By Tony Hockley, London School of Economics & Political Science (LSE)
In his article for the journal the former head of the UK civil service, Gus O’Donnell bemoans “the challenge of making nudge attractive to politicians”. In the same piece he recommends that the leading article by Sanders et al should be be read carefully by readers of his own response. A careful reading shows very clearly that enthusiasm for “nudge” came not from civil servants, but from the politicians themselves. So much for having to make nudge attractive to politicians.
Similarly, Richard Thaler’s foreword to David Halpern’s 2015 book “Inside the Nudge Unit” recalls that it was David Cameron’s political advisors who approached him, prior to Cameron becoming Prime Minister of the UK in 2010, to discuss the incorporation of nudge into the policy process. In the USA, it was the politician Barack Obama, who chose to bring his former Chicago Law School colleague (and Thaler’s co-author of “Nudge”) Cass Sunstein into the White House. Around the World it has been the politicians who have championed the adoption of behavioural public policy. It is strange, therefore, that O’Donnell should suggest a lack of interest.
After all, the attractiveness of nudges to politicians is easy to understand. Governments are mostly judged by simple, measurable changes in outcomes and in their capacity to contain administrative costs. The publication of “Nudge” showed that highly desirable (and persistently elusive) changes in outcomes can be achieved, and at minimal public cost. Furthermore, the toolkit of libertarian paternalism meant that this could be done with minimal risk of intrusion beyond the fuzzy boundaries of acceptable state paternalism.
Making nudge attractive to politicians seems, therefore, not to the primary challenge faced in the development and delivery of behavioural public policy. The real challenge might actually lie in containing politicians’ enthusiasm for this new toolkit. Nudge faces a serious mis-selling problem that could easily swing the pendulum back to more traditional approaches, as outcomes fail to match up to promises. Firstly, many of the policies inspired by “Nudge” go far beyond the parameters set by Thaler and Sunstein, and secondly, these policies will often struggle to live up to the claims made by the politicians who sell them to the electorate.
It is reassuring to read in the article by Sanders et al that behavioural insights teams in government are beginning to reflect on governments’ own vulnerability to biases and pathologies, after a decade of rather excited pursuit of “low hanging fruit”. It is debateable, for example, whether the UK Behavioural Insights Team’s early win on timely payment of income tax serves any purpose other than the pursuit of arbitrary government target dates for the payment of some major taxes, beyond which many individuals and businesses seem content to take the option of paying the exchequer interest and penalties for late payment. The dramatic spike in public revenues in January and July each year is created by these targets, fortunately balanced by the receipt of other taxes with a smoother calendar profile. The letters designed by the Behavioural Insights Team appear to tip-toe the line on the honesty of the claims being made within them, and of their net effect on the public finances. The more this is done, the more the potential risk of reactance or immunity to future behavioural public policies when scaled up.
Additionally, the competitive nature of party politics lends itself to hyberbole on the potential impact of policies, and nudges are unlikely to be immune from this behaviour. In an earlier blog, Richard Thaler ascribed some of this general tendency to overconfidence bias and concept stretching amongst practitioners and advocates to the decision by him and his co-author Cass Sunstein to adopt the suggested book title “Nudge”, over “The Gentle Power of Choice Architecture” or similarly expectation-dampening alternatives. Thaler observed that: “With one of our original titles the book would have been lucky to sell a hundred copies. But a one-word catchy title for a book is not meant to be a complete statement about the span of a policymaker’s toolkit”.
The “silver bullet” tendency has been particularly evident in the UK of late in the shift of the organ donation default from an opt-in to a system of presumed consent, first in Wales, and later in England. In England ministers presented a government “ambition” of 700 extra lives saved as its headline estimate of how many lives could be saved by the policy. The pursuit of a good headline was prioritised over accuracy. In 2011 the First Minister of Wales announced the seemingly assured success of presumed consent: “A soft opt-out system will not change the medical treatment patients are given, up to and including the time of death, but it will increase the number of organs and tissues available.”
Within six months of implementing the policy ministers in Wales were already declaring its success. Health minister Vaughan Gething declared: “I fully expect that the new system will create a step change in consent for organ donation in Wales. The early indications are that this certainly is the case”. At the end of the first year the Welsh Government had to acknowledge in its impact evaluationthat: “The routine data analysis shows no consistent positive trend in terms of organ donation from either Welsh hospitals or from Welsh residents”, now declaring that: “a longer period for comparison will be needed to draw firmer conclusions about the impact of the law on donation rates”.
Thankfully, in democratic states there is a firm electoral check on politicians’ hyperbole. Journalists and political foes will be quick to highlight failings of outcomes against the claims made. Successive UK prime ministers over the past 25 years have felt the impact of this. John Major will always be remembered for the clash between the rhetoric of “back to basics”and the “cash for questions” scandal; Tony Blair for the declaration of a new “ethical foreign policy” and his headlong pursuit of the “Iraq War”; Gordon Brown for announcing “no more boom and bust” followed by the 2008 crash; and David Cameron for promising the “greenest government ever” only to shift to a “gas first” policy for energy security from a priority on climate change. Politics is, eventually, tough on hubris in the end, even though it is temporarily beneficial and largely unconstrained.
O’Donnell appears to lament the constraints of democracy upon the application of behavioural science to public policy, so that civil servants were not: “incentivised, or even encouraged, to work out what outcome the communication was trying to achieve and then come up with a way or ways that could be tested in order to get the best outcome. Instead the incentives were to draft an accurate letter that could not be criticised by Select Committees or lobby groups”.
His argument is that this discipline upon behavioural public policy has acted as a barrier to scaling up. Yet the policies that have scaled up have been those which have, over time, won broad support. He cites, for example, the ban on smoking in public places, as an example of a bold step forward in an area within which people did not appreciate in advance that they would turn out to have an implicit preference for smoke-free pubs and bars. This fails to acknowledge the forty years of incremental tobacco control policies that preceded this step and, most importantly, that smoking had declined to such an extent that only a minority of the population were smokers by the time of the 2007 ban on smoking in public places. If the story of tobacco control tells us anything, it is that behavioural approaches work well within the comprehensive use of the full range of public policy tools, and that timing really does matter.
We can now witness the same, incremental process taking place in relation to single-use plastics. Over a decade the UK has shifted from a voluntary 5p charge on single use shopping bags, to a mandatory charge, and a widening of the policy focus onto other everyday plastic items as majority acceptance of the problem builds and solution aversion diminishes. The consistency of incremental action in both of these policy areas conflicts with a general assumption that politicians will not pursue small policies against long-term goals, given that electoral judgement comes around every few years. Exceptional present bias amongst politicians may be more imagined than real. Indeed, in order to avoid accusations of doing nothing on any particular policy problem amongst the panoply of problems faced, small steps do, at least, give the impression of action. Behavioural science offers a means to make these small steps that contribute towards progress, rather than amount to little more than shuffling deckchairs.
O’Donnell’s article demonstrates well the benefits of a public policy process that involves both technocrats and politicians. It is too easy to underestimate the practical value of politicians who are daily accountable to their electors, alert to the risks of spillover effects or unintended consequences and the risk of reactance to poorly-designed policies. By the same token the natural resistance of civil servants to hyperbole is of considerable value to political reputations and to the credibility of behavioural public policy.