A new way of thinking about individual choice has taken the political landscape by storm. America’s new president Barack Obama and the leader of the British Conservatives, David Cameron, (just to drop a couple of names) have shown an interest in it. Its intellectual and academic pedigree is impeccable. It is said to be effective, evidence-based, and cheap to implement. Most of all, it stakes a claim to a degree of philosophical coherence of which the various “third ways” of the last decade could only dream.
The novel idea, elucidated in Cass Sunstein’s and Richard Thaler’s book Nudge, is that skilfully controlling the way alternatives are presented to us can “nudge” us toward making the choices that our own “better selves” would make. “Libertarian paternalists” like Sunstein and Thaler claim that we have two different ways of making decisions: one “from the gut” (called System I), and the other more deliberative and far more effective (called System II).
But, while System II choices may be more effective than System-I decisions, they are more “expensive” to make: one needs data, analysis, and concentration. Only when the importance of the task warrants the effort do we shift gear and deploy the System-II heavy artillery. This division of labour between the System-I and System-II mechanisms would work well were it not for the fact that our lazy and cheap decision-making mode tends to take over in situations that should command our fullest attention: choosing a pension plan or a healthcare scheme, for example. As one can imagine, the results of this System-I coup d’etat are not pretty.
Old-fashioned paternalists have always been very well aware of this. In these situations, liberal paternalists feel no qualms about taking control and forcing choices upon us (“Wear seat belts and enrol in the pension plan, and in the end you’ll thank me”). The criticisms of this attitude boil down to a simple question: “Who is better equipped than I am to make choices about my welfare?”
Libertarian paternalists are different. To make us choose what is good for us, they avoid fines, compulsion, and prohibition in favour of “nudges” — institutional arrangements that we could, in principle, easily override, but that, given our tendency to rely on System-I, we end up going along with. With a neat twist of logic, our cognitive imperfections are turned on their heads, and made to work to achieve System-II choices. And, indeed, well-chosen nudges have been shown to be extremely effective in altering choices that make a substantive difference to the lives of many (say, enrolment in pension plans).
But here lies the rub: if the architecture of choice is truly so important in determining outcomes, does it really matter if manipulation rather than coercion is used to make us choose what someone else has decided is ultimately good for us? What is gained by tricking rather than forcing people into making choices they wouldn’t have made? Isn’t the power of the manipulator more insidious, and perhaps to be more greatly feared, than that of the policeman? In short, what is so libertarian about “nudging”?
Libertarian paternalists offer a novel answer to these questions. As long as choice engineering tricks us into making choices that our own more deliberate self would make, they say, manipulation is justifiable. Nudges should be so chosen as to prod us toward those choices that we ourselves would make, if we would only sit down and think carefully about the matter at hand.
There are, unfortunately, some logical problems with this apparently elegant solution. Is it really always so reasonable and effective to listen to System-II preferences? According to neo-classical economists, the answer is a resounding “yes”, and, in many cases, they can point to substantial evidence to validate their claim.
Unfortunately, however, hyper-rationality does not always produce attractive outcomes when social, rather than individual, choices are concerned. Indeed, sometimes it leads to results that are inefficient, “nasty”, or both. The benefits of tax evasion may far outweigh the risk of getting caught, but imagine if everyone decided to evade taxes at the same time. In these situations, the System-II decision maker has no tools to resort to in order to escape the sub-optimality (and social nastiness) of the rational decision.
It gets worse. Consider the following example. I am an organ donor. Thaler and Sunstein also believe that donating organs is a good thing. Therefore, one of the “nudges” they recommend is to make organ donation as the default option in the case of a fatal accident. But there is nothing irrational in having a very strong preference not to have one’s body used for spare parts after one’s death. We cannot use System-II to fault this preference. How would we feel if we used the System-I decisional laziness about changing defaults to “trick” someone into a course of action to which she would have violently objected, perhaps on religious grounds?
Ultimately, there is a fundamental and unresolved accountability deficit in the way libertarian paternalists go about nudging us. The Enlightenment idea that our System-II rationality can always point to an optimal choice about which every reasonable individual would agree clashes with the widely accepted modern idea that there is a plurality of reasonable choices.
If this is the case, who is to decide in what areas individuals can be fairly nudged by social engineers? Whose and which choices are the nudgers suppose to encourage? And who will nudge the nudgers?
These questions, I am afraid, still await a convincing answer. —DTPS/IHS
Riccardo Rebonato is Visiting Lecturer in Mathematical Finance at Oxford University, Adjunct Professor at Imperial College and Tanaka Business School, and Global Head of Market Risk and Quantitative Analytics at the Royal Bank of Scotland