Going Beyond Conventional Wisdom: Designing Executive Pay to Balance
Risk and Performance

A fundamental principle of modern financial theory is the inherent trade-off between risk and reward. To be attractive, riskier projects or investments must hold out the prospect of bigger rewards. At the corporate level, this risk premium is reflected in lower share prices for stocks and higher interest rates for bonds. But risk premiums and credit spreads are linked to attitudes toward risk and returns. When investors and companies are more willing to accept higher risks in pursuit of greater rewards, risk premiums and credit spreads tend to fall and aggregate risk levels rise.

As the recession has deepened, a backlash to what is perceived as excessive risk taking has made risk a four-letter word. For now, the pendulum of public opinion has swung in the direction of safer investments, debt reduction, delayed purchases and other risk-reducing strategies. This backlash against “excessive” risk is affecting corporate America as well. Companies are reacting much like households — borrowing less, sharply curtailing expenses and scaling back investments to conserve cash flows.

The impact on executive compensation design

Just when we thought we had the answers …
Until recently, analysis of executive compensation focused on two questions, which companies have developed tools to address:

• Are executives overpaid?
• Is there a meaningful relationship between executive pay and firm performance?

As executive pay rose rapidly over the last 10 years — primarily through long-term incentive (LTI) plans heavily weighted with options — there was no shortage of executives receiving outsized pay. However, empirical evidence suggested that executive pay levels at publicly traded companies — while quite high — were consistent with rewards for top-performing entertainers, lawyers and executives at private equity firms.

Corporate directors use peer group analysis and salary surveys to ensure that executive pay reflects market levels and their pay philosophy. These efforts have been facilitated by greater disclosure in proxy statements — beginning with the reforms made by the Securities and Exchange Commission (SEC) in 1992 through the more recent introduction of the Executive Compensation Summary tables and the Compensation Discussion and Analysis section.

There are always a few executives at poorly performing companies who earn a lot, sometimes through severance packages that pay out large lump sums even to chief executive officers fired for poor performance. Nevertheless, empirical evidence has consistently demonstrated that while the link between pay opportunity and performance varies in strength, there is a strong correlation between firm financial performance and realizable pay (measured as salary and bonuses earned, as well as changes in the value of LTIs during that period).¹

Companies use tools such as pay-for-performance or performance scenario and termination analyses to determine the proper level of executive pay. A scenario analysis measures the expected payout for a proposed set of executive compensation programs under various performance scenarios, ranging from very good results to results that would lead to termination.

Figure 1
Testing the Alignment Between Executive Pay and Financial Performance

Source: Watson Wyatt Worldwide.

A pay-for-performance analysis evaluates executives’ performance and pay level relative to their peers (see Figure 1). A good pay-vs.-performance relationship creates a strong alignment between the composite performance and the realizable pay percentiles, which tend to be along the 45-degree line (in the middle band of Figure 1). A company with consistently lower performance and higher pay than peers is likely overpaying its executives. On the other hand, an executive at a high-performing company who is paid less than his or her peers is likely to perceive the compensation as unfair, which can adversely affect motivation and retention — and eventually corporate performance.

… they changed the questions
The growing concern about risk is affecting executive compensation as well. Companies are still expected to reconcile executive pay with performance. But a third question has taken on increased importance:

• Does the executive compensation program engender excessive risk for the company?

In the past, analysts looked at the amount of pay at risk for the executive. But the recessionary landscape has shifted the focus to how the executive compensation program affects the risk to the company.

Participants in the Troubled Asset Relief Program (TARP) are legally required to perform an analysis to ensure their programs do not encourage senior executive officers to take unnecessary and excessive risks. Compensation committees are also required to “discuss and evaluate employee compensation plans in light of an assessment of any risk posed to the TARP recipient from such plans” twice a year.

But non-TARP companies might soon have to take the same steps. In an April 2009 address to the Council of Institutional Investors, SEC Chairwoman Mary Schapiro said:

I want to make sure that shareholders fully understand how compensation structures and practices drive an executive's risk taking. The commission will be considering whether greater disclosure is needed about how a company — and the company's board in particular — manages risks, both generally and in the context of setting compensation.

Lacking more formal empirical analysis, companies have relied on the conventional wisdom that putting the executive’s pay at risk imposes comparable risk for the company. Unfortunately, the tools used to analyze executive compensation in terms of risk for the executive do not measure risks to the firm.

An empirical approach
Watson Wyatt conducted an empirical analysis of the executive compensation architecture at more than 1,000 firms from the S&P 1500 from 2005 to 2007. We examined each program element in terms of one fundamental question:

• Does this element of the compensation architecture strengthen the link between the executive’s realizable pay and the change in the organization’s risk?

To measure the risk to the organization, we used the Z-score.² Empirical studies have confirmed the Z-score helps predict bankruptcy as early as 18 months beforehand and closely relates to many other risk measures, such as stock price and earnings volatility or interest coverage and leverage ratios. Many financial experts in business and academia believe companies that manage their credit risk poorly are also unlikely to manage other risks well, and vice versa. As shown below, companies that have kept their credit risks low have been rewarded with superior shareholder returns in the current economic environment.³

Our empirical analysis helped us divide the elements of the executive compensation architecture into risk mitigators and risk aggravators:

• A risk mitigator strengthens the relationship between realizable pay and risk reduction, encouraging executives to manage risk more effectively.
• A risk aggravator weakens the relationship between realizable pay and risk reduction, potentially encouraging managers to take excessive risks.

As many critics of executive pay have suggested, “overpaying” executives — measured here as providing pay opportunity that is considerably higher than the norm for their industry and size — is a significant risk aggravator. But for the most part, the results of the empirical study contradicted conventional wisdom (see Figure 2).

Figure 2
Conventional wisdom misses the most effective tools for managing risk

Source: Watson Wyatt Worldwide.

The empirical data provide guidance on key risk mitigators and risk aggravators in executive compensation architecture. These empirical relationships form the basis for a new tool in executive compensation analysis.

Watson Wyatt PayRiskScore
The Watson Wyatt PayRiskScore is a single number that companies can use to evaluate their executive compensation architecture in terms of avoiding excessive risk. We evaluated the executive compensation architecture at more than 1,000 firms and assigned each a score. The score reflects the numbers of risk aggravators and risk mitigators and their relative impact on the relationship between realizable pay and changes in the company’s risk. So, for example, overpaying an executive has a larger negative impact on the score than does using return-based metrics in the bonus formula. To simplify, we converted the scores into percentile rankings from 0 to 100, such that the typical company has a Watson Wyatt PayRiskScore of 50. Companies with the most favorable (highest) scores are those with the most risk mitigators and fewest risk aggravators, which creates a very strong link between realizable pay and risk reduction. These companies provide greater financial incentives for reducing risk. Companies with more risk aggravators and fewer risk mitigators have low (unfavorable) scores.

Executive compensation and TRS: The Watson Wyatt pay-for-performance score
Shareholders want companies to reduce excessive risk while continuing to deliver greater shareholder returns. The Watson Wyatt pay-for-performance score measures the impact of executive compensation features on the relationship between realizable pay and firm financial performance (in this case, total returns to shareholders or TRS). This score is based on the same features that make up the Watson Wyatt PayRiskScore and, for ease of comparison, uses the same 0-to-100 scale. The executive compensation architecture at companies with higher scores creates a stronger link between realizable pay and TRS.

Putting it together: The Pay-for-Performance–PayRiskScore Matrix
Effectively managing a company often comes down to a choice between riskier investments that hold out the promise of greater rewards and lower-risk investments with less potential. The design of executive compensation is a key governance tool for shaping those decisions. The Pay-for-Performance–PayRiskScore Matrix helps identify how the company’s executive compensation architecture affects these trade-offs. Plotting the company’s position, as well as that of its peers and the industry norm, enables decision makers to identify the link between executive compensation architecture and incentives to take on or reduce risk and to deliver superior shareholder returns (see Figure 3). The score can help decision makers determine whether the company’s executive compensation design reflects its compensation philosophy and business objectives. The “ideal” score will vary according to a company’s strategy and current circumstances. Companies with a more egalitarian pay philosophy would likely choose a different target than companies facing significant default risk on their bonds or those with little debt and significant growth potential.

Figure 3
The Watson Wyatt Pay-for-Performance–PayRiskScore Matrix

Source: Watson Wyatt Worldwide.

Compensation design, performance and risk
Over the last year, many stocks dropped by 30 percent or more, and some large companies are now in severe financial distress. In many ways, this financial environment has been a “stress test” of compensation designs. Unsurprisingly, our research and that of others show that companies that managed their risks better through 2007 outperformed their peers in 2008. And distinguishing features of these companies include more favorable PayRiskScores based on their 2007 executive compensation design, and more company stock owned by executives at year-end 2007.

Our research also shows that companies with greater executive ownership and higher PayRiskScores managed risks more effectively in 2008 as well. If they started the year with high risks, these companies were more likely to have lowered their risk by year end. If their risks were low to begin with, they were more likely to remain low-risk. Moreover, in 2008, total shareholder returns were 7.5 percent higher at companies with favorable PayRiskScores, more executive ownership and higher pay-for-performance scores.

Conclusion
Companies are facing many challenges in this recession, including reexamining and possibly revamping their executive compensation programs. Current TARP participants are already required to show that their programs do not encourage excessive risk taking, while others are looking at the issue as a matter of good corporate governance. But empirical data contradict conventional wisdom, which many companies are relying on to make executive compensation decisions.

Watson Wyatt believes companies should examine their executive compensation programs from two perspectives: first, whether pay programs are paying executives for better performance and higher return to shareholders; second, the extent to which these programs might encourage executives to take excessive risks.

Many companies already undertake the first analysis in response to demand from shareholders and pay critics alike. But the need for the second analysis has only recently emerged, and most companies are basing their pay decisions on the conventional wisdom that the following executive pay architecture leads to better risk management:

• Lower annual incentive leverage
• Higher proportion of fixed to variable pay
• Fewer overall equity incentives
• Restricted stock instead of stock options

Rather than relying only on widely accepted theory, Watson Wyatt encourages companies to conduct an empirical analysis that demonstrates a link between pay architecture and a lower risk profile. Companies must also balance program elements that mitigate excessive executive risk with elements that continue to pay for performance to identify the executive compensation architecture most appropriate for them.

¹ See, for example, Watson Wyatt’s 2008/2009 report on executive pay, “Executive Compensation in Uncertain Economic Times,” www.watsonwyatt.com.
² The Z-score is a measure widely available through Standard & Poor’s or Bloomberg, which was developed by New York University Professor Edward Altman and has been used for more than 40 years to measure credit risk.
³ Effective risk management maximizes the expected return on the firm’s investment for a given level of risk or minimizes the risk for a given level of expected return — in short, achieves an optimal risk-return trade-off. In the current economy, financial returns have been higher for companies with lower credit risk than for those that took on higher risk, so effective risk management has been equated to reducing overall risks. But in periods of robust economic growth — or for firms with strong growth prospects — the market may reward companies that assume greater investment risks. Z-scores tend to increase along with a higher market-value-to-book-value ratio but to decrease given certain accounting factors, such as greater leverage or declining working capital ratios. In this case, effective risk management might lead to lower Z-scores and higher risk in the short term. However, in the aggregate, the relationship between Z-score and probability of bankruptcy has been robust in both economic growth periods and recessions.