The Role of Defined Benefit Pension Plans in Executive Compensation

Abstract

Although defined benefit (DB) pension plans constitute a significant portion of both annual compensation and firm-related wealth for many CEOs, prior studies of CEO compensation contracts generally exclude these plans from their analyses due to lack of data. Taking advantage of recently increased disclosure requirements, we analyse the role of DB pension plans in these contracts. We find that firms with CEO DB pension plans grant the CEO annual compensation that is larger than predicted by economic determinants. We also find that more powerful CEOs, identified by their extraordinary pension benefits, receive higher total pay in addition to the pension benefits. We find no evidence that CEO pension benefits contribute to the pay-for-performance sensitivity of the annual pay.

Notes

1. Frydman and Jenter (2010) claim that the evidence on CEO pensions is ‘even sparser’ than the limited evidence on executive perquisites.

2. Securities and Exchange Commission Item 407 of Regulations S-K and S-B mandated increased disclosure requirements for executive and director compensation effective for fiscal years ending after 14 December 2006.

3. According to the Employee Benefit Research Institute, of those private sector US employees with a retirement plan, 84% had a DB plan in 1979 either alone or together with a defined contribution plan. In 2008, the comparable percentage was 33%. Nonetheless, approximately 70% of the S&P 500 firms still provide at least some portion of their employees a tax-qualified DB pension plan and more than 80% of these firms provide a nonqualified plan to their CEO.

4. McKesson Proxy Statement, SEC Schedule 14A, dated 21 June 2013.

5. Bebchuk and Jackson (2005) predict that ‘the new SEC disclosure rule … can be expected to reveal large amounts of performance-insensitive compensation via executive pensions’.

6. In 2003, the European Union Commission recommended that all EU member nations require their companies to disclose individual executive compensation (Murphy, 2013).

7. This is according to the ‘2011 Manifest and MM&K Total Remuneration Study’.

8. Say on pay rules have been adopted by Australia, Belgium, Canada, Denmark, Germany, Italy, the Netherlands, Norway, Portugal, Spain, Sweden, and the UK.

9. Several EU countries (including the UK Ireland, Belgium, France, Germany, Italy, Netherlands, and Sweden) as well as Norway have adopted more extensive CEO pay disclosure requirements reflecting increased interest by both investors and regulators in the issue (Murphy, 2012). The UK has imposed even stricter disclosure regulations for executive compensation, including pensions, effective from October 2013.

10. See Gustman, Mitchell, and Steinmeier (1994) for a survey of the literature.

11. If the non-qualified DB plan is funded and the funding is protected from the firm's creditors, then the benefits are taxable to the beneficiary at the time of funding.

12. However, Bebchuk and Jackson (2005) cite evidence that ‘firms going through Chapter 11 reorganization often assume, in full, the company's obligations toward executives promised a defined benefit pension even when paying only part of the claims of financial creditors’ (p. 831).

13. Examples of recent work that excludes executive DB pension plans in their analyses of executive compensation include Murphy (2013); Fernandes et al. (2013); and Conyon et al., (2011).

14. Edmans and Liu (2011) show analytically that forms of compensation such as pensions represent inside debt rather than equity and are an optimal form of contracting.

15. Bebchuk and Fried (2004b) maintain that executive DB pension plans constitute ‘stealth compensation’ due to the lack of disclosure.

16. The number of firms for 2006 is smaller than the numbers for the other years because the 2006 COMPUSTAT year is defined as June 2006–May 2007 and the disclosures were not required until December 2006 resulting in a partial year for 2006.

17. Results are similar when we substitute sales as the measure of firm size.

18. For example, Dechow and Sloan (1991) find that firms grant greater proportions of annual compensation in equity as the CEO approaches retirement to extend the CEO's investment horizon in the firm.

19. These results contrast with those of Gerakos (2010) who finds evidence that CEOs with pension benefits are compensated less on other dimensions of pay. Our study differs from his because we focus on a period when the pension benefits are disclosed in the proxy statement, whereas his sample relies on estimates of value as in Sundaram and Yermack (2007), which have been shown to have measurement error.

20. Because the annual increment in most DB plans is tied to either three- or five-year average cash compensation, we expect some association with salary by construction.

21. Please refer to the Online Supplement for the tabulated results.

22. Please refer to the Online Supplement for the tabulated results.

23. Not all firms refer to their executive pension plans with these terms, but the conditions and provisions are consistent with the following descriptions.

24. The term Supplemental Executive Retirement Plan (SERP) is inconsistently applied to both restoration and supplemental retirement plans. SERPs can take many forms and the term is used rather loosely. SERPs technically refer to nonqualified plans that apply to a limited group of senior executives (they are often referred to as Top Hat plans) and are used for broader purposes than the restoration plans. Some proxy statements refer to the restoration plans as SERPs. We avoid using the acronym SERP because of its lack of precision.

25. For example, Home Depot CEO Robert Nardelli was the only person in the firm to be granted a DB pension plan. As another example, OfficeMax terminated its DB plans in 2004 for all employees except the CEO.

26. Please refer to the Online Supplement for tabulated results and further discussion.

27. Kalyta and Magnan (2008) and Goh and Li (2010) assume that executive DB pension plans provide discretionary compensation for Canadian and UK companies respectfully.

28. Gerakos (2008) reports that based on analysis of his proprietary database, only 10% of firms with either type of trust for their CEO DB pension plan reported such publicly. However, the great majority of firms in our sample report that the plans are unfunded regardless of whether they specifically have a trust.

29. ‘Closed’ DB plans are those that do not permit employees hired after a specified date to participate in the plan; employees hired prior to the cut-off date continue to accrue benefits. ‘Frozen’ plans no longer accrue benefits for participants after a specified date, but continue in existence in order to pay out the benefits already accrued. These terms are not consistently used by firms, but the concepts apply to all firms.

30. The SEC issued its proposal for the increased proxy disclosures in January 2006 and adopted the changes in July 2006. The proposal was anticipated for a number of months prior to its release. We consider the period 2005–2007 as appropriate for gathering data on firms closing or freezing their executive DB plan potentially in response to the new disclosure requirements.