Gambling with House Money

In the United States, defined benefit pension plans total more than $2 trillion in liabilities. Recent research shows that the size of defined benefit plans and managers' autonomy in characterizing them to capital markets make pension accounting a fertile area for earnings manipulation.

In the recent study "Earnings Manipulation, Pension Assumptions, and Managerial Investment Decisions," University of Chicago Graduate School of Business professor Joshua D. Rauh, along with Daniel Bergstresser and Mihir Desai of Harvard Business School, examine how manipulating earnings using pension accounting is linked to CEO incentives.

The authors find that some managers adjust pension assumptions in order to boost earnings at opportunistic times for the firm and for themselves. This manipulation has a significant impact on pension fund asset allocation.

"Increasing the assumed rate of return on pension assets is a very precise tool for manipulating earnings, because you can immediately create income," says Rauh. "The sheer simplicity of this method of altering reported income is very striking."

Although defined contribution plans such as 401(k) plans are the dominant form of retirement savings today, before the 1980s the only pension plans available to a firm's employees were defined benefit plans. Under a firm-sponsored defined benefit plan, a company sets aside retirement funds on behalf of an employee and makes investment decisions about those funds. For firms with defined contribution plans, employees choose their level of investment.

A firm sponsoring a defined benefit plan has a liability equal to the present value of all future payments due an employee's usually a function of the employee's age, years of service at the firm, and salary. This liability is funded with pension assets, which are to be managed in the interest of the employee beneficiaries.

A firm's pension assets are a mixture of stocks, bonds, and alternative securities. The firm reports an assumed return on pension plan assets on its income statement rather than a realized rate of return, because firms want to insulate annual earnings from year-to-year fluctuations in the marke t. Managers enjoy significant discretion in setting the assumed return used for the calculation of pension costs on the firm's income statement. Theoretically, the assumed and actual rate of returns will converge in the long run.

Consider a firm with $1 billion of operating assets, generating $50 million per year in reported income. If the firm has $100 million of pension assets, and the expected rate of return on these assets is 10 percent, the firm can book $10 million worth of income. As long as the pension fund earns a 10 percent average return over the long run, such reporting is fine. However, if firms adjust the assumed rate of return in response to their operating profits and investment opportunities, the situation becomes tenuous.

The authors find that the assumed rate of return is higher and more likely to be adjusted upward at firms whose earnings are more sensitive to changes in the assumed rate of return.

The results paint a reasonably consistent picture: managerial opportunism is a factor in determining assumed returns on defined benefit pension plan assets. In settings where the return assumption has a larger impact on earnings, managers make more aggressive assumptions. This finding is consistent with firms balancing the costs of reporting aggressive return assumptions with the benefits that come from increased reported earnings.

Managers are more aggressive with their assumptions as they acquire other firms, issue equity, and exercise stock options,indicating a link between managerial reporting decisions and managerial investment decisions.

Pension Sensitivity

Given that the assumed rate of return on pension assets can affect reported income, the authors needed to determine if such assumptions are opportunistic. The authors used a data set from Compustat Executive Compensation that provided information on firm nonpension income, nonpension assets, pension fund size, pension liability size, and assumed returns on pension assets from 1991 to 2002. The final data set consisted of 2,172 firms.

The authors obtained pension fund asset allocation data from an annual survey conducted by Pensions and Investments covering asset allocations of the largest U.S. pension funds from 1991 to 2002 as well as IRS 5500 filings. Firm acquisition data are taken from the Securities Data Company database. Compustat provided information on whether CEOs exercise their stock options between 1992 and 2002.

"It was necessary to first build a statistical model that would measure how sensitive a firm's earnings are to changes in the assumed rate of return for pension plans," explains Rauh.

If the firm's pension assets are large relative to its operating earnings, the manager has an economically meaningful opportunity to manipulate reported earnings. The authors measured pension sensitivity using a ratio of pension plan assets to operating earnings with the goal of understanding how sensitive reported earnings would be to the assumed rate of return. This measure of "pension sensitivity" indicates how much managerial incentive there is to raise assumed returns in an opportunistic manner.

Rauh, Bergstresser, and Desai identified situations when changing the assumed rate of return is particularly advantageous to managers, such as mergers, periods when the firm is near critical earnings thresholds, and when managers are exercising or being granted stock options.

Managers may want to raise reported earnings to boost the price of stock that might be used as currency in the acquisition of other firms in addition to generating bargaining power in the bidding process. The authors find that firms make particularly high return assumptions in periods leading up to the acquisition of other firms. If a firm is able to raise the value of its shares by increasing the assumed return on pension plan assets, it will be able to acquire other firms at less cost.

Changes to assumed returns are more likely in situations where increases to the assumed returns can be pivotal in helping firms cross these critical earnings thresholds. Taken together, the evidence indicates that managers employ assumed returns opportunistically and this opportunism interacts significantly with individual and firm financial and investment decisions.

Increasing Risk

Rauh, Bergstresser, and Desai tested whether firms increased their risk exposure,the amount of stocks held versus bonds, in their pension funds to justify reporting higher expected rates of returns. A firm is supposed to choose an expected return on its pension plan assets that reflects the actual risk balance in its portfolio. Auditors also want to see that assumed returns reflect the risk level of pension fund assets. If a firm changes its asset allocation from a ratio of 50/50 for stocks and bonds to 60/40, for example, one would expect to see a higher rate of return.

The authors wanted to determine if firms were increasing the portion of stocks in their asset portfolio to justify stating higher earnings, or if the stated earnings reflected greater risk-taking in the firm's pension fund investments. The authors find that firms chose riskier assets to justify stating higher earnings.

The authors find that a 0.25 percentage point increase in assumed returns is associated with a 5 percentage point increase in stock allocation.

"It is surprising that firms are making decisions about the asset allocation of their pension plans in order to justify manipulating earnings," says Rauh. "The investment decisions of the firm become distorted by this desire to report higher earnings even though the firm is supposed to invest on behalf of its employee beneficiaries."

To further assess the role that opportunistic managerial behavior plays in setting assumed returns, the authors focused on CEO option activity. Using data on executive compensation, pension assets, and assumed returns, the authors investigated CEO option exercise and grants. The authors find that managers use high assumed returns if high returns are particularly efficient ways of managing earnings as they approach critical financial and investment decisions. Such opportunistic behavior also may be more evident when changes in return assumptions can produce earnings that would help managers meet critical earnings thresholds.

These results illustrate how managerial actions can redistribute value among current shareholders, managers, and potential shareholders. If managers inflate profits and thus stock prices and use this inflated stock as currency in the acquisition of other firms, then current shareholders could benefit from a redistribution of wealth to them from future shareholders. However, the authors also note that assumed returns on pension assets are substantially higher at firms where current shareholders have weaker control over managers. Managerial opportunism in pension fund decision-making does not appear to be guided by shareholder interest.

Who Pays the Price?

Manipulating pension plan earnings benefits some constituents of the firm and hurts others. Benefits will accrue to people who gain from short-term increases in the stock price, including investors with short-term horizons and managers who want to exercise their stock options. Investors who intend to hold on to the stock for the longer term are likely to bear the cost of this activity.

Ultimately the employee beneficiaries also share in the ownership of the pension assets. If a firm takes excessive risks with its pension plans in order to report higher earnings for the benefit of managers, and there is a large equity market downturn and potential bankruptcy for the firm, that will have major effects on the pension of a worker who has been with the firm for many years.

"While the U.S. government does guarantee employee pensions up to a certain amount, this insurance would not cover a pilot or auto worker who has been on the job for 25 or 30 years," says Rauh.

"There are thousands of firms in the United States that have not been able to meet the obligations of their defined benefit plans, and as a result, firms have gone bankrupt and dumped their pension obligations on the government," notes Rauh. " If a company takes large risks with its pension funds and ends up in bankruptcy at a time when those assets have performed poorly, then employees will get much smaller pensions."

Rauh adds that it is important to learn about the consequences of this form of earnings manipulation in order to b e tter understand conflicts of interest between short-term and long-term shareholders.