Counterpoint to Six Potential Arguments Against Regulating Internalities
April 20, 2015 04:43 PM
Hunt Allcott and Cass R. Sunstein
Article first appeared online May 2015
Click here for the online article
Here we consider six potential counter-arguments to our paper, “Regulating Internalities.” This is a general rebuttal, not a specific counter-argument against Dudley and Mannix (2015), which appears in this same edition; but we believe that the arguments that we rebut include the major claims set out in that paper.
Argument 1: Internalities don’t exist.
Such an argument is difficult to square with a large and growing body of research showing that consumers can be present-biased, over-optimistic, inattentive, and misinformed in some settings; see DellaVigna (2009) for a review. Of course, there are many other settings where we do not make systematic mistakes, so policymakers need contextually-relevant empirical evidence. And it is important to note that the magnitude of the internalities, when they exist, may be either small or large.
Argument 2: As a general principle, governments should never interfere in individual decisions unless there is harm to others.
This idea raises deep philosophical questions (see Sunstein, 2013; Conly, 2012; Rebonato, 2011), which we cannot engage here. Many policies (seat belt laws, product safety regulations, etc.) enjoy broad support even though a paternalistic justification is hard to avoid, which suggests that most people do not think that the government should never act paternalistically (Conly, 2012; Sunstein, 2013). In markets where there are internalities, numerous papers have shown theoretically that taxes or other forms of government intervention can increase welfare (Allcott, Mullainathan, & Taubinsky, 2014; Heutel, 2011; O’Donoghue & Rabin, 2006, etc.), and our simple model (shown earlier in our Point article for this section) gives the basic intuition for this result. Of course, the fact that policy can increase welfare does not mean that any policy will increase welfare—each specific proposal requires cost-benefit analysis.
Argument 3: Policymakers should not impose their preferences on constituents.
We agree with this argument. The central feature of the approach we advocate is that policymakers are not imposing their own preferences on constituents. Instead, policymakers should set and evaluate policy to match individuals’ own preferences from the subset of contexts when they are fully informed, making active choices, and fully considering present and future costs and benefits.
Argument 4: Policymakers should maximize behavior change instead of maximizing welfare.
While economists hardly ever make this argument, it frequently arises in environmental and health policy debates. For example, one reading of some provisions of the Clean Air Act is that the EPA must impose economically and technically feasible requirements instead of maximizing net benefits. In the health context, one might argue that some foods be taxed or banned because they are unhealthy, without regard for how sugar, salt, and fat can make food taste better. We think that welfare maximization is the right criterion (for one explanation, see Sunstein, 2014; on the notion of welfare, see Adler, 2011). Energy conservation and healthy eating can produce large gains, but their socially optimal quantities are not infinite, so we must use cost benefit analysis to determine the social optimum.
Argument 5: Policymakers should not regulate internalities without solid evidence that consumers are biased.
Taken in one sense, this statement is important and unobjectionable; policies should always be supported by solid evidence. Taken in another sense, however, the statement is more provocative; it suggests a strong presumption against regulation based on internalities. We believe that policymakers should use the best available evidence, and should not adopt any such presumption.
Indeed, key features of the rational model such as the discounted utility model originated as mathematically-convenient modeling tools, not as accurate descriptions of human behavior (Frederick, Loewenstein, & O’Donoghue, 2002). Even as he formalized the discounted utility model, Samuelson (1937) wrote that “any connection between utility as discussed here and any welfare concept is disavowed,” adding that “it is completely arbitrary to assume that the individual behaves so as to maximize an integral of the form envisaged.” It is not clear why a model introduced for mathematical convenience should become the “null hypothesis” if a modified version could be more descriptively accurate in a given situation.
If a policymaker is risk-neutral and uncertain about the average marginal bias, uncertainty about internalities enters the optimal subsidy formula just like heterogeneity in bias. For example, imagine that the policymaker believes that there is a 50 percent chance that there is no bias and a 50 percent chance that the average marginal bias is 2b1. The optimal subsidy is still (2b1+0)/2=b1, just as in the case with heterogeneous bias when the policymaker knows with certainty that the average is b1. Thus, being uncertain about bias is not a reason to avoid regulation. Of course, being uncertain about whether there is a bias at all (e.g., suggesting that the average marginal bias is 0 when it was previously thought to be b1) does imply that the optimal corrective policy is less stringent, or even that the optimal policy is to do nothing.
This is just a special case of a broader issue: policymakers almost always act with imperfect information, whether regulating internalities, providing public goods, or acting in any other traditional function. When maximizing expected welfare under uncertainty, uncertainty does not generically imply that the optimal policy is to do nothing.
Argument 6: Regulators make mistakes, so they shouldn’t regulate in response to consumer mistakes.
Glaeser (2006) provides three models under which bounded rationality reduces the quality of government decisionmaking more than it reduces the quality of private decisionmaking, implying that bounded rationality strengthens the case for limited government. His three models show that (1) regulators have less incentive to make the right decisions for consumers than consumers do for themselves; (2) if errors come from the influence of firms or interest groups, it is easier to “buy” a small number of regulators than a large number of consumers; and (3) individuals have more incentives when making private decisions than they do when voting, so elected politicians are less likely to have the right views on how to interfere.
We would make two notes about Glaeser’s argument. First, while the models focus on paternalistic policies, the results would also generalize to regulation in traditional spheres such as regulating externalities or providing public goods. Thus, while bounded rationality strengthens the case for limited regulation in these models, this is true for all kinds of regulation, not just paternalistic regulation.
Second, and more importantly, these are not the only three models, and it’s easy to formulate alternative models that strengthen the case for intervention. Glaeser (2006, p. 144) proposes one such model that could reverse his main results, under which there are returns to scale in information acquisition and decisionmaking. In this case, it is more likely to be optimal to centralize decisionmaking. This model is broadly relevant to our consumer protection and energy efficiency examples, where it can be difficult to gather and weigh information about chemical composition, future energy use, and obscure but relevant safety risks. Turrentine and Kurani (2007) document how calculating the present discounted value of energy costs can be difficult, even for accountants, bankers, and financial analysts.
HUNT ALLCOTT is an Assistant Professor in the Department of Economics at New York University, 19 West 4th Street, New York, NY 10012 (e-mail: email@example.com).
CASS R. SUNSTEIN is the Robert Walmsley University Professor at Harvard, Areeda Hall 225, Harvard Law School, Cambridge, MA 02138 (e-mail: firstname.lastname@example.org).
Adler, M. (2011). Well-Being and Fair Distribution: Beyond Cost-Benefit Analysis. Cambridge, UK: Cambridge University Press.
Allcott, H., Mullainathan,S., & Taubinsky, D. (2014). Energy Policy with Externalities and Internalities. Journal of Public Economics, 112, 72–88.
Conly, S. (2012). Against Autonomy: Justifying Coercive Paternalism. Cambridge, UK: Cambridge University Press.
DellaVigna, S. (2009). Psychology and Economics: Evidence from the Field. Journal of Economic Literature, 47, 315–372.
Dudley, S., & Mannix, B. (2015). The Limits of Irrationality as a Rationale for Regulation. Journal of Policy Analysis and Management,
Frederick, S., Loewenstein,G., & O’Donoghue, T. (2002). Time Discounting and Time Preference: A Critical Review. Journal of Economic Literature, 40, 351–401.
Glaeser, E. (2006). Paternalism and Psychology. University of Chicago Law Review, 73, 133–156.
Heutel, G. (2011). Optimal Policy Instruments for Externality-Producing Durable Goods under Time Inconsistency. NBER Working Paper 17083.
O'Donoghue, T., & Rabin, M. (2006). Optimal Sin Taxes. Journal of Public Economics, 90, 1825–1849.
Rebonato, R. (2011). Taking Liberties: A Critique of Libertarian Paternalism. London, UK: Palgrave McMillan.
Samuelson, P. (1937). A Note on Measurement of Utility. Review of Economic Studies, 4, 155–161.
Sunstein,C. R. (2013). Why Nudge: The Politics of Libertarian Paternalism. New Haven, CT: Yale University Press.
Sunstein, C. R. (2014). Valuing Life. Chicago, IL: University of Chicago Press.
Turrentine, T., & Kurani, K. (2007). Car Buyers and Fuel Economy? Energy Policy, 35, 1213–1223.