How to regulate CEO pay (and how not to do it)

In the good old days political campaigns were about silly things like taxes, government spending, and foreign policy etc whereas today there about important things like CEO compensation and how government can regulated it to make the world a better place. In a column at VoxEU.org Alex Edmans considers the various arguments for regulating CEO pay and questions whether it is a legitimate target for political intervention. Some arguments for regulation are shown to be erroneous, and some previous interventions are shown to have failed. While regulation can address the symptoms, only independent boards and large shareholders can solve the underlying problems.

A obvious first question to ask is, Is regulation necessary? Our politicians don't ask this question they just assume it is. After all there are votes in say regulation is needed.

A contract is a private agreement between an employer and an employee. It is shareholders who bear the direct costs of the contract, and so it’s unclear whether the government should intervene.

A common argument is that pay has indirect costs – it causes CEOs to take actions (such as risk-taking) that affect society. However, there is very limited causal evidence that pay does have these effects. Even if it did, it is not clear why the government should regulate pay rather than these actions themselves. Firms take many other decisions that have far larger effects on society – firing workers, restructuring, or making large investments – which are rarely regulated. Thus, calls to regulate pay may be driven by envy, rather than pay actually being a critical social decision (Murphy 2012).

Moreover, we have remedies for bad decisions – private equity firms and hedge funds take large stakes in underperforming companies. They are certainly not in the CEO’s pocket, and they’re not afraid to make major changes – they even fire the CEO in many cases. But, they very rarely cut CEO pay (Brav et al 2008, Cronqvist and Fahlenbrach 2013). Thus, while large investors see many things to fix in a firm, the level of pay doesn’t seem to be one.

If we do what politicians don't and look at all the evidence, what do we find?

Public opinion is typically informed by a few high-profile examples of egregious compensation. The media will only report the worst cases because these are the most newsworthy. But, it is important to assess all the evidence – and that’s the role of academic research. Indeed, the large-scale evidence is that pay is highly linked to performance:

• A 1% fall in the stock price reduces CEO wealth by $480,000 (Murphy 2013);
• CEOs with high stock compensation deliver superior long-run stock returns of 4-10% per year (von Lilienfeld-Toal and Ruenzi 2014);
• Firms with high income inequality exhibit superior operating and stock performance (Mueller et al 2016).

So what is the role of the policymaker?

A policymaker should exhibit similar caution to a doctor. The Hippocratic oath is "first, do no harm", i.e. to ensure that any intervention doesn’t worsen the problem. Murphy (2012) describes how the entire history of compensation regulation is filled with unintended consequences. The forced disclosure of perks in 1978 increased perks as CEOs could see what their peers were receiving; the 1984 law on golden parachutes catalysed the adoption of golden parachutes by alerting some CEOs to their existence; and Bill Clinton’s $1 million salary cap led to CEOs below the cap raising their salaries to above it, and those above merely reclassifying salary as a bonus.

Indeed, regulation is often driven by political agendas – by politicians’ desire to be seen as tough, rather than to create social value. Far more important than taking action is taking the right action.

What to do? Two common ideas are pay ratios and employee voting. But

The motivation behind capping pay ratios is sound – to ensure the CEO is not paid for poor performance, and to improve equality. But, to ensure the CEO is paid for performance, we should compare pay to her performance, not the pay of the median employee. One argument is that a well-performing CEO should share the spoils with her workers – she wasn’t the only one responsible for her firm’s success. But, the flipside of bonuses for good performance is they allow punishments for poor performance – former JC Penney CEO, Ron Johnson, suffered a 97% pay cut when the stock price tumbled. Workers pay didn’t fall – and it shouldn’t have.

Moving to equality concerns, a focus on pay ratios can actually increase inequality. A CEO can lower the pay ratio by firing low-paid workers, converting them to part time, or increasing their cash salary but reducing their non-financial compensation (such as on-the-job training and working conditions).

and

Employees rarely wish to weigh in on a pharma company’s R&D policy, since this should be left to the experts – the scientists. But everybody believes that they are an expert on CEO pay, even though economics and finance requires as much expertise as science.

An employment contract is an extremely complex issue and cannot be whittled down to a simple number such as a pay ratio. How should we best filter out industry performance? Indexed options? Indexed stock? Options on indexed stock? Stock with indexed performance vesting thresholds? Confused? Well, so might the general public be – as they were during the UK’s EU referendum, so the debate was narrowed to simple dimensions and driven by misinformation.

To get past a vote, a board may focus on the ‘optics’ of pay (e.g. a low ratio) and ignore more important dimensions, such as performance targets being long-term rather than short-term. The median size of a Fortune 500 firm is $20 billion, and median pay is $10 million. Thus, even if a CEO was paid double what she deserves, that costs 0.05% of firm value. If a CEO’s pay is tied to short-term performance, she might take myopic actions that reduce firm value by several percentage points (Edmans et al. 2015).

Employees often come up with ingenious ideas. But, companies already have incentives to consult workers – without any need for regulation – and many often do. In Edmans (2011), I find that firms with high employee satisfaction – for which consultation is key – beat their peers by 2-3% per year. Even if boards only cared about shareholder value, they should consult employees. We want to push the message that employees and executives are in partnership, rather than employee consultation destroying value so we have to pass laws to ensure it happens.

And, there is a big difference between consulting employees and putting them on the board. Firms do market research by consulting customers, but don’t put them on the board. Otherwise, every corporate decision would become a political process. Indeed, Gorton and Schmid (2004) found that worker representation on German boards is associated with lower profitability and firm value.

So the big question, What should be done?

Should we do nothing? Far from it. But we should leave the decisions to major shareholders, who have the expertise and incentives to get these decisions right. High CEO pay comes straight out of shareholder returns, and if the contract causes the CEO to take bad decisions – or demoralises employees and customers – shareholders suffer the consequences. Unlike regulation, which is one-size-fits-all, shareholders can decide what the optimal pay package is for that particular firm.

And things are being done. Eleven countries have passed ‘say-on-pay’ legislation since 2002; Correa and Lel (2016) show that it reduces pay and increases pay-performance sensitivity. Interestingly, advisory votes are more effective than binding votes – in contrast to politicians’ desire to take the toughest possible action. We have also seen innovation in other dimensions of pay, such as lengthening vesting horizons to encourage the CEO to think long-term (Gabaix and Edmans 2009), and paying with debt rather than just equity to dissuade excessive risk-taking (Edmans 2010).

Moreover, when pay is inefficient, it is often a symptom of an underlying governance problem brought on by conflicted boards and dispersed shareholders. Addressing pay via regulation will solve the symptoms; encouraging independent boards and large shareholders will solve the problem. That will improve not only pay, but other governance issues.

What about equality? Kaplan and Rauh (2010) showed that high CEO pay has actually not been a major cause of the rise in inequality – it has risen much more slowly than pay in law, hedge funds, private equity, and venture capital. If inequality is truly the concern, it may be better addressed by income tax. This will address inequality resulting from all occupations (including sports, entertainment, and trust funds); it is not clear why CEOs should be singled out.

The bottom line is that, to the extent that pay is a problem, it should be shareholders, not politicians or employees, who fix it. The electorate will be more impressed by a politician halving pay than extending the vesting horizon from three to seven years, even though the latter will have greater impact. The goal of policy shouldn’t be to write headlines, but to create long-term value for society.

We can conclude that one size does not fit all and the shareholders of a firm have the best information and incentives to get CEO pay right. So why not let them?