≡ Menu

A Note on the Knowledge Problem

The excellent political philosopher Eric Mack e-mailed me recently with a few questions about Hayek’s knowledge problem. Here (slightly modified) is my response to Eric:


Your assessment that Hayek’s criticisms of socialism are part of his larger, deep understanding of the complex nature of social order is spot-on correct. I recommend, if you haven’t read it, Hayek’s 1964 New Individualist Review paper, “Kinds of Order in Society.” It’s a gem that should receive more attention than it does.

Also correct is your understanding that Hayek’s objections to state-directed allocations of resources do not become relevant only when the state practices full-on socialism. The knowledge problem affects all attempts to allocate resources in opposition to how those resources would be directed by following signals sent by competitively set market prices.

Of course, as you recognize, the amount of damage done to the economy is less the more limited are the attempts to override the price system. But “less damage” isn’t no damage. The (say) industrial-policy mandarin who is instructed (say) to create within the country more capacity to produce semiconductors has no idea just what particular goods and services his or her interventions will cause to not be produced.

More specifically to Mises’s and Hayek’s core economic objection to socialism: If non-market direction of resources is sufficiently limited (with market direction being dominant), Mises’s and Hayek’s core objection to socialism doesn’t apply. Inputs are still priced in competitive markets and, thus, each of these prices reflects the opportunity cost of each input. The industrial-policy mandarin charged only with expanding domestic capacity to produce semiconductors can, unlike under more extensive socialism, look to market prices for information on what is the least costly way of carrying out his task. (Public-choice considerations make the mandarin unlikely in reality to follow these prices, but that’s a different criticism. A second different criticism of industrial policy is that it necessarily stifles innovation, as creative ideas for reallocating resources away from the channels into which resources are directed by the mandarin must be squelched or at least ignored.)

BUT a knowledge problem remains even when non-market allocations of resources aren’t widespread. To wit, this mandarin cannot possibly know what he presumes to know, namely, that the economic value of the extra semiconductor-production capacity that he brings about – even when he does so at the lowest possible cost – will prove to be greater than that cost (that is, greater than economic value of the goods and services that his interventions ensure will not be produced). In fact, the mandarin’s presumption is almost surely mistaken. If the mandarin’s resource-allocation decisions are ones that will bring about a net increase in the market value of the economy’s output, private entrepreneurs would almost surely have already done what the mandarin is doing.

This particular knowledge problem doesn’t disappear if the mandarin is charged with overriding the market’s allocation of resources in an attempt to supply a “public good” – that is, a good (or service) that private-sector entrepreneurs cannot be relied upon to supply. From whence comes the knowledge that the value of the public good is greater than is the value of the outputs that production of the public good makes impossible?

As long as inputs are privately owned and exchanged – and, hence, priced – mostly in private markets, finding the least costly ways to produce some particular interventionist economic outcome is in principle doable. Note, though, that the meaning of “least costly ways” subtly changes as mandarins replace markets. Input prices reflect the costs to their owners of being used in particular ways. As mandarins replace markets, these prices will more and more reflect the costs to their owners of being used in the mandarins’ projects (as opposed to in projects that successfully are tested in markets). Because the economic value of the mandarins’ outputs cannot legitimately be presumed to be greater than the value of the foregone outputs, the prices of inputs – as mandarins increasingly replace markets – more and more reflect private costs that are disconnected to the ‘true’ costs of those inputs (that is, to the values of the outputs that those inputs would have produced in an unfettered market).

Note that economic calculation even by private actors is distorted by these not-very-informative input prices.

As for particular industry studies along these lines, none occur to me off the top of my head. I’ll ponder. But let me also recommend the late Don Lavoie’s brilliant 1985 book National Economic Planning: What Is Left? It is, I think, the most important economics tract written on this topic.

Hoping the above is at least somewhat helpful.