"Arguing in My Spare Time," No. 5.06
by Arnold Kling
February 24, 2002
...Within 10 or 15 yearsí time, practically every computer and every handheld device will be online all the time. What many people donít realize, however, is that this visionary network is increasingly up and running today. And it doesnít even require any new technology, business models or significant investment...the real wireless Internet doesnít cost $50 a month--itís free. All thatís required, really, is openness.
--Simson Garfinkel, The Internet Amenity, Technology Review, March 2002
In the Internet era, many of the most bitter struggles over commerce and policy are being fought because of the potential to offer valuable goods and services for free.
The Microsoft anti-trust case centers around the decision by Microsoft to distribute the Internet Explorer browser at no charge.
Napster became the flashpoint in a war over music distribution by offering music-swapping service for free.
As Garfinkel, David Reed, and others point out, new protocols and receiver technologies make it possible to deliver high-speed wireless Internet service at essentially no marginal cost.
Beyond these immediate cases loom several other issues of products and services where research and development costs are high, but the marginal cost of the final product or service is low.
The marginal cost of manufacturing prescription drugs is low. Our traditional policy has been to allow drug companies to recoup their research and development costs by giving them patents that enable them to charge prices for drugs that are far above marginal cost. This policy has come under fire with the AIDS crisis in Africa and more recently with the Anthrax scare in the U.S.
The bioinformatics revolution promises to bring about medical treatments that have high benefits and extremely low marginal cost, but only after difficult and extensive research.
The cost of data storage, search, and retrieval keeps falling.
The cost of food production, which has been falling for centuries, could approach zero with bio-engineering of crops.
As the cost of providing and distributing various types of goods and services is falling, some other costs remain high. For example, Garfinkel writes,
One of the most surprising things we learned from launching our Internet startup was that providing wireless Internet service is really cheap. What ended up bankrupting the company were all the ancillary services we had to develop--credit card billing, technical support, the corporate Web site and the various security measures we had to put in place to prevent unauthorized use of the network by nonsubscribers.
Garfinkel is suggesting that it costs more to maintain an infrastructure that allows you to charge for the service than it does to provide the service itself. If that is the case, then what these asymptotically free goods and services represent is nothing less than a breakdown in the economic system.
I want to try to offer a precise definition of an asymptotically free good. I believe it differs from other concepts that have been used by economists.
A traditional economic good is one for which the marginal cost is above the average cost. This means that adding another consumer will raise the price required to cover the cost of providing the good. Another characteristic of traditional goods is I cannot consume the good without taking it away from someone else. If I eat a hamburger, then you cannot eat that same hamburger.
A natural monopoly is a good where the marginal cost is below the average cost. Natural monopolies typically involve heavy investments in capital equipment. This means high costs for ongoing maintenance in addition to expensive initial investment.For example, consider a ride at an amusement park. The marginal cost imposed by the next rider may be close to zero, even though the average cost is high. If the natural monopoly charges its marginal cost, it will lose money. To recover its average costs, the amusement park might charge an admissions fee. Once you are in the park, your marginal cost of a ride may be small, perhaps zero. (This example is due to Walter Oi, "A Disneyland Dilemma: Two-part tariffs in a Mickey Mouse Monopoly.")
Other examples of natural monopolies include utilities, such as electricity. Much of the cost of providing electricity is fixed costs associated with plant and equipment. The marginal cost tends to be much lower.
A public good is one that I can enjoy without paying for it. National defense is an example. In the absence of government provision of such goods, individuals will have an incentive to hold back and let others provide them, which means that collectively the amount provided will be too little.
Clean air is another example. When I pollute, I suffer only a negligible portion of the social damage that I cause. Thus, I face little disincentive to pollute. Every individual treats his or her own pollution as costless, even though it is costly in the aggregate.
An asymptotically free good is a good where almost of all of the cost involved consists of research and development. It differs from a natural monopoly in two ways.
In contrast with an amusement park or a utility, the cost of maintaining the capital for an asymptotically free good is relatively low. Once the research is complete and the idea is proven, the costs are trivial. In the absence of patent protection, there is nothing to stop a competititor from taking the idea and driving the price close to zero.
For example, once you have undertaken the research to produce a new miracle drug, the marginal and average costs of producing it are low. To take another example, once devices have been designed and protocols established for a high-speed wireless network, the cost of providing and maintaining the equipment for a network may be low relative to the number of users.
Asymptotically free goods are like public goods in that it is costly to exclude someone from enjoying the benefit of an asymptotically free good.
It is costly to hook someone up to the electric grid. It is costly to keep someone off a wireless network.
The cost of setting up and maintaining a gate at an amusement park is relatively low. The cost of policing the Internet to stop music swapping is enormous.
Liberals consider the role of government in an environment with asymptotically free goods and see opportunities.
In the absence of government involvement, one person's research can be appropriated by someone else. This would deny any reward to researchers. The only economic ways to encourage research and development are (a) direct government funding and (b) defining and enforcing patent law. The better job the government does with these functions, the better will be the outcomes.
It is not efficient for prices to be set far above marginal costs. When the marginal cost of drug therapy or wireless Internet service is set artificially high, there will be pressure on government to do something to correct the mispricing.
Conservatives consider the role of government in this environment and see threats.
Government is traditionally slow to recognize and shut down failures. The private sector is better at the sort of trial-and-error learning that is required for applied research.
Private-sector competition is what allows cost-saving innovations to deliver benefits to consumers.
Asymptotically free goods are disruptive. In a disruptive situation, government tends to become aligned with the opponents of change. For example, recent legislation concerning intellectual property has strongly favored traditional music and movie distributors at the expense of innovation and consumer benefit.
Asymptotically free goods are a new economic force. Problems are being solved not by throwing capital and labor at them, but by undertaking research and development which, when completed, leads to solutions that cost relatively little in terms of traditional factors of production.
I would not want to own stock in a company that generates its revenue from music distribution on CD's and tapes, from phone service, from selling information retrieval services, or from traditional health care services. For these companies, the present value of future earnings has to be calculated under the assumption that at some point they will be crushed by the steamroller of asymptotically free goods.
Asymptotically free goods also raise issues of public policy that may exacerbate the polarization between liberals and conservatives. For those who tend to view government as an instrument of the public good whenever the free-market outcome may be flawed, asymptotically free goods provide an excuse for more government intervention. For those who tend to see government as providing an instrument by which status quo interests can impede change, asymptotically free goods are a reason for keeping government hands off.