Peter Blau, "A Formal Theory of Differentiation in Organizations," 1970

This article was written in the halcyon days when Peter Blau taught here at Chicago. The basic argument of the article can be summed up in a single sentence: organizational size (in absolute number of employees) affects organizational structure in significant, measurable, predictable ways. Blau goes step by step through this premise, showing you that he is right. (Talk about proving the obvious. Nevertheless, he makes some interesting points.)

First, note that Blau is only concerned with companies with paid employees. He is not making statements about voluntary associations, community structure, or any other form of human association. We are talking people with "bosses" who get cash payment for a certain amount and type of labor.

In classical prolix style, Blau states his objective thus: "The endeavor of this paper is to construct a systematic theory of differentiation in organizations consisting of two basic generalizations and nine [total] propositions derivable from them, which can acount for a considerable number of empirical findings." This paper is not about the environment, nor is it about transaction costs; it is an introspective paper, examining orgs as separable, unique, bounded events that are made up of countable bodies at different levels in the organizational structure.

Generalization 1: increasing org size generates differentiation along various lines at decelerating rates. rephrased, as orgs grow, they differentiate quickly at first, and then more slowly. The term "differentiation" here means breaking the org up into separate divisions or subunits (marketing, R&D, etc.).

Prop 1.1: as the size increases, its marginal influence on differentiation decreases. Blau admits that this is "merely a restatement of one part of the original generalization." For example, in a factory, production can be increased by adding workers, but the marginal increment of output resulting from adding more and more workers without changing plant size and equipment eventually declines (obviously).

Prop. 1.2: the larger the org is, the larger the average size of its structural components of all kinds. E.g., large agencies have more and larger offices than small agencies, more and larger headquarters divisions, and the same holds true for every one of their structural components.

Prop. 1.3: the proportionate size of the average structural component (as distinguished from its absolute size) decreases with increased org size. Hence, most groups or categories of employees in large orgs are larger in absolute numbers, but constitute a smaller proportion of the total personnel than in small orgs. Ceteris paribus, change expectations are that the proportionate size of any personnel complement decreases with increasing org size.

Prop. 1.4: The larger the org is, the wider the supervisory span of control. That is, the number of subordinates per manager (this is what "span of control" means) must expand with increasing size.

Prop. 1.5: Orgs exhibit an "economy of scale" in management. Large scale operations reduce the proportionate size of administrative overhead; the relative size of administrative overhead decreases with increasing size. Again, we are talking proportions: the absolute number of managers increases, but these proportions are diminishing ones with org size.

Prop 1.6: The economy of scale in administrative overhead itself declines with increasing org size. This is "administration" as opposed to "management," and the article does not make explicit what the difference is, but I am willing to take Blau's word for it.

That wraps up the propositions under general concept number 1. The bottom line (something which is a good summary statement for the entire article, and is all you probably need to know) is this: "The structure of formal organizations seems to undergo repeated social fission with growth." If you get nothing else from Blau's article, this should be it.

Props under generalization number 2: structural differentiation in orgs enlarges the administrative component.

The problem of coordination drives this generalization: more people means more differentiation means more problems to sort out and control means more managers and supervisors are necessary to do the sorting and to take responsibility for it.

Prop 2.1: the large size of an org indirectly raises the ratio of administrative personnel through the structural differentiation it generates. Here, size moves differentiation, and differentiation moves administration, therefore size moves administration via differentiation. Got it?

Prop 2.2: the direct effects of large org size lowering the administrative ratio exceeds its indirect effects raising it owing to the structural differentiation it generates. In other words, we know that as orgs increase in size, they require more administrative personnel, but they require fewer per unit of employee. These two effects counteract each other, and the power of differentiation supercedes the power of mere size increase.

Prop 2.3: the differentiation of large orgs into subunits stems the decline in the economy of scale in management with increasing size, that is, the decline in the decrease in the proportion of managerial personnel with increasing size. Frankly, this proposition is hard to understand, and Blau admits as much. He also admits that he has no empirical support for it, and states that it "should be regarded as mere conjecture."

Blau repeats all of these props in his conclusion, without relating it extensively to his supporting data (of which there is much in his article- it is very much an empirical research work, and all of his propositions except 2.3 are strongly supported by a comprehensive data set).

Again, there is one main concept to take from this piece: size affects org structure. From this basic idea, you can derive many of his propositions: size creates differentiation, differentiation creates problems of coordination which creates a need for more managers and administration, etc. etc. I doubt that memorizing specific propositions is going to be useful for the exam, though if you feel the urge, go for it.