Paternalistic Arguments Advocate Forcing Or Manipulating Individuals to Change Their Behavior
Taking Internalities Seriously:
A Coasean Perspective on Intrapersonal Economics
DRAFT COPY, June 2005
by Douglas Glen Whitman
Department of Economics
California State University, Northridge
18111 Nordhoff St.
Northridge, CA 91330
818-677-4542
Taking Internalities Seriously:
A Coasean Analysis ofPerspective on Intrapersonal Economics
By Douglas Glen Whitman
ABSTRACT: Economic arguments for government intervention are usually justified by pathologies of social choice, not self-regarding choice. Externality theory is the paradigmatic case. But the growing “internalities” literature suggests individual choice may display pathologies similar to social choice, justifying similar interventions. Internality theory currently stands where externality theory stood in 1960, before Ronald Coase’s seminal work on the subject. This article applies post-1960 externality theory to the internality literature, emphasizing: the reciprocal nature of internalities, the least-cost avoider principle, property rights, exchange, transaction costs, and comparative institutional analysis. The lessons are applied to demonstrate potential adverse effects of Pigovian sin taxes.Economic arguments for government intervention are usually justified by pathologies of social choice, not self-regarding choice. The theory of externalities is the paradigmatic case. But the growing “internalities” literature suggests individual choice may display pathologies similar to those of social choice, thereby justifying similar interventionist cures. Internality theory currently stands about where externality theory stood in 1960, just prior to Ronald Coase’s seminal work on the subject. If internality theory is to be taken seriously, it must be exposed to the same scrutiny to which externality theory has been subjected over several decades. This article will apply post-1960 externality theory to the internality literature, with special emphasis on: the reciprocal nature of internalities, the least-cost avoider principle, property rights and exchange, transaction costs, and comparative institutional analysis. The overall perspective cautions against legal and legislative solutions to alleged internality problems.
Paternalistic arguments advocate forcing or manipulating individuals to change their behavior for their own good, as distinct from the good of others. Economists have tended to resist paternalistic arguments for two reasons. First, most economists espouse subjectivism, the notion that preferences are individualized and not subject to objective refutation or comparison. Second, they often endorse revealed preference, the notion that individuals’ behavior always reflects – more or less perfectly – their own subjective preferences. Given these two premises, it follows that interventions designed to change an individual’s behavior for his own good, to the extent they affect his behavior at all, will change it for the worse by the individual’s own standards.
But the conventional wisdom no longer commands the support it once did. Some economists have rejected revealed preference in favor of more fallibilist models of choice. Simply put, individuals may have a systematic tendency, particularly in situations involving intertemporal choice, to pay too little attention to costs and benefits of decisions to themselves. Or, to use a common phrasing of the problem, to their other selves. If so, then interventionist policies could potentially make individuals better off even according to their own subjective preferences.
To put it another way, the old paternalism rejected by most economists said, “We know what’s best for you, and we’ll make you do it.” The new paternalism says, “You know what’s best for you, and we’ll make you do it.”[1]
The approach is clever for two reasons. First, it apparently retains the axiom of subjectivity (while rejecting revealed preference). Second, and perhaps more importantly, it draws on economists’ greater willingness to countenance interventions designed to deal with interpersonal problems, especially in cases where choices create costs and benefits to people other than the decision-maker. By treating the individual as a multiplicity of selves, the new paternalism invites economists to import the theory of externalities into the analysis of individual choice. This strategy is reflected in the popular use of the term “internality” for the failure of decision-makers to take full account of the costs or benefits of their choices to other (usually future) selves. Herrnstein, et al. (1993) coined the term and defined it as “a within-person externality … which occurs when a person underweighs or ignores a consequence of his or her own behavior for him- or herself” (150).
As readers of this journal know well, the theory of externalities has undergone a radical transformation in the decades since the publication of Ronald Coase’s (1960) “The Problem of Social Cost.” Prior to Coase’s article, externality theory followed the analysis of A. C. Pigou (1932). Pigou argued that (a) private decision-makers will engage in inefficiently high (low) levels of activities that create external costs (benefits), and (b) that an appropriately chosen tax (subsidy) of the activity could correct the inefficiency. Coase’s analysis threw a mighty wrench in Pigou’s analysis, forever altering externality theory.
Yet the theory of internalities continues to proceed on mostly Pigovian lines. Exhibit A: Gruber and Koszegi (2001), in the Quarterly Journal of Economics, contend that cigarette smoking produces negative internalities, argue that government policies “should depend not only on the externalities that smokers impose on others but also on the ‘internalities’ that smokers impose on themselves” (1261), and calculate that “there are sizable optimal ‘internality’ taxes on the order of $1 per pack or more” (1294). No references to Coase or the post-“Social Cost” literature appear in their references; little or no attention is paid to Coasean insights.[2]
Exhibit B: O’Donoghue and Rabin (2003a), in AEA Papers & Proceedings, discuss “optimal sin taxes” designed to correct self-control problems, using the specific example of overeating. In their not-yet-published extension of that article (O’Donoghue and Rabin 2003b), they note that “Since people with self-control problems impose negative externalities on their future selves – dubbed ‘negative internalities’… – the role that sin taxes play in our analysis is much like a Pigovian tax to correct negative externalities” (2, note 3). References to Coase and his insights are again absent.
Other authors, like Camerer, et al. (2003) and Thaler and Sunstein (2003) have offered more tentative policy prescriptions, preferring instead to weaken the prima facie case against paternalism. Recent research in behavioral economics, they say, “potentially broadens the scope of situations in which paternalistic policies could usefully be developed” (Camerer et al. 2003, 1214). Still, they invoke the term “internalities” (1221, note 50) along with its hefty market-failure baggage: “When consumers make errors, it as if they are imposing externalities on themselves because the decisions they make do not accurately reflect the benefits they derive” (1221). The model deployed is straightforwardly Pigovian, again with no reference to the post-1960 law-and-economics literature.
A very few authors, usually often working outside economics, have come closer to applying Coasean insights to internalities. George Ainslie (1992, 2001), in particular, comes tantalizingly close. He regularly refers to an “internal marketplace” within the person (2001, 39), and he argues that choice emerges from a process of “bargaining” among the competing interests within the self (e.g., Ainslie 2001, 43-44). Both phrases are suggestive of the transactional solutions to externality problems that figure so prominently in the Coase Theorem. Yet Ainslie never invokes Coase’s more specific insights. (Ainslie’s approach will be considered in more detail in section III.)
The intent of this article is to take internality theory seriously, at least for argument’s sake. To do so involves exposing it to the same scrutiny to which externality theory has been exposed for the last several decades. If the unified individual is a myth or an exaggeration – if a person really does consist of competing preference sets that can usefully be characterized as distinct selves – what follows? Is there a prima facie case for paternalist intervention? I propose to examine these questions through a Coasean lens, deploying the following concepts: the reciprocal nature of internalities; the least-cost avoider principle; property rights and exchange; transaction costs; and comparative institutional analysis. The lessons will be specifically applied to demonstrate the possible ill effects of a Pigovian sin tax. In the process, I hope to demonstrate that the case for internality-correcting interventions is substantially weaker than it appears.
This article is intended as an internal critique. It takes the concept of internalities as given (except when tensions within the literature require further clarification) and explores its logical and economic implications. Accepting the article’s conclusions does not necessarily mean endorsing internality theory. Readers who find internalities intuitively plausible may regard the article as a natural continuation (or correction) of the internality literature. Others may regard the article as further evidence that internality theory is untenable. Either conclusion would be consistent with the paper’s intent.
I. The Reciprocal Nature of InExternalities
“The question is commonly thought of as one in which A inflicts harm on B and what has to be decided is, How should we restrain A? But this is wrong. We are dealing with a problem of a reciprocal nature. To avoid harm to B would be to inflict harm on A. The real question that has to be decided is, Should A be allowed to harm B or should B be allowed to harm A? The problem is to avoid the more serious harm.” (Coase 1960, 2)
Let us suppose the individual consists of more than one self. One common approach says the individual contains many different selves, one present-oriented self for each moment in time and a single future-oriented self. E.g., Thaler and Shefrin (1981) refer to many present “doer” selves and a single forward-looking “planner” self.[3] An alternative (and not always clearly distinguished) approach imagines just two selves, a single present-oriented self and a single future-oriented self. For now, I will simplify by considering just two selves, a short-run self and a long-run self (which may be interpreted as the agent of future short-run selves).
The typical analysis observes that the short-run self can take actions, like smoking or overeating, that will harm the interests of the future long-run self. He may also take actions, like exercising or visiting the dentist, that will benefit the future self. Insofar as the short-run self does not concern itself with such future costs and benefits, they constitute negative and positive internalities. The short-run self therefore engages in too much of the harmful activities and too little of the beneficial activities, at the expense of the long-run self. For simplicity, I will focus on the harmful activities.
The approach just described is essentially Pigovian. It regards one actor, the short-run self, as the sole cause of the harm, and the other actor or actors, the long-run self or all future short-run selves, as the passive victims of that harm. As Coase observed in the quotation above, that analysis is one-sided. True, allowing the present self to smoke or overeat means harming the future self. But by the same token, preventing smoking or overeating on behalf of the future self means harming the present self.
To take the notion of multiple selves seriously, the analysis must consider both sets of interests or preferences. We may not assume that the long-run self’s interests somehow supercede those of the short-run self, as that would involve choosing one preference set over another – thereby violating the axiom of subjectivity. Thus, it would not be appropriate to adopt policies solely on grounds that they advance the interests of the long-run self.
As Coase observes, the real problem is to avoid the more serious harm. But nothing about the situation, certainly not the fact that the short-run self may impose harm on the long-run self, shows that the long-run harm exceeds the short-run harm. This becomes apparent if we consider that the long-run self might also have the capacity to harm the short-run self by adopting self-control devices – e.g., flushing cigarettes down the toilet, refusing to allow ice cream in the house, checking into a clinic, and so on. The future long-run self may also impose the cost of guilt on the present self.[4] Such actions help the long-run self at the expense of the short-run self. Given the reciprocal nature of the problem, and having no further information, we could just as easily conclude that the long-run self imposes internalities on the short-run self that require correction. Perhaps we should tax weight-loss clinics.
Going a step further, we could observe that future-oriented selves sometimes induce behaviors that, at least to outside observers, appear excessive. In contrast to the obese and the profligate, whose short-run selves constantly trump their long-run selves, we might – like Cowen (1993) and Ainslie (2001, 115) – point to the misers, workaholics, and anorexics for whom the reverse appears to be true. Even among “normal” individuals, Heath and Soll (1996) observe that excessive self-control efforts can lead to underconsumption of desirable things. Perhaps we should subsidize Krispy Kreme.
Or, following Coase, we could recognize that harm is a two-way street. The existence of an interactive effect does not, in itself, tell us that an inefficiency exists; nor, given an inefficiency, does it tell us the direction of its effect.
Internality theorists often – e.g., O’Donoghue and Rabin (1999), Ainslie (1993, 2001) – tie internalities to hyperbolic or quasi-hyperbolic discounting. Unlike exponential discounters, who apply a constant rate of time-discounting between time periods, hyperbolic discounters apply an additional discount to time periods other than the present.[5] This phenomenon can be captured simply in the following quasi-hyperbolic[6] utility function:
where β < 1 and δ < 1. From this utility function, it can easily be shown that the agent’s marginal rate of substitution between utility in any two successive future periods is δ, but the marginal rate of substitution is βδ if one of the two periods compared is the present. The parameter β represents the agent’s degree of present-bias.
Hyberbolic discounting dovetails nicely with internality theory, because the agent’s present-bias can be attributed to the present self’s inattention to costs and benefits experienced by future selves. However, the same analysis can be used to demonstrate the reciprocal nature of externalities. The hyberbolic utility function could be corrected to make it exponential in two distinct ways. First, β could be raised to equal one (or the agent could be forced or manipulated to behave as though it were). Then the agent’s effective utility function would be: