Date of Award

Summer 7-14-2014

Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Mattias Nilsson

Second Advisor

Sanjai Bhagat

Third Advisor

Roberto Pinheiro


Using a large panel data set, this paper studies the relationship between corporate governance and defined benefit plan management in the United States. I show that poorly governed firms, proxied by log of median director ownership and five other governance variables, are more likely to sponsor defined benefits (DB) plans. Using a Heckman selection model I show that these poorly governed firms have more pension assets and more pension liabilities both on aggregate and on a per employee basis, and they fund their defined benefit plans better. Poorly governed firms also tend to assume higher rates of return on their pension assets, as well as higher discount rates on their future pension obligations. Poorly governed firms are also more likely to sponsor DB plans that are funded better than the median plan. These firms tend to operate in union-intense industries and usually in more concentrated markets. Better governed firms, in turn, have more growth in their pension assets and liabilities as a result of path dependencies. I further examine CEO characteristics related to these firms and find that firms are more likely to sponsor defined benefit plans when their CEO salaries are higher and the CEO total compensation is higher. The data also shows that lower CEO ownership in the firm increases the likelihood of sponsoring a defined benefit pension plan, in line with previous literature (Pagano and Volpin (2005)) showing CEOs with lower ownership and rights on cash flows direct more cash towards employees. This evidence is in line with Bertrand and Mullainathan's (2003) Quiet Life argument, and shows that entrenched CEOs value non-pecuniary benefits as theorized by Jensen and Meckling (1976) and supported empirically by Cronqvist et al (2009).