Do Managers of U.S. Defined Benefit Pension Plan Sponsors Use Regulatory Freedom Strategically?

• Date: 01 December 2017
• Author: MICHAEL KISSER,JOHN KIFF,MAURICIO SOTO
• DOI: http://doi.org/10.1111/1475-679X.12182
• Published date: 01 December 2017

DOI: 10.1111/1475-679X.12182

Journal of Accounting Research

Vol. 55 No. 5 December 2017

Printed in U.S.A.

Do Managers of U.S. Defined

Benefit Pension Plan Sponsors Use

Regulatory Freedom Strategically?

MICHAEL KISSER,

∗JOHN KIFF,

†AND MAURICIO SOTO

†

Received 7 June 2016; accepted 13 July 2017

ABSTRACT

We use historical particularities of pension funding law to investigate whether

managers of U.S. corporate defined benefit pension plan sponsors strategi-

cally use regulatory freedom to lower the reported value of pension liabili-

ties, and hence required cash contributions. For some years, pension plans

were required to estimate two liabilities—one with mandated discount rates

and mortality assumptions, and another where these could be chosen freely.

Using a sample of 11,963 plans, we find that the regulated liability exceeds

the unregulated measure by 10% and the difference further increases for un-

derfunded pension plans. Underfunded plans tend to assume substantially

∗Norwegian School of Economics; †International Monetary Fund.

Accepted by Douglas Skinner. The views expressed herein are those of the authors and

should not be attributed to the IMF, its Executive Board, or its management. We thank an

anonymous referee, Stephen Boyce, Benedict Clements, Frank Eich, Daniel Folkinshteyn,

Alan Glickstein, David Gustafson, Allison Hodge, Gregorio Impavido, Laura Kodres, Olivia

Mitchell, Theo Nijman, S. Erik Oppers, Mike Orszag, Konrad Raff, Otto Randl, Rina Ray,

Francisco Santos, Irina Stefanescu, Ian Tonks, and Amy Viener for invaluable feedback. Fi-

nally, we are grateful to seminar participants at the 8th Longevity Risk Conference, the Amer-

ican Economic Association, the Austrian Working Group for Finance and Banking, Bocconi

University, the Conference on Financial Economics and Accounting, the Eastern Finance As-

sociation, Eri-Ces at University of Valencia, the European Finance Association, the Netspar

International Pension Workshop, the Norwegian School of Economics, the Southern Finance

Association, the Swiss Finance Institute, and the Vienna Graduate School of Finance. An

online appendix to this paper can be downloaded at http://research.chicagobooth.edu/

arc/journal-of-accounting-research/online-supplements.

1213

Copyright C, University of Chicago on behalf of the Accounting Research Center,2017

1214 M.KISSER,J.KIFF,AND M.SOTO

higher discount rates and lower life expectancy. The effect persists both in

the cross-section of plans and over time and it serves to reduce cash contribu-

tions. We further show that plan sponsor managers use the freed-up cash for

corporate investment and that credit risk is unlikely to explain the finding.

JEL codes: G23; G39; J32; M48

Keywords: defined benefit pension plan; managerial decisions cash man-

agement; investment

1. Introduction

The Employee Retirement Income Security Act of 1974 (ERISA) estab-

lished minimum sponsor contribution standards for private industry de-

fined benefit (DB) pension plans. Despite subsequent rounds of legislation

aimed at further ensuring adequate plan funding, plan sponsor managers

still kept considerable leeway in calculating contributions. That is, funding

law contains provisions that allow DB pension plan sponsor managers to

use less stringent actuarial assumptions and thereby reduce funding gaps

and required contributions.

Giving more freedom to plan sponsor managers involves a difficult trade-

off. Temporary funding relief for underfunded pension plans provides

breathing room that might restore the long-term viability for plan sponsors.

But allowing for lower pension contributions means that both credit risk

and longevity risk (the risk of retirees outliving their financial resources)

are shifted from shareholders to pension plan participants and ultimately

to the Pension Benefit Guaranty Corporation (PBGC) and the taxpayer.For

well-funded and financially healthy pension plans, this risk may be negligi-

ble. However, it may become more relevant if underfunded pension plans

are more likely to opt for such a funding relief.

In this paper, we investigate whether managers of U.S. corporate DB pen-

sion plan sponsors strategically use regulatory freedom to their own bene-

fit. We tackle this question by exploiting historical particularities of pension

funding law, which we describe below. We first provide detailed evidence at

the pension plan level, before moving to an analysis of the plan sponsor

where we account for potential conflicts of interest between shareholders

and management.

We focus on a large sample of pension plans using data provided by the

U.S. Department of Labor (DOL), which contains detailed information on

various actuarial assumptions, including the mortality tables and discount

rates used to estimate pension plan liabilities. Our main analysis is based

on a sample of 11,963 U.S. corporate DB pension plans over the period

from 1999 to 2007. In addition, we provide a detailed investigation of 657

plan sponsors as well as a further robustness check of the Moving Ahead for

Progress in the 21st Century Act (MAP-21) funding relief episode in 2012,

focusing on another sample of 5,452 plans.

PENSION PLANS AND REGULATORY FREEDOM 1215

During the 1999–2007 period, the Internal Revenue Service (IRS) re-

quired plan sponsor managers to employ two different liability concepts

for calculating required sponsor contributions: a current liability (CL)and

an accrued liability (AL) measure. These liability measures coexisted over

the sample period and they affected pension contributions differently: ALs

were used to compute the normal level of contributions to the pension

plan, whereas the CL measure was the basis for additional top-up contri-

butions for significantly underfunded plans. Discount rates and mortality

tables used for the CL calculation were imposed by legislation, yet plan

sponsors were given more discretion for ALs. These historical particulari-

ties allow us to investigate whether differences between the two liabilities

and their actuarial assumptions vary across plans. Furthermore, because we

compare two measures from the same pension plan at the same point in

time, we are able to control for many plan-specific factors that would be

unobservable otherwise.

We find that, on average, ALs would have to be increased by 10% in

order to keep up with the regulated CL measure. Most of the difference

stems from using higher discount rates. On average, corporate DB pension

plans employed AL discount rates that exceeded the regulated measure by

approximately 170 basis points. In addition, we find that a subset of pension

plans made substantially lower and outdated life expectancy assumptions.

While these results suggest that managers employ regulatory freedom

when setting actuarial assumptions, the mere existence of such descrip-

tive differences is insufficient to also claim strategic behavior. Our analy-

sis further shows that plan funding is systematically related to the actuarial

assumptions chosen. Underfunded plans employ substantially higher dis-

count rates (relative to the mandated rate) and are more likely to assume

lower life expectancy (relative to the regulated mortality table). These re-

sults hold when controlling for several plan-specific factors as well as year-,

industry- and plan-fixed effects. We document a similar time-series effect as

changes in funding levels are negatively correlated with changes in relative

discount rates. Taken together, these findings suggest that underfunded

plans seem to stretch actuarial assumptions in order to reduce the reported

value of pension liabilities.

Crucially, this appears contrary to the regulations under which actuarial

assumptions should not be related to a plan’s funding status. The results

suggest a degree of opportunistic behavior by pension plan sponsor man-

agers, as the lower reported value of pension liabilities translates into a

substantial reduction in cash contributions. We show that using regulatory

freedom to report low AL values (relative to the CL measure) reduces cash

contributions by 6%–8% of plan assets.

These results complement the documented opportunistic behavior relat-

ing to pension accounting decisions, earnings management, and reporting

requirements (Blankley and Swanson [1995], Amir and Benartzi [1998],

Godwin [1999], Bergstresser, Desai, and Rauh [2006], Chuk [2012], Jones

[2013]), and indicate that attempts to take advantage of the leeway to set