How Virtually Every Pay Regulation Has Backfired

Few topics make the public as angry as CEO pay. In the UK, the average FTSE 100 CEO earned £5.4m in 2015, 148 times the median worker. For US S&P 500 CEOs, these figures are even more extreme: $12.4m and 335 times.

This pressures politicians to do something about CEO pay. Indeed, regulating pay might nowadays be an even better way of winning the public’s approval than tax cuts and spending increases. In the US, both Donald Trump and Hillary Clinton have been unusually united in blasting the high levels of CEO pay. The UK’s new Prime Minister, Theresa May, has proposed putting workers on boards, making say-on-pay votes binding rather than advisory, and forcing firms to disclose the ratio of CEO pay to median worker pay.

But, let’s take a step back. First, it’s not actually clear that CEO pay is a problem to begin with. In an earlier post, I explained that many accusations are based on myths, which simply don’t hold up when you look at the evidence. Second, even if it were, regulation may not be the best way to fix it.

In an excellent article, “The Politics of Pay“, Kevin J. Murphy of the University of Southern California (one of the world’s leading experts in executive pay) argues that the entire history of U.S. compensation legislation is littered with unintended consequences. Simply put, 80 years of evidence shows that regulating pay doesn’t work.

How can politicians keep getting it so wrong? Because their motivation for regulating isn’t to increase social welfare or curb inequality, but to appear tough. When a pay scandal breaks out- even if it’s a single isolated incident – the public demands that politicians do something. For a politician, taking a bad action is better than taking no action. Here are three examples:

1. Golden Parachutes

In 1982, William Agee, the CEO of Bendix, was given a $4.1 million golden parachute after his firm was taken over. The public was outraged, and demanded that Congress do something about it. It responded in 1984 by imposing severe tax penalties on golden parachutes that exceeded 3x salary. Sounds reasonable – surely taxes would deter such excessive payments? But in fact, it encouraged them.

  • Companies Introduced Golden Parachutes. Back in 1982, golden parachutes were fairly rare. But, the new law alerted CEOs to this attractive clause. So, many CEOs who previously hadn’t heard of golden parachutes now demanded them.  By 1987, 41% of the largest 1,000 firms had golden parachutes, which rose to 70% by 1999.
    • Murphy shows that this problem isn’t specific to golden parachutes. Most laws forcing disclosure of perks led to CEOs receiving more perks, as they now saw what their peers were getting and demanded to get the same.
  • Companies Increased Golden Parachutes. The “3x salary” cap implicitly suggested that a golden parachute is OK as long as it doesn’t exceed 3x salary. So firms who had modest golden parachutes prior to 1982 suddenly increased them to 3x salary. By 1991, 47.5% of golden parachutes were at this level; by 1999, this was 71%.
  • Companies Introduced Tax Gross-Ups. Some in-demand CEOs got the company to pay the tax themselves. And, even though these tax gross-ups were introduced in response to taxes on golden parachutes, they subsequently spread to company cars, club memberships, and corporate jets. Congress unleashed a monster.

2. The Clinton $1 million cap

In his 1992 election campaign, Bill Clinton (foreshadowing his wife 14 years later) lambasted CEO pay as being excessive. A year after his election, Congress changed the tax code to prevent companies enjoying tax deductions on salaries exceeding $1 million – unless they were performance-based. This again backfired:

  • Companies Increased Salaries to $1m. The new code implicitly suggested that a high salary is OK as long as it doesn’t exceed $1m. Sound familiar? Clinton could have learned from history and seen what had happened with golden parachutes. But, his incentives were not to create social value; instead to appear tough.
  • Companies Paid Bonuses For Mediocre Performance. Let’s say a firm needs to pay a CEO $1.5 million to attract her, because she’s uniquely qualified. How can it do so without suffering tax penalties? It pays her $1 million, and reclassifies the other $0.5 million as a bonuses with a easy-to-meet target, so it’s effectively guaranteed. This defeats the whole point of bonuses – they should only be given for good performance.
  • Stock Options Increased. Since stock options are performance-based, their use started to significantly increase. Murphy points out other tax and accounting rules that also led to the explosion of stock options. So, if people complain about excessive stock options, it’s regulators they should be complaining about.

3. The EU Banker Bonus Cap

The problem isn’t confined to the US. A separate Murphy article, “Regulating Banking Bonuses in the European Union: A Case Study in Unintended Consequences“, addresses the European Union’s cap limiting banker bonuses to two times salary. There are several problems:

  • Increased Salaries. To remain competitive (with non-EU banks), a fall in bonuses must be accompanied by an increase in base salaries. Thus, if the bank fails and the banker gets no bonus, he gets a higher salary (for non-performance) than he would have before.
  • Increased Bankruptcy Risk. The advantage of bonuses is that banks can cut them when times are hard, and so remain solvent. You can’t cut salaries. This increases the risk that banks go bankrupt – ironic since the main driver for regulation was the financial crisis.
  • Reduced Incentives. The reduction in bonuses decouples the banker’s pay-for-performance. This reduces his incentive to manage risk as, even if the bank performs poorly, he still receives a high salary. It also reduces his incentive to create value by innovating, as he captures less of the upside. Just following the status quo is enough to be well-compensated.

Perhaps you might think – there’s nothing wrong with regulation, but it’s sneaky firms responding to regulation (e.g. by reclassifying salaries as bonuses). This argument is false on two grounds. First, responding to regulation may not be devious. It may well be that the firm truly needs to pay $1.5m to attract the CEO, and so get around the regulation legally. This is little different to a restaurant getting around liquor license laws by allowing diners to bring their own alcohol. Second, even if the response is devious, it’s how firms and CEOs actually behave. You might say – in an ideal world, CEOs shouldn’t demand perks when they find out about their peers’ perks. But, we don’t live in the ideal world, we live in the real world. A textbook solution only works in textbooks; any real-world solution should take into account how people behave in the real world.

So, what is the solution? To do nothing? Far from it. But, to leave the decisions to major shareholders, who have the incentive 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 demoralizes employees – shareholders suffer the consequences. And, things are being done. 11 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 – lengthening vesting horizons (to encourage the CEO to think long-term) and paying with debt rather than just equity (to dissuade excessive risk-taking).

Moreover, when pay is inefficient, it is often a symptom of a more 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.

The bottom line is that, to the extent pay is a problem, it should be shareholders, not politicians, who fix it. Leaving it to shareholders is like a doctor choosing self-cure rather than amputating a limb. The amputation is the most dramatic action, but may not be the most effective. The goal of policy shouldn’t be to write headlines, but to create long-term value for society.

4 replies
  1. Sam
    Sam says:

    Thanks for this post, Alex. On Golden Parachutes and the Clinton salary cap, is it clear beyond doubt that these trends wouldn’t have happened without the legislation? I mean, it’s possible (in fact, quite likely) there was going to be an increase in the use of golden parachutes regardless. I understand the logic behind the arguments, but the figures presented don’t prove the assertions. Are there more detailed figures which show a sudden break in a trend, for example?

    As for whether CEO pay is a political matter, I would say that high pay – more precisely, pay inequality – is definitely an issue which requires political solutions due to the corrosive effect it has on society. This is quite distinct from looking at CEO pay in relation to company performance, which you have covered thoroughly.

    Reply
    • Alex Edmans
      Alex Edmans says:

      Thanks very much Sam for the thoughtful comment. You’re correct that time trends are insufficient to show causality; they can only show a correlation. One could argue that causality is the other way – pay and parachutes were rising anyway, and that’s what caused the legislation. However, the following evidence makes me think that regulation did pay some causal role – the fact that the changes were “sharp” and precisely in line with the regulation. A general increase in parachutes would not suggest causality, but pensions rose precisely to 3x salary. Similarly, a general increase in salaries would not suggest causality, but salaries rose precisely to $1 million.

      You’re totally right that inequality is a very important social issue. However, this inequality arises not just due to CEO pay, but also bankers, lawyers, consultants, trust fund babies, pop stars, actors, athletes. (There’s actually evidence that CEO pay is not responsible for the rise in equality versus other professions). E.g. one of the best paid “employees” of Nike is not an executive, but Joseph Schooling ($30 million/year, largely unrelated to performance). He has a great story, but so do many executives who grew up on council estates etc. A high income tax (e.g. for everyone earning £1 million) would address inequality from all of these sources and be far more effective.

      Reply
      • Sam
        Sam says:

        Thanks Alex! I don’t suppose there are graphs showing this effect clearly, are there? I’m certainly inclined to believe it, but I’m a little uncomfortable about the lack of killer evidence on something that has a petty big message for policy makers.

        As for taxation being a better means of addressing high pay than targeting specific cases, I’m totally with you on that. The danger is, of course, that with no signs that this is at all likely, we’re kind of stuck, and the implications are an increase in political polarisation and the rise of Donald Trump and his ilk.

        Reply
        • Alex Edmans
          Alex Edmans says:

          There are no graphs, but the original Murphy paper (which is surprisingly readable for an academic paper) does specifically refer to parachutes rising to precisely 3x (rather than any other amount) and salaries rising to $1m.

          It is pleasantly surprising that a Tory government seems to want to tackle inequality. However, them going after CEOs in particular, rather than the wealthy in general, makes no sense. Unfortunately, sometimes politicians are motivated by envy or scoring political points, rather than true social welfare creation. (A CEO is much more high profile than a rich trust-fund baby, so an easier target, even though the latter is less deserving).

          Reply

Trackbacks & Pingbacks

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *


*