Pension Assumptions and Earnings Manipulation

12/01/2004
Featured in print Digest

The evidence indicates that as managers prepare for acquisitions and for exercising their options, they have an increased incentive to produce higher earnings and share prices -- and increase their assumed rates of return in order to do so.

In Earnings Manipulation and Managerial Investment Decisions: Evidence from Sponsored Pension Plans (NBER Working Paper No. 10543), co-authors Daniel Bergstresser, Mihir Desai, and Joshua Rauh identify a simple way to manipulate firm earnings: by manipulating the assumed rates of return on the firm's pension assets. They further show that such manipulation is linked to CEOs' incentives, and that firms change investment decisions both to justify and to capitalize on this type of earnings manipulation.

Many firms have pension plans that are large enough to allow them to substantially increase reported earnings in the short run by changing the assumed long-term rate of return for the pension assets they manage for their workers. Those managers who determine that manipulating the rate-of-return assumption can boost their firms' stock price, as was apparently the case during the 1990s, have strong incentives to set this long-term rate of return assumption opportunistically, particularly as they undertake mergers and approach option vesting periods.

In their study, Bergstresser, Desai, and Rauh investigate the degree to which managers are opportunistic with these assumed returns. The researchers evaluate the extent to which choices on assumed returns intersect with the managers' own option exercises and with their firms' merger activities. Bergstresser, Desai, and Rauh create a measure of the sensitivity of a firm's overall profits to the assumed long-term rate of return on pension assets. They show that this sensitivity measure is an important determinant of the levels of, and the changes in, assumed rates of return. Specifically, a firm whose pension assets are twice as large relative to its operating income as the median firm in the sample makes a long-term rate-of-return assumption that is, on average, approximately 10 basis points higher than the median. A firm in the 90th percentile of sensitivity, on average, has a long-term rate-of-return assumption that is 40 basis points higher that a firm in the 10th percentile. Such differences in return assumptions can have a significant impact on reported earnings for these firms.

The researchers further show that firms make particularly high return assumptions in periods leading up to the acquisition of other firms. This relationship is especially strong for firms whose reported income is the most sensitive to pension assumptions. Indeed, assumed long-term rates of return are approximately 30 basis points higher for firms that are acquiring other firms.

Asset allocation within pension plans is another investment decision that may reflect earnings manipulation. Instrumental-variables analysis suggests that managers increase equity allocations to justify their higher assumed rates of return on pension assets. Large equity allocation in most firm pension plans remains a persistent puzzle. The authors note that their analysis suggests that the interaction of managerial opportunism and pension accounting may help to explain part of this puzzle, as managers increase equity allocations to justify a rate-of-return assumption. The study concludes by showing that managers who are the least constrained by their shareholders appear to be the most aggressive with their rate-of-return assumptions. Indeed, the evidence suggests that the earnings manipulations being examined do not benefit shareholders.

The researchers allow that other factors likely may affect firms' long-term rate-of-return assumptions. A variety of tests for these alternative explanations do not detract from the basic results concerning the opportunistic nature of these changes. To the degree that pension earnings are capitalized into market prices, they say, the opportunistic use of assumed rates of return led to aggregate levels of overvaluation. Their data on asset allocation add another mechanism by which pension accounting could have contributed to market overvaluation, as higher assumed rates also appear to be associated with higher equity allocations. While market participants were capitalizing pension earnings, the researchers point out, firms were increasing equity exposures to justify those very pension earnings.

-- Matt Nesvisky