Agency Conflicts and Short- vs Long-Termism in Corporate Policies

Should firms target short-term objectives or long-term performance? The question of the optimal horizon of corporate policies has received considerable attention in recent years, with much of the discussion focusing on whether short-termism destroys value. The worry often expressed in this literature is that short-termism—induced, for example, by stock market pressure—may lead firms to invest too little (see Asker, Farre-Mensa, and Ljunqvist, 2015; Bernstein, 2015; Gutierrez and Philippon, 2017, for empirical evidence). Another line of argument recognizes, however, that while firms must invest in their future if they are to have one, they must also produce earnings today to pay for doing so. In line with this view, Giannetti and Yu (2018) find that firms with more short-term institutional investors suffer smaller drops in investment and have better long-term performance than similar firms following shocks that change an industry’s economic environment.

While empirical evidence relating short- or long-termism to firm performance is accumulating at a fast pace, financial theory has made little headway in developing models that characterize the optimal horizon of corporate policies or the relation between firm characteristics and this horizon. Our paper titled Agency Conflicts and Short- versus Long-Termism in Corporate Policies attempts to provide an answer to these questions through the lens of agency theory. To do so, we develop a dynamic agency model in which the agent controls both current earnings and firm growth (i.e., future earnings) through unobservable investment. In this multitasking model, the principal optimally balances the costs and benefits of incentivizing the manager over the short or long term. As shown in the paper, this can lead to optimal short- or long-termism, depending on the severity of agency conflicts and on firm characteristics. Additionally, we show that the same firm can find it optimal at times to be short-termist (i.e., favor current earnings) and, at other times, to be long-termist (i.e., favor growth).

The paper starts by formulating a dynamic agency model in which an investor (the principal) hires a manager (the agent) to operate a firm. In this model, agency problems arise because the manager can take hidden actions that affect both earnings and firm growth. Specifically, the manager can stimulate current earnings via short-term investment and firm growth via long-term investment. Investment is costly and the manager can divert part of the funds allocated to investment.

The moral hazard problem requires the manager’s compensation contract to provide sufficient incentives. The optimal contract uses performance-sensitive deferred compensation and generates short- and long-run incentives by conditioning the manager’s compensation on earnings and asset size. In practice, executive compensation is commonly linked to stock prices (via stock and option grants) and to accounting results (via performance-vesting provisions of these grants and via performance-based bonuses). Consistently with these practices, we show that the optimal contract implied by our model can be implemented by exposing the manager to stock prices and earnings. This characterization of the optimal contract highlights an important implication of our model: while stock prices account for both short- and long-run shocks to firm value, exposing the manager solely to the firm’s stock price cannot, in general, provide a right mix of short- and long-run incentives. To achieve optimal incentives, the manager needs to be additionally exposed to short-run accounting performance metrics such as earnings.

The paper demonstrates that the optimal incentive contract generates short- or long-termism in corporate policies, defined as short- or long-term investment levels above the levels attainable in the absence of agency frictions. In other words, short- or long-termism can be an optimal response to the dual agency problem over the short and long run. Long-termism arises because the investors find it optimal to expose the manager excessively to long-term shocks. This can be necessary because, with low exposure, the manager’s stake in the firm is diluted by positive shocks to firm value: the manager’s value changes little relative to firm value, so her stake is diluted and, under some conditions, becomes insufficient to run a bigger firm. Our analysis demonstrates that long-termism is more likely to arise when cash flows are more volatile or when the investment technology is less efficient.

A key result of the paper is to show that short-termism is more likely to occur when the agent’s stake in the firm is low, and the risk of termination and agency costs are high. Indeed, in such instances, the benefits of long-term growth are limited. By contrast, stimulating short-term investment increases earnings and reduces the risk of termination and agency costs. Interestingly, a recent study by Barton, Manyika, and Williamson (2017) finds using a data set of 615 large- and mid-cap US publicly listed companies from 2001 to 2015 that “the long-term focused companies surpassed their short-term focused peers on several important financial measures.” While our model does indeed predict that firm performance should be positively related to the corporate horizon, it, in fact, suggests the reverse causality.

A distinctive feature of our model is that the optimal contract introduces exposure to permanent shocks (via exposure to the firm’s stock price) that is not needed to incentivize investment. In particular, the agent is provided minimal long-run incentives when the firm is close to financial distress and higher-powered long-run incentives after positive past performance when sufficient slack has been accumulated. In this region, incentives have option-like features and increase after positive performance. In other words, positive permanent shocks lead to additional pay-for-performance, and negative permanent shocks eventually eliminate this extra sensitivity to performance implied by the optimal contract. Our model, therefore, provides a rationale for the asymmetry of pay-for-performance observed in the data (see, e.g., Garvey and Milbourn, 2006; Francis et al., 2013).

The complete paper is available for download here.