Two (lay) dogmas on externalities

From whose (Quine, 1951) I have shamelessly appropriated, not only my title, but also my opening motif.

In this context it may be worth noting that my present project has something of its own pedigree. Writing just over fifty years ago, S. E. Holtermann (1972, p. 78) complained of “a tendency in the literature on public goods to identify the two concepts [that is, public goods and externalities] with each other,” and advertised that “[t]he view taken here is that they are two distinct concepts”—even if it remains the case that “many externally produced commodities have the character of public goods.” I am clearly not alone, then, in taking note of this conflation, and evidently it was common enough among professional economists in the 1960’s and 1970’s. Half a century has elapsed since Holtermann published these remarks. While perhaps they had some effect in clearing up economists’ thinking on this score, apparently they did not suffice to completely banish the confusion—and hence the need, still, for the present paper.

See, e.g., the discussion of Costello and Kotchen (2022) and Jacobsen et al. (2017) in footnote 10 below.

For example, see the horizontal dimension of Fig. 1 in Paniagua and Rayamajhee (2023) elsewhere in this volume, which offers an explicit attempt to taxonomize externalities with respect to their scope in precisely this fashion.

Or, in a previous era, on “indivisibility” rather than non-rivalry.

The laity also exhibits a regrettable tendency to improperly deploy the term “public good.” Oftentimes the phrase seems to be understood to mean roughly “anything that’s generally in the public’s interest”; other times it seems to be conflated with “any good provided by a public (i.e., state) entity.” Analysis of this “lay dogma” on public goods lies beyond the scope of this paper, however.

Elsewhere in this volume, Paniagua and Rayamajhee similarly emphasize the importance of not eliding the distinction between goods and services, on the one hand, and the impacts of their production and exchange, on the other—i.e., the importance of distinguishing analyses of the goods that generate externalities from analyses of the externalities themselves. As they state in the fifth paragraph of their (2023), one of their “points of departure” is that they “focus on externalities themselves rather than on associated goods or services.”

Alex Tabarrok (1998) has proposed the use of dominant assurance contracts as one potential market mechanism for surmounting some of the market failures that preclude public good provision.

Other formulations of this dogma might be cited as well, but Friedman’s is worth highlighting because it is especially clear and forceful in this regard. For good measure, though, I will share here another reference which, though perhaps expressing the position less directly, bears a slightly more “canonical” status. Cornes and Sandler (1996) describe public goods as a “special case” of externalities in their Preface (at page xvii; cf. also the definition offered on p. 6). Shortly thereafter—and also on p. 6—they clarify that externalities and public goods are both instances of the more general phenomenon of (market-failure-generating) “incentive structures.” 

One also often encounters this view in the applied economics literature. A couple recent examples drawn from the Journal of the Association of Environmental and Resource Economists will illustrate. Costello and Kotchen (2022, p. 952) clearly betray their coextensivist commitments when they write that “Our starting point is one where an environmental externality exists (creating an environmental public bad) …” Meanwhile, Jacobsen et al. (2017) title their paper “Public Policy and the Private Provision of Public Goods Under Heterogeneous Preferences,” but it seems clear from the examples they employ that the “public goods” whose private provision they are investigating are more aptly characterized as garden-variety external effects. Their central example, in fact, centers on literal gardens. (This is especially literal for speakers of British English, for whom “garden” is sometimes used to mean roughly what Americans mean by “yard” or “lawn.”) On p. 245 we read that “landscaping around private homes” is a “local public good.” (It is, in fact, an instance of what they label “individually-provided public goods.”) And later they write (on p. 261) that “Privately provided neighborhood amenities such as well-kept lawns, sidewalks, and home exteriors are local public goods to which the policy implications of our model may be applied directly.” (A few sentences later they further elaborate their analysis by noting that “manicuring and fostering an extensive garden” is a more costly way of providing this public good, relative to merely “mowing the lawn and trimming.”) Furthermore, it seems clear that they are not modeling such behaviors as individual contributions to a broader public good, such as “neighborhood aesthetics”; rather, the outcomes of such homecare initiatives are themselves the public goods. Later, in the context of analyzing individual households’ decisions regarding energy efficiency, they invite readers to “Imagine a public good like the one coming from household energy efficiency and the associated reduction in climate change impacts for the globe” (p. 277). Again: that household’s increased energy efficiency is itself regarded here as the public good.

See, again, note 8 above, citing Paniagua and Rayamajhee’s insistence on drawing this same distinction.

For more on the “packageability” of public goods and services, see Rayamajhee and Paniagua (2021), especially Sect. 3.3.

Heath (2022, p. 7), emphasis my own. Other formulations of this claim might be cited as well, but Heath’s is worth highlighting because it is especially clear with regard to the fact that a public good is, in an interesting sense, a “limiting case” of an externality-generating good. For good measure, though, I will share here another reference which, though perhaps expressing the position less directly, bears a slightly more canonical status. Section 3.5 of Cornes and Sandler (1996) comprises a taxonomy of “special types of externalities”; they write there that “The special case [of an externality] that we call the ‘standard’ pure public good is perhaps the most common one encountered in the literature” (1996, p. 53).

Samuelson (1954) is widely regarded as the origin of the contemporary literature on public goods.

This example involves a positive externality, but of course the point could also be illustrated by cases involving negative externalities. The real-world case of brush fires on private lands, caused by errant sparks from locomotive steam engines—something which clearly represents a negative externality visited upon the relevant landowners, but which is in no wise a public good (or, as the case may be, “public bad”)—has been made famous in the economics literature via its extensive analysis in, e.g., Pigou (1920) and Coase (1960).

Obviously, the local/global distinction I’ve been deploying here is a matter of degree, not a firm distinction. The spillover effects from one party’s consumption of a private good can impact just one other party (a purely local externality), or they can impact every single member of the relevant community (a wholly global externality), and of course anything in between. This continuum is nicely captured by the scalar nature of the horizontal dimension of Fig. 1 in Paniagua and Rayamajhee (2023). (This is what they term the “scale/size of externalities” dimension. While their Fig. 1 depicts this variable as discrete—cleaving into the categories of “small-”, “medium-”, or “large-scale” externalities—it is clear from their main-body text (e.g., Sect. 2 of their paper) that this tripartite division is simply an artifact of their diagram, and that really they regard this variable as continuous.) We can also express this observation in terms of my formal model from above. Where the spillover impacts only party “i + 1”, we have the limit case of a (very) local externality. When the spillover impacts every single other member of the relevant public—the case where j = 0—then we have the limit case of a wholly global externality. In reality, of course, most cases will fall somewhere between these polar extremes, and spillover effects will be more-or-less local or more-or-less global, to varying degrees.

A Wikipedia entry supplies an overview of this affair, and includes the text of the memo itself: https://​en.​wikipedia.​org/​wiki/​Summers_​memo. For their part, Summers and Pritchett maintain that the memo was in crucial respects intended as sarcasm or parody; Summers’ side of the story is succinctly presented in a contemporaneous article in the Harvard Crimson: https://​www.​thecrimson.​com/​article/​1992/​2/​25/​summers-memo-ironic-pa-harvard-professor/​.

Indeed, as Steven Cheung (1973) has taught us, orchard owners and beekeepers in Washington state (and presumably elsewhere) have learned to package, and to trade, the (beneficial) external effects of their commercial enterprises: orchard owners contract to have beekeepers transport their hives into their orchards for purposes of promoting pollination. (Obviously, the beekeepers benefit likewise, from their bees’ access to nectar.)

Once again, additional sources might be cited here as evidence of the pervasiveness of this dogma. But once again, I shall rest content with buttressing the case with citations from a single source—one which, whatever it might lack in forcefulness of expression, it compensates for in its “canonicality.” Cornes and Sandler (1996) first define externalities as any “uncompensated interdependencies” (p. i), then later as “the action of one economic agent [that] influences the utility or production function of another, and [for which] no mechanism for compensation exists” (p. 5). Eventually, they cite with approval a definition offered by Meade (1973): “An external economy (diseconomy) is an event which confers an appreciable benefit (inflicts an appreciable damage) on some person or persons who were not fully consenting parties in reaching the decision or decisions which led directly or indirectly to the event in question” (cited at Cornes and Sandler (1996, p. 39)).

Though the linguistic device of “economic vs. social externalities” is not commonly employed in the literature, the distinction it marks does seem to be increasingly taken note of. For example, Cornes and Sandler attest to the historical reality of a sort of “conceptual inflation” with respect to externalities when they write (1996, p. 6) that, subsequent to Pigou, “the notion of externality encompassed an ever-increasing variety of economic situations.” And for two contemporary examples of authors who do use the term “social externality” in approximately this way, see Vogler (2024) and Fleurbaey et al. (2020). The former argues that, at least in certain analytic contexts, it is useful to have a “definition of externalities … broader than the traditional perspective because it is not limited to voluntary economic exchange, but also applies to many other forms of social interaction.” (Vogler then proceeds to define “economic exchange” as “all market transactions of labor, goods, services, and currencies that two or more parties voluntarily engage in and that are associated with prices,” and “social interaction” as “voluntary and involuntary encounters of two or more people in a narrow space that involve oral communication or physical contract or both.”) (All quotations taken from p. 2.) The latter propose to “further enlarge the [traditional] perspective [on externalities] and start thinking about a broader framework in which any pattern of influence of an agent or a group of agents over a third party, which is not mediated by any economic, social, or psychological mechanism guaranteeing the alignment of the marginal net private benefit with marginal net social benefit, can be attached the ‘externality’ label and be scrutinized for the likely negative consequences that result from the divergence” (page 6, emphasis in the original).

To say that relatively few social externalities merit resolution through standard economic policy-making tools, though, is not thereby to imply that all economic externalities are the proper subject of Pigouvian intervention. To cite just one important example, pecuniary externalities are no doubt properly classified as negative economic externalities, but few economists would endorse government intervention to mitigate these effects of market competition.

At this point, though, it is important to consider a rival view of things. For it may well be the case that a citizenry’s facility with the notion of “social externalities” could lead to behavior that is more thoughtful and considerate. Pervasive recognition of the fact that personal behavior generates uncompensated external effects might manifest in a stronger desire not to (unintentionally) harm others in one’s immediate environment. Widespread uptake of the notion of social externalities might thereby substantially reduce the need to ask for third-party enforcement. (I am indebted to an anonymous reviewer for recognizing this possibility, and in particular for this way of expressing it.) While this perspective undoubtedly has merit, I remain unpersuaded that the net effect of widely adopting such an economic, transactional, “market-ized” conception of all social relations would be pro-social.

Daniel Halliday and John Thrasher (2019, p. 249) make a similar point when they discuss the idea (which they attribute to some unpublished work of political philosopher Jonathan Wolff) that even the most enthusiastic partisans of markets are likely to wish for some spheres that remain insulated from market forces. They use the example of passengers on crowded, standing-room-only subway car, buying and selling seats on their daily commutes, as an illustration.