Three essays on macroeconomic and econometric analysis of business cycle

This dissertation is a collection of three self-contained papers on macroeconomics and econometric analysis of business cycle. The first paper deals with an optimal monetary policy issue when there exists a preferential tax treatment for owner-occupied housing. The finding of the paper is as follows: when the Ramsey government considers the tax discrimination as given, the Friedman rule of setting the nominal interest rate to zero is not necessarily optimal, even though preferences are homothetic and separable. That is, if the government cannot impose the same tax rate on owner-occupied and rental housing, which is the optimal policy in a complete tax system with homothetic preferences, an inflation tax can minimize the welfare loss.; The second paper focuses on statistical inference and errors in variables problem when economists make use of the Business Survey Index (BSI). The BSI is used as a leading indicator of the business cycle and as a direct measure of the expected value in testing the rational expectation hypothesis. However, since the BSI is an outcome of survey data, the users face a judgmental problem given a small change in BSI between periods and the errors in variables problem in the econometric analysis. This paper suggests two methods for statistical inference: the sign test and the bootstrap test. As an alternative to dealing with the errors in variables problem, a method combining the non-parametric bootstrap method and the Consistent Adjusted Least Squares (CALS) estimation method is proposed.; The third paper explores a methodology applicable to generating a regional economic activity index. There have been some attempts to develop a comprehensive index using either the coincident composite index model or the dynamic factor model. This paper suggests a modified principal components regression method: using the Chicago Federal Reserve Bank National Activity Index (CFNAI) and local economic indicators, local and national factors are extracted by applying the partitioned regression and the principal components method. From these results, a local economic activity index is generated through combining national components with local components by a regression method. Three experimental local indices for the Chicago Metropolitan area are produced.