When I first discovered Friedrich A. Hayek’s work on information and markets, it struck me as brilliant. The simple observation that practical information is often widely distributed across societies and markets struck me as not only brilliant, but something underappreciated in philosophy. When we consider the specialized information and expertise is takes to make a peanut butter and jelly sandwich, not merely by assembling a few premade ingredients from the supermarket, but making the sandwich from scratch. Each individual simply lacks the information and expertise do make a peanut butter and jelly sandwich from scratch. As I wrote in a past Substack post,
There are many ingredients to a peanut butter and jelly sandwich, not to mention the components needed to process, pack, and ship those ingredients. Start with the bread: to make the bread one would need to know things like how to domesticate wheat, how to produce farm equipment to grow and harvest the wheat, how to produce fertilizer—along with the equipment required to manufacture farm equipment and tools. Just think about the many inputs needed to produce the rubber that makes up the tires on the tractor one may use to harvest what. And that’s just some aspects of producing wheat, not to mention the processes needed to produce wheat, the machinery needed to produce the bread, the energy needed to run the machine(s), and so on and so forth. The same point applies to peanut butter and the jelly too. So, really, in a fundamental sense, no individual knows how to make a peanut butter and jelly sandwich. And that is a relatively simple object.
As Hayek himself writes,
[It] is almost heresy to suggest that scientific knowledge is not the sum of all knowledge. But a little reflection will show that there is beyond question a body of very important but unorganized knowledge which cannot possibly be called scientific in the sense of knowledge of general rules: the knowledge of the particular circumstances of time and place. It is with respect to this that practically every individual has some advantage over all others because he possesses unique information of which beneficial use might be made, but of which use can be made only if the decisions depending on it are left to him or are made with his active cooperation.
And that is right to a point. Hayek does hold that a great deal of practical information is widely dispersed throughout the population such that no individual knows how to do much of anything by themselves. In modern societies, it is often better for people to specialize in producing or offering goods or service as a means of increasing the productivity of workers, and the prosperity of a society.
However, this isn’t the deep insight for which Hayek deserves credit. Hayek aims to argue against the centralized planning of the economy. And often this is mistakenly taken to be Hayek claiming that the problem of widely dispersed information makes central planning too difficult for a group of people on a committee pull off. And here, critics will often argue that such a problem can be solved by more computing power. If only we had a sufficiently powerful supercomputer or AI, we design the economy having solved the problem of gathering widely dispersed information with the aid of a very powerful computer or AI.
However, appealing to supercomputers and AI won’t solve the Hayekian challenge to the feasibility of a centrally planned economy that must rely on widely distributed information to plan a prosperous economy. The reason for this is that the problem Hayek diagnosed isn’t simply that information needed to effectively plan an economy is dispersed through the economy and society. Although that is a problem too. Instead, it is that markets generate the information that central planners would need to plan a prosperous economy. No supercomputer or AI could solve the problem of generating information in the forms of prices. As the economist, Michael Munger, explains,
Suppose that our society has perfect information about the amount, quality, and location of all the resources available in every barn, closet, nook, and cranny of the economy. All this information is stored in an infinitely fast computer, meaning that computations are all performed instantly, with zero lag. One thing we have to decide is how to produce a widget. There are two production processes, one that requires ten pounds of iron, and the other requires an ounce of gold. Which should we select? … The limiting factor is not calculating power, it is the capacity of a system to generate information
Why? Munger explains:
There’s the problem. Prices are not given; they are not inputs to allocation decision. Prices are the result of millions of people, all over the world, simultaneously choosing and reacting to the feedback that dynamically adjusting prices reveal about the consequences of choices by others […] Having the amounts of resources in a database tells you very little. Only the fact that an ounce of gold costs more, in terms of value foregone, answers the question. But that information is not calculated by markets, it is generated by markets
My misunderstanding of Hayek was that the issue with centralized planning wasn’t simply that the salient information is often, though not always, is widely distributed in a society, but also that market interactions between consumer and producers, buyers and sellers are what generate the price information that one would need to plan an economy. However, since one must rely on the market to generate the information needed to plan a prosperous economy, it would appear that the challenge of planning a prosperous society, without the aid of markets, is deeper than even I appreciated.
I used to vaguely think along the supercomputer line, thanks for this