¹ Kochen, Manfred and Deutsch, Karl W., “Toward a Rational Theory of Decentralization: Some Implications of a Mathematical Approach,” this Review, 63 (09 1969), 734–749 Google Scholar.

² Goodwin, Richard, “Reflections,” The New Yorker, 44 (01 4, 1969), p. 48 Google Scholar.

³ Kochen and Deutsch, op. cit., p. 734.

⁴ In economic theory, profit maximization and cost minimization are two sides of the same problem. In this paper we use the terms interchangeably.

⁵ By the term economic good we do not intend to suggest that such goods are or should be provided primarily by the private sector; we mean to include in this category all goods which are primarily produced and distributed according to traditional profit maximizing rules of behavior. By the term government good we mean to include goods and services which have some degree of publicness in the economist's traditional sense of the term. (The classic articles on this subject were written by Samuelson, Paul A., “Diagrammatic Exposition of a Theory of Public Expenditure,” Review of Economics and Statistics, 37 (11 1955), 350–356 CrossRefGoogle Scholar; and “The Pure Theory of Public Expenditures,” Review of Economics and Statistics, 36 (11 1954), 387–389 CrossRefGoogle Scholar. We also mean to include goods which while provided by the government have, at least potentially, a distinct private consumption element. This private element may exist for one of two reasons: (1) the physical impact of the goods is not spread evenly across the political jurisdiction of the government or (2) the goods are divisible, panied by considerable variation in the private element. (This apparent quarrel with the strong distinction between public and private goods implicit in the original Samuelsonian analysis has been recognized by many authors. See, for example, Buchanan, James M., “Breton and Walden on Public Goods,” Canadian Journal of Economics and Political Science, 33 (02 1967), 111–115 CrossRefGoogle Scholar.

⁶ We recognize, of course, that not all personnel work to gain advancement in their organizations. But for the personnel with whom we are primarily concerned, managers and school principals, the assumption is probably justified. A good listing of the goals of organizational personnel can be found in Downs, Anthony, Inside Bureaucracy (Boston: Little, Brown and Company, 1967)Google Scholar, especially Chapter 8.

⁷ Kochen and Deutsch, op. cit., p. 739.

⁸ It is interesting to observe that Kochen and Deutsch do not find it necessary for their analysis to include a center or headquarters. This omission is trivial in their model, but not in a model of an organization producing government goods.

⁹ With today's increased concern over consumer protection this statement is somewhat of an oversimplification. The government may step in and order a firm to alter the nature of its product with or without receiving a complaint from the clients.

¹⁰ These different kinds of behavior by disappointed clients facing different types of organizations are described in some detail in Hirschman's, Albert recent work Exit, Voice and Loyalty: Responses to Declines in Firms, Organizations and States (Cambridge, Mass.: Harvard University Press, 1970)Google Scholar. “Staying and fighting” corresponds to Hirschman's “voice” and “taking one's business elsewhere” corresponds to his “exit.”

¹¹ See footnote 5. Without this or some other public good aspect of education there would be no reason for education to be supplied by the public sector. Instead it could be supplied by the private sector according to the rules of an economic organization. See Friedman's, Milton discussion of this issue Capitalism and Freedom (Chicago: Chicago University Press, 1962)Google Scholar, especially Chapter 6.

¹² A good example of this attitude was the deep Protestant suspicion of the initial attempts by Catholics to set up parochial schools. See, for example, Stauffer, Alvin P., “Anti-Catholicism in American Politics, 1865–1890” (Ph.D. dissertation submitted to the Department of History, Harvard University, 1933)Google Scholar.

¹³ This possibility exists even without interdependent utility functions. See Kenneth Arrow's discussion of the paradox of voting in Social Choice and Individual Values (2nd edition; New York: John Wiley and Sons, Inc., 1963)Google Scholar. See also Down's, Anthony comments on the paradox in An Economic Theory of Democracy (New York: Harper and Row, 1957), pp. 60–62 Google Scholar.

¹⁴ If we were to permit some differences to exist between the type of education received by one child and that received by another, arguing that the public component can be separated from the private component of the government good, then it is possible that such decentralization or differentiated responses to different tastes can only be achieved with an increase in money costs. This is due to the fact that some economies of scale would be lost from having different children follow different programs.

¹⁵ Because the model presented by Kochen and Deutsch does not explicitly contain a center, they do not consider the question of information flows between facilities and the center. In their model this omission is not serious, but it is important when comparing economic and government organizations.

¹⁶ On the ambiguity of verbal information, see McLuhan, Marshall, Understanding Media (paperback edition; New York: Signet Books, 1966)Google Scholar, especially Chapter 9.

¹⁷ An analogous distinction between types of information is discussed in some detail by Kochen and Deutsch, op. cit., pp. 742–745. What they fail to recognize, however, is that different types of organizations lend themselves to different types of information flows where the cost per unit of information content, as they do point out, may vary by several orders of magnitude.

¹⁸ We examine prices in more detail in the following section.

¹⁹ In conversation Professor Sandy Muir has commented that the difference in information flows can be attributed to the difference between an organization primarily concerned with distribution (the mail-order house) and an organization primarily concerned with production (the school). This suggests that production may play a relatively larger role in government goods than in economic goods.

²⁰ Economic organizations also experience the interdependence of requests, but this interdependency is likely to be of an extremely simple additive form. This occurs, for example, when several facilities request the shipment of quantities from a scarce, centrally stockpiled resource. For a mathematical treatment of this problem as it arises in centrally planned economies, see Danzig, George, Linear Programming and Extensions (Princeton: Princeton University Press, 1963)CrossRefGoogle Scholar, Chapter 23. Typically, the form of interdependence in a government organization will be more complex than simple additivity.

²¹ This is a broader formulation of the problem addressed by Kochen and Deutsch in their paper.

²² Strictly speaking, this characterization overstates the case. The center has no way of determining whether low profits mean a poor manager or a manager placed in an impossible situation. Nevertheless, the observation of low profits at least narrows the center's decision to two alternatives: fire the manager or close down the facility.

²³ There are some policies which are professionally acceptable but not politically acceptable. A good example is sex education for children in elementary schools.

²⁴ Merton, Robert K., “Bureaucratic Structure and Personality,” Merton, et al. (eds.), Reader in Bureaucracy (Glencoe: The Free Press, 1952), pp. 361–371 Google Scholar.

²⁵ Conflict over hiring procedures was one of the elements of the 1968 dispute over the Ocean Hill-Brownsville school district in New York. (See Mayer, Martin, “The Full and Sometimes Very Surprising Story of Ocean Hill, The Teacher's Union and the Teacher Strike of 1968,” New York Times Magazine, February 2, 1969 Google Scholar). At a different level in the educational system, the resumption of final authority over hiring by the Regents of the University of California was in response to the “unorthodox” hiring practices on some campuses. Such lessons are quickly learned.

²⁶ Hofstadter, Richard, The Age of Reform (New York: Alfred A. Knopf, 1955)Google Scholar.

²⁷ Banfield, Edward C. and Wilson, James Q., City Politics (New York: Random House, 1963), p. 115 Google Scholar.

²⁸ We leave to one side the issue of whether there is an empirical basis for Banfield and Wilson's middleclass ideal. Its existence and operational significance has been challenged by, among others, Wolfinger, Raymond E. and Field, John O., “Political Ethos and the Structure of City Government,” this Review, 60 (06 1966), 306–326 Google Scholar.

²⁹ One piece of evidence supporting this hypothesis is provided in James Q. Wilson's recent book on different police forces. The old-line middle-class ideal corresponds in Wilson's analysis to a “legalistic” style of police force, where policemen regard the law as an absolute and make arrests whenever they see a violation. The ethnic ideal is characterized by the “watchman” style, where the policeman is more interested in maintaining order than in making arrests and is willing to judge each case on the basis of special circumstances. Wilson, however, reports on a third type of police force with a “service” style, where small middle-class communities have opted for a police force which judges cases on their merits and would rather, if feasible, send a drunken driver home than arrest him. This evidence suggests that where circumstances allow, the middle class would rather not be governed by uniform rules but have judgments tailored to individual circumstances. Wilson, James Q., Varieties of Police Behavior (Cambridge, Mass.: Harvard University Press, 1968)Google Scholar.