CoCom Kept Western Technology from Moscow for Forty Years, a Model for China

In the summer of 1987, a group of members of Congress gathered on the Capitol grounds, picked up sledgehammers, and beat a Toshiba radio to pieces for the cameras. It was theater, and it was also the best-publicized moment in the history of an obscure committee that most Americans have never heard of. The radio died for the sins of a different Toshiba product altogether, one that had helped the Soviet Navy get quieter. How a machine tool ended up as a national argument is the story of CoCom, and it is a story worth telling now, because the same question is back: can the free world keep a hostile power from buying the technology it needs to threaten us?

The claim I want to defend is simple. Export controls work, imperfectly and at a price, when the allies who make the technology agree to hold the line together and have the will to punish cheaters. CoCom proves it. It also shows exactly how such a system fails, which makes it a better guide to the China problem than any slogan.

Let me define terms. Export controls are legal limits on selling certain goods, software or know-how to certain countries. Dual-use means a technology with civilian and military applications, like a computer that can run a payroll or a missile guidance system. Diversion is the practice of buying a controlled item through an intermediary in a third country, so the final buyer stays hidden.

Why 1949, and why Paris

The Cold War had barely begun when the West faced a puzzle. America had emerged from World War II as the world's leading industrial power. Its allies in Europe and later in Asia were rebuilding, and they wanted customers. The Soviet Union, which had just taken control of Eastern Europe and would soon test an atomic bomb, wanted Western machines. If the United States alone refused to sell, the Soviets would buy the same goods from France or Britain or Japan, and America would lose both the sale and the argument.

So the answer had to be collective. In 1949, the United States and its principal allies set up the Coordinating Committee for Multilateral Export Controls, soon known by its acronym, CoCom. It was not a treaty. It was an informal body that met in Paris and had no formal charter, no enforcement powers, and a small staff. Its members were the NATO countries other than Iceland, and Japan joined in the early 1950s. Australia came in later, in 1989. It kept three lists of controlled items: one for munitions, one for atomic energy goods, and a large industrial list for dual-use technology such as advanced computers, precision machine tools and electronics.

The core rule was unanimity. Every member held a veto, so any member could block an exception that another wanted to grant. That made the system slow, and it also made it durable. A rule everyone had signed on to could not be undone by a single government under commercial pressure without a fight in the open.

Congress supplied the American side of the machinery. The Export Control Act of 1949 gave the executive power to license exports for national security, and later laws refined it. The pattern was consistent: the committee set the lists, each member enforced them through its own laws, and the American Congress and the Pentagon pressed the allies to keep their enforcement honest.

What CoCom was for

The goal was never to starve the Soviet economy. That would have been impossible, and no serious official claimed it. The goal was narrower and more realistic: to deny the Soviet bloc the specific technologies that would improve its military, and to make everything else they wanted cost more and arrive later.

It helps to look at the Soviet side. The Soviet Union ran a system of technology acquisition that the West came to know in detail. A branch of the KGB, called Line X, worked through spies and front companies to obtain Western designs. Soviet engineers copied what they found. The Ryad series of Soviet mainframe computers, for example, was modeled on the IBM System/360. Copying has a price, because a copy is always behind the original, and CoCom's work was to widen that gap. Every year a Soviet design lagged by was a year in which Western forces held an edge in the air, at sea and on the battlefield.

Economists will tell you that is a fair way to judge export controls. They do not have to stop everything. They have to impose costs and delays that exceed the cost of the controls themselves. By that standard, the record is better than the skeptics admit.

The cheating, and the catch

It would be dishonest to claim the system ran smoothly. Allies cheated, and the best-known case is the one with the sledgehammers.

In the early 1980s, a Japanese firm, Toshiba Machine, sold the Soviet Union advanced computer-controlled milling machines between 1982 and 1984, and a Norwegian firm, Kongsberg Vaapenfabrikk, supplied the numerical control systems that ran them. The machines could shape propeller blades with great precision. The Soviet Navy's submarine propellers had been noisy, and noise is what sonar listens for. Better machining gave Soviet boats a quieter propeller, and a quieter propeller made them harder to find. The sales broke CoCom rules, and by the accounts later published, the companies had worked around the restrictions through paperwork that hid what they were doing.

The affair became public in 1987. American officials said the sales had damaged the West's anti-submarine advantage. Toshiba's top executives resigned. Congress moved to punish the company, and the United States imposed restrictions on Toshiba Machine. Japan and Norway acted against their own firms too: a Tokyo court fined Toshiba Machine and gave two of its executives suspended prison terms in 1988, and Norway sanctioned Kongsberg.

This is the central point about CoCom that critics skip. The scandal is usually told as a failure, the proof that controls leak. I read it the other way. It was a case in which a violation was detected, attributed and punished, which is precisely the cycle a working system must have. A rule that is never broken has never been tested. A rule whose breakers pay a price is a rule.

There were other cases. Smuggling through neutral and third countries, such as Sweden and Switzerland, which were not CoCom members, was a constant worry, and the allies spent years closing those gaps with bilateral understandings. Customs officers on both sides of the Atlantic investigated diversion networks. Some succeeded and some did not. Nobody claims it was perfect. The claim is that it was worth doing.

The skeptic's best case

A fair argument against controls, stated at its strongest, goes like this.

One, controls hurt our own companies. American firms lose sales, foreign competitors fill the gap, and the lost revenue funds their research.

Two, technology leaks anyway. Spies, copying and reverse engineering mean the other side gets the technology eventually. The Soviets did copy the IBM mainframe.

Three, controls push the target to build its own industry. A country locked out of imports develops domestic capability, so the restriction makes the rival more self-reliant in the long run.

Four, allies will not cooperate. The Toshiba case shows that even close friends cheat when the money is good.

Those are real points, and they come from people who understand the trade-offs.

The answer

One: the cost to our firms is real and should be weighed. It is a reason to make the lists narrow, targeted and reviewed, not a reason to drop them. CoCom itself ran a continuous argument about what belonged on the lists and what to take off, and the best version of the system trimmed items that had become common. Controls on a technology that is already everywhere are pointless. Controls on the leading edge are not.

Two: leakage is a reason for delay, not for surrender. If a copy arrives five years late, the five years are the prize. The Soviet mainframes that copied IBM's design ran a generation behind. And a copy tells you something about the copier: it can follow, but it cannot lead. What it cannot do is leapfrog, and leapfrogging is what controls are built to prevent.

Three: the self-reliance argument is the most interesting, and the record cuts both ways. A closed economy that must build everything itself pays heavily for it. The Soviet system's weakness was not simply that it lacked Western machines. It lacked the open competition, free flow of ideas and consumer pressure that make an economy innovative. Isolation made that weakness worse. But I grant the point has force for a modern China, which has a much larger economy, huge domestic investment and a serious engineering base. A control strategy that ignores that would be naive, which is why the lessons below matter.

Four: allies cooperate when the cost of cheating is visible. After the Toshiba episode, the cheaters faced penalties at home and abroad. The fact that cheating was exposed, that Congress and the press paid attention and that the cheaters paid a price is what made the system hold. This, to me, is the lesson that a free society can govern itself even in this: sunlight and consequences.

When the system ended

CoCom formally ended on March 31, 1994. The Soviet Union was gone, and many in Washington and Europe believed the reason for the committee had gone with it. In 1996 the Wassenaar Arrangement was set up as a successor, but it works differently: it is a voluntary regime for exchanging information, it has no veto, and its members decide for themselves what they will license. It lists what members have agreed to watch, and it does not stop anyone from selling.

There is an irony in that. The device built to contain an enemy was dismantled just before a new competitor started to grow. In the 1990s, many people believed trade would turn China into a liberal society. That is a hopeful theory, and a reasonable one to have held at the time. It did not work out that way, and the wider habit of that decade, harvesting a peace dividend before the peace was secure, is examined in Peace Dividends Always Come Due, the 1990s Drawdown Reconsidered.

Lessons for China

What does the CoCom record teach about technology controls today? Five lessons follow from the history.

First, you need the allies. The key technologies are no longer made only in America. They depend on firms in Japan, the Netherlands, South Korea, Taiwan and elsewhere. A unilateral American rule only works if others do not fill the gap. CoCom's central insight was that unanimity among suppliers beats a single country's ban.

Second, keep the lists narrow and refresh them. Control the real chokepoints and let the rest trade. A list that tries to cover everything protects nothing.

Third, enforcement matters more than the list. The Toshiba episode proved that a rule is only as good as the penalty for breaking it. That means customs investigators, prosecutors, and the will to name companies and punish them.

Fourth, expect diversion. Controls on China will be tested by transshipment through third countries, shell companies and front businesses, just as the Soviets used neutral states. Plan for it, close the routes and watch the ports.

Fifth, do not mistake a control for a strategy. CoCom worked because it was one part of a larger effort that included NATO, nuclear deterrence and a growing American economy. It worked alongside the policy described in NSC-68 and the 1950 Memo That Told America the Price of Staying Free, not instead of it. Taking technology away from a rival works best when you are also out-building him.

The security side of the story ties together here. The theft of Americans' personal files, covered in The OPM Breach Gave a Rival Government the Files on Millions of Americans, shows that a rival does not only buy technology. It also steals it, and a control regime that ignores espionage leaves a door open. The industrial question is also central, since a country that cannot build its own ships and chips has little to deny anyone. The Defense Production Act of 1950 Is a Wartime Tool with Peacetime Habits and The Shipbuilding Gap with China Is an Industrial Story Before a Navy Story cover the capacity side. And there is the argument about whether a rising power must be accommodated, which The Thucydides Trap Is a Bad Excuse for Giving Ground to Beijing answers directly. Finally, Textbook Cold War Tales the Soviet Archives Quietly Overturned is a reminder that the record of the Cold War deserves to be read in the original, not in the textbook summary.

One wry observation. Every generation of free traders discovers, with some surprise, that the other side does not read the same economics textbooks. The theory that commerce makes friends is a sound one among free peoples. It has a poorer record when the customer is a regime that treats its purchases as military procurement and its own citizens as a resource. It is worth remembering when someone tells you that selling advanced equipment to a rival is just business.

The case in brief

1. CoCom was an informal allied committee, founded in 1949 and dissolved on March 31, 1994, that denied the Soviet bloc Western dual-use technology through a unanimity rule.
2. Its goal was realistic: delay and raise the cost of Soviet military modernization, not starve the Soviet economy.
3. The 1987 Toshiba-Kongsberg propeller scandal showed allies cheating, and also showed a system detecting and punishing the cheating.
4. The skeptics' points about lost sales, leakage and self-reliance deserve weight, and they argue for narrow, refreshed lists and not for abandoning controls.
5. For China, the lessons are allied unity, narrow lists, serious enforcement, vigilance on diversion and a wider strategy of building and deterrence.

What you can do: look up which export control rules apply to advanced chips and equipment through the Commerce Department's published rules in the Federal Register, then ask your representative where they stand on enforcing them.