DISCLAIMER: The information and views set out in this article are those of the author(s); and do not necessarily reflect the views of the Centre for Policy Studies or the Indian Institute of Technology Bombay.
Note: This article is the second part of a two part series. Click here for Part 1.
Over the past few weeks, you may have stumbled upon the words ‘nudge’ and ‘behavioural economics’ more often than you had in the past few years. There is a serious attempt to mainstream nudge in policy design. The Economic Survey 2019 in the abstract of its second chapter, titled ‘Policy for Homo Sapiens, Not Homo Economicus: Leveraging the Behavioural Economics of “Nudge”’, notes: ‘Drawing on the psychology of human behaviour, behavioural economics provides insights to “nudge” people towards desirable behaviour’. Following the aspirations to leverage behavioural insights for ‘path-breaking change’ as formulated in the Economic Survey 2019 as well as the Finance Minister’s references to the potential of using behavioural insights in policy design, NITI Aayog (the substitute for the defunct Planning Commission as a think tank, not a central planner), has promptly advertised the setting up of a ‘Nudge Unit’. In what follows, I argue why in an attempt to nudge citizens to change their behaviour, policy makers are being nudged into falling prey to the very problems that they aim to solve. Furthermore, drawing from behavioural economics as discussed in Part-I of this series, I argue that the recent emphasis on behavioural economics and nudge theory is a ‘decoy’ and is merely an attempt to climb onto a popular bandwagon, and has no real clarity in policy design.[1]
It is important to understand that behavioural economics and nudge, while related, are not the same and should not be used interchangeably.[2] Nudge, as popularised in the book by Thaler and Sunstein, is considered to be ‘soft paternalism’ or ‘libertarian paternalism’ – a philosophy where the choice architecture or context of decision making is slightly altered in a manner that makes the decision maker better off in a nearly costless manner, and without limiting freedom or autonomy of the individual to make the choice. The ‘paternalism’ here is associated with the justification on the ground that the one making the choice will be better off, ‘as judged by themselves’[3]. The ‘libertarian’ relates to people having the freedom of choice and the freedom of opting out.[4] Nudge preserves the freedom of people and does not force them or make it harder for them to make a choice. Simply put, nudge attempts to nudge people in the right direction without bans or intrusions.
Nudge works by exploiting cognitive biases and heuristics that people tend to use in judgment and decision making – the very limitations of human decision making that behavioural economics has brought to light. From a libertarian perspective, a nudge does not limit nominal freedoms but limits effective freedoms. Although one might wonder about the inevitability of nudge in policy making, and believe that a nudge by a government is preferred to that by a non-government actor[5]; accepting the principle of nudge for policy raises certain normative issues and ethical concerns that are critical, mostly from the point of view of interference with liberty and autonomy.[6] At the same time, considering an ‘anti-nudge position’ as a ‘non-starter’[7] is problematic.
Now, to draw attention to the government’s intention to leverage behavioural insights for policy design, it is a step in the right direction to acknowledge the soft paternalism central to nudge and identify the relevance of incentives and disincentives to regulate behaviour. According to the Economic Survey 2019 (p.30) ‘nudge’ policies ‘lie between laissez faire and incentives’ in the influence spectrum of public policy.[8] However, it is disappointing to see an illustration of an improper understanding of the concepts and non-acknowledgment of any concerns about the philosophy of nudge in policy design. Instead, it goes on to commit the error of ‘optimism bias’ along with other errors, by highlighting the positives of nudge applied in different contexts and focusing on apparently successful applications in India. For instance, the logic followed in assessing the successful use of behavioural insights in the Swaccha Bharat Mission and Beti Bachao, Beti Padhao (BBBP) schemes are good examples of ‘self-serving bias’, ‘attribution bias’, and ‘confirmation bias’.[9] Moreover, it is surprising to see along with a shallow treatment of ‘norms’, an obsession with ‘mythological role models’[10]; and religious references.[11] One is likely to experience cognitive dissonance while reading through the strategy of how the avoidance of Mahatma Gandhi’s ‘Seven Social Sins’[12] can shape the destiny of the nation.
A careful reading of the arguments put forward in support of how behavioural insights played out in the two schemes would make it clear that there is not only a disregard for the socio-economic context of persistence of the problems but also a pseudo-scientific attribution of causality. The Economic Survey seems confused in its treatment of information and rational persuasion as nudge. The fallacy of this line of reasoning is that every message pertaining to government programmes or schemes is amenable to ‘framing effects’; but without any thought to the design of how variations in framing may influence outcomes of interest or a theory of change, attributing causality to behavioural insights is not prudent. For nudge to be an effective tool for evidence based policy, evaluation frameworks should be built into the implementation of nudge. It is in this context that aspects of data privacy, informed consent, and legality need serious thinking. If experimental evidence is expected to be scaled up for policy design, concerns about external validity need to be anticipated and articulated.
Going beyond the Economic Survey’s articulation of the vision to mainstream nudge in policy design, I argue that the challenges to mainstreaming nudge as a tool for effective public policy are manifold. I will focus on four specific problems.
First, it is necessary to clarify how it has been established what is ‘beneficial’ or ‘good’ for the people for whose welfare the nudge policy is designed, and what is the ‘correct’ direction in which people are being nudged. Merely arguing that an anti-nudge position is a non-starter is in fact a non-starter. If any degree of paternalism is justified on the ground that people cannot judge by themselves what is beneficial for them, it is illogical because behavioural insights are justified on the ground that people make poor decisions. Therefore, people’s judgement of what is good for themselves may not be rational nor consistent. This is important because nudge policies fit into the political ideology of a government that sets down its agenda to manipulate the choice architecture of people. Whilst a socialist planner may be abhorred by libertarians for limiting individual freedoms to arrive at socially optimal allocations, a ‘nudge’ oriented think tank that replaces the socialist planner also limits individual freedom, albeit to a lower degree. Nor is there any evidence that the choice architect in one form or the other knows better about the welfare of the citizens. Nudging people to make the correct food choices in a cafeteria and nudging people to save more are fundamentally different and not comparable in terms of the associated ethics.
Second, there are questions about the legitimacy of the practice of implementing nudge oriented policies that need to be addressed. If a government coerces people to the highest extreme of the influence spectrum by mandating policies such as demonetisation and goods and services tax (GST) implementation, trying to nudge people into changing behaviour two and a half years later reveals the time inconsistent preferences of the choice architect. On a related note, coexistence of mandatory as well as nudge policies are likely to confuse citizens, as would attempts at nudge by benevolent autocrats. In setting the agenda for nudge policy making, it is important to have ex ante evaluation criteria rather than ex post rationalisation arguments in favour of nudge. It is also critical for the government to spell out the concerns about agency and autonomy in a specific nudge policy. There is a need for greater deliberation on the justification of nudge by a government that clearly had an option to nudge people into changing behaviour as desired but chose to implement a coercive law instead. Since nudge works when people are irrational, nudge policies have the unfortunate disadvantage that they can easily be used as a ‘decoy’ to cover up the negative effects of prior policies and distract citizens from policy paralysis.
Third, overplaying the administrative costs or choice related costs involved in nudges without delving into the true costs of the decisions made under a behaviourally motivated choice architecture is troubling. In a bid to steer people into doing what the government thinks is right for them, the focus shifts on the importance of the choice rather than the outcomes of the choice. For instance, more workers with poor financial illiteracy committing ‘default bias’ in a retirement savings plan may be good for the metrics of the policy makers when subscription rate or national savings rate is the metric.[13] However, a focus on such an outcome removes the need for safeguards and accountability. Furthermore, a lack of concern about the opportunity cost of the choice and the dynamic implications on consumption and investment for the illiterate worker raises serious concerns.
Fourth, a ‘one size fits all’ nudge policy assumes that the level of irrationality is homogeneous: a deviation from what would be required in contexts where individuals have different levels of irrationality, that is, ‘asymmetric paternalism’ or applying soft paternalism depending on the level of irrationality of people.[14] Such a nudge centric policy design also runs the risk of asking narrow questions about what works and what does not. For example, replacing an ‘opt-out’ option in forms for voluntary organ donation by ‘opt-in’ systems may increase levels of consent for organ donation in a society, but there should be clear articulation of how such an increase in the supply of organs should be managed or how it would affect the pricing of organs, rather than just a central focus on using insights to increase organ donation in isolation.
While behavioural insights have the potential to aid in effective public policy, and the popularity of using behavioural insights for policy design is expected to grow, there is a need to be mindful about the limitations of behavioural economics in bringing about desirable behaviour change, as well as the impact of the decisions that the choice architecture enables. Furthermore, it is necessary to clarify whether the government considers behavioural economics as antithesis of standard economics or believes the two sub-disciplines to be complementary. Equally important is the need to deliberate upon the philosophical and ethical concerns about nudge policies. For instance, there is confusion in academia and policy circles over whether acts of nudge pertain exclusively to state paternalism (narrow paternalism) or in a broad paternalism, where non-state actors such as a teacher, a parent, a restaurant, or a retail store also engaging in nudging.[15] There are deep concerns about the definitional and normative aspects of nudge that policy makers cannot wish away.[16]
Having an ‘ostrich effect’ along with ‘self-serving bias’ will be detrimental to creating an environment for democratic and lively policy debate. It is not surprising that bureaucrats and policy makers demonstrate ‘herd behaviour’, because even bureaucrats and policy makers are prone to FOMO (fear of missing out), when other countries have already started using nudge policies to ‘steer’ people to make ‘right’ choices. The adoption of behavioural insights as a policy tool, while welcome, calls for a critical evaluation of philosophical, ethical, and pragmatic dimensions.[17] However, it is surprising that the justification for mainstreaming behavioural insights in policy design is driven by the premise that the choice architect has the ability to ascertain the true preferences of individuals in helping them avoid welfare reducing decisions.
[1] The irony is that the definition of ‘anchoring bias’ in the Economic Survey 2019 (p.30) is faulty. It confuses ‘default bias’ with ‘anchoring bias’. Furthermore, the reference to Kahneman and Tversky’s (1974) article titled ‘Judgment under Uncertainty: Heuristics and Biases’ published in Science (Vol.185, No.4157) in this context is misleading.
[2] This is applicable to behavioural insights used in a broad sense; beyond the economics discipline. I confine the discussion to behavioural economic insights in particular.
[3] Thaler, Richard, and Sunstien, Cass. (2008). Nudge: Improving Decisions About Health, Wealth, and Happiness. USA: Yale University Press. (p. 11)
[4] See Thaler and Sunstien’s 2008 book Nudge: Improving Decisions About Health, Wealth, and Happiness (see Note 3 for full citation) and Cass Sunsteins’s book Why Nudge? The politics of Libertarian Paternalism, for a nuanced discussion on paternalism and its relationship with welfare and autonomy as well as discontents of soft paternalism. The distinction between ‘soft paternalism’ vs. ‘hard paternalism’ as well as ‘weak paternalism’ vs. ‘strong paternalism’ is also of relevance in ethical considerations of policy making, an elaboration of which is beyond the scope of this article. Full citation: Sunstien, Cass. (2014) Why Nudge? The politics of Libertarian Paternalism. Yale, USA: Yale University Press.
[5] Thaler and Sunstein 2008, pp.9-11. See Note 3 for full citation.
[6] Sugden, R. (2016). Do people really want to be nudged towards healthy lifestyles? International Review of Economics, 64(2), 113–123.
[7] Thaler and Sunstein 2008, p.11. See Note 3 for full citation.
[8] At the lowest extreme of influence is ‘laissez faire’ whereas the highest extreme is ‘mandate’.
[9] Self-serving bias refers to the human tendency of evaluating the self in an over favourable manner: positive events are attributed to intrinsic characters whereas negative events are attributed to external factors. Fundamental attribution error pertains to the tendency of explaining the behaviour of others based on internal causes. Confirmation bias refers to the tendency to favour outcome or information that confirm one’s prior beliefs. For example, if you believed that a scheme is successful then you tend to favour positive evidence as it confirms your belief.
[10] Government of India. (2019). Economic Survey of India, 2018-2019. (p. 45)
For online access, see: https://www.indiabudget.gov.in/economicsurvey/
[11] Economic Survey of India (p.54). See Note 10 for full citation.
[12] Economic Survey of India (p.42). See Note 10 for full citation.
[13] Default bias or ‘status quo bias’ pertains to the tendency of people to choose the default option or behaving as they have always done.
[14] Camerer, C., Issacharoff, S., Loewenstein, G., O’Donoghue, T., Rabin, M. (2003). Regulation for Conservatives: Behavioral Economics and the Case for “Asymmetric Paternalism”. University of Pennsylvania Law Review, 1151(3), 1211-1254.
[15] Hansen, P. G. (2016). The Definition of Nudge and Libertarian Paternalism: Does the Hand Fit the Glove? European Journal of Risk Regulation, 7: 155–174.
[16] Mongin, P., and Cozic, M. (2018). Rethinking Nudge: Not One but Three Concepts. Behavioural Public Policy, 107-124.
[17] See Kuehnhanss, C.R. (2019). The challenges of behavioural insights for effective policy design. Policy and Society, 38:1, 14-40, for a good review of the history of use of behavioural insights in policy design including the behavioural law and economics development as well as fresh perspectives on the problems of using nudge in public policy.
Sarthak Gaurav is an Assistant Professor at SJMSOM, IIT Bombay. He is also an Associate Faculty at the Centre for Policy Studies, IIT Bombay. He teaches Game Theory, Economics, and Behavioural Economics. He can be contacted on email: sgaurav@iitb.ac.in
Image: Consumer by Nick Youngson [CC BY SA-3.0]