Can Artificial Intelligence Fix The Fatal Flaws Of Central Planning?

Or do computers face the same problem as man

Long before communism’s dissidents revealed its trampled rights, mass killings, and recurring famines to the world, economist and philosopher Friedrich Hayek had explained the fatal critique now known as the “calculation problem.” His pithy, 2,500-word critique showed that to be successful, central planners require complete information on every individual’s preferences and capabilities, along with an efficient way to respond to new information. These twin requirements have been dubbed the “calculation problem” and while they are easy to understand intuitively, they are impossible to achieve in practice.

“The fool who sold wisdom,” from A Hundred Fables of La Fontaine, illustrated by Percy-Billinghurst, published by John Lane, the Bodley Head, 1900. TheCMN/Flickr.

An apparatchik in a command economy responsible for wheat production, for example, needs to know proper responses to a drought that decreases the expected harvest. On the demand side, he must also know the proper response to an immigration boom that causes unexpectedly high demand. To determine the correct responses for the demand and supply shocks, the central planner would require omniscience. The vast amount of society’s dispersed knowledge about innumerable numbers of goods and services would need to be centralized into one database. The society’s wants, abilities, and efforts would need to be uploaded into a single brain which would then need to solve for the optimum resource allocation. Since this is beyond mortal reach, the demand and supply curves are never at equilibrium in a command economy, leading to perpetual surplus or shortage. Gosplan plans; Mammon laughs.

In contrast, a market system has exponentially shorter latency when responding to market shocks. If the price of wheat rises to the point at which another grain is more economical, then the individual will substitute. If there is no substitute, producers will pay the higher wheat price and pass that cost onto consumers or accept lower profit. If the price is too high for consumers and too low for the producer to turn a profit, the producer will shut down production. Prices are an extraordinarily compact datum that transmit the only relevant information. A market economy compresses the disaggregated knowledge of all individuals to produce an executive summary in the form of one number.

“The miser and the monkey,” from A Hundred Fables of La Fontaine, illustrated by Percy-Billinghurst, published by John Lane, the Bodley Head, 1900. TheCMN/Flickr.

Prices are the practical, quantifiable output when theoretical, amorphous supply and demand curves intersect. It is the unidirectional transmutation of curves into a point, the dispersed into the concentrated, and the unknown into the known. If we set the price by diktat, the supply and demand curves are impossible to reverse engineer. To do so would require knowing the inner workings, thoughts, and desires of every single economic actor. Central planners have tried to achieve this feat to no avail. In practical terms, prices set an exchange rate between every single good and service in an economy, across time. A market based price system allows you to see how many apples a computer is worth, whereas a central planner would just have to guess. This kind of guessing can leave you open to ridiculous decisions like paying poets by the line.

A.I. Enters the Scene

Every advancement in computing power and artificial intelligence (A.I.) reinvigorates the case and—for some—the hope that a centrally planned economy is finally within grasp. As early as 2016, Alibaba boss Jack Ma said:

Over the past 100 years, we have come to believe that the market economy is the best system, but in my opinion, there will be a significant change in the next three decades, and the planned economy will become increasingly big. Why? Because with access to all kinds of data, we may be able to find the invisible hand of the market.

Large corporations such as Google, Amazon, and Walmart, already gather huge amounts of customer data, then use that data to predict consumer behavior, tailor individual consumer experiences, and suggest products to purchase. The big tech corporation’s business model, coupled with computing power’s ever-increasing caliber gives policy analysts a pollyannaish, misguided confidence in central planning.

“The oyster and the litigants,” from A Hundred Fables of La Fontaine, illustrated by Percy-Billinghurst, published by John Lane, the Bodley Head, 1900. TheCMN/Flickr.

Central planning enthusiasts are right in asserting that the calculation problem’s complexity is easier to handle with plentiful computing power. But they miss the fact that that complexity is itself a function of abundant computing power. If companies and individuals start using A.I. to make decisions, the A.I. Overlord responsible for the whole economy would necessarily have to account for all the other A.I.s’ decision making calculations. The central planner A.I. would have to stand one intellectual rung above all the other A.I.s. But no A.I. can be so far advanced relative to the others. Just as no human being is so far advanced compared to other human beings that he could become a central planner, no A.I. is so far advanced the others to be an A.I. Overlord. In short, as A.I.s become more powerful the economy also grows more complex.

The challenge of perfectly organizing an economy does not become any less intractable with greater computing power precisely because greater power also makes the economy more complicated. In other words, the calculation problem’s complexity is both easier to handle and grows more complex due to abundant computing power. A.I. is not the solution to the calculation problem and we should continue to repose our faith in the bedrocks of decentralization and price driven revelation. For the unquantified element in the economy is like that of nature, which each fresh advancement in science merely reduces but does not abolish.

Akshat Patel is a former submarine warfare officer in the U.S. Navy, a graduate of the Warton School of Business, and an alumnus of the Common Sense Society Carolina Fellowship, class of 2021.