The Information Content of Corporate Pension Funding Status in Japan

• Author: Shingo Goto,Noriyoshi Yanase
• Published date: 01 July 2016
• DOI: http://doi.org/10.1111/jbfa.12197
• Date: 01 July 2016

Journal of Business Finance & Accounting

Journal of Business Finance & Accounting, 43(7) & (8), 903–949, July/August 2016, 0306-686X

doi: 10.1111/jbfa.12197

The Information Content of Corporate

Pension Funding Status in Japan

SHINGO GOTO AND NORIYOSHI YANASE∗

Abstract: This paper tests if a firm’s pension funding ratio (pension assets/PBO) reveals the

management’s private information about the firm’soperation when the firm can exercise discre-

tion in pension funding. The lax enforcement of pension funding rules and the prevalence of

management forecasts make Japanese firms an ideal testing ground. We show that, among firms

with large business uncertainty, large accruals, or high effective tax rates, the pension funding

ratio predicts the firm’s management forecast errors significantly beyond conventional control

variable and the effects of pension accounting management. However, the stock market does

not appear to incorporate this information immediately.

Keywords: corporate pensions, pension funding, management forecasts, market efficiency,

cross-sectional predictability,japanese stock market, pension accounting management, measure-

ment errors

1. INTRODUCTION

A firm sponsoring a defined benefit retirement plan (‘DB plan’ or ‘pension plan’ for

short) has a contractual obligation to pay a stream of post-retirement benefits to its

employees. To ensure the payments of the pledged benefits, the firm is responsible

for contributing liquid assets to its pension plans (pension assets) that may only be

used for the plan beneficiaries. In practice, pension funding status – the value of

pension assets relative to the expected present value of pledged benefits (called the

‘projected benefit obligations’ or ‘PBO’) – varies widely across firms and pension

plans, depending on the pension assets’ historical investment performance and the

sponsoring firms’ pension funding policies. We call a pension plan underfunded

(overfunded) when the value of pension assets is lower (higher) than PBO.

Recent literature has shown that a firm’s pension funding status can have large

impacts on the firm’s growth and hence on its equity value. For example, a large

loss in pension plans, often resulting in large pension underfunding, has negative

effects on the firm’s future earnings through loss amortization. Moreover, firms with

large pension deficits are typically required to make mandatory contributions to the

∗The first author is from College of Business Administration, University of Rhode Island and the second

author is from Faculty of Business Administration, Tokyo Keizai University; address: 1-7-34 Minami-cho,

Kokubunji, Tokyo1858502, Japan. (Paper received October 2014, revised revision accepted February 2016).

Address for correspondence: Shingo Goto, College of Business Administration, University of Rhode Island,

7 Lippitt Road, Kingston, RI 02881, USA.

e-mail: shingo goto@uri.edu

C

2016 John Wiley & Sons Ltd 903

904 GOTO AND YANASE

underfunded pension plans, which can inhibit their growth by constraining invest-

ment opportunities (e.g. Rauh, 2006). Mandatory contributions can also increase their

cost of external financing for firms facing financial constraints (e.g. Campbell et al.,

2012). Although these effects are known to market participants, pension accounting

information is notoriously complex and difficult to interpret.1Several recent studies

demonstrate that even professional equity analysts in the US have difficulty in

processing pension funding information efficiently (e.g. Coronado and Sharpe, 2003;

Picconi, 2006; Coronado et al., 2008; and Chen et al., 2014). Franzoni and Mar´

ın

(2006) further find that small firms with large pension deficits tend to be overvalued in

the US, and attribute this result to investors’ lack of attention to underfunded pension

plans.2Knowing this difficulty, firm managers also take advantage of complex pension

accounting items to manipulate earnings (e.g. Bergstresser et al., 2006).

This paper proposes a new channel through which a firm’s pension funding status

predicts its operating performance and stock returns. While most existing studies focus

on the effects of large pension deficits on the firms’ prospective growth through loss

amortization and mandatory pension funding requirements (e.g. Franzoni and Mar´

ın,

2006; Rauh, 2006; Franzoni, 2009; and Nakajima and Sasaki, 2010), we hypothesize

that, in the absence of strictly enforced pension funding requirements, a firm’s

pension funding status signals the management’s private information about the firm’s

operation. This is because DB plan sponsors face a trade-off between tax benefits and

financial flexibility associated with pension funding (e.g. Harrison and Sharpe, 1983;

and Bicksler and Chen, 1985). While some firms aim to fund their pension plans as

fully as possible to maximize the tax benefits, other firms have incentives to keep their

pension plans as underfunded as possible to secure financial flexibility. Consequently,

a firm’s optimal pension funding decision depends on the firm’s ability to fund its

projects through business operations or lack thereof (e.g. Stone, 1987; Petersen, 1992;

and Cooper and Ross, 2002). We therefore posit that a firm’s pension funding status

reveals the management’s inside forecast of the firm’s operating performance.

The implications of optimal pension funding theories depend crucially on pension

funding regulations and firms’ disclosure practice. The implications are not neces-

sarily easy to test in the US for two reasons. First, a firm’s pension funding policy is

severely constrained by a stringent system of mandatory contributions (Rauh, 2009).3

Second, it is not easy to separate the effects of the management’s ‘private’ information

from those of publicly-available information, especially when the management has

incentives to manage its operating performance measures. This concern can be

potentially mitigated if firm managers disclose their own forecasts of operating

performance measures to the public, but only a small fraction of US firms release

management forecasts (e.g. Mensah et al., 2008).

On the other hand, Japan provides an ideal ground to test the proposed implica-

tions for the following reasons. First, pension funding rules are enforced much less

1 Early literature documents that the stock market treats pension assets/liabilities quite fairly as corporate

assets/liabilities (e.g. Daley, 1984; Dhaliwal, 1986; Landsman, 1986; and Gopalakrishnan and Sugrue,

1993). Meanwhile, Landsman and Ohlson (1990) document that the US stock market displays a general

underreaction to underfunded pension liabilities.

2 Nakajimaand Sasaki (2010) report similar results in a sample of Japanese firms between 2000 and 2003.

3 In the US, the system of mandatory contributions is established by the Employee Retirement Income

Security Act of 1974 (ERISA) and strengthened in subsequent legislation, the Pension Protection Act of

2006 (PPA) in particular.For US firms, the effects of mandatory contributions have been the main focus of

the recent literature (e.g. Franzoni and Mar´

ın, 2006; Rauh, 2006, 2009; and Franzoni, 2009).

C

2016 John Wiley & Sons Ltd

CORPORATE PENSION FUNDING STATUS IN JAPAN 905

strictly in Japan than in the US. For example, Japanese firms can keep their pension

plans underfunded (or even ‘unfunded’ in book reserves) for a long time without

being penalized. Japanese firms can even forfeit the pledged retirement benefits (even

the vested ones) under certain conditions (‘pension cuts’). In addition, unlike in the

US where only a quarter of public firms have DB plans, almost all firms traded on

the Tokyo Stock Exchange (our sample universe) maintain DB plans, which mitigates

concerns over sample selection bias. Second, the corporate disclosure practice of

Japanese firms is also very different from that of US firms. In particular, management

forecasts are ‘effectively mandatory’ in Japan, and almost all Japanese public firms

release point estimates of sales and earnings of the on-going fiscal year, among others.

In principle, the release of management forecasts should mitigate the information

asymmetry between firm managers and outside investors. If management forecasts

were fully rational and truthful, there would be no information asymmetry be-

tween the managers and outside investors about the firms’ business prospects, and

the firm’s pension funding status would not reveal incremental information about

the firm’s business operation beyond the management forecasts. The degree of the

resolution of information asymmetry, however, depends on the degree to which the

management forecasts reflect the managers’ true private information. For example,

Xu (2009) shows that management forecasts often fail to incorporate information in

prior forecast errors. To the extent that management forecasts do not fully reflect

its true inside information, variables that are correlated with the inside information

should predict management forecast errors. If the pension funding ratio signals the

management’s inside information, it may help predict management forecast errors.

The prevalence of management forecasts in Japan allows us to test this empirical

prediction.

To test our empirical prediction, we first define the ‘pension funding ratio’ – the

value of pension assets minus PBO, scaled by PBO – as a measure of the firm’s pension

funding status, following Francis and Reiter (1987) and Rauh (2006, 2009), among

others. Our evidence from multivariate cross-sectional regressions (with the procedure

of Fama and MacBeth, 1973) indicates that the pension funding ratio makes the

management’s private information effectively public by front-running management

forecasts, particularly within the group of firms with large business uncertainty, low

disclosure quality (large accruals), or high effective tax rates.4We obtain three major

findings to support this interpretation.

First, among the firms with large business uncertainty, large accruals, or high

effective tax rates, the pension funding ratio significantly predicts the management’s

forecast errors of sales and earnings in the following two years, after controlling for

conventional financial characteristics, labor unions, industry fixed effects, and the

management’s choice of pension accounting items. Second, the predictability of man-

agement forecast errors by the pension funding ratio carries over to the predictability

for stock returns. Third, the stock return predictability of the pension funding ratio

4 To examine the significance of the proposed informational effects of pension funding, we explicitly

control for the effects of pension accounting management (e.g. Bergstresser et al., 2006; Hann et al., 2007).

Specifically, we include the following three pension accounting choice variables: (i) the assumed discount

rate for the calculation of PBO; (ii) the assumed expected return on pension plan assets; and (iii) the

ratio of unrecognized pension liabilities to PBO. These variables may also reflect the management’s private

information about the firm’s business prospects. In this paper, however, we use these pension accounting

choice variables purely as control variables.

C

2016 John Wiley & Sons Ltd