Traffic jam - problem with involving the federal government in the technological and logistical development of information technology and computer networks

National Review, August 15, 1994 by William Letwin

When Vice President Gore unveiled his scheme for an Information Superhighway, to be built with aid from the Federal Government, applause broke out all over. Everybody likes information and superhighways, so what could be better than a combination of the two? Since its birth in January, Gore's scheme, charmingly described as a "new policy vision," has attracted less criticism than any of the Administration's big initiatives, and much less than it deserves.

For one thing, the scheme silently and wrongly assumes that the U.S. as yet lacks an information superhighway. In fact the country is already covered by a network of information superhighways that carry great volumes of sound, data, and images from anywhere you can mention to anywhere else. The core of that network is made of optical fiber (which transmits information by pulses of light), augmented by co-axial cable, radio links, and satellite links. It belongs to private companies, which built it without government subsidies. So two questions spring to mind: Why should the existing information superhighways be upgraded to super-superhighways? And why, if upgrading is desirable, should the government invest taxpayers' money in it?

The argument in favor of upgrading is that the new network will be able to move information faster and more widely. Moving information faster presents no technological problem. Optical fiber is like a pipeline: the broader it is, the faster stuff will flow through it. But the problem with a speedy information superhighway is just like the problem with a real superhighway: your car belts along the beltway until it hits the inevitable traffic jam on the side streets.

In the information network, the crowded side streets are the so-called local loops, the pairs of thin copper wires that connect the phones in most homes and workplaces to the telecom superhighways. Local loops can carry voice and digital data but cannot carry information that comes in larger lumps, such as television or high-resolution images. The obvious way to cure that difficulty is to replace the copper-wire pairs with optical fiber. That is quite a project for a country with a hundred million or so residential units and another hundred million offices, factories, and places where people work or gather. At a conservative estimate, laying fiber to all those places would cost about $50 billion. The less obvious cure is to attach a clever computer to each copper-wire pair; this solution is in the experimental phase and nobody yet knows whether it will cost any less than installing cable to the house. Be that as it may, widening the main pipelines will be largely wasted if in the end the flow is blocked by tiny funnels.

Now, assuming as I do that suppliers of information services will make the necessary investment if they believe that users will pay the cost, why should the government invest large sums to subsidize research and development? The clear answer is that it should not. When a government provides R&D grants, it cannot avoid choosing among competing applications. But there is no reason to suppose that government officials are especially skilled at "picking winners," and there is solid evidence that they often pick big losers. The British government's enormous investment in the supersonic airliner, Concorde, is one example; the Japanese government's investment in developing bigger mainframe computers just when they were being driven into extinction by minicomputers and personal computers is another [see "Let Freedom Ring," p.45]. In the absence of government intervention by selective funding, competition among companies filters out the less satisfactory solutions, and the cost of failure does not fall massively on taxpayers. The sweetener of public investment in the Gore package should be rejected as worse than useless.

Gore's superhighway would be able to deliver to every American something like a million times more information than she or he gets now. That would certainly solve the Freedom of Information problem. Every evening every citizen could get a copy of every piece of paper generated that day in every government office. In a flood like that, people would drown. Being informed is a matter not of having loads of information but of getting just the relatively few bits one wants; quality rather than quantity is what counts. Even with the data banks that are easily accessible today, the art of constructing them and using them is the art of selection, or, as the experts say, of "retrieval." Roughly speaking, the more information you get, the harder it is to use. Besides, a lot of what passes for information is either misinformation or disinformation. So Gore's package should carry a surgeon general's warning: "Excessive information can injure your brain."

But, of course, Gore's package is not really about information. It is really intended to usher in the brave new age of "interactive multimedia services." Described pedantically, a multimedia facility can transmit sound or pictures or text, or any combination of them, down the same line into a television receiver or computer. Described poetically, multimedia will mean video-phones, shopping by phone (with the TV screen showing you models wearing the clothes or sitting on the chairs you're thinking of buying), and any number of similar marvels. Of course multimedia is really a new name for products as old and common as talking pictures (formerly "talkies," now "movies") or television.