determining the dynamics of economic variables. We can hardly expect to accurately analyze the forward-looking economy without realistically modeling agents' expectations. This dissertation seeks to provide a meticulous investigation into the inflation dynamics and the effect of monetary policies when a cognitively consistent model is employed for agents' expectations.;The first chapter (co-authored with Wei Xiao) considers two information limitations on learning agents: they cannot observe fundamental shocks and they only possess finite samples of past endogenous variables. They use statistical methods to decide their optimal belief function as well as to estimate the optimal parameters. Under this assumption, we show that the economy goes to an equilibrium which is generally very different from the RE equilibrium (REE) and that the NK model exhibits intrinsic inflation persistence that would not exist in the REE.;The second chapter traces the source of inflation persistence. We define model learning and parameter learning as two separate elements in agents' expectation formation process, and we analyze how much each element contributes to inflation persistence. We quantify such contributions using a standard general equilibrium model with nominal rigidity.;The last chapter explores desirable monetary policies particularly on the aspect of inflation persistence. Under Taylor-style interest rate rules, we search for the policy parameter range in which the economy converges to an equilibrium and the level of inflation persistence is maintained to be low. We find that, for some realistic belief functions, obeying the Taylor principle is no more sufficient and far more aggressive monetary policies, in response to inflation or inflation expectations, are required. |
