Research note. Regulation As Delegation

Reseach note by Oren Bar-Gill and Cass Sunstein
based on the article Regulation as Delegation, Journal of Legal Analysis


In diverse areas – from retirement savings, to consumer credit, to prescription drug use, to fuel economy and energy efficiency rules, to tobacco consumption, to food and beverage consumption – government makes decisions for us or endeavors to help us make better decisions. In other words, government serves as our agent. Principal-Agent Theory (PAT), broadly applied in economics and political science, can serve as a useful framework for considering the optimal scope and nature of this assistance that our agent, the government, provides.

It is quite common to talk about government as the agent of the People in a democratic society (Ackerman 1993).[1] Our focus is not on the People, but rather on an individual person. Accordingly, we are thinking about personal decisions – decisions whose primary effect is on a single principal, an individual.[2] To this extent, we are excluding cases in which people’s decisions affect others. This is not to say that these personal decisions do not have external effects. Often they do. When people eat unhealthy food, they might affect others as well, especially if they become sick. Whenever nations have welfare systems, making sure that people are healthy, save enough, and borrow prudently may prevent third-party effects.[3] But our focus in this Article is on the well-being of the individual, seen as principal.

Our central claim here is that read in light of PAT, a great deal of modern regulation can be understood and evaluated as a hypothetical delegation, through which sensible principals delegate authority to those who can make decisions on their behalf. This claim helps cast a fresh light on some objections to apparent paternalism – as, for example, where government requires people to obtain a prescription before using certain medicines, or forbids workers from running certain risks in the workplace. The use of PAT helps to discipline discussions that might otherwise be far too abstract. Adding a behavioral lens, Behavioral PAT (BPAT) helpfully enriches the basic analysis, suggesting that boundedly rational principals will be prone to both insufficient and excessive delegation.[4]

One of our principal goals is to enlist PAT and BPAT to distinguish among several distinctive kinds of hypothetical delegations, involving information, default rules, incentives, precommitments, mandates, and prohibitions. Focusing on the benefits and costs of delegation, which depend on its type, we identify the circumstances in which one or another approach makes sense.

In the domain of personal decisions, we argue, it is helpful to think about the individual as principal and the government as agent. A different set of regulatory problems – public goods problems – can be conceptualized as a reverse delegation, with the government as principal and the individuals as agents. Here the government-principal, as representative of the People, sets a public objective – a clean, sustainable environment, financial stability, higher educational attainment – and enlists individuals-agents to help attain this objective.

Our central argument is that the idea of regulation-as-delegation provides a useful frame for the evaluation and design of regulation. It brings to the fore personal decisions as a central object of regulation. It offers a unified model for studying disparate regulatory tools that are usually considered in isolation, highlighting important interactions between them. It captures both hypothetical and actual judgments and decisions by informed but boundedly rational agents, who often favor delegations of this kind (Sunstein 2015a, Sunstein 2015b, Bar-Gill and Sunstein 2015). It embeds this unified model of regulatory powers in an even more general framework that exposes the necessary links between the design of regulatory agencies – their structure, goals, and incentives – and the powers afforded to these agencies.

We acknowledge that a sophisticated analysis of welfare-maximizing regulation could produce many (all?) of the results that we describe, without recourse to the delegation frame. It would be possible to dispense with that frame and simply to ask: What kinds of regulation actually promote social welfare? We view regulation-as-delegation not as a substitute for welfare analysis, but rather as a vehicle for pushing welfare analysis further and deeper, and for identifying features of the analysis that might otherwise be ignored. And independent of the direct welfare question, it is valuable to ask about the kinds of delegations that people would, and do, support. We approach that question theoretically, by trying to specify when and why such support would be forthcoming. In ongoing work, we test the theory against survey evidence (Bar-Gill and Sunstein 2015; see also Sunstein 2015a, Sunstein 2015b).

In particular, PAT and BPAT help to identify more clearly the proper scope of regulation. Focusing on personal decisions, the delegation frame emphasizes deficits of information and of rationality on the part of regulated individuals as the main reason for regulatory intervention.[5] PAT and BPAT also help to identify and design the appropriate regulatory approach. For example, we conceptualize default rules as veto-based delegation – as a course of action that the government-agent suggests to the individual-principal. We show that the desirability of this regulatory approach depends on the alignment (or misalignment) of interests between the principal and the agent and on the ability of the individual-principal to effectively exercise her veto power, i.e., to opt-out of the default when it is appropriate to do so.

The unified regulation-as-delegation framework also highlights important interactions between different regulatory tools. In the literature, disclosure mandates and default rules are usually discussed separately. We argue that often, they are best understood as complements: default rules will not work well unless individuals have sufficient information effectively to exercise their veto-power and choose whether to opt-out of the default. Sometimes, of course, individuals will already have that information, or the market will generate it, but sometimes disclosure is desirable or necessary to make opt-out rights something more than a formality. Finally, as a normative matter, conceptualizing regulation as delegation, in the personal decisions domain, bolsters the legitimacy of some regulation, which might otherwise seem questionable as a form of paternalism – but only when the relevant regulation conforms to the optimal delegation contract.[6]

Another contribution of this Article is in employing the regulation-as-delegation frame as a bridge between two important literatures that have developed independently, yet have much to learn from each other. On the one hand, the behavioral economics literature studies policy interventions that respond to the bounded rationality of individuals, but sometimes without fully considering the ignorance, the motivations, or the misalignment of interests of regulators. On the other hand, the political science and administrative law literature pays much attention to regulatory agency costs, but much less attention to the bounded rationality of individuals and how it affects the optimal scope of regulation. Standard treatments of “market failures” do not devote much discussion to the kinds of “behavioral market failures” that, in our view, provide the strongest justification for prominent regulatory regimes.

Read the full-text article:

  • O. Bar-Gill, C. Sunstein. Regulation as Delegation, Journal of Legal Analysis (Spring 2015) 7 (1): 1-36. doi: 10.1093/jla/lav005


About the Authors:



Oren Bar-Gill is Professor of Law and Economics at Harvard Law School. Oren Bar-Gill’s scholarship focuses on the law and economics of contracts and contracting. Personal webpage




Cass Sunstein is Robert Walmsley University Professor at Harvard Law School. From 2009 to 2012, he was Administrator of the White House Office of Information and Regulatory Affairs. He is the founder and director of the Program on Behavioral Economics and Public Policy at Harvard Law School. Personal webpage



[1]For a critical review, see Vermeule 2015. PAT has also been used in political science to characterize and analyze relationships between different government actors, e.g., President and administrative agencies or Congress and administrative agencies (See: Cook & Wood 1989; Gersen 2010).

[2] Saul Levmore’s work (2014a; 2014b) on regulation responding to internalities overlaps with our “personal decisions.” Like us, Levmore writes of individuals soliciting the help of government in dealing with their internalities problems.

[3] At the same time, the issue of third-party effects has to be investigated, rather than asserted, when people run risks. For example, premature mortality might reduce, rather than increase, costs for the welfare system, taken as a whole.

[4] A behavioral lens has been added to PAT in other contexts (See: Gomez-Mejia & Wiseman 1998; Pepper 2015; See also: Leymour 2014a; 2014b).

[5] Relevant discussion can be found in Thaler & Shefrin (1981). Influenced by principal-agent models, Thaler and Shefrin explore the possibility of dual-self models in which the Planner, acting as a kind of principal, constrains the Doer. One of the relevant strategies involves the issuance of binding rules. With modest adjustments, the same approach can be used in cases in which regulation operates as a delegation.

[6] For complementary discussion, which could be adapted to incorporate BPAT, see Chetty (2015) in particular at 25: “Further work is needed to determine whether and how subjective well-being metrics can be used to reliably measure experienced utility, but they appear to offer at least some qualitative information on ex post preferences than can help mitigate concerns about paternalism in behavioral welfare economics.”

List of References

  • Ackerman, Bruce. 1993. We the People Vol. 1: Foundations. Cambridge, MA: Belknap Press.
  • Bar-Gill, Oren, & Cass R. Sunstein. 2015. Regulation As Delegation: Survey Evidence. Unpublished Manuscript.
  • Chetty, Raj. 2015. Behavioral Economics and Public Policy: A Pragmatic Perspective. 105 American Economic Review 1-33.
  • Cook, Brian J., & B. Dan Wood. 1989. Principal-Agent Models of Political Control of Bureaucracy. 83 Am. Pol. Sci. Rev. 965-978.
  • Gersen, Jacob E. 2010. Designing Agencies. In Daniel A. Farber and Anne Joseph O’Connell eds., Research Handbook on Public Choice and Public Law.
  • Gomez-Mejia, Luis R., & Robert M. Wiseman. 1998. A Behavioral Agency Model of Managerial Risk Taking. 23 Academy of Management Review 133-153.
  • Levmore, Saul. 2014a. From Helmets to Savings and Inheritance Taxes: Regulatory Intensity, Information Revelation, and Internalities. 81 University of Chicago Law Review 229-249.
  • Levmore, Saul. 2014b. Internality Regulation Through Public Choice. 15 Theoretical Inquiries in Law 447-470.
  • Pepper, Alexander. 2015. Behavioral Agency Theory: New Foundations for Theorizing About Executive Compensation. 41 Journal of Management 1045-1068.
  • Sunstein, Cass R. 2015a. Do People Like Nudges? Unpublished manuscript, available at
  • Sunstein, Cass R. 2015b. Which Nudges Do People Dislike? A Nationally Representative Survey. Unpublished Manuscript.
  • Thaler, Richard H., & H.M. Shefrin. 1981. An Economic Theory of Self-Control. 89 J. Polit. Econ. 392-406.
  • Vermeule, Adrian. 2015. The Administrative State: Law, Democracy, and Knowledge. In Mark Tushnet et al eds., Oxford Handbook of the United States Constitution.


Photos credits: