Arnold Kling, "Arguing in My Spare Time", No. 3

"So complex, indeed, will be the job of organizing specialists that there will be specialists in organization. More even than machinery, massive and complex business organizations are the tangible manifestation of advanced technology."

--John Kenneth Galbraith, "The New Industrial State," p. 16.

In large corporations, computer systems are developed in an environment whose absurdity is captured in the famous "Dilbert" comic strip. In later essays, I will want to argue that the Dilbert-style bureaucracy may disappear. To make this argument, one has to suggest that some factor has changed that affects the economic viability of bureaucracy.

The purpose of this essay is to try to explain how the Dilbert Sector arose in the first place. To understand the role of bureaucracy in large organizations, we turn to Galbraith, who focused on large corporations during the post-WWII period when they seemed most dominant in our society.

John Kenneth Galbraith began writing his most famous books, "The Affluent Society" and "The New Industrial State," in the late 1950's. Its Cold-War context did much to shape his work.

The Cold War pitted as adversaries the believers in two competing ideologies, one represented by the American right wing and the other by Soviet Communism. Each protagonist was convinced that the other was morally bankrupt as well as on the wrong side of history.

During this period, there were intellectuals, including Galbraith, who sought to stand above this ideological fray. Their point of view could be represented by a joke which was popular at that time: "The difference between Capitalism and Communism is that in Capitalism man exploits man, whereas in Communism it's the other way around."

Those who viewed the Cold War conflict as over-stated usually would describe the Soviet system as a "planned economy." They tried to reduce the apparent ideological differences between Communism and the West to a difference of economic organization between planned economies in the former and market economies in the latter.

Galbraith carried this difference-minimizing analysis one step further: he argued that the modern American economy is as much a planned economy as the Soviet economy. In what he called "the new industrial state," the planning that takes place within the largest corporations serves the same function as central planning in a Communist system. To quote Galbraith directly,

"The enemies of the market [are] not socialists. It is advanced technology and the specialization of men and process that this requires and the resulting commitment of time and capital. . .planning is essential. The modern large corporation and the modern apparatus of socialist planning are variant accomodations to the same need." (ibid, p. 33)

Galbraith argued that traditional economics applies only to an ever-diminishing share of the economy characterized by small firms producing products that can be developed using simple processes requiring little lead time. He expected that as this textbook small-firm sector continued to shrink in actual economic significance, economists would have to study the corporate sector as a form of planned economy. His expectation was that classical economics would come to be seen as anchronistic, persisting only as a historical intellectual curiosity.

As it turned out, Galbraith failed in his projects to equate Capitalism with Communism and to overthrow conventional economics. It is Galbraith who today is little more than an intellectual curiosity. Still, there are valuable insights embedded in his writing.

The insight that matters to us here is the relationship between technology and bureaucracy. Galbraith characterized the complexity of modern business processes by describing what he called the "imperatives" of technology. These included (all pages refer to "The New Industrial State"):

"An increasing span of time separates the beginning from the completion of any task." (p. 13)

"There is an increase in the capital that is committed to production."

(p. 14)

"the commitment of time and money tends to be made ever more inflexibly" (p.14)

"Technology requires specialized manpower."

"coordination will be a major task" (p. 16)

". . .the necessity for planning. Tasks must be performed so that they are right not for the present but for that time in the future when, companion and related work having also been done, the whole job is completed. And the amount of capital, that meanwhile, will have been committed adds urgency to the need to be right. So conditions at the time of completion of the whole task must be foreseen as must developments along the way." (p. 16)

For Galbraith, the prototypical economic outputs are the complex products of heavy industry. Examples he gives are "jet aircraft, nuclear power plants or even the modern automobile" (p, 34). His thesis is that large firms with bureaucratic organizations are required to produce these goods. The number of components that must fit together, the number of requirements that must be met in order to satisfy consumers and regulators, and the large commitment of resources needed to produce even a single unit all suggest the need for extensive planning and management.

Yet we observe large bureaucratic organizations also in the non-industrial sector. In banking, insurance, retailing, and other services, the same large planning and management staffs emerge. What do these businesses have in common with Galbraith's industrial producers?

The suggestion being made here is that since the late 1960's, the element that has produced a similarity in size and organizational structure between service businesses and heavy industry is the need for computer systems. For a bank, a computer system is the counterpart of a jet aircraft, a nuclear power plant, or a modern automobile.

Like a jet aircraft, a computer system has many components that must be integrated, including hardware, communications, user interfaces, data storage, and processing. Like a nuclear power plant, a computer system's design is costly to change once construction is underway. And as with automobiles, uncertainty about what users may want creates risks and challenges for designers.

For the past quarter century, computer systems have required planning and management along the lines of heavy industrial machinery. Studies reportedly show that a large percentage of computer systems efforts were abandoned prior to implementation, at great expense to the companies that attempted development. The high failure rate is indicative of the management challenges that these systems impose on organizations.

If one could take away the need for computer systems, service industries would be much simpler. They would not require extensive planning, management, and bureaucracy to design and implement new products, develop innovative work processes, or form new partnerships.

Of course, the benefits of computer systems appear to exceed their costs. Companies that have been successful at implementing computer systems have tended to survive. The list of banks, insurance companies, etc. without major computer systems is rather short.

We may summarize as follows:

1. To deliver the best services, businesses require computer systems.

2. Computer systems, as they have been developed for the past 25 years, pose complex challenges of integration, planning, and management.

3. In order to address the imperatives of planning and management in businesses dependent on computer systems, service industries have adopted the organizational bureaucracy that Galbraith argues is needed for planning in the heavy industrial sector.

This summary explains the rise of the "Dilbert sector." Service industry giants know only one way to deal with the need for computer systems: enlarge the bureacracy, and add more layers of management. This places the software engineers who ultimately must deliver computer systems in an ever-more remote and difficult position. The plight of the software engineer surrounded by ineffectual management is perhaps the main theme of the "Dilbert" comic strip.

Computer systems function as capital assets in the Dilbert sector. In future essays, it will be argued that the way computer systems are assembled will change.

Today, we have a situation in which computer hardware is flexible and inexpensive. It is relatively simple to string together processors, storage devices, printers, and so forth. However, computer software continues to be expensive and inflexible. Once software architecture conforms to hardware architecture, the computer systems that support business applications will bear less resemblance to large capital equipment. This in turn has implications for the size and structure of firms.