Winner of the New Statesman SPERI Prize in Political Economy 2016

Friday 27 May 2016

Bonus culture

Diane Coyle has an excellent article in the FT about an apparent puzzle. Why do executives get incentive bonuses (extra pay on meeting some target), but most workers do not? Her article is based around a classic paper by Bengt Holmstrom and Paul Milgrom. Their basic argument is that incentive pay linked to specific targets works (it increases effort) when tasks are simple and effort can be easily measured. However if tasks are complex, and only some aspects of performance can be accurately measured, incentive pay can distort the allocation of effort between those tasks, leading to undesirable outcomes. As Diane says “pay structures not only incentivise effort and direct risk-taking, they also determine the worker’s allocation of effort between different tasks.”

So target related bonuses make sense for workers conducting simple tasks where effort can be easily measured, but are a bad idea for workers undertaking complex tasks where only some aspects of performance can be measured. To quote Diane:
“Indeed, the best arrangement would seem to be the opposite of the pattern we observe now. Corporate executives and senior bankers doing complex jobs involving many impossible-to-monitor activities are the last people who ought to be paid via an incentive scheme; while bonuses for fast-food workers or shop-floor employees make more sense.”
The implication she draws is straightforward: the bonus culture for corporate executives and senior bankers should end. But this leaves us with a puzzle: why did this bonus culture arise in the first place? Perhaps bonuses created something beneficial that we are missing.

Here is a simple conjecture, based on another paper by Piketty, Saez and Stantcheva which I discussed here. They note that increases in executive pay are strongly correlated with reductions in the top rate of income tax. Their explanation notes that executive pay is the result of bargaining between the executive and the firm. The executive has a lot of bargaining power (what successful firm wants their CEO to quit), but whether they choose to use it depends on the reward from doing so. If top tax rates are low, the rewards are high.

The executive still has to convince their firm to pay them more. What better way to do this than to suggest they get paid a lot more only if the company is successful. In the climate of the 1980s and 1990s in the UK and US (when the income share of the 1% took off) that argument would have seemed pretty convincing. My conjecture therefore is that bonus pay became endemic among executives and senior bankers not because it was more efficient for the firm, but because it was a useful tool in a bargaining game. [1]

This argument completely reinforces Diane’s conclusion. Executive bonuses are a way for senior management to extract rent from their firms, which is a quick way of saying that these high salaries redistribute money from everyone else to themselves. A consequence was that they reduced efficiency by diverting the executive’s attention to just hitting specific targets. One final thought, as we in the UK are obsessed with Brexit right now. It was the EU that passed laws limiting bonuses in the financial sector, and it was George Osborne that spent the UK public’s money trying to stop that law coming into effect.

[1] Bonuses can play a useful role in small firms where revenues are volatile, as Chris Dillow notes. That argument hardly applies to the CEO of a large multinational.  


  1. Regarding the point you have buried in a footnote, actually large multinationals are the very epitome of firms with volatile revenues - you only need to see the daily news items about how an individual large company's revenues have dropped/increased by more than 10%-20% from one year to the next.

    1. I don't think that the point of the footnote is that the revenues of small firms are necessarily more volatile than those of large firms, rather that bonuses are a helpful remuneration tool for firms that meet both the criterion of being small and that of having volatile revenue streams. A small firm with stable revenues, that is to say, or a large one with volatile revenues, will both, though for different reasons, have much less need of bonuses as a means of responding to fluctuations in revenue.

  2. I don't really understand the argument. Isn't CEO's bargaining power related to the value they bring to the company? If a CEO can harm the company by leaving it, doesn't it mean that his contribution is actually valuable? Are you arguing that we should ignore every market signal since it's entirely inefficient?
    I agree that rewards plays a big part, but what happened in the past with high tax rate was that CEOs got all kinds of alternative perks which didn't pay taxes and probably costed a very similar amount to the company anyway.
    By all means argue that CEO pay is too high. I actually agree with that. By variable vs fixed pay is a red herring. Plus isn't always the complaint that bonus are not actually variable enough: CEO's bonus still gets paid when performance is crappy.
    The only thing bonus directives achieve is higher base salaries for bankers and I guess more fired bankers (to reset the base salary).
    If you want, argue that bankers should be paid with the instruments they actual sell or create (to align incentives), I could get behind that. If you think they should be paid less, start or invest in a bank with a different model. Lobby the government to try a different model in the nationalized banks. Raise the fee banks pay for deposit insurance. There are better ways to achieve what you want.

    1. I am not an economist and I don't have any data, but I believe CEO compensation is way out of line with the value they bring to the company. In fact, I think a lot of executive compensation is. We know that wages of the lowest earners are not based on their marginal profitability (sorry, don't remember the link). I think labor economics has been badly served for at least 70 years.

      The point that the bonus always gets paid even when performance is terrible is right on.

  3. Diane's piece is indeed excellent. I am not aware of any academic research which finds that performance related pay actually "works", except perhaps in straightforward sales environments ie commissions, rewards for hitting sales targets.
    The perversity of top pay is also seen in this bizarre use of language: to most of us a bonus is a trivial and unexpected thing, whereas at the top end bonus has come to mean something vast and expected, perhaps even guaranteed. This linguistic confusion symbolises, it seems to me, our problems with excessive pay at the top.
    Stefan Stern, High Pay Centre

  4. A very interesting post. Thank you. Bonus culture is also strong now in large law firms, which is one of my areas of interest. Those bonuses are often built around inputs (hourly time recorded) and billing (which often/usually derives from hourly inputs). One of the arguments that is increasingly made (although the claim that it drives profitable behaviour is also made) is that such bonuses are 'fair' because they reward effort and success.

    The claim that bonuses reward productivity is not so dumb for lawyers operating on an hourly fee model, but it is not very comforting for their clients. Bonuses reward hourly inputs. Complex (professional) work is reduced to simple units of time, and clients have to cope with (and the more sophisticated try and manage for) poorly aligned incentives that encourage waste, gold-plating and a reluctance to innovate.

  5. Revenue volatility has something to do with it, and the lack (AIUI) of a conventional career structure. But at the end of the day I think people who routinely work with large sums of money are overpaid for much the same reason that people who work in bakeries eat a lot of pies. Not that they're crooks, any more than the person taking home a couple of sausage rolls is a crook - just that there's a lot of it about & nobody minds if some of it sticks to your fingers.

  6. Revenues in investment banking are also very volatile, which is why bonuses are such a large part of remuneration. And this applies to everyone, not just the senior bankers. This is why the bonus cap in banking is such a stupid idea: banks just increase salaries to compensate so remuneration is less aligned with the success of the firm and earnings volatility increases.

  7. Finaly that someone hit it on the head about the effects of Marginal Tax Rate. There is more to it.

    Ever since i saw the chart of productivity and income separating the path in around 1968, long before oil shock of 70', i was puzzled by it untill i connected JFK's lowering MTR from 91% to 75% in 1963.

    High MTR is not needed for funding a government but for the following reasons;
    i) it prevents menagement from taking income share from workers
    ii) it prevents them from looting a company they manage by going into more debt for operating expenses in order to get higher profits
    iii) it prevents them from cuting on investments for developement of a firm
    iv) it prevents them raiding cash on hands and recomending going into more debt which makes firms vulnerable to recessions. Even posting shares as colateral for loans becomes risky when shares fall in value even tough there is no recession. Fall in share values given as colateral will imidiatelly require more colateral which can banckrupt a firm.
    v) it discentivises management from cutting benefits to employees (for USA), like health insurance, vacation time and sick pay whie enlarging benefits for themselves.
    vi) high saving rate of management forces rest of the population to go into debt in order to fuel demand

    High MTR prevents management from bribing politicians to further cut MTR and puts more taxes on middle class.

    Yes, today's GFC have its roots in JFK lowering the MTR to 75% from 91%.
    The highest prosperity achievemnt of human history was achieved under MTR above 90%.

  8. 2 points.
    I read a piece many years ago that said that actually bonus pay is a very ineffective way of increasing efficiency /productivity. It works in the short term, but the effect tails off. The really effective way to encourage (yes even the erks like me) productivity is to praise, affirm, and recognise effort, (and also, probably, to point out failure, non-judgementally). i.e. we are a team, thank you for your contribution.
    2. I have thought for a long time that high profile jobs should be allocated on a Dutch Auction basis. So for example the next head of say the BBC should be advertised, and say 5 perfectly capable people are called for interview. The lowest bidder gets the job. It's called competition.

  9. PS. and please forget pensions top ups, share options, golden hellos and goodbyes. Just a straight salary, even an hourly rate, just like the rest of us.

  10. <a href=">This comments thread</a> is quite interesting and (given the subject matter) reasonably short on acrimony. I'd call attention to the comments by Phil (i.e. me) and Tom Slee.

  11. I think one other reason for the growth of the bonus culture may be the transition in areas such as law and investment banking from a partnership structure to something closer to a limited company. When the firms concerned were small and required little capital (this applies particularly to law and accountancy firms), pay for senior personnel was naturally linked to profits and, importantly, those who gained in good times would be liable to cover losses in bad ones.
    This type of structure became hard to maintain once the firms grew (often through merger), and needed significant amounts of capital to finance modern IT equipment etc. Also the requirement that partners should carry unlimited liability, and should put substantial funds of their own into the partnership, made it hard for these firms to recruit able staff who did not come from 'old money' or the right family. So just as nineteenth-century capitalists persuaded Governments to change corporate law and introduce the principle of limited liability, late 20th century lawyers and bankers persuaded Governments to introduce limited liability partnerships, or to allow them to convert their firms to conventional corporations.
    But this restructuring separated ownership from control, and raised the problem (which economists had been aware of since the 1930's) of how limited companies could continue to motivate their senior employees (many of whom would previously have drawn partnership dividends). Like the stock options offered to corporate executives, the bonus system was seen as a way of doing this.
    Another point which may be relevant is that making a substantial part of the remuneration paid to relatively junior members of the firm a bonus, the size of which depends on the good opinion of the employee's immediate boss, satisfies the manager's urge to control his or her subordinates and ensures an adequate supply of sycophantic acolytes.

  12. Note. In the US, the tax code was changed in the early 90's to limit the tax deductibility of salaries over 1 million dollars. But this limit does not apply to "performance based" compensation.

  13. It may be worthwhile to consider where exec comp is going. These top dogs think they're as good or better than anyone else. So when one of them is managing a $40 billion revenue corp and takes home $100 million on a regular basis (think Oracle) for a take of a quarter percent, and a hedge fund guy with $40 billion in assets takes 2% or $800 million in a bad year, you can be sure the pressure among the execs is to equalize these ratios.

    We ain't seen nuttin' yet.


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.