Monday, 22 November 2010

Conformity and rewards

One of my books currently on the go is Why Societies Need Dissent by Cass Sunstein, which is rather ace. There is a lot of very useful material in there is you're interested in a more behaviourally-informed view of governance. Just a quick excerpt below about one of my favourite subjects - the impact of financial incentives on behaviour. In this instance Sunstein is looking at how incentives affect conformity of opinion.
When people stand to make money if they are right, the rate of conformity is significantly decreased if the task is easy. People are less willing follow group members when they stand to profit from a correct answer. But there is a striking difference when the experiments are changed to make the underlying task difficult. In that event, a financial incentive for correct answers actually increases conformity. When the question is hard, people are more willing to follow the crowd if they stand to profit from a correct answer.
It's just a snippet, and there are plenty of other bits in the books that give pointers to how people can, and do, overcome the pressure to conform. Nonetheless it does seem to provide grounds for another argument against the merits of performance-related pay. Unless we think running a company is a simple task. And maybe it tells us something about benchmark-hugging in the investment world too.

PS. As I thought the book lends a bit of weight to my argument that it would be better if more investors went public with their concerns.


Aled said...

A slightly different but related comment from something I've been reading recently on evidence-based management:

In the chapter on financial incentives and when they do & don't work it uses examples of very simple tasks (e.g. the no of replacement windshields an employee can fit in a day) vs complex tasks (delivering against corporate strategy & objectives...).

It argues that for simple tasks with no dependent activities (ie from other employees) bonuses can work very well. For complex tasks or those involving team work or the reliance of various departments on one another for the comlpetion of a task, incentives are close to useless as individual behaviour has limited scope for influencnig dependent parts of the process. At least that's what I understood when I read it...

ho hum

Tom Powdrill said...

Hi Aled

Yeah that message seems to come through from a lot of the stuff I have read about incentives. provided the task is simple they can have a positive impact. But if the task is complex or requires some degree of creativity incentives don't work or have a negative impact.

John Kay came out with a good line that carrots are good if you're employing donkeys and you know exactly what you want the donkeys to do.

There's an interesting twist to all this in the educational sector. Roland Fryer's research on using financial rewards for school kids found that you could improve performance - if you tie the money to inputs (books read, classes attended etc) rather than the outcome (grades). So if you can identify a simple. measurable input that positively correlates with the output you're after maybe there's a way round the non or negative effect of rewards.