Comment and Discussion

Abstract

Mark J. Warshawsky:I Pension accounting is a timely and important topic, because of the specific issue of measuring pension funding adequacy and because of the broader issue of financial accounting accuracy. (Of particular interest is the subissue of the appropriate discount rate to be used in calculating the pension liability.) Knowledge of these subjects has spread beyond the experts and into the broader public. The importance and relevance of this paper's results are further heightened by potentially even broader implications going beyond pension accounting to government budget accounting and Social Security reform. My comment will summarize the paper and offer some questions and suggestions to challenge the analysis. The paper's main result is that the market mistakenly focuses on smoothed pension earnings, or expenses (as argued by the opaque hypothesis), rather than on net pension assets, or obligations (as argued by the transparent hypothesis), in establishing the effect of a defined-benefit pension plan on a firm's value. The analysis is based on Compustat and I/B/E/S data from 1993 to 2001 on the firms included in the S&P 500 from December 1996 through 1998. The authors primarily use a residual income model with both nested and nonnested tests of the transparent versus the opaque hypothesis. Pension earnings are defined on an after-tax, per-share basis net of pension service cost, where the after-tax measure is hard-wired, that is, set by the authors' own formula. Net pension assets are defined as the market