CEO Compensation and Fair Value Accounting

The following post comes to us from Ron Shalev of the Department of Accounting at New York University, Ivy Zhang of the Department of Accounting at the University of Minnesota, and Yong Zhang of the School of Accounting and Finance at Hong Kong Polytechnic University.

In our paper, CEO Compensation and Fair Value Accounting: Evidence from Purchase Price Allocation, forthcoming in the Journal of Accounting Research, we investigate the influence of bonus intensity (i.e., the relative importance of bonus in CEO pay) and alternative accounting performance measures used in bonus plans on the allocation of purchase price post acquisitions. Upon the completion of an acquisition, the acquirer is required to allocate the cost of acquiring the target to its tangible and identifiable intangible assets and liabilities based on their individually estimated fair values. The remainder, namely, the difference between the purchase price and the fair value of net identifiable assets, is recorded as goodwill. The recognition of goodwill has different implications for subsequent earnings than that of other assets. Tangible and identifiable intangible assets with finite lives, such as developed technologies, are depreciated or amortized, depressing earnings on a regular basis. In contrast, goodwill is unamortized and subject to a periodic fair-value-based impairment test. As write-offs of goodwill impairment are infrequent (Ramanna and Watts, 2009), recording more goodwill generally leads to higher post-acquisition earnings.

Considering the implications of earnings for CEO pay, we posit that CEOs are motivated to overstate goodwill, with their earnings-based bonus plans providing stronger incentives than other forms of compensation. This prediction builds on the findings that bonus plans often include mechanical earnings-based formulas (e.g., Murphy, 1999), whereas other forms of compensation do not directly contract on earnings and are therefore less likely to increase with earnings inflation. Consistent with this conjecture, existing research provides evidence that bonuses are more correlated with accounting-based performance measures than other forms of pay (Core et al., 2003), and that the relative importance of accounting-based pay is positively associated with the extent of earnings management (Larcker et al., 2007). Cohen et al. (2007) and Carter et al. (2007) suggest that the importance of cash compensation has been increasing in the years following the Sarbanes-Oxley Act of 2002 and the mandatory expensing of employees stock option grants, which overlaps with our sample period.

Furthermore, the cost of overstating goodwill—an increase in the likelihood of future impairment write-offs—is less likely to affect bonuses than equity-based pay. Dechow et al. (1994) and Gaver and Gaver (1998) suggest that CEO cash pay is shielded from non-recurring losses. Thus, goodwill impairment, a non-recurring charge, could have a minimal impact on bonuses. In contrast, goodwill impairment likely leads to reductions in CEO equity-based pay. As goodwill impairment write-offs trigger significant negative market reactions (Li et al., 2010; Bens et al., 2007), CEO equity-based pay likely decreases given the connection between stock performance and equity-based pay documented in the extant literature (e.g., Core et al., 2003). Thus, we expect that CEOs are more likely to overstate goodwill as the relative importance of bonus increases in their pay package. Although earnings are the most commonly used accounting performance measure in bonus plans, other accounting-based measures, such as cash flows and sales, are also used in bonus plans (Murphy, 1999; Perry and Zenner, 2001; Hui et al., 2008; Huang et al., 2010). Compared to bonus plans based exclusively on earnings, bonus plans relying more on alternative performance measures provide different managerial incentives for purchase price allocation. We focus on the most commonly used alternative accounting performance measures and classify them into two groups: measures that are not affected by the overstatement of goodwill (cash flows and sales), and a measure that implies ratcheting up of the earnings target in future periods (earnings growth), thereby making the overstatement less beneficial and possibly costly for future periods by overstating the base on which growth is measured. We predict that the positive association between bonus intensity and the allocation to goodwill becomes weaker when bonus plans include these two types of performance measures. We also test whether performance measures that are affected by the overstatement of goodwill but to a lesser extent than earnings can mitigate the incentives to overstate goodwill (ROA/ROE).

We use the setting of purchase price allocation to study accounting consequences of bonus intensity and bonus performance measures for several reasons. First, acquisitions are significant business decisions that have a long-lasting impact on financial reporting. As a result of its economic significance, the purchase price allocation is likely a powerful setting for detecting the influence of various reporting incentives. Second, different bonus performance measures likely create different incentives for the allocation, since the recognition of goodwill versus other assets influences subsequent earnings but not other performance measures such as cash flows and sales. The setting therefore allows us to sort out the differential implications of alternative accounting performance measures. Third, the fair-value-based purchase price allocation is an interesting subject in and of itself. The SEC has been concerned with potential misallocation of the purchase price. However, so far, there has been little academic evidence on the motivation and form of misallocation. More generally, the allocation process is similar to the fair value accounting for non-financial assets and liabilities being considered by standard setters worldwide. While this move towards fair value has been controversial, there is limited empirical evidence on fair value measurements for a wide range of assets and liabilities under the current historical-cost-based reporting framework. The purchase price allocation provides a unique opportunity for examining how CEO incentives, in particular those derived from bonus plans, affect fair value measurement for assets without active markets.

We collect the fair value allocation data for a sample of acquisitions completed between July 2001 and April 2007, and use the proportion of acquisition price recorded as goodwill as a summary measure of allocation. We then test our predictions by examining the impact of bonus intensity and the use of alternative accounting performance measures on this summary measure, after controlling for other determinants of the allocation. As CEO bonus intensity could be determined in anticipation of the acquisition, we adopt a two-stage least squares (2SLS) approach using industry median bonus intensity as the instrumental variable to estimate the predicted value of bonus intensity. Consistent with our first prediction that CEOs exploit their discretion in unverifiable fair value measurement to increase their bonuses, we find that the proportion of acquisition price recorded as goodwill increases with the predicted bonus intensity.

Consistent with our second prediction, we find that the association between bonus intensity and the allocation to goodwill diminishes when cash flows, sales, or earnings growth is used as a performance measure, suggesting that the use of alternative accounting performance measures in addition to earnings can significantly mitigate CEOs’ incentives to overstate goodwill. We do not find the use of ROA/ROE to mitigate the incentives, possibly because it does not reduce the benefits of overstating goodwill as much as the other alternative performance measures and/or our tests lack power.

In all of the above tests, we control for economic determinants of the allocation and other incentives CEOs may have in the allocation. Specifically, we control for target characteristics that are related to the recognition of major identifiable intangible assets, fixed assets, and goodwill. We also control for synergies from the combination. These economic factors along with the target industry fixed effects are in fact the most significant determinants of the allocation, explaining about 59% of the total variation of the dependent variable. Finally, we consider the cost of overstating goodwill and find some evidence that acquirers’ ability to avoid future goodwill impairment also affects the allocation to goodwill.

This paper makes several contributions to the literature. First, it extends the research on bonus plans and managers’ accounting choices. Performance measures are one of the three key elements of bonus contracts (Murphy, 1999). However, it has largely been absent in the literature how various performance measures in bonus plans affect management behavior. Our paper fills this void by providing direct evidence on the differential implications of various performance measures for management accounting choices in purchase price allocation.

Second, this study furthers our understanding of fair value measurement when it is applied to a wide range of assets and liabilities, contributing to the debate on fair value accounting for non-financial assets and liabilities. Given that procedures similar to the purchase price allocation were proposed for measuring fair values of internally developed intangibles (AASB, 2008), our findings are particularly relevant to the regulatory debate on fair value accounting for intangible assets.

Third, this study adds new evidence to the regulatory assessment of the accounting for acquisitions. Aboody et al. (2000) find a positive association between the importance of bonuses and the use of pooling to account for acquisitions under the pre-SFAS 141 regime. While SFAS 141 is expected to correct for the ‘gaming’ incentives associated with the choice of pooling by abolishing the method, we show that incentives similar to those affecting the choice of pooling continue to cause distortions in accounting despite the regulatory effort, consistent with the conjecture of Ramanna (2008). Our results thus provide supporting evidence for the SEC’s concerns about the purchase price allocation. Interestingly, we find that the opportunistic incentives are mitigated by the use of alternative performance measures in bonus contracting.

Finally, unveiling an approach of earnings management that is largely missing in the literature, our findings highlight the notion that earnings management is a dynamic process. The extant earnings management literature focuses on manipulation of concurrent accruals (e.g., Jones, 1991) or real activities (e.g., Roychowdhury, 2006). We study an accounting procedure that has a long lasting effect on earnings and differs from the typical accrual or real activity choices. Our results indicate that earnings management can result from past opportunistic actions and be hard to detect using the existing models of discretionary accruals.

The full paper is available for download here.

Both comments and trackbacks are currently closed.