Arnold Kling     Essays | Short Book Reviews | Favorite Links | Internet Bubble Monitor | Home
May not be redistributed commercially without the author's permission.

Index

From Allocating Capital to Allocating Talent

Arnold Kling, "Arguing in My Spare Time", No. 12

April 24, 1998

As an economist, I believe that institutions evolve in order to solve economic problems. In an agricultural economy, land is the scarce resource. In an industrial economy, capital is a scarce resource. The institutions that evolved in an industrial economy differ from those in an agricultural economy. This raises the question: what is the economic problem today, and how might institutions evolve in order to address it?

Capitalism Without Capital

In 1994, it was difficult trying to explain the Internet to people in the mortgage business. I remember a mortgage banking company (one of the top ten) sending me a draft contract that read in part, "Whereas Arnold Kling is the owner of a service known as the Internet. . ."

One of the lines that I developed to try to explain the Internet was the following:

"With the interstate highway system, the concrete and steel provided most of the value, and the rules of the road were almost incidental. With the information superhighway, the reverse is true."

This statement is intended to make the point that the Internet cannot be described adequately as an economic phenomenon solely by referring to the physical capital used to provide it. Indeed, one could argue that in dollar terms the physical capital that provides the Internet backbone is relatively trivial.

Many observers have noted that in the "information economy" ideas have become relatively more important, and physical capital (buildings, machinery, etc.) has become somewhat less important. If this is true, then it would seem that institutions would be affected.

Economic theory as I learned it put a great deal of emphasis on capital. Capital accumulation affected the distribution of income and the rate of economic progress. Business firms were modeled as making decisions about what types of capital to deploy and how much capacity to build. Financial markets were viewed as mechanisms for shifting capital from one firm or industry to another.

As I understood it, capital helps to determine the size of firms. In an industry where a lot of capital is required (electric power generation or automobile manufacturing, for example), firms will be large. In an industry where less capital is required (the legal profession or a restaurant, for example), firms will be smaller.

In a previous essay, "The New Industrial State and the Rise of the Dilbert Sector," I have suggested that computer systems developed between the 1960's and the 1990's were like large capital assets. They took years to plan and to build, and once built they were meant to be used for a long time. During this period, the development of computer systems required service firms such as banks to adopt some of the large scale and bureaucracy needed in the heavy industrial sector.

Bureaucratic institutions are functional in allocating capital because the best decision often is to postpone capital investment. As several economists have written, there is an "option value" in putting off the investment decision.

Bureaucratic delay is dysfunctional in software development. Software needs to be developed quickly. More importantly, it needs to be discarded quickly. Many experienced programmers believe that the best time to re-write software is just after it is finished--while the problem is still fresh in the developer's mind. In a bureaucratic environment, it is unthinkable to have a software development lifecycle that calls for immediate and continuous rewriting.

What is Today's Scarce Resource?

In today's economy, capital may no longer be the primary scarce resource to be allocated. What has supplanted it?

I shrink from saying that "information" now replaces capital. There are daunting issues of measurement, particularly if one tries to distinguish quality from quantity. Moreover, information does not strike me as something to be allocated, in the sense that it needs to be taken away from someone to give to someone else.

To me, it is easier to understand the economic challenge today as one of allocating talent to solving problems. Furthermore, when we consider the nature of software as a quasi-public good, the problem is one of allocating talent to producing quasi-public goods.

If the central economic problem today is one of allocating talent to producing quasi-public goods, how will this problem be solved? What sorts of institutions will evolve?

The problem of allocating talent is more complex than the traditional job market. In a classic job market, there are slots (job descriptions, such as "junior accountant" or "level II programmer" or "drill press operator") in which to place people. The function of the job market is to slide people with the right skill set into the right slots.

For jobs to be defined, the work must be predictable. To reach that point, the interesting problems must already be solved. Only after the complex problems were solved could the jobs be defined. Therefore, if the economic issue is one of allocating talent to solve problems, a job market will not suffice.

Talent allocation can take place top-down. For example, the leaders of Microsoft decided that as soon as Windows 95 was out the door the next focus needed to be the Internet. They re-allocated talent toward that focus.

The optimum size for a firm may depend on how it solves the talent allocation problem. If senior executives can allocate talent wisely over a wide set of problems, then the firm can be large. If senior executives only can allocate talent wisely over a narrow set of problems, then the firm will need to be small.

Venture capitalists can allocate talent. Today, a company may require funding more for salaries than anything else. In choosing which business plans to support, a venture capital firm is allocating talent to those business plans.

Another complicating factor in the talent-allocation problem when there is a need for software is the fact that software is a quasi-public good. There is no natural pricing mechanism, as was pointed out in the 9th essay in this series. In particular, charging a cost per license is artificial.

Eric Raymond, in "The Cathedral and the Bazaar," http://sagan.earthspace.net/~esr/writings/cathedral-bazaar/cathedral-bazaar.html, argues that software developers can organize themselves in order to solve problems. The examples of Linux and Apache appear to illustrate this. Recently, Web Review (www.webreview.com) ran an issue devoted to the "open source" model of software development, in which the source code is made available to anyone who wishes to try to enhance the product.

The challenge with the open source code model of software development is that when the source code is public, charging for licenses becomes meaningless. New institutional arrangements become necessary in order to pay software developers. Partisans are trying to sort out the distinction, if any, between open-source software and free software.

I continue to believe that software is more naturally a quasi-public good than a free good. That is why I believe that institutions that I call "software clubs" are a natural evolution. (See the previous two essays in this series.) The club can obtain revenue from membership dues, and this revenue can be used to pay software developers and designers/referees who work in an open-source environment. This may not be the optimal institutional arrangement, but it at least provides a mechanism for getting money to flow from users to developers without requiring per-license pricing.

As a software development model, open source code modified by individual developers overseen by "referees" may be very powerful. However, it needs an economic model to go along with it. Pure altruism will not last long as an energy source for productive activity. Any institutional solution to the central economic issue of allocating talent must ensure that talented people are paid to work on important problems.