Essays in financial econometrics

Greiner, Christopher (2024) Essays in financial econometrics. PhD thesis, London School of Economics and Political Science.

Text - Submitted Version Download (4MB)

Abstract

This thesis contains three chapters developing and applying novel techniques in financial econometrics. In the first chapter, I show that to construct factor models for the cross-section of expected returns, principal component factors should be selected based on risk premia or Sharpe ratios – rather than eigenvalues, as is predominantly done in the literature. This is because principal component factors’ prices of risk are given by their risk premia divided by eigenvalues. I show that selection based on risk premia (Sharpe ratios) minimizes the sum of squared pricing errors (Hansen-Jagannathan distance) for a factor model and demonstrate empirically that the proposed selection methods lead to substantial in- and out-of-sample improvements. Further, I devise a test to determine the number of factors to approximate the stochastic discount factor. In the second chapter, co-authored with Tommaso Mancini-Griffoli, Christian Julliard, and Kathy Yuan, using an equilibrium network model and a large international panel of cross-border trade, we analyse empirically the drivers of foreign currency invoicing. First, we find evidence of strategic complementarity in currency invoicing across countries. Second, key players for a given currency are also countries that are central to the international trade network. Third, we find evidence of natural hedging, between the choices of export and import currencies. Fourth, in counterfactual analysis, we find that the position of the USD is inherently fragile. In the third chapter, I estimate risk preferences through nonparametric methods from option data. The proposed estimator is shown to be consistent and asymptotically normally distributed. The estimated risk preferences are more in line with preferences implied by classical utility functions than other studies suggest. Specifically, formal statistical tests suggest that there is no statistically significant evidence supporting the pricing kernel puzzle. In contrast, constraining estimated risk preferences to be monotonically decreasing improves the associated beliefs’ forecasting performance substantially.

Item Type:	Thesis (PhD)
Additional Information:	© 2024 Christopher Greiner
Library of Congress subject classification:	H Social Sciences > HF Commerce H Social Sciences > HG Finance
Sets:	Departments > Finance
Supervisor:	Julliard, Christian and Martin, Ian
URI:	http://etheses.lse.ac.uk/id/eprint/4783

Actions (login required)

Record administration - authorised staff only