15 April 2019 by Kirsty-Anne Jasper
The chief executive of the Behavioural Insights Team talks about how psychological insights and small changes can help companies to make big gains.
David Halpern is a psychologist and Chief Executive of the Behavioural Insights Team, unofficially known as the Nudge Unit. The team was originally set up as part of the Cabinet Office in 2010 and uses nudge theory (behavioural economics and psychology) to try to improve government policy and services and to save the UK government money.
Our particular focus is across all kinds of public policy issues. So that can range from what can you do on productivity; either you make businesses more productive, or the long tail of low productivity, to almost anything else.
Sometimes businesses will know about our work because we’re intervening on some other issues such as auto-enrolment of pensions, which some businesses may or may not thank us for, but it’s certainly increased the UK savings rate rather dramatically. Or sometimes it’s that we’re thinking about it in relation to what businesses can do to improve that productivity or improve in some other kind of way. The Behavioural Insights Team is a social purpose company co-owned by British government. We occasionally work for individual companies, but it has to be for a social purpose. We tend to be thinking about business as a whole.
honesty. Honesty is a human thing. We’ve basically built governance systems for lots of reasons; to make wiser and better decisions. And to encourage people to be honest and so on in reporting terms. When you look at it in those terms your see that it’s quite core to what behavioural scientists think about.
There can be lots of aspects. There are a lot of interests around where NED directors don’t ask questions. We have certain biases even at the top level and people very rarely ask for counter-factuals.
Even at the top level of boards this applies. And there are some very powerful illustrations of this. Sometimes it’s what do you not notice as a board. It’s those kind of dynamics that apply. Apart from individuals there are group effects. Most famously, so-called group think where you get a group together and you hope the group will be wiser, but in fact what happens generally is, constructed groups particularly, converge to more extreme positions.
There could be a strong case for making sure, even in a group around a board, that you assign someone to be the devil’s advocate. Ideally you don’t make it the most genial person. This is really important. Because groups have a tendency to reinforce, including boards, which is you what you often see when you have corporate failure. You think, well why the hell did they all go along with it?
It’s because they move in their own little bubble. And instead of taking an average position they move to an extreme position. So there are literally group dynamics, things which you play in.
Suddenly you have a process which is supposed to make sure that shareholders of companies know what’s going which is actually very easily subverted and doesn’t perform well.
Basically whenever you introduce human beings, a lot of human beings interacting with each other, you will see that we are full of biases and shortcuts - even the most smart of us - and these come out
“At the top of organisations,it is the most senior people who are often the most biased”
One warning sign you might have at the top of organisations is that the most senior people often are the most biased in certain ways. So for example, the most senior people are the most likely to be over-confident. They’ll think they’ve got a good idea, that it will really work and it’ll succeed, and they greatly overestimate the probability that this is the case.
Good governance is supposed to correct some of these faults, and sometimes it does, but they’re very, very powerful. And if you’re not even aware of what those likely kinds of errors are, it’s hard to correct them.
Well you can obviously read about it. Future generations I think will be much more trained in behavioural science and will be more aware of those kinds of issues. We recently did a bit of work on behavioural biases in government rather than corporate. Governments are in the governance business, which lays bare some of those types of biases and errors and then you can start to rehearse the kinds of things that you can do in governance terms to reduce them.
The assignment of someone else to make sure that you’ve got the oppositional position. You should make sure that the most senior person in the room is not the first person to talk. If the most senior person has spoken first then, they don’t want to contradict the boss, right? It can be those kind of micro habits which make a real difference.
A lot of it otherwise goes into design principles. A lot of biases that occur, such as in recruitment and elsewhere, it’s very hard to think your way out of them, for obvious reasons. Diversity-type training doesn’t work. Whereas you can redesign processes to reduce biases. So that’s where good governance meets good behavioural sciences. It is generally around the design of processes to reduce biases.
That’s a potentially very long list, but recruitment would be a simple example. There are ways of recruiting using certain processes which we built a platform ourselves and apply, which is debiased.
We ourselves were doing a lot of recruiting and we know that it is a process full of biases. For example you want to get a good employee, just like if you want to get a date actually, it’s almost the same effect, make sure the person before you is really bad.
You want to get a date, make sure you go on a date with someone who looks quite a lot like you but is a bit uglier. Because it’s all judged in relativity. The same is true in jobs. So the person before you does badly, it greatly increases the positive effect. Or if you muff up the first question it’s very hard to recover from it.
Of course, there are issues about diversity and appearance in other ways too. People think in many recruitment processes that they’re doing a very sophisticated thing but it’s very hard to escape a lot of biases, which is why most processes, particularly in the interviews, are very low predictability.
So, we design a process initially for internal use which does stuff like you would expect. You ask a series of questions, certainly in wide screening of people, you don’t see their name. You don’t see the questions even together. If you’d applied to BIT you might have had to answer six or ten questions. Your question one would go off to three people, your question two would go to three people. It’d all be divided back up. They’re scored separately and then they’re all put back together. You’re much more likely to choose the right candidate, as the process is, by design, taking out biases.
That’s available, anyone can use it. Any company can use the applied platform that’s out there now. That will be one example but it’s true of many things. When you’re making a decision about difficult things, you think about how you’re going to be biased: Good governance should be doing all these things frequently but it doesn’t.
You see certain systematic biases, particularly with group effects. A particularly notorious one is over-confidence. Senior people who make it to the top of organisations, are very prone to being over-confident. They think their idea is really a good idea and it will work, when often actually their estimate is way too high.
We’re human beings. We’re kind of making guesses all the time using our biases and sometimes that can work well and we stop each other, and good governance I’d say definitely can do that. But they’re very powerful and they’re very hard to see in ourselves.
Well you definitely don’t want it all from government. We do a lot of stuff for government but you want some sense of what good governance looks like.
To some extent firms are exposed to the discipline of markets, which means if you really screw up badly enough there will be some kind of sanction. But hopefully that’s not where you want to end up. Firms can learn and evolve these practices. The best business schools now start to talk about it. You go to Harvard Business School, there’s a lot of stuff on behavioural science.
One of the puzzles is using something like behaviour economics, which even though last year’s Nobel Prize for economics was won by a good friend and colleague of mine, Richard Thaler, it isn’t widely understood. One of the ironies is that behavioural economics hasn’t been used much with economics.
“If you’re not even aware of what kinds of errors there are, it’s hard to correct them”
So yes, we’ve been working with the Bank of England. It would seem quite a dry issue, but the main point about our work is that basically it’s about real human beings, like you and like me.
We aren’t economic maximisers. We use mental shortcuts to figure out what’s going on in the world. That’s true for company workers, or the boss, or the customers, and sometimes this is used in very strange ways.
It has big impacts, one of which is, even at the bigger scale, sometimes known as animal spirits or volatility in economies, when they bounce up and down. And so if a firm’s trying to decide on the next move, the boss may ask shall I invest in that more plant or equipment? Or shall I take on those extra workers? The real question is how do they make that judgement?
And some of that is by weighing the data, however, a lot of it is gut feeling, or some other view, or looking at what you think other firms are doing. It’s subject to the same kind of human frailty that everything else is.
Coming back to the stuff about Bank of England, one of the key issues if you are running a central bank is that, particularly nowadays, with somebody like Mark Carney, that you’re kind of telling firms what you think is going to happen. Is what overtly Mark calls forward guidance.
That itself then changes the behaviour of firms. If you think the economy is about to go off a cliff, you think, well I’d better hold onto my cash. But if you think it’s about to surge you’d think, oh well, I’m going to invest. So one of the things that banks do now is they put out a lot of information. which one of the driving and important bits is the, so-called, inflation reports.
“Human frailty is woven into everything; from what firms do to what whole economies do”
So, we’ve been working with the bank to say; what do people understand from inflation reports? You’re putting it out there and you’re hoping that you will interpret in a certain way and make a wise judgement. Maybe you’re not even understanding it yourself. And certainly that will effect what you do about it.
We work with the bank to test alternative ways of presenting and shaping inflation reports, and then we test levels of comprehension amongst people who read them, and would it change in some ways their judgements or the behaviour. And it turns out it makes a big difference.
If you write up an inflation report in a so-called relatable form, people are far more likely to correctly interpret what it’s saying and what it means the economy is likely to do, and the bank’s judgement, it also interestingly increases trust in the bank.
It might seem pretty dry there but it’s incredibly consequential because a lot of what actually makes whole economies move up and down is to do with sentiments and what you think everyone else
Human frailty is woven into everything in terms of from what firms do to what whole economies do. And we found, writing an inflation report in a more relatable form results in at least a 40% increase in comprehension.
It’s quite a recent phenomenon. Richard only got the Nobel Prize last year, but I think it’s very interesting. It’s kind of an interesting paradox that something called behavioural economics or insights being used quite heavily in other areas. In health or really getting back to work faster or so on.
I think one of the arguments is that economists are very smart and they have well-entrenched theories about how economies and people work. And so in some ways that knowledge is a big barrier to change. It’s therefore economists who sometimes find it hardest to make the change.
The wellbeing of your workforce and the quality of your management. To me that’s a game changer. And ideally firms would only do it internally. If we were really serious about it we would publish on it just like you have published accounts. Firms would have a standardised way, publish information about the wellbeing of their workforce.
You can use the Mind’s Index but it needs to be comparable enough. Good firms that are doing it as they know that as well as looking after their work forces, they’ll get the best workers.
When you’re trying to decide on which company to work for. You can see the pay but now I can also see how happy are the people who work there.
This would be pretty useful to know when you’re choosing your next job. Most of them are so-called shrouded variables. You can’t find out about them. So the firms that are doing this well, they should publish their data. That itself, even if other firms don’t, will create a pretty strong market.
Well what would be your conclusion? You don’t necessarily need governance to gather this kind of data. But since so many firms often are doing staff surveys, they should use some standard metrics to start to publish it. At the moment when you go on Glass Door you’ll hear about the really disgruntled employees, the ones who have taken the time to post online, but you don’t hear about the experience of most people.
What I love about this idea is that it fundamentally changes the labour market because these firms start to report on the basis of the quality of their work. It also has an echo factor at board level because you can then say how well do you do compared to others.
Then instead of it being a vague aspiration it becomes real, and it’s meaningful because it affects who comes to work with you. And a firm that does badly, of course, then has to pay a market premium to get anyone to come and work there.
One of the most famous series of studies done on government, the Whitehall Studies shows that every grade you move up people’s mental health gets better. What a surprise! They’re more in control and have more power.
If you’re a firm with very large numbers of low paid employees , then your wellbeing will be tracked at that. It’s not fair to compare that firm to a firm which has a very small number of low paid people. So you have to make comparisons meaningfully and fair, and ideally see it broken down by grade. But that’s something doable.