Essays on Financial Econometrics: Analysis of Classical Problems with Novel Econometric Methods

The dissertation consists of two essays that apply nonparametric econometric tools to studying the financial market. The first essay, “Sentiment and Volume-Volatility Elasticity” is presented in Chapter 2 and explores the relationship between sentiment and market participants' trading activities around public news announcements. We develop a theoretical model under the short-sale constraint and the model predicts a nonlinear relationship between sentiment and the volume-volatility elasticity. We estimate parametric regression models and carry out nonparametric series estimation around important macroeconomic news announcements with high-frequency intraday trading volume and transaction price data of the S&P 500 E-mini futures contract. Empirical results not only corroborate predictions of the theoretical model, but also suggest varying effects of sentiment on the volume-volatility elasticity around announcements of different importance.

Chapter 3 presents the second essay, “Conditional Superior Performing Assets”. We utilize a novel functional test for conditional moment inequalities to select conditional superior performing assets (CSPA). We apply the CSPA test to evaluate the performance of U.S. mutual funds. The test is carried out with gross as well as risk-adjusted returns of domestic equity mutual funds bearing various investment objectives. By inverting the CSPA test for sets of benchmark assets, we obtain confidence sets for the uniformly most superior asset. Empirical results indicate superior performance of funds with various investment objectives over certain conditioning states of measures from various aspects of the economy. Additionally, the CSPA confidence sets can serve as criteria for fund selection. The usefulness of the CSPA test on fund evaluation is demonstrated through identifying significant conditional superior performance from assets that are indistinguishable unconditionally.