Files in this item

FilesDescriptionFormat

application/pdf

application/pdfKe_Chen.pdf (25MB)
(no description provided)PDF

Description

Title:Essays in financial economics
Author(s):Chen, Ke
Director of Research:Bernhardt, Daniel
Doctoral Committee Chair(s):Deltas, George
Doctoral Committee Member(s):Bernhardt, Daniel; Bera, Anil K.; Bengtsson, Ola
Department / Program:Economics
Discipline:Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):semi-parametric competing risks models
corporate defaults and mergers
financial analyst recommendations
Abstract:My dissertation consists of two essays that investigate dynamic financial phenomena using semi-parametric competing risks models. The first essay analyzes the determinants of corporate defaults and mergers. The second essay investigates how sell-side analysts make recommendation revisions facing various incentives at different points in time. The first essay, Contagious Corporate Default, analyzes contagious corporate default - default by one firm that raises the likelihood of defaults by similarly-situated firms, while possibly lowering defaults of competitors. We use a semi-parametric competing risks model that incorporates non-linear relationships between corporate defaults/mergers and risk factors. We use two-dimensional penalized tensor-product splines to flexibly model default correlations among firms, including their time dependence. The direct impact of one firm's default on default intensities of other firms is quantified by the financial distance between them, as captured by the 12-month trailing correlation in their equity returns. We also allow for non-linear impacts of observable firm-specific and macroeconomic default risk factors and unobservable frailties that capture common unobservable macroeconomic risk factors. We document large intra-industry contagion effects in defaults and mergers. The contagious impact of a default takes time to cumulate, peaking at lags of about 90-120 days, before declining. The impacts of many financial and macroeconomic variables are highly non-linear: many only exhibit significant impacts on default once they are above or below some critical level. Our model better estimates the clustering of defaults and more precisely predicts firms' credit risks than existing parametric hazard models, delivering massive improvements in out-of-sample predictive power. My second essay, Hazardous Analysts: Reputation Management and the Duration of Recommendations, investigates how incentives to manage appearances for their customer audiences appear to drive analyst recommendation revisions, using a semi-parametric competing risks hazard model that accommodates both unobserved heterogeneity and the time-varying effects of covariates on analysts’ recommendations to uncover the determinants of sell-side analyst recommendation revisions. We find that analysts tend to keep past good “accurate” recommendations for too long, and to drop past bad “inaccurate” recommendations too quickly. That is, incentives to maintain a ‘good’ reputation - a good-looking recommendation list in front of customers - appear to aggravate rather than alleviate conflicts of interest. More generally, we exhaustively characterize the impacts of covariates on recommendation changes. For example, we find that the information content of earnings announcements is fully incorporated into recommendations within one week, and that the likelihood of revisions falls sharply with experience over an analyst’s first three years, but then plateaus. Using the predictive power of future returns as the criterion, we show that, overall, analysts appear prescient: downgrades (upgrades) are more likely when future returns are bad (good). However, the investment values of downgrades decline when analysts make their decision driven by past performance pressure.
Issue Date:2014-05-30
URI:http://hdl.handle.net/2142/49595
Rights Information:Copyright 2014 Ke Chen
Date Available in IDEALS:2014-05-30
Date Deposited:2014-05


This item appears in the following Collection(s)

Item Statistics