Econometric analysis of heterogeneity in financial markets using quantile regressions

Abstract

This dissertation studies heterogeneity in financial markets using quantile regressions.

The first chapter develops uniform confidence bands for the linear quantile regression estimator in a time series setting. The confidence bands are important for estimating the precision at different quantiles of the conditional distribution. The inference procedure is carried out through bootstrapping and allows for serially correlated error terms. An empirical application to the relationship between stock returns and investor sentiments suggests the method can be informative.

The second chapter analyzes the heterogeneity of firm characteristics on returns to capital. It develops a theoretical model under a utility maximization framework with imperfect insurance and credit markets constraints. From the model, the returns to capital are derived as a function of the parameters, which affects the production function of the firm and the entrepreneur utility form. Quantile regression is applied to analyze the field experiment data from the Sri Lanka Micro Enterprises Project (2005-2010). Empirical evidence shows that returns vary across different quantiles of firm profits. Further, the ability/risk aversion of entrepreneurs affect the returns differently at different quantiles.

The third chapter examines capital account liberalization and its effect on price volatility in the Chinese housing market. The chapter assesses the extent to which: a) short-term capital flows and foreign direct investment may have impacted prices and volatility in the Chinese housing market; and b) whether 2006 Capital Account Regulations on foreign purchases of Chinese real estate were effective in reducing the level and volatility of prices in the housing market. The results show that hot money magnified the impacts of capital flows on housing prices during upward surges in housing prices. Quantile regression provides quantitative evidence that the more volatile the housing market was, the larger the impact short-term capital flows had on accentuating such volatility. Furthermore, the 2006 CAR continued to have a strong impact on reducing volatility in the Chinese housing market during the period under study.