The Omniscient Planner · Nik Khosravipour

The world stands ready to move on from such crude things as markets and property. An omniscient overlord has arrived to prove that humanity can do better, and that an era of perfect efficiency is one decision away. He knows what we want, he knows what we have, and he knows how to combine all of this in perfect unison. He is all-knowing and all-powerful. He wants nothing but to make us happy and satisfied. Because he is benevolent, there are not the usual worries of corruption or wrong incentives. There is nobody to bribe, no official taking a cut, no office faking its numbers to please a superior. Every objection ever raised against the planners of the last century has been answered in a single stroke. No inefficient rent-seekers or profiteers. Only pure mechanical efficiency.

With full knowledge of every consumer's unspoken wish, every warehouse, every hectare, every skill and every technique, the only thing left is the final equation that solves economics forever. Why should we rely on the mercy of businessmen and entrepreneurs to produce what we want, if a mind like this can do it without taking something from the top? The old books said planning fails because no office can gather the knowledge. The office has been replaced. The knowledge is here, all of it, updated every second. Planning should now be trivial.

The first request lands on his desk within the hour. California wants a new rail line, and the governor wants it soon. Steel or aluminium?

The overlord begins the way an all-knowing being begins.

INSERT MATERIALS. Steel: tensile strength, corrosion rate under coastal air, every deposit from Minnesota to the Pilbara, every ton sitting in every yard, every furnace and its capacity. Aluminium: lighter by two thirds, five times the energy to smelt, every smelter, every bauxite mine, every order book on earth.

CONSIDER LABOR. Human crews or tunneling machines. Every welder in the state, every wage ever paid, every back that will ache.

CONSIDER ENERGY. Smelters demand power that could heat homes through the winter. Furnaces demand coal that could fire cement kilns instead.

CONSIDER TIME. Six years of construction that could be spent on hospitals, on housing, or on nothing at all, leaving the workers their evenings and their holidays.

The recitation is flawless. It is also useless. Every ton of steel is wanted in a thousand other places, and so is every hour of labor and every unit of power. Choosing the rail means giving up the hospital, the housing, the holidays. To choose, the overlord must know which path is cheapest, and cheapest turns out to be a strange word. Tons do not convert into hours. Hours do not convert into kilowatts. A holiday refuses comparison with a corrosion rate. Calculation is addition, and addition requires that every factor first arrive as a magnitude of the same kind. Prices are the only device that has ever performed that conversion. Inside a household the problem never appears, because a family ranks its wants directly and consumes what it ranks: the roof before the vacation, dinner before dessert. An economy produces for others, with factors that must be weighed against each other, and no list of any length supplies the unit to weigh them in. The overlord could survey every possible combination before breakfast. The ranking is what escapes him, and size has nothing to do with it.

He knows everything, and he holds no numbers. Where does a number like that come from?

Cut to a world without the overlord, to the barbaric spectacle of a free market. The same ton of steel sits in a yard in California, and three men want it. A rail contractor bids because he believes the coastal cities will keep growing and freight will follow. A hospital developer bids because an aging county has been promised a new hospital wing. A machine-tool maker bids highest of all: he is convinced that within five years every factory on the coast will want his automated presses, and the presses are built almost entirely out of steel. Each of them holds a different picture of a future nobody has seen. Each backs his picture with his own money, and whoever is wrong will carry the loss himself. The tool maker, as it happens, is wrong. His presses will find no buyers, the loss will be his, and next year his bids will be gone from the yard. The prices that survive are prices that have been filtered, every wrong forecast paying for its own removal.

The judgment these men perform has an old name: appraisement. It sits between knowing and computing. Knowing is the inventory, the corrosion rate, the order book. Computing is the arithmetic any clerk or engine performs once the numbers exist. Appraisement is the step that creates the numbers. A man reads everything he knows, forms his guess about tomorrow, and turns that guess into the one figure he is willing to stake his property on. The stake is what makes the figure honest. A bid backed by nothing is an opinion. A bid backed by a man's own capital is information the whole world can use.

Prices do carry information, and entrepreneurs need it. The information exists because a contest manufactured it. When the contractor, the developer and the tool maker bid against each other, their three private pictures of the future collapse into one public number: the price of steel. That number sat in nobody's head before the bidding, and it lives nowhere else afterwards. Knowledge alone, even complete knowledge, produces no prices. Knowledge must be scattered across many rival owners for a price to come into being at all, because a price is the result of settled disagreement.

The overlord holds every fact the three bidders hold, and more. What he lacks is the contest. He owns everything, so he bids against no one. One mind cannot hold an auction.

Through the night the same request stands unanswered. AWAITING INPUT: PRICE PER TON, STEEL. PRICE PER TON, ALUMINIUM. The fields stay empty. The only machinery that ever filled them has been abolished.

Grant the overlord one more gift and freeze the world. If wants, techniques and resources stood still forever, an economy already running could repeat yesterday's proportions, and repetition needs no calculation. But the rail line is new, and for a new choice the stillness supplies nothing: the missing item is the common denominator, and a frozen world holds no more bidders than a spinning one. Steel or aluminium stands exactly where it stood. The world spins anyway, and with it spin the wheels of production and consumption. Techniques shift, harvests fail, consumers change their minds by the season, and every shift reprices everything else. Every shift asks him the question again, and he could not answer it the first time.

Suppose the overlord had picked the steel and sent the crews out. By the second year a lab in China has perfected an alloy built for the strain of high-speed trains twice the old weight, and the population that was supposed to live along the route has drifted to the coast, where new building methods made land cheap. Six years is a long time in an economy. The line would be priced for a world that stopped existing before the first train ran.

Of course the overlord can forecast too. Give him the finest forecasts ever produced. However, a forecast turns into a price only when rival owners back competing forecasts with their own property, and his forecast turns into nothing. There is no one to compare his appraisals with. He does not profit and he does not lose, so no corrective ever tells him whether his allocation was wisdom or waste.

The obvious reply is to let reality grade him: build the line, count the passengers, adjust next time, like a weather model learning from every storm. But the reply fails on a detail. The steel he committed to the rail was steel the hospital and the presses never received, and those unbuilt alternatives produce no passengers to count and no revenue to miss. They leave no trace at all. Reality grades the road taken, and even that grade says only that the trains are full, never that they were worth their cost. Only a price, set before the choice, weighs every alternative he did not take, all at once.

He never chooses on empty ground either. The overlord wakes owning the remains of every past decision: a roadbed built toward towns the people have left, tunnels bored for trains half the new weight, yards of alloy nobody will ever use. Complete, scrap or convert? Each answer is a comparison of salvage against completion against some third use, and every term in it is a price he cannot summon. A market would long since have priced the used steel and the half-finished tunnels, bid on by men who pay for their own mistakes. He inherits all of the mistakes and none of the numbers.

The dream of replacing the contest is old, and its best minds have already confessed. In 1936 the economist Oskar Lange offered socialism its most famous defense: let the planners order their managers to play market, posting trial prices and adjusting them against shortages and surpluses, until the books balance the way real markets balance them. The profession declared the debate settled. Then Lange withdrew the offer, in an essay published in 1967, two years after his death: "The Computer and the Market." The market, he wrote, was a computer of the pre-electronic age, and the electronic computer would now perform its function even better. Nobody needed to play market anymore. It was a remarkable confession. The man who had sold the pseudo-market revealed that he had always seen the market as a slow calculator, a placeholder waiting for better hardware.

The one great experiment ran on the same quiet substitution. Soviet planners priced their capital goods by copying prices from the capitalist world market and translating them into rubles. Underneath the official plan, a black market of bribes and barter moved everything the plan itself could not deliver. The largest planned economy in history outsourced its numbers.

The hardware has improved ever since, exactly as Lange hoped. Differential equations, cybernetics, linear programming, big data, and now artificial intelligence: each generation announces that the machine is finally large enough. Each generation is right about the machine. A modern computer solves the planner's equations in moments, and every decade faster. But the equations themselves stand empty. Their coefficients are prices, and prices are what died with the market. A machine owns nothing and risks nothing, so it can simulate the bidding and never perform it. A simulated bid stakes no property, and a bid that stakes nothing reveals nothing. One can order a manager to play market. Nobody has ever ordered a man to stake his fortune, and meant it. The newest version of the dream splits the planner into a million artificial bidders, each with its own budget. Either those budgets are play money, and the bids reveal nothing, or the agents genuinely own their resources and carry their own losses, and the planner has rebuilt the market and lost the plan. Without market actors there is no input for any of this technology to use. The dream keeps buying bigger calculators, and the missing column stays missing.

On the overlord's desk the two reports still wait, steel and aluminium, complete to the last decimal and silent on the only question. The governor of California calls again. If the answer does not come by spring, he says, he will offer the project to the open market. He does not notice what he has just admitted. Whenever the planned world needs a number, it borrows one from the free world next door, the way the Soviets borrowed theirs from the world market they despised. His threat gives the whole game away: when the plan needs a real answer, it asks the market for one.

The new era was promised as an escape from the crude machinery of markets and property. That machinery turns out to be the thing doing the work. Free markets already are the technology the dreamers keep waiting for: a standing contest in which thousands of planners stake their own property on rival pictures of tomorrow, working to satisfy our wants before we know them ourselves, and every bid and every loss creates the numbers that make rational choice possible. Grant a planner every fact in the world and a machine that never errs, and he still stands where the overlord stands, waiting for a column that no amount of knowledge can fill. The numbers come from the contest. The contest is the market, and there has never been another factory for them.

This essay draws on Ludwig von Mises's "Economic Calculation in the Socialist Commonwealth" (1920), Joseph Salerno's postscript to the 1990 edition and his 1994 reply to Leland Yeager, and Murray Rothbard's "The End of Socialism and the Calculation Debate Revisited" (1991).