One big problem in the financial system are the short-term incentives in place for managers, which leads them chasing the wrong objectives and excessive risk taking. Here is a way to fix that..
The public outrage over bonuses paid to AIG executives, and resulting government proposals to cap pay at companies receiving federal bailouts, illustrate rising concerns about executive compensation. Indeed, some analysts contend that ineffective compensation structures encouraged Wall Street executives to take on excess risk in the hopes of winning huge payouts, which in turn contributed to the continuing financial crisis and recession.
Bonus backlash aside, new Wharton research suggests that a compensation structure based on long-term escrow accounts could be better — at least for a company’s future — than the common practice of rewarding short-term changes in share price. Wharton finance professor Alex Edmans and a group of colleagues propose linking an executive’s compensation to the performance of the firm over a longer horizon. Their method also takes into account changing conditions within the firm.
In a working paper titled, “Dynamic Incentive Accounts,” Edmans, along with Xavier Gabaix and Tomasz Sadzik of New York University, and Yuliy Sannikov of Princeton University, outline a system that escrows compensation for a set period of years stretching into the executive’s retirement. The longer time frame is designed to prevent the executive from taking short-term actions that may enrich the manager at the expense of the firm’s future profits. The plan also provides a rebalancing mechanism to maintain a constant percentage of compensation in cash and stock, so that the executive always has sufficient equity in the firm to provide performance incentives — even if the stock price falls.
Linking compensation to share price has been viewed as a way to make sure the interests of the CEO and other top managers are aligned with those of shareholders. In recent years, however, the system has encouraged executives to take actions that boost share price in the short term but hurt shareholders and other stakeholders later, after the executive has cashed out.
Sell the Stock, Kill the Company
Existing compensation schemes typically have short vesting periods that allow executives to reap the rewards of their actions before their full effect may be realized. As an example, Edmans points to Angelo Mozilo, the former CEO of Countrywide Financial, who made $129 million from stock sales in the 12 months prior to the start of the subprime crisis, which sent Countrywide shares tumbling and led to the firm’s acquisition by Bank of America.
Other executives have manipulated corporate accounting to boost share price, then sell at the peak, as did Enron executives through the creation of shell subsidiaries in which they hid losses. Or, he says, executives could take other actions, such as cutting investment in research and development which would lead to short-term earnings spikes and an increase in share price, but cripple the firm in the future. Placing compensation in escrow accounts with a longer-term horizon of around five years, the paper suggests, would help prevent destructive short-term actions.
Another problem with the current system is that it fails to adapt to changing conditions in the firm and its share price. If the firm is in trouble and the share price falls, stock options are worth little. Consequently, the executive loses an important incentive to perform just at the time when it is most important to be working at full throttle. So, to keep managers working hard at troubled times, boards typically give the executives more shares for free or they re-price existing options. “That’s highly controversial because, in effect, the CEO is being rewarded for failure,” Edmans says.
The authors’ new system addresses that by maintaining a constant balance of stock and cash in the executives’ escrow accounts, a practice they call “the constant percentage principle.” (This principle was first introduced in an earlier paper by Edmans, Gabaix and NYU’s Augustin Landier titled, “A Multiplicative Model of Optimal CEO Incentives in Market Equilibrium.”) The optimal proportion will vary across firms depending on their industry and maturity.
According to the authors’ research, motivating a manager requires rewarding a given percentage increase in firm value, say 10%, with a sufficiently high percentage increase in pay, say 6%. Using those hypothetical figures, the company would put 60% of the executive’s compensation in stock. Edmans explains: “If the CEO earns $5 million, he would be given $3 million in stock and the remainder in cash. If firm value rises by 10%, his stock will now be worth $3.3 million and the cash remains at $2 million. Total pay rises from $5 million to $5.3 million, an increase of 6%.”
The authors’ incentive account proposal would place top managers’ compensation in escrow accounts invested in company stock and cash. The account would be rebalanced each month to maintain the percentage of compensation in cash and stock. For example, if the percentage was set to maintain 60% in stock, cash would be drawn down in the account to buy more stock when the share price falls, to keep the proportion at 60%.
“The [account] is reloaded when the stock price goes down, and it makes sure the CEO has enough skin in the game,” says Edmans. Critically, the reloading is not for free, unlike the current practice of re-pricing options or granting additional shares after stock price declines. Instead, the additional equity is purchased with cash in the account. On the other hand, if the stock price goes up and the CEO has a lot of stock, he can sell it for cash to make sure he is not subject to unnecessary risk. However, the cash proceeds cannot be immediately withdrawn. The account vests gradually. This vesting continues until a number of years after the executive’s retirement. Since the proceeds from stock sales must remain within the account, this deters the CEO from manipulating the share price upwards to trigger a sale and immediately withdrawing the proceeds. Instead, if the stock price subsequently falls, the cash proceeds are converted back to stock.
Other Long-term Plans
Credit Suisse, Goldman Sachs, UBS and Morgan Stanley have all introduced mechanisms to deter short-term manipulation. These have come in the form of clawbacks for bonuses, which allow them to be rescinded if long-term performance is poor. Non-financial firms are also looking into new compensation systems to replace the current reliance on stock, options and bonus-based incentive plans.
However, while these systems are similar in spirit, Edmans, Gabaix, Sadzik and Sannikov argue that clawbacks are a less effective way to deter manipulation than escrow accounts. “If you have paid out bonuses and the company then tanks, trying to claw back the bonus seems like shutting the barn door after the horse has bolted,” says Edmans. “Rather than paying out the bonus and then clawing back, just don’t pay it out in the first place.”
The authors note that extending the vesting period may require added compensation to make up for the CEO’s greater risk. However, they calculate that the benefits are likely to outweigh the added costs. For example, if an optimal contract induces the CEO to increase firm value by only 1%, that would add up to $100 million for a $10 billion firm, which swamps any required increase in CEO pay. In addition, research underlying the working paper indicates that for a set vesting period and target incentive level, the proposed accounts are always less costly than other schemes such as stock options, restricted stock, clawbacks and bonus banks, which hold back bonuses for a set period of time.
The incentive account model is valuable not only for compensating the CEO and top executives, according to Edmans. He notes that while AIG’s problems have been attributed to senior management, some of the blame has been directed at rank-and-file traders. When applied to lower-level employees, the model can be modified to take into account measures other than share price that more accurately reflect whether an individual is performing effectively. “For a trader, perhaps the best measure is not the stock price because he’s just a cog in a big wheel. Perhaps it is the profit of his own division,” says Edmans.
Despite the increasing role of government in overseeing compensation, at least for financial firms that have accepted bailout funds, Edmans says government regulators should not play a part in setting up compensation plans. The problem with government regulation is that it tends to be “one size fits all.” He says the incentive account model must be constructed with enough flexibility to adapt to unique qualities in various firms. An innovative pharmaceutical firm or start-up would need to account for different time horizons and levels of risk than a mature manufacturer of concrete.
Government involvement in compensation could also lead to unintended consequences, according to Edmans. Indeed, he says that when the Clinton administration imposed a cap of $1 million on cash payouts to be eligible for corporate tax deductions, many firms shifted compensation to guaranteed bonuses, such as the controversial AIG payouts.
The role of enacting new incentive structures falls to boards of directors, Edmans states. “The scheme should be up to the board of directors who should be deciding the compensation scheme. It should be important to set the scheme in the long-term interest of the shareholders.” He says an individual board of directors is best suited to determine what “long-term” means at their firms. In general, he says, it should be as long as it would take for key actions to be reflected in operations and share price. The time frame, he emphasizes again, would be different at the research-based pharmaceutical firm than at the concrete company.
As for the furor over the bonuses for bailout firms such as AIG, Edmans notes that many commentators have criticized the response as being ad hoc, politically motivated and piecemeal. “We wanted to do something that could apply to all firms, not just bailout recipients, and [that is] grounded in economic analysis.”