Essays on monetary and macroprudential policies with different models of expectation

[摘要] This thesis evaluates the design of monetary and macroprudential policies. Different models of expectation are examined to get a comprehensive understanding about the work of monetary and macroprudential policies. It is begun by assuming agents are boundedly rational under Recursive Least Square (RLS) and Stochastic Gradient (SG) learning, followed by fully rational agents under rational expectation (RE). When agents follow RLS learning, both determinacy and E-stability criteria are required to find preferred policies. We focus on the effect of habit in consumption in the design of preferred policies. We found the presence of habit in consumption enlarges both determinacy and E-stability region. The same methodology is then used in another model that features housing market and financial constraint. The result showed that a response to the growth of housing prices via the LTV rule may increase determinacy and E-stability. We also conduct a refinement in the design of preferred economic policies by incorporating SG-stability criteria, in addition to determinacy and E-stability. The result showed that central bank’s task gets more difficult since the Taylor Principle is insufficient to ensure a robust learnability of REE. In other works, we deviate from the assumption of boundedly rational agents and consider fully rational agents (RE). We examine the issue of monetary policy, banks’ lending decisions and business cycles in Indonesia. This thesis completes its analysis by evaluating the role of news in the formation of agents’ expectations.

[发布机构] University:University of Birmingham;Department:Birmingham Business School, Department of Economics

[关键词] D History General and Old World;DS Asia