The lead article in Volume 21, Issue 5 of the Review of Finance is “How Important Are Risk-Taking Incentives in Executive Compensation?” by Ingolf Dittmann of Erasmus University Rotterdam, Ko-Chia Yu of National Chung Cheng University, and Dan Zhang of BI Norwegian Business School.
Should CEOs Receive Stocks or Options?
Few topics in finance have grabbed the attention of the general public as much as executive compensation. Almost everyone seems to have an opinion on it, and most people believe that the current system is broken and must be fixed. In the 2016 US election campaign, Donald Trump and Hillary Clinton agreed on very little, but one of the few opinions they shared was that the level of executive pay is far too high (although Trump expressed his opinion a bit more aggressively than Clinton!).
While the level of pay receives most scrutiny, the structure of pay is also under fire. One particularly controversial element is the rapid rise of stock options, from 19% of pay in 1992 to 49% by 2000. Critics accept that it’s valuable to tie the CEO’s wealth to performance, to provide incentives – but claim that stock would do so more effectively than options. For example, Bebchuk and Fried (2004) argue that options give CEOs a one-way bet – they gain if the share price rises above the strike price, but don’t lose if it falls below it. In contrast, a CEO paid with stock would suffer from falls in the share price. Murphy (2013) argues that options became popular because they didn’t need to be expensed prior to the FAS 123R accounting rule (as long as they were not in-the-money) – unlike stock. Thus, options were effectively free, encouraging firms to use them even if they were less effective than stock.
But, the arguments aren’t all one-way. There’s a trade-off. Consider a model in which a risk-averse CEO takes a single action – he exerts “effort” to improve the stock price. Then either stock or options may be better:
- On the one hand, $1 of options provide stronger incentives than $1 of stock, because options are riskier than stock.
- On the other hand, since options are riskier than stock, a risk-averse CEO values $1 of options less than $1 of stock. Thus, he will demand more options to work for the firm.
- A second, more technical, disadvantage is that options pay off only if the stock price rises. The stock price rises in economic upswings, where the rest of the CEO’s portfolio is performing well, so the extra payoff from options is worth little to him. In downswings, stock is still worth something but options become out-of-the-money and have zero value – precisely when the rest of the CEO’s portfolio is performing poorly. Thus, options provide little insurance and are less valued by the CEO – just like high-beta companies have lower prices.
What Did Prior Research Teach Us?
Which forces dominate? For this, we need a model. Hall and Murphy (2002), Jenter (2002), and Dittmann and Maug (2007) calibrate a standard agency model, where the CEO has standard risk-averse preferences, and find that stock is superior – suggesting that the use of options is inefficient, exactly as argued by the critics. Moreover, when Dittmann and Maug (2007) drop the restriction that the contract must be piecewise linear (i.e., consist of salary, stock, and options), they find that the optimal nonlinear contract is concave. In contrast, options are convex.
However, as with many economic questions, the answers depend critically on the model used. Dittmann, Maug, and Spalt (2010) calibrate a model in which the CEO is loss-averse. (Loss aversion is an alternative type of preferences to risk aversion. It’s commonly used in behavioral economics, and assumes that the pain of losses outweighs the pleasure of gains.) They show that the observed use of options can be rationalized by realistic levels of CEO loss aversion, since options provide downside protection.
This Paper’s Contribution
Ingolf, Ko-Chia, and Dan make a different important addition to the standard model. They argue that CEOs don’t just take “effort” actions that affect the mean of the stock price, but also “risk” actions that increase the variance of the stock price. Contrary to some critics who assume that risk-taking is bad, they point out that risk-taking can be good and that – in the absence of sufficient incentives – CEOs will take too little risk. Finance 101 teaches us that executives should ignore idiosyncratic risk when making investment decisions, because shareholders are diversified and so are unaffected by it. But, a CEO is undiversified, as his wealth, firm-specific human capital, and reputation are all tied to the firm’s performance. Thus, left to their own devices, CEOs may coast and preserve the status quo rather than innovating, investing, or restructuring. If so, the seemingly “unfair” feature of options – that the CEO gains from any upside but doesn’t lose from the downside – is actually desirable as it encourages risk-taking. These are not just hypothetical textbook considerations, but real-world ones. For example, Google’s executives could have preserved the status quo and remained in charge of a very successful firm. But, potentially due to their option compensation, they went further and took risks by restructuring into Alphabet, taking on the iPhone by launching Pixel, and investing in self-driving cars.
Their optimal solution to the model has three features: First, the optimal contract is convex for medium outcomes – i.e. should encourage risk-taking. Second, it protects the CEO from bad outcomes – not because the CEO is loss-averse (as in Dittmann, Maug, and Spalt (2010)) but to encourage him to take risk. While these two features encourage options, the third discourages them. The optimal contract is concave for good outcomes – as explained earlier, in good times, the rest of the CEO’s portfolio is doing well, so there’s less need to pay the CEO generously.
What The Data Say
Other scholars have written models in which options are useful to encourage risk-taking, but the authors’ contribution is to calibrate such a model to find out which forces dominate in the real world. They take a sample of the observed contracts of 1,707 U.S. CEOs and find that their model matches them much better than the standard model without risk-taking incentives. Simply put, contracts look the way they do – CEOs receive options as well as stock – because CEOs should be incentivized to take risk.
The authors then go further and ask what the optimal strike price of the option should be. Rappaport (1990) argues that options should be issued out-of-the-money, so that they only pay off if the CEO delivers exceptional performance. In contrast, the authors find that options should be issued in-the-money (when assuming piecewise linear contracts). However, in-the-money options are tax inefficient, since they are taxed as soon as the option is awarded. Taking tax considerations into account, at-the-money options are optimal (since they are as in-the-money as possible without being actually in-the-money) – exactly as awarded in practice.
A final, more methodological contribution is to propose a new way of measuring the CEO’s incentives to take “good” risk. The literature commonly measures such incentives using the CEO’s vega – the change in his wealth for a change in stock return volatility. However, the authors point out that risk-taking incentives provided by vega can be mitigated by delta – the change in his wealth for a change in the stock price. Simply put, the more the CEO’s wealth is tied to the stock price, the more risk-averse he will be. The authors show that risk-taking incentives should not be measured by vega alone, but vega scaled by delta. (A secondary contribution is that both vega and delta should be utility-adjusted – what matters for the CEO is how the stock price and stock return volatility affect not his wealth, but his utility, because he is risk-averse). This measure is potentially valuable not only for researchers, but also boards, compensation consultants, and shareholders – to ensure that CEOs are sufficiently tied to their stock price to motivate them to increase effort, but not tied so strongly that they become too risk-averse. In some firms, options may have a role in achieving the best of both worlds.