Impossibility of a theory of externalities

Time of India

17-05-2025

If the same analytical 'tool', when placed in the hands of different people, tells us two radically different things, such as that lockdowns are good and that they are bad, it means we are not dealing with science but with quackery. As I have shown earlier, ethics and epidemiological models fall in that category. In this piece, I show that the standard theory of externalities in economics falls in the same camp.
For long, like any other student of economics, I thought of externalities as an inefficiency, a market failure, that might require government intervention. But when I tried to frame my review of public health in the context of externalities, I arrived at a dead end.
My empirical analysis had conclusively shown that quarantine fails in almost all cases even as it causes great harm. But many economists supported quarantine on the basis of the externalities theory. In 2020, Thomas Firey at the Cato Institute wrote: 'Given the seriousness of the Covid-19 negative externality, legally mandating masks and distancing is appropriate'. In 2020 also, Guillaume Vandenbroucke at the St. Louis Federal Reserve justified an 'early social-distancing mandate' by citing the lack of 'internalisation' of negative externalities. Hideo Hayakawa, a Japanese economist wrote in 2021 that 'there is a powerful economic case to be made for government intervention, including full or partial lockdowns'.
Just by mentioning the term 'negative externality', lockdowns can apparently be justified.
Likewise, while my empirical analysis had conclusively rejected vaccine mandates, many economists support them based on the theory of externalities. Arguing that those who don't take a vaccine are 'free riding' on the positive externality created by those who take a vaccine, Nobel prize winner Joseph Stiglitz justified vaccine mandates in 1988. He never bothered to check if anyone was actually free riding or whether there were other reasons why someone might choose not to take a vaccine. Dr Jay Bhattacharya's 2014 health economics textbook argued that 'many governments have adopted an alternative Coasian solution in which property rights are reassigned: people do not have the right to stay unvaccinated'. In law, property is either physical or intellectual; humans are not anyone's property, so such an 'assignment' is intrinsically unlawful.
In 2020, Nobel prize winning economist Paul Romer wrote that 'this [negative externalities] is one problem economists know how to solve'. His solution to the covid issue: mass testing. But empirical evidence shows that mass testing fails at multiple levels and causes much harm.
Fortunately, in 2011 I had an extensive email discussion about externalities with one of the world's greatest economists, Harold Demsetz. In his 2011 paper: 'The Problem of Social Cost: What Problem?' he had destroyed the roots of the theory of externalities: 'The literature of externality problems, from Pigou to Coase, makes it seem as if they are associated with inefficiencies, but they are not'. Not only are Pigouvian 'externalities' not inefficient, the so-called Coase theorem is trivially true even as the idea that positive transaction costs lead to inefficiency is wrong.
I'm writing a book (available freely online) on externalities which goes into details, but in this piece I will summarise why a theory of externalities is impossible. And why, therefore, we must abandon this Quixotian enterprise and focus on empiricism to determine any external effects and their 'solution'.
1. The concept is undefinable, possibly ideological
It is important to start by noting that the theory of externalities is not a theory of external effects. It is a theory of the failure of markets to (allegedly) allocate resources efficiently because of unpriced external effects. No one is questioning the existence of subjectively felt or observed external effects. The dispute is purely with those who claim that (a) external effects are inefficient, and that (b) Arthur Pigou's remedies (tax or subsidy) can allocate resources optimally.
Pigou was the source of the idea of inefficient 'externalities'. In the 1932 fourth edition of his 1920 book, Economics of Welfare, he wrote about costs and benefits that 'are thrown upon people not directly concerned … All such effects must be included—some of them will be positive, others negative elements —in reckoning up the social net product'.
This definition has turned out to be impossible to understand. After intensive doctoral studies under Amartya Sen, Andreas Papandreou wrote in his 1998 book, Externality and Institutions that: 'Given the importance of externality in economic theory … it is surprising how hazy a concept it has remained. …[N]o precise and agreed-upon meaning of the term seems yet to have emerged.' After over 250 pages of dense analysis, Papandreou concluded: 'there cannot be a unique good characterization of externality' and that '[a]ttempts to characterize externality [have] stumbled over attempts to find a good criterion demarcating 'external' from 'internal''. If there's no clear definition, we obviously can't proceed in a scientific or logical way. Instead, as Carl J. Dahlman wrote in 1979: 'This is not science; it is metaphysics: value judgments and political goals … enter into the determination of whether externalities occur in our world'.
2. Impossibility of measurement
Every moment we face trade-offs and associated opportunity costs. If we know that traffic will be congested between 8 am and 9 am, we can start for work at 7 am (and thus save 10 minutes of driving time) but that means we will need to reduce our sleep by an hour. Instead, we might decide that the inconvenience of 10 minutes spent in peak traffic is less than the loss of an hour's sleep. The point being that all costs are subjective: only we ourselves know what's best for us, and we signal this through our choices.
What about air pollution? 'We' may think there is air pollution in our city but who are 'we'? Are 'we' a new migrant from a village, looking for a job, or are 'we' a wealthy, fashionable economics professor who lives in a posh locality but finds the air 'unbreathable'? All external effects are subjective – a matter of perception. Many costs become visible to external observers when we interact with the price system, but with external effects, there is no such interaction.
The concept of 'social marginal product', which Arthur Pigou wrote about, is about the sum of all such individual subjective valuations of external costs or benefits across an entire society. How can we possibly measure this? All we can observe (in the case of the air pollution example) is new migrants moving into polluted cities and abandoning their clean villages. As newcomers, they are signalling that the net external effect of polluting factories is positive due to jobs and opportunities for their children, and overall agglomeration effects that offset pollution. The Chernobyl nuclear meltdown did impose a negative social cost, with people fleeing; but for most air pollution, the net external effect is indeterminate.
Does that mean there is no air pollution? This is an intensely local and empirical question, linked with affordability. The externalities theory can tell us nothing about it. Typically, only a rich society that has gone through a phase of heavy air pollution while it was poor, will be capable of abating pollution without significantly reducing overall output. There are no costless Pigouvian solutions.
Facebook Twitter Linkedin Email Disclaimer
Views expressed above are the author's own.