Does paying big bonuses result in higher performance?
The issue of bonuses is quite a divisive issue (divisive in the sense that most people think they have got way out of hand, while a small minority think they are needed to attract the best people). George Osborne placed himself in the latter camp when he flew to Brussels recently to argue against imposing a cap on bankers bonuses of 200% of their annual salary. But is there any evidence that paying high bonuses results in those incentivised by them achieving higher levels of performance?
Now most people who have taken any sort of course in economics will have had it drummed into them pretty early on that people (referred to as consumers) always act rationally and seek to maximise their utility. Often economists use models to help them predict what will happen in certain situations that are based, in part on these assumptions. But there is a branch of economics known as behavioural economics which does not just accept these assumptions as facts, and actually try to set up scientific experiments (a novel idea I know) to see how people actually behave in certain situations.
A recent series of blog posts on Robert Nielsen’s site led me to a book by behavioural economist Dan Ariely called “The Upside of Irrationality“. His earlier work establishes the evidence that people are not in fact rational in the way many economists assume they are, and actually act irrationally in many situations. In The Upside of Irrationality (which I recommend to anyone), Ariely outlines some of his experiments into this irrationality. Chapter 1 of the book is called “Paying more for less: Why big bonuses don’t always work”, and its conclusions are very interesting.
Ariely and his team set up an experiment where volunteers would be asked to perform a series of tasks requiring varying levels of skill in return for payments based on performance (examples of tasks included remembering and repeating a sequence of flashing lights and throwing balls at a target). Each task had thresholds for performance. There was a good standard and a very good standard which would results in a bonus if achieved, but if participants failed to achieve the good standard, they would receive nothing.
Participants rolled a dice to determine the size of the bonuses they would be playing for – either a maximum of an equivalent of a days pay, two weeks pay or five months pay (the experiments were done in India where wages are very low).
So what were the results? Did those playing for the highest bonuses perform the best as we might expect? After all their incentives to do well were the greatest. Actually no. What Ariely’s researchers found was that although there was not much difference in performance between those playing for the low and medium-sized bonuses, those playing for the largest bonuses performed considerably worse. Such was the pressure they were under, they choked when they needed to perform. Note though, that the people in this group actually received the most money, but the “bang for the buck” was much lower than for those receiving lesser incentives.
Another interesting element to the research was that initially, they wanted to give the participants the money up front and then take it back if they failed to reach the required levels. Arguably, this reflects the reality of banker’s bonuses in that they have come to expect these large payments regardless of performance. For example, RBS recently announced it was paying bonuses of £600m despite making a loss of £5.2bn. The researchers had to abandon this early on though after those playing for the big bonuses either failed utterly, or cheated by running off with the money after failing to perform well!
What does this show then? It seems that the evidence shows that paying very large bonuses is unlikely to maximise the performance of staff, and that they can in fact worsen it significantly, particularly if they those receiving them come to expect them in all situations. What was found in subsequent experiments though was that big bonuses did result in improved performance when the tasks being performed were mechanical in nature.
Research was conducted with MIT students where they were asked to perform two tasks, two times each, once for a small bonus, and once for a large one. The first task was just clicking keys on a keyboard, the second involved solving maths problems. The results showed that in the key clicking task, students performed better when a high bonus was offered, but in the other task, where some brain-power was required, as above, the higher bonus had a negative impact on performance. As Ariely writes:
“The conclusion was clear: paying people high bonuses can result in high performance when it comes to simple mechanical tasks, but the opposite can happen when you ask them to use their brains – which is usually what companies try to do when they pay executives very high bonuses. If senior vice-presidents were paid to lay bricks, motivating them through high bonuses would make sense. But people who receive bonus-based incentives for thinking about mergers and acquisitions or coming up with complicated financial instruments could be far less effective than we tend to think – and there may even be negative consequences to really large bonuses.”
The results then seem clear, those currently receiving large bonuses would probably perform better if their bonuses were much lower, while those currently receiving no bonuses (low-skilled workers) would perform much better if the were. The sky will not fall in if bonuses within the EU are restricted, and shareholders of these big companies would do well to appraise themselves of Ariely’s research before approving large payouts for company employees.
I’ll end with this quote Ariely cites in his book from US Congressman Barney Frank, who in a speech to in 2004 to an audience of bankers said:
“At the level of pay that those of you who run banks get, why the hell do you need bonuses to do the right thing? Do we really have to bribe you to do your jobs? I don’t get it. Think what you are telling the average worker – that you, who are the most important people in the system and at the top, your salary isn’t enough, you need to be given an extra incentive to do your jobs right”
Does this mean that capping bankers bonuses will make them even more shit?
Banks are a cartel that skim a percentage from every transaction. They generate so much cash by doing it that the value of it loses meaning and they splash it around among themselves with impunity and congratulate themselves on how they earned it. The truth is that they are parasitic on the rest of us and it would be impossible for them to fail.
Hi,DID NOT GET PASSED YOUR FIRST paragraph because you question is put BACKWARDS. Also in a bad form. Bonuses generally should be paid as reward for outstanding performance resulting in higher PROFIT or Quality of Operations
In the case of BANK BONUSES the higher Bonuses are paid for TAKING RISKS outside the norm.
Risks that no other person would be prepared to take. there are many other types of Bonus reasons, so your question needs to be reframed.
Your article could be taken as an argument for raising the basic salaries of bankers, but if most of casino banking is a socially useless activity, as many economists argue (Turner Kay et al), then we just don’t want it, bonuses or not. The arguments for performance related pay, other than in menial work as you say, is not well supported by the evidence, but such laboratory experiments with MIT students is very poor evidence. There are reams of much better evidence from ‘in the field’ studies.
You misrepresent economists, but its a very common sin. Some ‘freshwater’ (e.g. Chicago) economists do indeed assert the near total dominance of individualistic utility maximisation, but I know no economist, and I do know many, who claims that people “always act rationally and seek to maximise their utility”. As you suggest in your next sentence, it’s just a modelling assumption, and good economists know all models are wrong but some are useful. We seldom see money blowing about our streets and it’s not because of litter conscious behaviour; utility maximisation, perhaps sadly, can often be a sufficient assumption for a useful explanation. You also need to distinguish between individual ‘rationality’ and market dynamics, there are many cases when markets can be maximising when individuals are not.
Incidentally, the survey evidence is that most economists are actually left-leaning.
Andy, whether economists believe in the primacy of utility maximisation or just use it as a modelling assumption, the end result can be the same. If you put garbage into your models, you are going to get garbage out. When you are building models based on human behaviour (whether group or individual), it’s pretty important to start with how they actually behave. So controlled lab experiments are actually very good evidence (one on it’s own not so much, but behavioural economists have done many and this is a blog post not a peer-reviewed article).
I’m not sure markets are ever maximising in the way economists assert either (firms don’t profit maximise for example).
As to economists being generally left-leaning, I would have to agree that that has been my experience too, but they are none too vocal at the moment (with notable exceptions).
Btw, are you Andy Ross of the GES?
Garbage in garbage out is an excellent maxim, but it applies to data inaccuracy. It is much more problematic to say that a behavioural postulate for a model is inaccurate. Kahneman, the father of the behaviourist revival for behavioural economics is actually very old indeed, accepts both ‘quick and slow’ thinking as relevant.
An analytical assumption does not have to be descriptively accurate, for example, sea level is descriptively nonsense, but it is very useful when navigating. There are loads of examples where a simple individual maximisation assumption yields revealing results, and Tim Harford makes a living from this in less formal analysis.
Laboratory experiments are heavily criticised for lack of context and assumptions being made about how representative the MIT student sample is. But my point is that there is much better evidence you could have used to support your contention that performance related pay does not generally increase performance in high level tasks. It is from actual participants in actual real world situations, for example, there’s a man in Aberdeen University who spends his life studying this. Though if the converse were true it still would not justify bankers’ bonuses.
Interesting you mention firms not maximising. I was actually taught by Robin Marris! Wilkinson, Simon and many others tried to displace profit maximisation as the usual assumption, but they largely failed as it does prove so useful, and then Solow showed that the alternatives are so constrained that there is usually little loss of explanation from just using profit maximisation anyway. Marx always did! But I do think BE will prove more durable than this, its already made enormous contributions, though it may be largely re-absorbed e.g. through neo-classical devices such as ‘rational inattention’. It is generally seen now by analysts as complementary to main stream rather than as a substitute.
But I fully agree with your main contention, rational maximisation taken to extremes is both unrealistic and produces garbage predictions and explanations e.g. efficient markets, real business cycles, rational expectation etc.
I am ‘Andy Ross of the GES’ no longer, free now to be impartial. So- yes, I think Osborne is very wrong, but I’m still passionate (stickler?) about good analysis and evidence from wherever it comes. Though I take your point about blogs contrasting with peer reviewed articles, civil servants don’t get to practise blogs, and so promise to think on your advice
It’s good to see you commenting on blogs like this Andy. Although you won’t remember, we’ve actually met a few years ago. I attended one of your intro courses in Sunningdale. Thought you were a good bloke who knew his stuff. Have been out of the civil service for a while now though. Didn’t know you had left. Have you ever considered blogging? Your inside knowledge of the Treasury and policy making in general would make fascinating reading. I’d be interested to know what the feeling is in the GES about Osborne’s economic policy for example. Have many people left?
Hi- Thanks for those kind words, and always delighted to hear from ex-inductees! I can’t identify you from ‘alittleecon’ but I do hope things are good for you. I really enjoyed those Sunningdale courses as the students were so bright and they really put so much in- v.good fun. I only left GES & Treasury this February. You’ll understand I am restricted in what I can say from inside knowledge, but I’m enjoying doing some teaching about economic policy making at Leeds Uni, I feel that because of my experience I can contribute things about how economics is used by practitioners in ways that most academics can’t. Good fellows though they are. On ex-GES, I think Jonathan Portes’ blogs are generally excellent. And there are many very respected economists saying the Osborne Merkel ridiculously anti-Keynesian policy would be and has been disastrous e.g. Simon Wren-Lewis, Ann Pettifer, ‘Danny’ Blanchflower, Vicky Pryce, Martin Wolf etc . It’s getting them heard in the largely right-wing popular media that has been the problem.
But now a confession, I do read a lot of blogs in keeping up to date, but frankly I have no idea how to do it. As my contributions here no doubt reveal!
Cheers- You take care.
Thanks Andy. Of the names you mention above, I guess I would identify most with Ann Pettifor, although all are on the right side of the argument at the moment. Would be interested in your views on the Post-Keynesian school? I think they have some interesting things to say about the Euro and money/government finance. I believe Malcolm Sawyer who is a leading PKer is also at Leeds Uni. Here’s some leading PK blogs just in case you’re interested:
I also think Chris Dillow’s blog is excellent, although I think he is more Marxist in outlook:
I think you forgot one of my favourite blogs:
Thanks, I knew some but not all of those. I listed only people who aren’t dismissed as ‘lefty political agitators’, but who are also well regarded by the mainstream as well as them being strongly anti-austerity in recessions. The range of interesting contributors is indeed much much wider, and includes ‘Marxists’ (which excludes Marx himself then) and contributors from the right as well. Yes Malcolm is at Leeds, I’d give a lot to be even a small fraction of the economist he is, but I’m never sure the post-Key/Post Kalecki tags are helpful. For example, ‘Heterodox’ is a real kiss of death to having any influence at all! For me, and I may be completely wrong, such tags smack of self-rightous impotent, which I never find appealing. Malcolm, Arestis, Harcourt, Chick, Tily etc. are impressive economists even in terms of mainstream ‘standards’.