Date of Completion

2-15-2012

Embargo Period

2-15-2013

Keywords

Executive Compensation; Agency Theory; Pensions

Major Advisor

Assaf Eisdorfer

Associate Advisor

Carmelo Giaccotto

Associate Advisor

John Phillips

Field of Study

Business Administration

Degree

Doctor of Philosophy

Open Access

Open Access

Abstract

This dissertation consists of three essays examining issues related to executive inside debt on firm risk, dividend policy, and compensation structure. In the first essay, we use a hand-collected executive pension database to study how both CEO and non-CEO executive compensation structures affect the overall risk of a firm. We extend the research of Sundaram and Yermack (2007) to non-CEO executives for the first time, demonstrate how the difference in compensation leverage between CEO and non-CEO executives is directly related to firm risk, and find that funding these pensions via a Rabbi Trust eliminates most of the risk-shifting effects.

In the second essay, we show that (i) dividend yield and dividend payout ratio are significantly lower when manager compensation relies more on pension payouts; (ii) given a general payout policy, managers will prefer the form of stock repurchase over cash dividend distribution; and (iii) the negative effect of pension on dividend is significantly weaker when the pensions are protected in a pre-funding rabbi trust. These findings provide support to the manager-owner agency theory.

In the third essay, we recalibrate the Dittmann and Maug (2007) principal-agent model with a pension factor to determine the new optimal structure of executive pay. Using a hand-collected data set of 828 executives from 141 firms, we calculate the optimal piecewise linear contract. This study provides a significantly refined answer to the original paper, and furthermore, finds little justification for high levels of pension compensation. Finally, we find that the pensions drive a substantial amount of contract mispricing among CEOs, but not for non-CEO executives.

Download
COinS