by Alex Edmans and Xavier Gabaix
Appeared originally at VoxEU.org, 24 June 2009
Many blame executive compensation for encouraging shortsighted risk-taking. This column argues that compensation should be structured so as to provide incentives consistent with the firm’s position and long-term interest. It proposes “incentive accounts” that it says would be superior to existing compensation schemes.
In an influential book, Bebchuk and Fried (2004) argued that executive compensation is set by managers themselves to maximise their own pay, rather than by boards on behalf of shareholders. Indeed, many commentators argue that executives’ pay schemes were major contributors to the financial crisis, encouraging them to take on too much risk and manage their company for short-term profit. In response, President Obama has proposed new executive compensation rules for firms seeking government aid. However, several critics have argued that the recent changes are politically motivated and focus on the level of pay, rather than the incentive structures (e.g. the relative amount of cash versus shares), which have the greatest economic impact.
Here we propose a systematic solution to address the economic issues that are at heart of the current crisis to prevent future value destruction. Moreover, it can be applied to all firms, not just those receiving bailouts. It thus may be relevant for President Obama’s ongoing discussions on broader changes in compensation across the economy.
Existing schemes have two main problems. First, stock and options typically have short vesting periods, allowing executives to “cash out” early. For example, Angelo Mozilo, the former CEO of Countrywide Financial, made $129 million from stock sales in the twelve months prior to the start of the subprime crisis. This encourages managers to pump up the short-term stock price at the expense of long-run value – for instance by originating risky loans, scrapping investment projects, or manipulating earnings – because they can liquidate their holdings before the long-run damage appears. Long-term incentives must be provided for the manager to maximise long-term value, which we call the “long-horizon principle.”
Second, current schemes fail to keep pace with a firm’s changing conditions. If a company’s stock price plummets, stock options are close to worthless and have little incentive effect – precisely at the time when managerial effort is particularly critical. This problem may still exist even if the executive has only shares and no options. Consider a CEO who is paid $4 million in cash and $6 million in stock. If the share price halves, his stock is now worth $3 million. Exerting effort to improve firm value by 1% now increases his pay by only $30,000 rather than $60,000 and may provide insufficient motivation. To maintain incentives, the CEO must be forced to hold more shares after firm value declines. Our research has shown that, to motivate a manager, a given percentage increase in firm value (say 10%) must generate a sufficiently high percentage increase in pay (say 6%). In the above example, this is achieved by ensuring that, at all times, 60% of the manager’s pay is stock. We call this the “constant percentage principle.” The appropriate proportion will vary across firms depending on their industry and life cycle, but we estimate 60% as a ballpark number for the average firm.
These two principles can be achieved by giving the executive a scheme we call an “Incentive Account,” which is based on our own prior research (Edmans, Gabaix and Landier, forthcoming) and ongoing work with Tomasz Sadzik of NYU and Yuliy Sannikov of Princeton. It contains two critical features – rebalancing to address the constant percentage principle and gradual vesting to satisfy the long-horizon principle. Each year, the manager’s annual pay is escrowed in a portfolio to which he has no immediate access. In the above example, 60% of the portfolio is invested in the firm’s stock and the remainder in cash. As time passes and the firm’s value changes, this portfolio is rebalanced monthly so that 60% of the account remains invested in stock at all times. In our example, after the stock price halves, the Incentive Account is now worth $7 million ($4 million cash and $3 million of stock). This requires the CEO to hold $4.2 million of equity, which is achieved by using $1.2 million of cash to buy stock. This satisfies the “constant percentage principle” and maintains the manager’s incentives after firm value has declined. Importantly, the additional stock is accompanied by a reduction in cash – it is not given for free. This addresses a major concern with repricing stock options after the share price falls – the CEO is rewarded for failure.
Each month, a fixed fraction of the Incentive Account vests and is paid to the executive. Even when the manager leaves, he does not receive the entire value of the Incentive Account immediately. Instead, it continues to vest gradually; full vesting will occur only after several years. By then, most manipulation or hidden risk will have become public information and affected the stock price and thus the account’s value. Since the manager has significant wealth tied in the firm even after his departure, he has fewer incentives to manipulate earnings in the short term.
While the Incentive Account may seem a marked departure from current practices, it can be approximately implemented using standard compensation instruments without setting up a special account. In each period, the board pays the CEO a mix of deferred (cash) compensation and restricted stock. If performance is poor, the next period the CEO’s salary is paid exclusively in restricted stock; upon strong performance, it is paid exclusively in deferred cash.
We note that gradual vesting is not without its cost. Compared to short-term vesting, it imposes some risk on the manager, and he may require a higher salary as compensation. However, the benefits of a high-powered incentive scheme are much greater than its costs. Even if an optimal contract induces the CEO to increase firm value by only an additional 1%, this is $100 million when applied to a $10 billion firm, which vastly exceeds any required compensation for risk. Similar to investing in a risk management system, the Incentive Account has a small cost, but pays off in sharper incentives and sounder risk-taking. Moreover, for a given vesting period and target incentive level, we demonstrate mathematically that Incentive Accounts are always less costly than other schemes such as stock options, restricted stock, clawbacks, and bonus-malus banks.
The Incentive Account is a basic framework that can be enhanced by additional features, such as benchmarking to market performance to ensure the manager is not rewarded for luck. Also, it may be applied to other critical employees, e.g. traders, and deter problems similar to those that afflicted AIG. In these cases, pay should be tied not to the stock price of the overall firm, but the profit of the worker’s department, as that is more closely under his or her control.
Our plan need not be imposed by regulators (although if regulators do wish to make prescriptions, Incentive Accounts are worth considering). Even in the absence of regulation, shareholders typically have sufficient incentives to implement any new scheme that is appropriate for their specific firm. Instead, we advocate that regulation should remove tax or accounting distortions that favour some forms of compensation over others. That will allow incentive schemes to compete on a level-playing field, and incentive accounts are well-positioned to win the market test.
Bebchuk, Lucian, and Jesse Fried, Pay Without Performance: The Unfulfilled Promise of Executive Compensation (Cambridge, MA: Harvard University Press, 2004).
Edmans, Alex, Xavier Gabaix, Tomasz Sadzik, Yuliy Sannikov “Dynamic Incentive Accounts”
Edmans, Alex, Xavier Gabaix and Augustin Landier “A Multiplicative Model of Optimal CEO Incentives in Market Equilibrium,” forthcoming, Review of Financial Studies.