Nudge, Nudge? Privatizing Public Policy
“Like all major changes to democratic accountability, it happened with a minimum of fuss. By the time we heard about it, it was already over.”
 |
| Photo: Illustration by Bill Butcher |
This week the government announced that the Behavioural Insights Team (BIT), commonly referred to as the ‘nudge unit’, has been ‘spun out’ of Whitehall into a mutual joint venture. The new “social purpose company” is now owned, in roughly equal shares, by BIT employees, the government, and Nesta (an independent charity established by the previous government using £250 million of National Lottery money). The privatisation deal has been described as “one of the biggest experiments in British public sector reform” (Financial Times), on account of this being the first time that privatisation has reached beyond public services and utilities to include an actual government policy team. My intuition, like many other people’s I would imagine, is that this marks a dangerous new precedent in the rise of private power over the public. But what precisely is it that is doing the work for this intuition?
Debates about the privatisation of government functions tends to give rise to instrumental questions concerning the desirable ways of providing these services, most often casting the matter in terms of the trade-off between private entrepreneurialism and public bureaucracy. Proponents of privatisation emphasise efficiency, innovation and flexibility; opponents of privatisation emphasise the loose fidelity of private entities on account of the lack of accountability. The case for privatising the BIT followed suit: Francis Maude, Minister for the Cabinet Office, has stated that “this government is determined to help Britain win the global race by setting free innovative public sector entrepreneurs like these”. This is part of its aim to build a “diverse ecosystem of providers… working alongside the public sector”, in order to drive innovation in government commercial models.* Similarly, the reaction from those who see this as “backdoor privatisation” focuses on the lack of accountability (for instance, the BIT is no longer subject to the Freedom of Information Act), a danger that is made more worrying in light of recent scandals involving services that the government has already outsourced (e.g., problems with private prison providers, and the referral of G4S and Serco to the Serious Fraud Office).
We can understand this debate, therefore, as taking place within a framework provided by a shared assumption of both the proponents and opponents of BIT’s privatisation, namely, that the governmental function in question (i.e., behaviour change policy) can in principle be performed by either public or private entities and that the choice between them must be based on addressing the empirical question of who is likely to perform this function better. However, my (and others’) intuitive dislike of BIT privatisation does not look like it can be explained in terms of instrumental concerns about the efficacy of function-provision. Rather, its underlying rationale is grounded in different normative considerations to those officially used to justify the opposition to privatisation: an oft-unarticulated conviction that the identity of the agent is crucial in providing the good.
It is the nature of ‘the good’ in the case of the private provision of nudge policy that seems to be doing a lot of the work for this normative constraints justification, which states that: while a privatised nudge unit can execute the task (and perhaps do it better than the state-owned nudge unit), there are normative constraints that preclude the performance of nudge policy being made by agents other than the state. “Nudging” is the controversial policymaking mechanism that implements policy on citizens’ automatic, subconscious behavioural responses in order to elicit targeted behaviour change. It works by applying Kahneman-style lessons from cognitive psychology and behavioural economics so as to promote behaviour that is in citizens’ own, as well as society’s general, interests; but nudging “works best in the dark” (i.e., without the ‘nudgee’ knowing it is happening), so has been charged with being a subtle type of psychological manipulation, which might look to be morally problematic as a behaviour change tactic even to those who are sympathetic to the goals that nudge is trying to achieve.
But if we put to one side issues surrounding the legitimacy of a state-owned nudge unit and focus on the public-private dimension of legitimacy, a non-instrumental anti-privatisation argument might maintain that some functions are “inherently governmental” and that nudging (as psychological manipulation aimed at behaviour change) is one such function. So the justification for the role of the state in nudging is based on who it is that is doing the nudging (rather than, say, the state’s superior nudging ability). This represents a normative consideration that has not been aired in the critical discussion of BIT’s privatisation. It is not clear that this line of argument could be fruitful; but it is worth exploring whether or not it could be, because if it does capture a relevant consideration, then it might be that we find ourselves in a strange situation: where the UK government has privatised (and wants to privatise other) inherently governmental functions – that is, functions that cannot in principle be privatized.
——
* In reality, it appears that a key motivation was “staff retention”: privatizing the BIT means that the 16-strong team are no longer bound by civil service pay grades; they are now policy consultants and shareholders in a for-profit company that has a great competitive advantage, and one that has already been linked with the use of success fees, meaning that if its advice leads to significant customer (governmental and/or private-sector) savings, the company gets a share.
Hi Fay, thanks for raising this topic. I’ve been wondering whether your argument that it matters who does the nudging is the end of the story, or can be analyzed further. One suggestion would be that the risk of abuse is taken to be greater if a private party gets involved in this, because they are motivated in the wrong way. For example, they might do their job to earn money (as your footnote indicates, this seems to have been an issue here), rather than for serving the public good.
An interesting parallel could be the privatization of prisons, on which one might have similar feelings.
Would you agree with this point, or do you think that the „inherent governmentalness“ is somehow a basic category (or that there could be other things behind it)?
Fay, thanks for the post. Really interesting and I think you are right that this dimension of the debate needs more attention. I would like to hear more, though, on exactly what you think is the inherent issue. Let us imagine that we really do set aside the two other debates you raise – whether it is permissible to nudge at all and what is most effective in doing so. Let us say that we are permitted to shape people’s choice for the own benefit and that a privatised organisation of some kind will do the most effective job of this goal. What, then, is crucially troubling? Is it, e.g., that this group has not been licensed by a democratic vote? There was an interesting recent paper by Dorfman & Harel in Philosophy & Public Affairs – ‘The case against privatisation’ (http://onlinelibrary.wiley.com/doi/10.1111/papa.12007/abstract) – which makes the case that some things (e.g., punishment) must ‘bear the mark of the state’ because it is an important feature of these things that they are a political act. Might that consideration be important? If so, does it definitely apply in the case of nudging?
Thanks Lisa. Yes, this was part of the thought that I was trying to capture; though I think that there is something else more fundamental too: the idea, raised by Andrew's comment below, that some things (e.g. punishment – which brings in your prison case parallel) "must ‘bear the mark of the state’ because it is an important feature of these things that they are a political act". [I will elaborate in my response to Andrew.]
This Dorfman/Harel-type consideration that you raise is definitely important; and the question of whether this applies to nudging I think gets to the heart of the issue. What I didn't have space to explain in my post was that my view that this category of "inherently governmental" might apply to nudging is premised on my understanding of there being a number of different 'nudge mechanisms', which differ in their moral permissibility. A recent Hansen and Jespersen paper (http://www.lexxion.de/en/verlagsprogramm-shop/details/3099/221) has distinguished four such mechanisms by distinguishing between System 1 and System 2 nudges (that is, those that seek to influence automatic behaviour and those that seek to influence our choice-making process) and between epistemic transparency and epistemic non-transparency re: nudges. This gives us: (i) transparent System 1 nudges (i.e., transparent influence [technical manipulation] of behaviour); (ii) non-transparent System 1 nudges (non-transparent manipulation of behaviour); (iii) transparent System 2 nudges (transparent facilitation/ prompting of consistent choice); and (iv) non-transparent System 2 nudges (non-transparent manipulation of choice). Hence, it looks like we might be inclined to hold that (ii) and (iv) are more morally objectionable (more intrusive) than the most intrusive traditional regulatory behaviour change interventions, which involve the elimination or restriction of choice (which have been democratically legislated). [It also shows that some nudges, category (iii), look to be morally unproblematic.] Of course, this analysis bears most on the question of whether (some types of) nudges are permissible at all (which you bracketed); but I think that it can speak to the 'inherent issue' too. When it comes to the psychological manipulation of citizens, it looks like we require that those involved can engage in certain forms of deliberation; and, according to Harel, only the deliberations of public officials can be characterised by what he calls "fidelity of deference" to the state. (This brings us round to Lisa's point about the motivations of for-profit nudge policy consultants.)
Thanks for the post, Fay. As with others, I think that the issue you raise is one that is best solved by thinking about public/private ownership more generally, rather than necessarily anything specific to the BIT.
Regardin the public/private ownership debate, I'm not so sure that I see the normative issue or, to be more precise, I don't think I was persuased by Harel's conclusion that certain activities 'must bear the mark of the state'. I acknowledge that there may be good instrumental reasons for preferring the state to perform certain activities, but I don't see the force of the other reasons that Harel mentions. More argument is need, I think, to explain why the BIT (or prisons or, for that matter, private schools) doesn't still bear the mark of the state in the sense that the state retains the right to re-nationalise if and when it see fits.
Thanks Tom. Yes, I think your comment is fair. I should probably state that I do not necessarily hold Harel's views… The post was merely an attempt to think through my intuition that there might be more than instrumental reasons to worry about BIT's part-privatisation (hence the tentative and underdeveloped offering in the final paragraph: "a non-instrumental anti-privatisation argument *might maintain* that some functions are "inherently governmental" and that nudging is one such function"). I wonder if there is any other way that I could cash out this intuition?
There is one distinction that might be relevant here, and I wonder how it relates to nudging. Some state activities are such that what matters are final results. Other state activities are such that what matters is also the way in which things are done. The latter can be the case for different reasons, e.g. because of the meaning of these things (which is Harel's claim, if I understand it correctly), or because of risks for other values.
One might hold that in nudging, the outcome (to get people to behave in certain ways) is really only one side of the coin. The other side of the coin is *how* this is done, and whether or not it is compatible with the ways in which we understand citizenship, respect, autonomy, etc.
Not sure how much this helps for cashing out Fay's intuition – what do you think?
Fay, is part of your objection perhaps that it the nudge unit is going to become a private company (presumably with profits and so forth). Would you be less concerned if it were to become an independent charity? My guess would be that this might be a marginal improvement (since we might be less worried about some of its incentives) but doesn't really deal with the lack of democratic accountability.
Thank you very much for your post, Fay. I realize the general permissibility of nudging is not at the center of your argument, but since you brought it up, I would be curious to hear a bit more about your view. I have not read the Hansen and Jespersen paper and so do not know how this would relate to the distinctions you mention, but one of the points often made by defenders of the permissibility of nudging is that it is misleading to assume that there is a "neutral" policy option against the backdrop of which particulars nudges have to be justified. (Taking the proposal of encouraging the consumption of healthy foods by placing them on the center shelf in a supermarket as an example, the point would be that the center shelf will always be occupied by *some* product that customers are then comparatively more likely to buy.) On the assumption that most psychological mechanisms that may be relied on for nudging purposes are public knowledge and will certainly be exploited by marketing experts in the private sector, why should we assume that the reliance on these mechanisms for public purposes faces a particular burden of justification?
So I think privatisation is an unhelpful term here – mainly because there are different aspects of privatisation which likely have different implications. Publicly owned companies with tradable equity are very different from a firm whose workers have a minority ownership stake (which I assume is untradeable).
I have deep problems with taking control and decision structures away from government (the problem with majority private ownership). Minority ownership however is almost purely about remuneration (for workers) and financing (equity vs debt funding (the latter of which we are fine with for government projects)).
My central issue with this policy is that, by offering large rewards and by using a bonus system which I assume will be the primary form of compensation, it both encourages employees to focus almost exclusively on their salary (as every action they take likely impacts it) and it undermines the idea that public sector workers should perhaps perform their jobs for non-monetary reasons. I suspect that this is an overly romanticised view of public sector work; however I do still feel that policies like this erode a public service ethos in society.
One question, would you find it acceptable to remove this reform and instead pay all members of the nudge unit £200,000 per year, while keeping the unit fully under government ownership? If you do object are they on any of the grounds related to your original objection?
Hi Lisa, I think that this distinction is really highly important. Sadly it is often overlooked. Two lines of thought may help here: a) the classic distinction between two kinds of paternalism that Mill already has put forth, i.e. a distinction between hard and soft paternalism; and b) some further considerations about how we want to respect freedom in liberal societies . On the one hand there is a procedural kind of paternalism (which includes waiting periods, duties to get informed etc.) that does not steer people towards a particular goal (such as loosing weight, being healthier etc.). This seems to be in a liberal spirit since it leaves people to decide about their goals, maybe in a more informed way. On the other hand, we can think of softer, but rather manipulative strategies to steer people in the right direction, such as making them choose healthier food, better health insurance etc. In the latter way one imposes a goal that one finds worthwhile on people,albeit there is a (theoretical) opportunity to opt out or choose differently. Still, I find the latter more questionable than the former. I think that there is a difference, which should be spelled out in more detail. Still, I am also not sure how this helps supporting the argument in Fay Niker's article. But anyway: does this in any way reflect what you meant?
Hi Rebecca, yes, I was thinking roughly along these lines, although I hadn't figured out the distinctions in as much details as you do. What you call "procedural" paternalism definitely looks more liberal to me than the other forms; it basically tries to help people to overcome rationality deficits such as impulsive action. The second form is more problematic, because it anticipates individuals' decisions. Relating this to Fay's distinction: maybe privatizing a state agency that is strictly limited to the former kind is less problematic – but then the question is who has the authority to decide which nudges are purely procedural and which ones are more substantive; it seems that this should still be decided by the state.