Knowledge and decisions, p.23
Knowledge And Decisions, page 23
Given categorical mandates and the law of diminishing returns, it is virtually inevitable that governmental agencies would eventually end up doing things which seem irrational as isolated decisions. The aggrandizement familiar in all kinds of human activities-from the dressing of babies to the spread of multinational corporations-applies as well to governmental agencies. But where other expansions are constrained not only by budget limits but also by incremental returns from other lines of activity, governmental agencies with mandated activities have every incentive to push those particular activities as far as politically possibleeven into regions of negative returns to society. This is especially apparent in preventive activities, designed to contain various evils. As those evils are successively reduced, either by the agency's own activity or by other technological or social developments, the agency must then apply more activity per residual unit of evil, just in order to maintain its current employment and appropriations level. If the agency is supposed to fight discrimination against minorities, it must successively expand its concept of what constitutes "discrimination" and what constitutes a "minority." Urgent tasks such a securing basic civil rights for blacks ultimately give way to activities designed to get equal numbers of cheerleaders for girls' high school athletic teams.28 A nongovernmental organization, such as the March of Dimes, could-as it did, after conquering polio-turn its attention to other serious diseases, but if it had a government mandate strictly limited to polio, it would have little choice but to continue into such activities as writing the history of polio, collecting old polio posters, etc., while children were still dying from birth defects or other maladies. The point here is not that the leaders of the March of Dimes were either more intelligent or morally superior to the leaders of government agencies. The point is that a non-governmental organization subject to feedback from donors or customers has incentives and constraints that lead to institutional decisions more attuned to rational social trade-offs.
More diversified government agencies-such as the Department of Health, Education, and Welfare-have opportunities to change the internal mixture of its activities in response to changing social priorities, but only to the extent that the HEW leadership is in a position to impose agency-wide considerations on the "warring principalities" under its nominal control. By the same token, private organizations supported by a narrow constituencysuch as the NAACP Legal Defense Fund supported by affluent white liberals-may pursue certain activities well into the region of diminishing returns from the viewpoint of its ostensible beneficiaries (blacks) or the society at large, however important its historic mission may have been in the past. In short, it is not the political versus the private control of organizations which is crucial. It is the scope of the organization's mandate, and what that implies about its likelihood of pursuing some activity past the point of negative social returns. The safeguards required for the use of massive government power and huge sums of government money often confine the decision makers' discretion to a given line of activity and contain numerous rules within that activity. Moreover, because taxpayers cannot monitor numerous government agencies the way donors, customers, or family members monitor fewer and closer activities, the feedback to nongovernmental organizations is usually faster and more effective in diverting their efforts into new areas as the most urgent needs in the original area are met.
Bureaucracies, by definition, are controlled by administrative or political decisions, not by incentives and constraints communicated through market price fluctuations.9° While an ordinary business enterprise is constrained to keep its costs of production below the value of the output to the consumerand has incentives to keep it as far below as possible-such incentives and constraints are not merely absent in a bureaucracy but are replaced by other incentives and constraints tending in the opposite directon. The rank and pay of a bureaucrat is determined by his degree of "responsibility"-in categories documentable to third parties judging a process rather than a result. He is paid by how many people he manages and how much money he administers. Overstaffing, "needless" paperwork, and "unnecessary" delays may be such only relative to social purposes-not relative to the incentives established. Every "needless" employee is a reason for his superior to get a higher salary; so is every "wasted" expenditure, and every "unnecessary" delay preserves someone's job. The more "channels" the citizen has to go through, the more work is generated for the organization. For a bureaucrat assigned a given task (result), the incentive is to require as many people and as much money as possible to achieve that result. What is politically possible depends upon how visible his costs are, not their magnitude in relation to the value of the result. Moreover, the bureaucracy can expand the demand for its services by simply pricing them below cost. There is no such thing as an objective quantifiable "need" for anything. When the price is lower, a larger quantity is demanded. Profitand-loss constraints mean that a private business can expand its sales this way only as long as its price covers its costs of production. A government bureaucracy, which can dispense its goods or services below cost-including at zero price, in some cases-can always demonstrate a large "need" for its output, and therefore a "justification" for a large staff and budget.
It has been claimed that bureaucratization in general cannot proceed to lengths that are counterproductive, either in terms of organizational efficiency or their limitations on individual freedom, in a democratic country. Otherwise, a "bureaucracy-wrecking" party could be elected,3' with the support of "every citizen who believed he was paying more to support wasteful bureaus than he was receiving from those minoritiesserving bureaus that benefitted him directly."32 This would be true if knowledge were costless. But one cannot destroy "bureaucracy" in general, but only specific and highly disparate bureaucracies. If government is not a zero sum game, there may be substantial benefits to avoiding anarchy, and these benefits shield specific inefficiency from a broad axe attack on government bureaus. More narrowly, each bureau's activities may produce some benefit, even if some bureaus as a whole produce no net benefits. For a citizen attack on wasteful bureaus to succeed requires knowledge of the point at which benefit turns to waste or counterproductive activity. Even the most bitter critic of the Food and Drug Administration's policies retarding the introduction of lifesaving drugs may hesitate to destroy the whole agency and allow all kinds of poisons to find their way into our food and water supply. As long as bureaucratic waste or restriction stays within broad limits, and shields itself from specific detection, it may persist indefinitely despite its incremental costs exceeding its incremental benefits-as the voters would judge these, if they knew. The contrary view is a special case of the democratic fallacy, which equates market decision making under explicit cost constraints expressed in price tags with vote casting on the basis of plausibility and with high knowledge costs per voter.
INSTITUTIONAL CHANGES The difference between incremental and categorical decision making has implications not only for the location of given kinds of decisions inside or outside government; it has implications for how and where government decisions can most effectively be located. Periodic campaigns to "reform" or "streamline" the government bureaucracy under some "rational" plan to "end duplication" look very different within this framework. Duplication, for example, means that similar processes or results in a given field are obtainable through different organizations, usually located within larger and more diversified organizations with ostensibly differing purposes. The Veteran's Administration and the Public Health Service both operate hospitals, for example. Often this means that a given citizen has the choice of where to go with the same problem, whether that problem be consumer fraud, antitrust violations, or cases of racial discrimination. When duplication means individual choice, a set of unpaid "unmonitored monitors" has been created, able to effectively constrain the behavior of each agency with the implicit threat of going to some other agency if the same service is not provided as well. The economies of scale that might (or might not) result from consolidating the activity must be weighed against the higher costs or lower quality that are apt to result when monitors become a captive audience for a government monopoly instead. Moreover, the location of similar activities within a variety of conglomerate government organizations means that the phasing out of the activity becomes more feasible within a decision making unit that has other activities which can absorb the people and the appropriations. A more rationalistic plan of gathering all like activities into an agency devoted solely to that activity means in fact creating incentives to keep that activity alive as long as possible and to pursue it as far as possible, with little or no regard for social costs and benefits. The costs of duplication at a given time must be weighed against these longer run costs of consolidation.
Political decision making tends to be categorical rather than incremental in another sense as well. The programs of government officials or political candidates tend to be expressed in categorical rather than incremental terms. The lifeblood of politics is popular emotion, and categorical declarations capture that emotion. No one is going to man the barricades for a little more of A and a little less of B. Nor are they even likely to ring door bells on cold election nights for such incremental considerations. Therefore political activity-whatever its substantive or ideological content-has built-in incentives for categorical presentation of alternatives. The competition among political groups does not therefore bring to bear more accurate knowledge, as in economic competition, but promotes exaggerated hopes and fears-and sometimes deeds. Nor is this a transient pre-election phenomenon. Once such categorical exaggerations have been set in motion, they become incentives and constraints on subsequent policy making, in even the most totalitarian regimes. The press in a free country is to some extent a constraint on the categorical rhetoric of politics in government, but the selling of newspapers to subscribers and news programs to advertisers also depends on maintaining a certain level of public excitement which is also promoted by categorical clashes. There is little incentive for any institution to promote an incremental approach to political decision making.
The government tends to categorical decision making not only because of the incentives it faces but also because of the incentives it creates for those outside government. By conferring a valuable right on some group at the expense of some other group(s), the government provides an incentive for expensive, internecine struggles to be the group that receives rather than gives. Naked group struggles, openly recognized as such, would provide the basis for incremental adjustments of competing claims. But in order to get more public toleration for private interest, the dispute is verbally or ideologically transformed into a clash of principles-which must then be resolved categorically. All-or-nothing decisions raise the stakes, and the resources devoted to being the winner, and lower the probability of a socially optimal result from this socially disruptive process.
There is clearly some optimal level of change and of the divisiveness that accompanies it. With everyone paralyzed by fear of divisiveness, no change would ever have taken place-politically, economically, or socially-and we would all be still living in the caves. But if every change immediately set off new struggles to change that change, the relative merits of each of the successive states might mean less than the incessant turmoil. Whatever the optimal rate of change for a given political entity as a whole, that optimal rate for a given political practitioner or party is likely to be greater, since he can gain as the ostensible champion of whatever group he selects or creates by his divisiveness.
SUMMARY AND IMPLICATIONS The government has been conceived of as a framework of rules within which other decision making units can make decisions without the high transactions costs of maintaining private force for the purpose of protecting their physical safety or of protecting their belongings or of maintaining threats to enforce the carrying out of agreed upon contracts. As a framework, the government simply delineates the boundaries within which other units determine substantive choices, the government making its own forces available to defend the established boundaries. But while the government sets the basic framework for others-narrowly or broadly, depending upon the degree of freedom in the country-it is also itself subject to incentives and constraints, institutionally and individually. Government is not simply "society" or "the public interest" personified. Indeed, in modern democratic government-especially in the United States-it is often not a consolidated decision making unit but an overlapping montage of autonomous branches, agencies, and power cliques-each of these responsive to different outside coalitions of interest groups or ideologists.
The simple fact that governments are run by human beings with the normal human desire for personal well-being and individual or institutional aggrandisement must be insisted upon only because of a long intellectual tradition of implicitly treating government as a special exception to such incentives and constraints. This tradition stretches from the impartial "philosopher king" of Plato to the exalted "statesman" of the mercantilist literature of two to three centuries ago to the public spirited government as conceived in modern tracts that bill themselves as "empirical social science and not value statements"" In this modern literature, as in their historic predecessors, governmental take overs of decisions from other institutions are treated as themselves sufficient evidence-virtually proof-that such actions are needed to "remedy deficiencies"34 of other decision making processes which are "irrational" in some way.36 A mere enumeration of government activity is evidence-often the sole evidence offered-of "inadequate" nongovernmental institutions," whose "inability" to cope with problems "obvi- ously"37 required state intervention. Government is depicted as acting not in response to its own political incentives and constraints but because it is compelled to do so by concern for the public interest: it "cannot keep its hands off" when so "much is at stake,"38 when emergency "compels" it to supersede other decision making processes.39 Such a tableau simply ignores the possibility that there are political incentives for the production and distribution of "emergencies" to justify expansions of power as well as to use episodic emergencies as a reason for creating enduring government institutions.
This ignoring of political incentive structures extends to the effects of government action as well as its causes, often "pretending that the effect of a law and appropriation will be what their preamble says it should be."'0 Much complaint about bureaucratic "inefficiency" or "stupidity" presupposes that bureaucrats are pursuing the goals stated in the preambles to the legislation authorizing their existence, rather than responding to the incentives created in the "details" of that legislation. Not even physical or engineering efficiency can be calculated without first defining a goal. Where bureaucrats are pursuing their own individual or organizational goals, they are hardly being "inefficient"-much less "stupid"-in terms of other goals that other people wish they were pursuing. This is not merely a matter of verbal fastidiousness but of practical policy: replacing the allegedly "inefficient" or "stupid" people with more intelligent people, or people with a record of efficiency in private industry, could not be relied upon to improve the implementation of the social policy described in preambles, as long as the structure of incentives and constraints remains the same.
The importance of actual institutional characteristics as a guide as to what to expect is obscured by the common practice of defining political institutions by their hoped-for results: the Environmental Protection Agency, the Equal Employment Opportunity Commission, the Defense Department, etc. The change of the latter name from "War Department," which describes what a military organization actually does or prepares to do, to "Defense De- partment"-presumably incapable of ever launching a military attack-was symptomatic of this pious obfuscation.
Incentive structures are important in explaining political behavior, not only in a static sense but in following dynamic changes of political patterns. Incentives operate not only by guiding the actions of given people, but by changing the mix of people drawn to particular activities. Very different kinds of people may be attracted or "selected"-in an impersonal Darwinian sense-by one set of incentives than by another. Used car dealers tend to differ from Red Cross volunteers. Movements for political change-that is, insurgents in general, whether moderate reformers or violent revolutionariesare essentially attempts to change incentive structures, however much they may choose to describe themselves in terms of their hoped-for results. But prior to the achievement of any success-whether reform or revolutionpeople who man insurgent movements are "selected" in a Darwinian sense under an entirely different pattern of incentive structures from the incentive structures that they are advocating. Insofar as the insurgency becomes successful, the new incentives tend to select a different mix of persons. For example, socialists under capitalism may differ from socialists under socialism.
A capitalist system, especially when it is actively defending itself, may offer few direct personal benefits for being a socialist and may impose various costs, ranging from social disapproval to jail, depending upon the condition of civil liberties in the particular country. Narrowly self-interested persons, or persons of weak will or timid disposition, are unlikely to be attracted to socialist movements under these conditions. But when socialism has become established, especially if in the form of a totalitarian orthodoxy, it is being a supporter of capitalism that now carries a high cost and being a supporter of socialism that offers higher reward. The mixture of people attracted to socialism should be expected to change accordingly.
