Contracting on the Stock Price and Forward-Looking Performance Measures

Abstract

We examine the use of earnings, forward-looking performance measures and stock prices in managerial compensation. When the firm's owner and its manager have identical time preferences, the stock price is not useful for motivating the manager, as it is a noisy aggregation of a forward-looking measure and future earnings. In contrast, when the owner and the manager have conflicting time preferences, the noisy stock price is useful for contracting. If the manager has no access to banking and cannot trade the firm's shares, the timeliness of the stock price dominates the extra risk imposed by its noise. At the same time, forward-looking performance measures (such as customer satisfaction) can induce a desirable allocation of management effort between the short term and long term more efficiently than the stock price can. Forward-looking performance measures and the stock price are thus not direct substitutes in rewarding farsighted effort.

Acknowledgements

For helpful comments on earlier versions of this paper, we thank Michael Alles, Tony Atkinson, Romana Autrey, Kate Bewley, Phelim Boyle, John Christensen, Wai Fong Chua, Masako Darrough, Alan Douglas, Steve Fortin, Yigal Gerchak, Raffi Indjejikian, Ken Klassen, Kari Lukka, Susan McCracken, Gord Richardson, Bill Scott, two anonymous reviewers, and workshop participants at the University of Waterloo, Curtin University, the AAA annual meeting, and the AAA Management Accounting Section conference. lgor Vaysman acknowledges funding provided by the INSEAD Alumni Fund.

Notes

1. Cash-based bonuses have attracted significant recent attention in practice, and firms are moving away from rewarding executives with stock options and toward cash bonuses. A recent survey finds that the percentage of overall bonus-based CEO pay has grown from l3% in 2001 to 19% in 2003 (Fink, 2004). During the same time period, long-term incentives in the form of stock options and restricted stock fell from 71 to 63%.

2. The principal and the agent are equally patient when they have the same preferences over future utility (i.e. they use the same rate to discount future utility).

3. This is similar to a result in Christensen et al. (2005), who show in a different setting, that performance measure timeliness can make a non-informative measure (in the sufficient statistic sense) valuable for contracting purposes.

4. Feltham and Wu (2000) extends the literature on contractible performance measures by analyzing determinants of the relative weights placed on a publicly available performance measure and a price-based measure in a single-period setting, where investors can affect the price-based measure by acquiring private information at a cost. Their work focuses on how contract weights are affected by the non-congruity of the principal's objective with the performance measures and the non-congruity of the two performance measures with each other. Dutta and Reichelstein (2003) specifically incorporate multi-period issues in their analysis of weights on publicly available, and price-based, performance measures. The present paper includes a forward-looking performance measure unrelated to price or earnings, allows the principal and the agent to have conflicting time preferences, exogenously specifies private information gathering and does not attempt to model performance measure congruity. These modeling choices isolate the value of forward-looking performance measures (in the presence of a stock-price contracting variable) that fundamentally distinguish this paper from both Feltham and Wu (2000) and Dutta and Reichelstein (2003).

5. An example of the profit signal is the firm's publicly disclosed accounting income measure.

6. See Hauser et al. (1994) and Dikolli (2001) for other work that investigates the effects of different discount factors (i.e. δ ≠ 1).

7. Recent trends in practice indicate that firms are increasingly using restricted stock (e.g. see Gabriel, 2005). For analysis of a setting where the agent trades in the firm's shares, see Baiman and Verrecchia (1995).

8. In our analysis, the market participants conjecture via the compensation contract, but do not directly observe the value of the forward-looking signal.

9. We analyze the case of δ ≠ 1 in Section 4.

10. Bushman and Indjejikian (1993) and Kim and Suh (1993) incorporate a rational expectations equilibrium in studying optimal incentive weights when the stock price and earnings are available for contracting. However, the focus in those studies is on the role that contractible earnings play in contracts, rather than on describing the conditions under which there is no contracting demand for the stock price. Those studies also do not consider the possibility that a forward-looking performance measure is available for contracting.

11. Consistent with a large literature on incentives, we make no claim that linear compensation contracts are optimal. Instead we focus on the optimal linear compensation. For examples, see Feltham and Xie (1994), Datar et al. (2001) and Dikolli (2001).

12. Note that the parameters of the wage contracts are either observable or can be inferred by capital market participants.

13. represents the aggregate variance of the idiosyncratic errors in the investors' signals.

14. Note that the independent distributions imply covariances of zero for the random variables, y, π _i and n.

15. Not surprisingly, the weight on the period-1 profit measure is identical to the case when δ = 1. We therefore omit discussion of in this section.