Equity Vesting and Managerial Myopia

The following post comes to us from Alex Edmans, Professor of Finance at the London Business School, Vivian Fang of the Department of Accounting at the University of Minnesota, and Katharina Lewellen of the Tuck School of Business at Dartmouth College.

In our paper, Equity Vesting and Managerial Myopia, which was recently made publicly available on SSRN, we study the link between real investment decisions and the vesting horizon of a CEO’s equity incentives. We find that research and development (“R&D”) is negatively associated with the stock price sensitivity of stock and options that vest over the course of the same year. This association continues to hold when including advertising and capital expenditure in the investment measure. Moreover, CEOs with significant newly-vesting equity are more likely to meet or beat analyst consensus forecasts by a narrow margin. However, the market recognizes such CEOs’ incentives to inflate earnings—the lower announcement returns to meet or beating earnings forecasts are decreasing in the sensitivity of vesting equity. These results provide empirical support for managerial myopia theories.

Many academics and practitioners believe that managerial myopia is a first-order problem faced by the modern firm. While the 20th century firm emphasized cost efficiency, Porter (1992) argues that “the nature of competition has changed, placing a premium on investment in increasingly complex and intangible forms,” such as innovation, employee training, and organizational development. However, the myopia theories of Stein (1988, 1989) show that managers may fail to invest due to concerns with the firm’s short-term stock price. Since the benefits of intangible investment are only visible in the long run, the immediate effect of such investment is to depress earnings and thus the current stock price. Therefore, a manager aligned with the short-term stock price may turn down valuable investment opportunities.

Despite its perceived importance, myopia is very difficult to test empirically. Standard measures of CEO incentives (e.g., Hall and Liebman (1998)) quantify the sensitivity of managerial wealth to the stock price stemming from his stock and option compensation. However, in myopia models, the driver of short-termism is not the overall level of equity compensation, but the weighting of this compensation towards the short-term as opposed to long-term stock price (Stein (1988, 1989)). Equity that does not vest until the long term will deter rather than induce myopia (Edmans, Gabaix, Sadzik, and Sannikov (2012)).

However, operationalizing this theoretical concept empirically is tricky. The ideal experiment would be for the CEO to be forced to sell some equity for exogenous reasons, and to be aware of this forced sale ahead of time so that this expectation affects his actions. (This is indeed how short-term concerns arise in the Stein (1989) model). However, identifying sales that are both exogenous and predictable by the CEO is difficult. Unexpected forced sales (e.g., due to sudden liquidity needs) are likely exogenous, but typically unobservable by researchers and unpredictable by the CEO. Actual discretionary sales are observable to researchers, but likely endogenous for two reasons. First, omitted variables may drive both actual sales and investment. For example, the manager’s negative private information on firm prospects may cause him both to sell equity and to reduce investment. Second, some actual sales (e.g., due to sudden liquidity needs) may have been unpredictable by the CEO. Thus, actual sales are a poor proxy for the ideal explanatory variable, predicted sales, leading to measurement error.

We measure a manager’s myopic tendencies using the sensitivity of his stock and options that are scheduled to vest over the upcoming year. We show that this sensitivity is highly correlated with actual sales: in the absence of private information, a risk-averse manager should sell his equity upon vesting. However, while a CEO’s actual sales are an endogenous decision, the amount of newly-vesting equity is largely driven by the magnitude and vesting horizon of equity grants made several years prior. [1] For the same reason, it is known to the CEO in advance. We identify the equity that is scheduled to vest in a given year using a recently-available dataset from Equilar. This dataset contains grant-by-grant information on an executive’s options, including whether they are vested or unvested. We can thus identify at an individual grant level the number of options that switch from unvested to vested in a given year. This grant-level information allows us to calculate the delta of the vesting options, which captures the manager’s incentive to inflate the stock price. Equilar also provides the number of vesting shares; since the delta of a share is 1, this number does not need to be decomposed into individual grants.

We use the sensitivity of newly-vesting equity in two ways. First, we employ it as the explanatory variable of interest, relating it to changes in several measures of long-term investment. Our primary measure is R&D scaled by total assets, but we also include advertising and two measures of capital expenditure. We control for determinants of investment opportunities and the firm’s ability to finance investment, firm and year fixed effects, and other components of CEO compensation—for example, the CEO’s unvested equity, his already-vested equity, and the standard components of salary and bonus.

We find a negative and significant relationship between nearly all measures of investment and the sensitivity of newly-vesting equity. For example, an interquartile increase in this sensitivity is associated with a 0.11 percentage point decline in the growth of R&D scaled by total assets, which corresponds to 37% of the average growth in R&D scaled by total assets, 2% of the average R&D-to-assets ratio, and an average decline in R&D of approximately $1 million per year. Our results suggest that firms reduce investment in years in which significant CEO stock and option holdings vest.

Newly-vesting equity is of interest as an explanatory variable, since boards may wish to take into account its link with investment when designing the optimal contract. Similarly, since boards know the amount of newly-vesting equity at the start of a year, they can predict the CEO’s incentives to cut investment, and if needed, counteract them. A broader question is how investment responds to expected equity sales in general. As in the Stein (1989) model, such sales can stem from channels other than vesting equity—a CEO may voluntarily hold already-vested equity as a long-term investment (e.g., in a family firm), subsequently sell equity to rebalance his portfolio, and anticipate such sales beforehand. Since actual equity sales are endogenous, our second use of newly-vesting equity is as an instrument for actual sales. The two properties of newly-vesting equity discussed earlier—its high correlation with equity sales and its determination by equity grants several years prior—are analogous to the relevance criterion and exclusion restriction. We find a negative relationship between instrumented equity sales and investment. An interquartile increase in equity sales is associated with a 0.25 percentage point decline in the growth of R&D scaled by total assets, 4.6% of the average R&D-to-assets ratio.

The negative association between investment and vesting equity (or instrumented equity sales) can arise from two channels. First, vesting equity could cause a decline in investment. Managers intending to sell equity reduce investment, to inflate earnings and thus the stock price. Second, there is no causal relationship but instead the link arises from an omitted variable—current investment opportunities—that our controls fail to capture. It may be that boards believe that vesting equity reduces the manager’s investment incentives, and thus schedule equity to vest precisely when they forecast that investment opportunities will decline. This explanation requires boards to be able to forecast investment opportunities several years in advance. Note that it is still consistent with myopia theories: boards ensure that options do not vest while investment opportunities are strong because they believe that vesting equity induces myopia.

To provide further evidence of the first channel, we show that newly-vesting equity is associated with a higher likelihood that a firm meets the analyst consensus earnings forecast or beats it by a narrow margin. In contrast, vesting is unrelated to the likelihood of beating the forecasts by a wide margin, consistent with earnings manipulation being more likely when earnings are close to the forecast. These results support the idea that vesting equity increases the CEO’s stock price concerns, but not the alternative hypothesis that equity vesting is correlated with changes in investment opportunities.

Finally, we study the market’s reaction to earnings announcements. Controlling for the earnings realization, the announcement return is significantly lower for firms with higher levels of vesting equity, perhaps because the market suspects that such firms have incentives to inflate earnings by cutting investment. This effect is especially strong for firms that meet or beat the forecast, suggesting that this outcome more likely stems from earnings inflation. These findings are consistent with the Stein (1989) “signal-jamming” equilibrium, where the market is efficient and recognizes managers’ myopic behavior.

The full paper is available for download here.


[1] Gopalan, Milbourn, Song, and Thakor (2013) show that most vesting periods are between three and five years.
(go back)

Both comments and trackbacks are currently closed.