As I expected, I’m greatly enjoying – and learning much from – reading Doug Irwin’s hefty new volume, Clashing Over Commerce  (2017). It’s a masterpiece of scholarship. I’m sure that I’ll quote from it, as well as write more about it, in the weeks and months – even years – to come.
In this post, though, I pick a nit. It’s a nit that I pick not only with Doug but with most other economists, for most other economists routinely commit the same sort of slip-up that I here flag (or, rather, what I regard to be a slip-up). Here’s Doug on page nine:
Economists generally believe that trade restrictions reduce national income, but there are theoretical reasons for promoting “infant industries” if certain conditions are met.
My gripe is with the second part of this sentence.
Although Doug immediately goes on, correctly, to explain that identifying these conditions is practically difficult, recognizing this practical difficulty doesn’t adequately capture the depth and breadth of the weakness of the argument for protecting so-called “infant industries.” Indeed, as I argue below, there in fact really aren’t any theoretical reasons for promoting infant industries. Here’s what I mean: to say that “there are theoretical reasons for promoting ‘infant industries'” is no different, in essence, from saying (for example) that there are theoretical reasons for believing that brown-eyed people can be made better off economically by authorizing blue-eyed people to restrict how brown-eyed people spend their own money.
It’s very easy to describe circumstances under which people with brown-eyes sometimes spend their money in ways that they each individually come later to regret. It’s also easy to describe theoretical conditions under which blue-eyed people, if given the power to superintend and override the spending decisions of brown-eyed people, will use that power in ways that improve the economic well-being of brown-eyed people (even as assessed, after the fact, by brown-eyed people themselves). And yet no one would take seriously the claim that there are theoretical reasons for giving blue-eyed people the power to superintended and override the spending decisions of brown-eyed people. Claims about promoting “infant industries” should be regarded with the very same lack of seriousness.
The theoretical case for government promotion of “infant industries” rests on what Harold Demsetz long ago identified as the “nirvana fallacy .” This case sneaks in the premise that the appropriate standard is a world of perfect knowledge – or, at least, a world in which government officials possess more knowledge of the future, as well as more knowledge of the manifold details of the present, than these officials actually do possess and than we have any good reason for believing that these officials can possibly come to possess.
Of course it’s incontestably true that if we have access to the insights of an honest and hyper-knowledgeable agent, then the economic case is quite strong for giving that agent the power to improve our future welfare by directing our present actions. But no such agent exists. Yet when people write or talk about a ‘theoretical case’ for this or that intervention, they too often simply assume that it is theoretically possible for the government to have the knowledge and information (and incentives) necessary to carry out the intervention in ways that improve human well-being. Those people who are disposed more favorably toward government intervention often believe (or act as if they believe) that government can in practice gather and process such knowledge, while those people (such as Doug) who are generally disposed to look with suspicion upon government intervention correctly point out that government cannot typically in practice gather and process such knowledge.
My point in this post is to note that whenever we use perfect (or super-human) knowledge as a standard, there is no more reason to suppose that such knowledge is, or can be, possessed by government officials than to suppose that such knowledge is, or can be, possessed by any other group of people that you care to name. That is, if, for example, someone writes that there are theoretical reasons for believing that government protection of infant industries will yield positive net results, someone else can write with equal (im)plausibility that there are theoretical reasons for believing that the suppression by blue-eyed people of the economic decisions of brown-eyed people will yield positive net results.
Put differently, if a case for government intervention contains, beyond the usually rather banal demonstration that real-world markets are ‘imperfect,’ merely the assumption that government has both the knowledge and the incentives to carry out the intervention in a welfare-enhancing manner, then it is either meaningless or untrue to say that this is a ‘theoretical’ case for government intervention. The reason is that there is here no plausible theory of how government will get the required information, will process it correctly, and will act on it productively. Without such a coherent and plausible theory of government action, nothing that deserves the name “theoretical reasons” or “theoretical case” for this or that government intervention exists.