Nudge theory. Neuroeconomics. Reputational game theory. Psychophysics. Behavioural economics. I love them all.
Why? Well, self-interest plays a part.
I work in advertising. And after a few years working in advertising, many people become just as frustrated with the neoclassical model of human behaviour as you probably are. And for a surprisingly similar reason.
Advertising, as the behavioural economist Daniel Kahneman himself has said, is a “System One” business– built on intuition and quick impressions. Like you, I suspect, people in advertising and marketing think that intangible, seemingly irrational things– instincts, feelings, perceptions– actually matter. And, also like you, we believe that it makes no sense to base all decisions on models of the world that are altogether blind to those very emotional or psychological considerations that drive the greater part of human behaviour.
In short, we are just as irritated as you are by the strange and unhealthy monopoly that neoclassical economic theory seems to enjoy among the social sciences when influencing business decisions and social policy.
Ironically, the reason for the near-unbridled power of economic thinking may itself arise from what you might call a cognitive bias. First of all, in decision-making, there is a known bias whereby any argument that contains mathematics is assumed to be more scientific, rational and credible than any argument that uses merely words (see Carl Bialik, “Don’t Let Math Pull the Wool over Your Eyes,” in the Wall Street Journal, for details). Second, executives or policy-makers making decisions may be far more eager to have seemingly rational bases for their decisions than to actually make effective decisions, since the semblance of rationality behind their recommendations protects them from being punished or fired.
Relying on mathematical models, however poor they may be, to lend a semblance of logic to a decision may thus be the product of a kind of loss aversion. Any decision influenced by nonmathematical, subjective or emotional factors comes with no such insurance.
Yet important emotions and desires — regret, uncertainty, trust, affection, identity, purpose and meaning — all play a huge part in our lives, for no less a reason than that it is to our evolutionary advantage as a social species to feel them. Some of these emotions and desires may be evolutionary hangovers we no longer need; many, however, continue to be essential in the functioning of real-world markets or societies. A species whose members casually returned to buy from shops that had short-changed them in the past, and who did not bother to share their sense of outrage with friends and neighbours, would not produce a healthy retail economy.
And so it doesn’t seem a good idea for government to proceed as though such emotions and desires don’t matter simply because they are unquantifiable and hence don’t fit neatly into a preconceived mathematical model. Because the price we pay for this methodological neatness is that we repeatedly introduce into the world government programs (and consumer products) designed for a kind of human being that doesn’t exist.
E.O. Wilson, the evolutionary theorist and the world’s leading expert on ants, once reacted to Karl Marx’s ideas with the comment “Beautiful theory. Wrong species.” If so, perhaps Marx can be forgiven for proposing a theory for an inappropriate species, since modern economists, it seems, have devised a model of economic activity for a species that does not exist at all.
In the imaginary world inhabited by Homo economicus, trust is universal, knowledge is perfect and the value of everything is already established in actors’ minds. There is no need for marketing, and no need for relationships, reputation, commitment devices or trusted intermediaries. The single goal of this system is efficiency.
We know the real world doesn’t work that way.
So enormous potential gains are there for the taking once the policy process can consider designing policies based on how people really are, and not on some strange, autistic assumption about human behaviour that is rigged to fit a preexisting economic model.
Doing this is easier than it first seems. It’s not necessary for policy to be perfectly right (in a complex world, perfectly correct answers rarely exist). All you have to aim to do is simply be less wrong than you were before.
The Save More Tomorrow pension, conceived by Richard Thaler and Shlomo Benartzi, which allows employees to allocate a portion of future salary increases to their retirement savings, is not a perfect pension. I am sure there are better designs for pension schemes yet to be discovered. But this idea for a pension is better by a factor of two or three at getting people to save for retirement than any other pension yet devised, despite the hours of work previously invested by thousands of well-paid people working for government departments and pension companies.
How did it achieve this spectacular uplift? Quite simply, it was brave enough to challenge one of the fundamental assumptions about human “rationality,” which is a product not of empirical observation but of a contrived economic model.
My contention is simple. When we challenge the assumptions I will list here, the results may not always be spectacular. But they might be. Sometimes we may fail. But, in any case, the cost of experimentation is relatively low, and the potential gains near limitless. And because these assumptions are so pervasive, you will find an amazingly large number of them have gone unchallenged for years.
Sometimes we might fail. But the cost of experimentation is low and the potential gains nearly limitless.
So my approach to deploying the new insights from behavioural science, game theory, neuroscience, evolutionary psychology and so forth to the world of business or government is very simple: Look for areas where these new findings conflict with conventional economic assumptions. Then test what happens if you assume that the new science is right and the old science is wrong.
Neoclassical economics assumes that people have consistent time preferences. All available evidence shows they don’t. All right, let’s assume that conventional theory is wrong and empirical evidence is right– what does a pension look like now? Well, it looks a lot more like the Save More Tomorrow pension than like a conventional pension, that’s for sure.
This isn’t rocket science. It doesn’t need to be.
So what are some more of these common assumptions? The first, a legacy of the neo-classical model, is the assumption that human actors are rampant individualists, making decisions based on their individual utility. We don’t. We are a herd species who have intelligently evolved the instinctive heuristic “When in doubt, I’ll feel safer doing what everyone else does.”
This is one of the reasons why I suspect the British government’s new opt-out occupational pension will work so well. Once most of your workmates have a pension, you feel less lonely and paranoid about having the selfsame pension.
A second assumption is that attitudinal change is a necessary and sufficient precursor to behavioural change. Actually it isn’t. In fact the process often works the other way round.
We recently deployed this insight in a major (currently confidential) water-saving program for the developing world. “Don’t even bother for a second to mention the water savings, except in passing,” we said. “Instead, simply make the environmentally friendly behaviour easier to adopt than the old behaviour. Let them think about the water savings for themselves.”
Often the language of environmental sustainability arouses an automatic feeling of “making compromises” and tends to frame the new behaviour as a task, not a pleasure. One of the great lessons of behavioural science is learning what not to say.
Another assumption is the idea that human beings act according to some fabulous sense of proportion– that in order to make them change their behaviour significantly, you need correspondingly massive interventions. In the policy world, that means grand schemes and usually millions if not billions of dollars.
Not necessary. If high spending equalled success, then we would not have health systems that strain to meet the demands upon them. Instead, we can move the behavioural needle by taking human nature into account. That’s already proven in the well-known example of how to encourage more people to agree to be organ donors. A simple behavioural cue, costing almost nothing, can move behaviour more than millions spent on advertising.
The slogan of OgilvyChange, the behavioural change arm of Ogilvy & Mather in the United Kingdom, is “Dare to be trivial.” Self-aggrandizement often leads policy-makers to seek solutions in proportion to the size of their available budgets. This may be a terrible mistake.
An important assumption is that people know why they act as they do, can predict their choices and have introspective access to all areas of their brains.
They don’t. Many of our instincts affect our behaviour in ways we don’t notice, don’t understand and can’t verbalize. Hence an excessive reliance on market research to design interventions can be dangerous. Remember that “Most people recycle their towels” message in hotel rooms, now famous as an early nudge experiment? That was the message people claimed would be the least motivating; in reality, it was the one that worked best.
I routinely choose a medium coffee because I want one ”kind of in the middle.”
Similarly, if you take the example of the “$300-million button”– the case of how changing a single clickable link on an online shopping site led to dramatic sales increases– you will see a psychological effect that is hard for anyone to explain. In this case, the practice of demanding that people register before they buy online violated some deep-seated human instinct around exchange, but one we don’t even have a name for yet. You can’t expect consumers or citizens to tell you of these instincts if social scientists haven’t even got a name for it yet.
Another assumption? That people choose using absolute rather than relative measures. I routinely choose a medium coffee without knowing how big “medium” is. I just want one that’s “kind of in the middle.” Starbucks sells a “short” version of its coffees, but it craftily does not display this option on its menus for fear (or knowledge) that the visible availability of this option will drag consumers towards smaller sizes. Perhaps New York Mayor Michael Bloomberg did not need to attempt to ban the sale of outsize sodas– he could have simply demanded that they not appear on menus and are available only on request. Sold under the counter, like dodgy porn.
Then there’s the assumption that people act purely in response to incentives or disincentives, rather than in response to cues and signals they find in their environment. In fact, people’s behaviour is far more driven by context than we can imagine. Lufthansa recently asked why people in airport lounges and in the air drink far more tomato juice than they do on the ground. We don’t know. But something in the environment is driving this.
Environmental design can be an inexpensive– and attractively libertarian– alternative to legislation or law enforcement. As a recent experiment involving OgilvyChange has shown, making a small aesthetic change to shutters in high-crime areas of a British city centre may have helped reduce crime in the surrounding area by almost 20 percent.
The area in southeast London had been badly hit by looters during the 2011 riots. In an attempt to discourage antisocial behaviour, a group of artists were enlisted to paint babies’ faces on the shutters of shops. Called Babies of the Borough, the experiment drew on research aimed at promoting more caring behaviour, with the facial images drawn from photographs supplied by local families. The numbers showing a subsequent drop in crime were supplied by the local police and may not be statistically significant (although as more evidence comes in, it seems the effect is enduring).
But let’s just imagine that this experiment had failed.
What’s the worst that could happen? You spent very little money. It is highly unlikely that there are adverse unintended consequences, but if there are, they will soon be evident. And even if the area is no safer than before, it at least has become more pleasant for passers by to look at. Graffiti, which had been a persistent problem when the shutters were plain grey, has not been scrawled on these pictures. The shopkeepers say they are more proud of their shops. Unlike crime statistics, these effects are unquantifiable. There is no government metric for retailer pride.
Now contrast that with traditional, more complex interventions: a heavy police presence, stiffer sentences. These are costlier, may have adverse consequences and may serve to make the area more dangerous once the police presence disappears. The shutters don’t demand overtime.
I am not disputing the value of larger interventions. But a general principle of all such actions should be to insist on trying simpler, cheaper, less authoritarian, more enduring interventions first. If these fail, then try something else.
But even though this approach seems to reflect common sense, institutions tend by nature to prize the big over the small, the rational over the emotional, the physical over the psychological and the heavy hand over the light touch.
Let me end by sharing a lesson from the direct-response advertising industry of the 1930s. When writing off-the-page advertisements, selling corn feed, piano lessons or whatever, young copywriters were taught to write and design the coupon first. Start with the behaviour you want, and work outwards. Only when you were happy with the coupon would you start to write the copy, and only at the very end would you write the headline.
This process seems to have much to commend it. Start by looking at the behaviour itself and perfect the design of the choice you want someone to make. Only then move away from the point of decision.
Too much policy is written by people who want to write a single headline, and then ignore the detail. But these old copywriters understood a vital lesson: if the coupon was no good, all your other effort, however commendable, was wasted.
How many noble and well-intentioned government programs fail because no one looked at the coupon?