摘要：
This dissertation consists of three essays in applied econometrics. Chapters two and three focus on atheoretical data analysis using new nonlinear time series techniques, and chapter four uses an application of rational expectations theory and the theory of cointegration to develop a test for the existence of a "corporate veil". Chapter 2 explores the applicability of general fractional ARIMA models to the study of aggregate labor market variables. Two frequency domain estimation techniques for these models are applied. Significant fractional integration is found for the series studied, suggesting that the imposition of integer integration assumptions for economic time series is misguided. In chapter three, an alternative nonlinear time series model is introduced, which is found to be useful for the study of the cyclicality of macroeconomic variables. This chapter demonstrates that much of the disagreement over time concerning the cyclicality of the real wage can be explained, using linear time series techniques and sample splitting, by the instability of the relationship between the growth rates of real wages and output. A Markov switching model, is developed which allows for formal statistical treatment of the question of cyclicality and allows for variables to experience both pro and countercyclical regimes. Application of the filter sheds new light on the cyclicality of employment, hours, and wages. Chapter four reexamines the implications of changing corporate savings, testing for the presence of a "corporate veil". There it is argued that previous tests for such a veil lacked a proper focus, identifying influences of corporate saving on private saving which are consistent with a complete piercing of the corporate veil. Traditional evidence suggesting the existence of a corporate veil is found to be sensitive to the particular assumptions researchers have made in the past and only weak evidence of the veil is found. This dissertation illustrates, how the applic
欢迎来到三峡大学图书馆！
