The Limits of Liberty: Between Anarchy and Leviathan

The Threat of Leviathan

Dictionaries define Leviathan as "a sea monster embodying evil." In 1651, Thomas Hobbes applied this term to the sovereign state. Three and one-quarter centuries later, we use the term only when we discuss government and political processes pejoratively, and then only when our purpose is to call attention to the dangers inherent in an expanding public sector of society. I have discussed the paradox of being governed in Chapter 6. In democracy, man considers himself simultaneously to be a participant in government (a citizen) and a subject who is forced to abide by standards of behavior that he may not have selected, including overt acquiescence in the confiscation through taxation of goods that he treats as "his own."

For late twentieth-century man, this bifurcation in his attitude toward the state is "natural" in the sense that it emerges directly from his post-Enlightenment, postsocialist cultural heritage. From our vantage point in the 1970s it is difficult to appreciate the importance of the initial change of vision which first enabled man to see himself as an independent will. I do not pose as an exegetical expert on ancient texts, but can there be much question that the conception of independent man, universalized over all persons, was largely foreign to Greek and Roman philosophy? Medieval Christianity introduces an ambivalence, in that individual salvation was stressed but almost always for the greater glory of God.*113 Only in the full emergence from the Middle Ages, only with Hobbes, Spinoza, and their contemporaries does man become possible independent of other men, of God, of state and city. In the Hobbesian jungle, the life of independent man was indeed described as poor, nasty, brutish, and short. But in Hobbes's ability to visualize, to conceptualize, such an existence at all lies the critical difference with earlier philosophers. Can we conceive of pre-Hobbesian anarchists?

Once independent man was set against the state, even in an argument that suggested rational bases for obedience, the potential for continuing revolution was guaranteed. The genie could not be put back into the bottle, no matter how logical the arguments of a Malmesbury philosopher. Man could now think himself into a role as king; in his mind's eye, man could now leap out of his estate or order, and some man or men would surely act out these dreams. Althusius, Spinoza, Locke, and, even more emphatically, Rousseau, commenced and continued to talk about a social contract among independent men, not a Hobbesian slave contract between men and a sovereign master. From contract among free men, all things might emerge, including basic law itself. For the first time, man seemed to be offered a prospect for jumping out of his evolutionary history. Man, in concert with his fellows, might change the very structure of social order.

The conception was as revolutionary as its consequences, the age of democratic revolution.*114 Repressed revolt, successful revolution, revolutionary terror, repressive reform, counterrevolution—these various stages in our spatially divergent modern history need not be discussed in detail here. We know that man failed to live up to the promise of his Enlightenment dreams. Hardly had some of the tyrants been overthrown and some elites vanquished, when others emerged. And once the political and social order was put up for grabs and was seen to be so, how could the economic basis of this order withstand assault? Locke's valiant efforts to erect a contractual superstructure over existing property rights were foredoomed to failure. If men, in concerted contract, are not bounded, need any limits be placed on collective action? Why need the economic order stand immune from fundamental structural rearrangements, especially when effective challenges were issued by Karl Marx? Socialism, in its varied guises, came to inform the consciousness of early twentieth-century man. The circle seemed almost complete; independent man once again seemed to have become submerged in an all-embracing collective will.

Once loosed, however, independent man could not be so readily destroyed. The Soviet Union was not the future, as the Webbs had proclaimed in ignorant joy. Even in Russia, where man had scarcely attained individualized independence before communist revolution, his innate stubbornness made efficient control impossible. In the West, where men have experienced freedom, where freedom itself has a history, democratic socialism was foredoomed. Collectivized governmental attempts to do more and more have been demonstrably revealed to accomplish less and less. Man finds himself locked into an impersonal bureaucratic network that he acknowledges to be of his own making. He begins to use the term "Leviathan" in its modern connotation, yet he feels personally unable to offer effective alternatives.

This difference between prerevolutionary and modern man must be understood if the latter's predicament is to be appreciated. Modern man cannot place himself in opposition to a government that is staffed and directed by an exterior elite, by members of a wholly different order or estate. To an American patriot, there was George III. To a member of the French bourgeoisie, there was the ancien régime. To the followers of Lenin, there was the Russian aristocracy. To modern man tangled in the web of bureaucracy, there is only himself, or others of his same breed.

This is not, of course, to suggest that imperfections in democratic process are absent or that all persons possess equal power of influencing governmental policy in the modern world, and in America in particular. I am suggesting that, even if all imperfections could be removed, even if all persons were placed in positions of equal political power, the central issues facing modern man would remain. When we speak of controlling Leviathan we should be referring to controlling self-government, not some instrument manipulated by the decisions of others than ourselves. Widespread acknowledgment of this simple truth might work wonders. If men should cease and desist from their talk about and their search for evil men and commence to look instead at the institutions manned by ordinary people, wide avenues for genuine social reform might appear.

Wicksellian Unanimity

Why need there be constitutional limits or controls over the scope and range of governmental activity? In order to understand this, we may first look at the idealized model which gives to the individual full power over his destiny. Consider a community that makes all collective decisions in accordance with a Wicksellian rule of unanimous consent. Let us further assume, this time quite unrealistically, that this rule is operative without major costs of reaching agreement. In such a model, each person is party to all collective decisions, no one of which can be taken without his express consent. How could the dynamics of such a decision model generate results that could be judged undesirable or inefficient by any one or by all of the persons in the community?

Because each person must agree positively to every decision taken, the flaw, if indeed one exists, must lie in the individual precepts for rational choice, not in the amalgamation of individual choices in producing collective outcomes. Analysis should, therefore, be concentrated on individual decision-making. Why would an individual agree to each one of a sequence of collective decisions, separately taken, only to find that the sequence generates an undesired ultimate outcome? Once the question is put in this way, numerous analogies from personal experience are suggested. Perhaps the one that is most pervasive is eating. In modern affluence, individual choice behavior in eating, on a meal-by-meal basis, often leads to obesity, a result that is judged to be undesirable. The individual arrives at this result, however, through a time sequence in which each and every eating decision seems privately rational. No overt gluttony need be involved, and no error need be present. At the moment of each specific choice of food consumption, the expected benefits exceed the expected costs.

The problem is not fully described as one of myopia in individual choice behavior, as a simple failure to take into account the future consequences of present action. Such myopia is, without doubt, one of the important bases for disappointment or regret when undesirable situations are recognized to be the result of a series of earlier choices. In this sense, all temporally related choices can be made to appear to be characterized by myopia. Consider saving and capital formation. From the vantage point of "now," a person may always wish that he had saved more and consumed less in earlier periods, and, in this vision, he may look on past behavior as having been myopic. More reasonable judgment might suggest, however, that each decision, when made, was based on some appropriately weighted calculation of costs and benefits in the "then" time setting. The decision to eat more than is dictated by the maintenance of some long-term weight standard is equivalent to a failure to save an amount sufficient to attain some long-term wealth objective. When this temporal interdependence among separate-period decisions is recognized, rational choice behavior at the "rule-making" level may internalize the interdependence through the explicit adoption of constraints on separate-period freedom of action. When he adopts a rule and insures its enforcement, the individual is exercising his freedom, at a more comprehensive planning stage of choice, only through restricting his own freedom in subsequent potential choice situations.

The person who recognizes his tendency to overeat may adopt a stringent diet. He deliberately imposes self-generated constraints on his own choice options. He locks himself into an eating pattern that he predicts to reduce the utility gains from separate-period behavior in exchange for predicted utility gains over an extended choice domain. The diet becomes the "eating constitution," the person's set of internally chosen rules that act to prevent overindulgence. It seems clear that individuals may want to impose comparable constraints on their separate-period and separate-choice behavior in undertaking joint or collective actions, even in the idealized setting of Wicksellian unanimity. That is to say, individuals might rationally choose to operate under a set of constitutional rules for taking collective actions even if each person knows that he is empowered, personally, to veto any specific proposal that might be presented. In this setting, however, we should note that such a set of rules might be made operative by the choice behavior of a single member of the group. The determination of a single person in the community to abide by some internal constraint on the range of collective action would be effective for the whole group. Collective action would be constrained in this strictly Wicksellian setting by the mere presence of one person who chooses to adopt internal rules for his own participation in collective choices.

Majority Voting under Benefit-Cost Constraints

We move somewhat closer to reality when we drop the assumption that collective action requires unanimous consent of all participants. As suggested, under a genuine unanimity rule, individual decisions can keep government under effective controls. Things become quite different, however, once any departure from unanimity is introduced. When the costs of securing agreement are acknowledged, departures from true government by consent become necessary if the political community is to function as a collectivity. In the conceptual constitutional compact establishing this community, some set of rules for making collective or governmental decisions is selected, and these rules, once made operative, are enforceable on all members, whether or not they belong to the decisive coalition which effectively makes particular choices under the rules.

The most familiar decision rule, both in the analytical models of political process and in existing historical structures that are appropriately classified as "democratic," is that of majority voting. We may assume that some constitutional structure exists, a structure that defines individual property rights and enforces contracts among persons and, further, requires that all collective or governmental decisions secure the majority of the representatives of citizens in some legislative assembly. Even in this formulation we have, by assumption, already bypassed a significant part of the issue being discussed. At the stage of constitutional contract, when individual rights are initially defined, few persons would conceptually agree to wholly unconstrained departures from a unanimity rule for collective decision-making. The reason is, of course, that once an individual's consent is not required for a decision that will be enforced upon him, the individual holds no protection of his own nominal assignment of claims, no guarantees that his rights will not be exploited on behalf of others in the name of governmental objectives. At the same time that a collective decision rule, say that of majority voting, is adopted, procedural limits on the exercise of this rule may be incorporated into the constitutional document or understanding. Experience indicates, however, that the procedural limits incorporated in constitutional structures historically have not been very effective in curbing the appetites of majority coalitions.

Nonetheless, it will be useful for analysis to develop the argument in two stages. In the first, we assume that an economically meaningful constraint on majority decision exists. Assume that a constitutional provision requires that all proposals for public or governmental outlay satisfy a benefit-cost criterion; gross benefits must exceed gross project costs, regardless of the array of votes in the legislative assembly.

We want to look at public-goods proposals that do not benefit all members of the group sufficiently to offset fully tax-costs, but which do, nonetheless, meet the benefit-cost criterion imposed. If, for example, in a three-person group there should be only two beneficiaries of a project costing $100, and if each of these beneficiaries expects to secure a value of $51, the proposal would meet the benefit-cost criterion no matter how costs are distributed. If the costs are equally distributed among all members, say, by a general tax, the proposal would secure majority approval. The effect would be to impose net losses on the minority. The benefit-cost constraint guarantees, however, that if compensation should be required, the majority could arrange to secure minority acquiescence with appropriate side payments. Another way of saying this is to state that the benefit-cost criterion insures that all spending projects are "efficient" in the strict economic meaning of this term. Still another version, and related to the preceding section, is to say that all projects could conceptually secure unanimous approval if the costs of making side payments are ignored.

If each and every proposal for spending funds governmentally is required to meet the efficiency criterion, how could the aggregate budgetary level fail to do so? How could the overall budget be too large or too small? Since each project, considered independently, meets the efficiency test, it would seem that the test could also be met by the aggregate of all projects. As the discussion of the preceding section may suggest, however, this result need not follow when there exists interdependence among the separate decisions.

Consider, as an example, two interdependent proposals for budgetary spending, Projects I and II. In the absence of, and independent of, the other project, each of these proposals is estimated to cost $100, of which $90 is for outlay on the purchase of resource inputs, and $10 is for outlay on collection and enforcement. For each project, similarly, estimated benefits are $103. Hence, regardless of the way benefits are distributed, each proposal is economically efficient. Suppose now that Project I is approved initially under these conditions and that it is included in budgetary plans. Project II is now considered independently, but subsequent to majority approval for Project I. Direct outlay on resource inputs is again $90, as with Project I. But, because more revenues are now required in total, collection-enforcement costs are now estimated to be $12, for a total project cost of $102. Benefits are estimated to be $103; hence, the project remains apparently efficient, and we assume that Project II is also approved by a majority. In adding Project II to the budget, however, collection-enforcement costs for Project I may also have been increased, from the $10 initially estimated to the $12 estimated for Project II. The external or spillover cost that the addition of Project II generates for Project I is $2, but this was wholly left out of account in the choice-making sequence that we have outlined.

Note that, in the numerical example, aggregate benefits of the two projects ($206) exceed aggregate costs ($204). Note, however, that gross fiscal surplus is reduced below that which is attainable on the approval of only one of the two projects; the surplus falls from $3 to $2 in the process of adding Project II, which, treated independently, is equivalent to Project I. The numerical example is, of course, illustrative only, and the totals need not be taken as at all descriptive. In terms that are familiar to economists, we can say that there exists a divergence between the direct or separable costs of a single project and the genuine social costs, which must include all external or spillover effects on other projects or components in the budgetary set. When stated in these terms, economists might suggest "internalization" through simultaneous consideration of all the interrelated budgetary items. Care must be taken, however, to insure that the appropriate maximand is selected. Taken as a two-part budgetary package, both projects in the numerical example would secure approval, even if they were jointly selected. Joint benefits exceed joint costs.

The more general phenomenon that the example represents has considerable real-world relevance in terms of widely acknowledged economic effects and of observed political institutions.*115 Collection and enforcement costs are always present, and these costs increase as budget size grows, possibly disproportionately beyond certain ranges. More important, taxation necessarily modifies incentives toward the earning of taxable incomes and accumulating taxable wealth in the private economy. These effects are directly related to budgetary size, and these are genuine social costs that incremental budget-making can scarcely incorporate.

Politically, budgets are made piecemeal.*116 Different legislative committees consider budgetary components independently, and possibly divergent majority coalitions are organized in support of each component. So long as benefits exceed costs, why should members of the effective supporting coalitions be concerned about spillover costs on components, past, present, or future? Political realism suggests the implausibility of achieving reforms at the level of incremental decision-making. Comprehensive budgeting, at either the executive or the legislative level or both, need not eliminate the inefficiency, as we have noted. Consider the position of a budget director or chairman of a legislative committee. By our restrictive assumption, any component must meet the overall benefit-cost constraint. But since this criterion is also satisfied for the budget in the aggregate, or may be, what incentive does this official have for reducing or eliminating particular components or line items so as to increase net fiscal surplus? Even if the official is ideally responsive to the demands of the citizenry, he will be led to incorporate too many components in the budgetary package. Consider again our two-project example. A budget director has overall coordinative responsibility; he must approve a project before it is submitted for a vote. If he eliminates one of the two projects, he incurs the displeasure of all direct beneficiaries. He pleases general taxpayers, but as we have assumed and as the real-world patterns suggest, taxes are more widely shared than the benefits. The indirect net costs that will be reduced by budgetary constraints are not likely to be sensed by the citizenry, and especially not in connection with specific budgetary choices.*117

The inefficiencies that emerge when there exists interdependence among the separate components of a budget can be reduced only if these are predicted at some planning stage of deliberation. Because of the tendency of budget-makers and of legislative majorities to approve budgets that aggregate to sizes beyond those which maximize fiscal surplus, explicit size limits or other constraints on revenues and/or outlays may be incorporated in the fiscal constitution with the expectation that such limits will be legally enforced.*118

Majority Voting without Benefit-Cost Constraints

If we drop the arbitrary requirement that all proposals for spending publicly collected revenues meet criteria of economic efficiency, it is evident that majority-voting rules for reaching collective or group decisions will produce at least some budgetary components that are inefficient in net. Some projects that will secure majority approval will yield less in total benefits than they cost. The minority will suffer net losses from these projects, and these losses will exceed the benefits secured for members of the majority. In a regime with costless side payments, the minority could bribe the majority so as to prevent the approval of all such projects. But when the absence of effective side payments is acknowledged, the existence of inefficient spending projects can hardly be questioned.

Consider again a very simple example, a three-person group that has organized itself collectively. Taxes are equal per head, and all spending decisions are made by majority voting. Suppose that there are three potential projects to be considered, each of which costs $99, financed by a tax of $33 on each person. We assume that these projects are wholly independent and that the externality effects analyzed in the preceding section do not arise. The benefits from each project are concentrated as indicated in the following:

Person	Project I Benefits	Project II Benefits	Project III Benefits
A	$35	$35	$0
B	35	0	35
C	0	35	35

Under the rules that we have postulated, each of these three projects would be adopted, so long as each project is considered separately. In the process, however, each person will have paid out a total of $99 in taxes and will have received only $70 in benefits. Each person will be worse off with the three-project budget than he would be with no budget at all. It is clear from this example that budgets will tend to be overexpanded under simple majority voting rules if budgetary components are considered separately in the legislative deliberations, and if benefits are more concentrated than taxes.

There is, however, a difference between this and the earlier model where we assumed projects to be interdependent. In this model, which we might call one of simple majority exploitation of the minority, "internalization" in the form of comprehensive or package consideration of the whole budget may eliminate some of the inefficiency. If the three-man group in this example should be forced, by institutional-constitutional requirement, to treat projects in a bundle rather than in isolation, and if members of the group accurately measure costs and benefits, projects that are demonstrably damaging in the net for all persons will not secure approval. Alternatively, constitutional restrictions might be imposed which dictate that only spending proposals that promise general benefits to the whole membership of the community can be considered.*119 Historically, procedural requirements have been interpreted to dictate tax uniformity or generality, at least over broad groupings. For the benefits side, however, no fully comparable requirements have been applied. As a result, there are relatively few effective limits on the fiscal exploitation of minorities through orderly democratic procedures in the United States.*120

Logrolling and minority benefits

The majority voting model discussed above suggests that inefficient budgetary projects may secure approval if considered separately, but that, at a minimum, the estimated value of benefits from any proposal to the members of an effective majority coalition must exceed the tax-costs borne by those members. Even this minimal constraint on budgetary inefficiency is not operative, however, when logrolling can take place among divergent minorities to produce effective majority coalitions on a subgroup of budgetary items. This procedure is familiarly known as "pork barrel" legislation in the American setting.*121

Political Income, Bureaucratic Rents, and Franchise

To this point, the models of collective decision-making examined have not allowed for the influence of politicians, governmental employees, or bureaucrats on budgetary outcomes. Implicitly, the models have contained the assumption that voters demand publicly supplied goods and services which, once approved, are made available to final beneficiaries or consumers directly. There is no intermediation by legislative representatives and no administration by bureaucratic agencies. Such models are useful for general purposes, and especially so when budgets are relatively small. In modern democracies, however, more than one-third of the national product is organized through the governmental sector. In these settings, neglect of the influence of politicians and bureaucrats on budgetary results may severely weaken the relevance of any analysis.

Politicians' preferences and budgetary bias

Collective decisions are rarely made directly by voters, by those persons who pay the taxes and who are supposed to benefit from the provision of governmental goods and services. Effective political organization requires that the roles of voters be limited largely if not entirely to the selection of representatives, persons from their own ranks, who will then participate in legislative and executive decision-making. These politicians are the men who make the direct and final choices on the quantities of public goods and services and on the size of the total budget along with its composition and financing.

It is unrealistic to assume that elected officials who occupy executive and legislative positions of responsibility have no personal preferences about the overall size of the public sector, its sources of revenue, and, most important, about the particular components for public outlays. A person who is genuinely indifferent in all these respects would not be attracted to politics, either as a profession or as an avocation. Politicians are likely to be those persons who do have personal preferences about such matters and who are attracted to politics precisely because they think that, through politics, they can exercise some influence over collective outcomes. Once this basic, if simple, point is recognized, it is easy to see that budgetary results will not fully reflect voters' preferences, even of those who are members of the effective coalition that achieves victory for its own candidate or party.

Once elected, a politician has considerable freedom for choosing his own preferred position on spending or tax issues. He is constrained by voters indirectly through prospects for reelection, for long-term party support, for generalized public acclaim. But even for the politician who is highly sensitive to these indirect constraints, there remains freedom of choice over substantial ranges of the political spectrum. Within what he treats as his feasible set, the politician will choose that alternative or option which maximizes his own, not his constituents', utility. This opportunity offers one of the primary motivations to politicians. In a meaningful sense, this is "political income," and it must be reckoned as a part of the total rewards of office.*122

The existence of opportunities for politicians to maximize personal preferences within constraints need not be of relevance to the subject matter of this chapter if the effects on budget-making could be predicted to be symmetrical or unbiased. If the "slippage" between the preferences of voters and the results emerging from the actual budgetary process should involve roughly offsetting differences on the up and down sides, no net influence on aggregate budgetary size would be exercised. Unfortunately, a unidirectional bias toward expansion in the fiscal accounts seems to be present. This direction of the political leader's preference bias involves several distinguishable elements. In the first place, those persons who place relatively high values on the ability to influence collective outcomes, and who do so in the genuinely incorruptible sense of desiring to "do good" for the whole community, are quite likely to be those who seek to accomplish their own preferred social objectives through collective or governmental means. By contrast, those persons who, ideologically, desire that the governmental role in society should be reduced to minimal levels are unlikely to be attracted to politics. Few natural anarchists or libertarians frequent capital cloakrooms.

Ideologues aside, persons may be attracted to politics because they intrinsically place high values on the power to make decisions affecting the lives of others. This characteristic is different from the first, where power to influence collective decision is instrumentally desired for the purpose of furthering social objectives. Some politicians may have very ill-defined objectives for social policy and those that they do have may seem relatively unimportant. They may seek political and/or elected office, however, because they enjoy positions of leadership and authority, positions that make it necessary for other persons actively to seek them out and solicit their assistance. This sort of politician secures utility more directly than his ideologue counterpart; his utility is increased by the emoluments of office that necessarily arise from public knowledge about the location of decision-making authority. If the list or menu for choice should be fixed in advance, the behavior of politicians of such nonideological stripe might produce results that are closer to the true preferences of voters. This correspondence would emerge from the desires to meet the demands of the largest possible number of constituents. In such case, no directional budgetary bias would be introduced by the necessary departures from pure democracy. When the list or menu for political choice is not predetermined, however, the directional bias toward expanded budgets again arises. The politician who secures his utility only because he chooses for and thereby pleases the largest number of constituents will find that favorable action on differentially beneficial spending projects offers more reward than favorable action reducing general tax-costs. The politician's bias, in this respect, is an additional institutional aspect of the asymmetry between the spending and taxing sides of the fiscal account. Because taxes cannot readily be lowered in a differential manner, there is a public-goods barrier which inhibits independent politician initiative toward tax reduction. By contrast, because the benefits from government spending may be differentially directed toward particular subgroups in the community, politicians are motivated to initiate the formation of coalitions that will exploit these latent demand opportunities. Given his degree of freedom to influence outcomes, the nonideological politician's behavior will tend to generate an exaggerated version of the nonpolitician model analyzed earlier. Because of the asymmetry in the effective fiscal constitution, aggregate spending will tend to be inefficiently large even if the ultimate demands of voters-taxpayers-beneficiaries could be accurately reflected in final outcomes. The introduction of politicians as the direct decision-makers will extend the results even beyond such limits.

To this point, we have assumed implicitly that both the ideological and the nonideological politicians are incorruptible and seek no pecuniary gain from political office over and beyond formal compensation. To these two types of officials it is now necessary to add a third, that of the politician who does seek pecuniary gains from his office. The direction of budgetary bias is the same as before. The prospect for profitable bribes, kickbacks, or by-product deals is directly related to the size and complexity of total government budgets, and, more generally, of the total governmental operation in the economy. With minimal governmental intrusion into the economy, with minimal and quasi-permanent spending components, the grasping politician may have little or no opportunities for graft. However, with a complex public sector, and one that involves new and expanding spending programs, there may be numerous opportunities. In a newly enacted program, one without established guidelines and procedures, politicians may find ample sources for direct and indirect kickbacks from the producers and producing firms whose rents are enhanced by the program. Such officials will, therefore, seek continually to enlarge budgets and, especially, to introduce new and different programs. On the other hand, the potentially corrupt politician would rarely press for general budgetary reduction. The direction of bias seems apparent, again under the institutional proviso that taxes are distributed more generally than the benefits of public spending.*123

Elected politicians may fall into either one of the three categories discussed, or a single politician may himself represent some mixture of two or all three of the types. The directional bias on budgetary size is the same for all types. Although their reasons may differ, the ideologue, the seeker after public acclaim, and the profiteer each will be motivated to expand the size and scope of the governmental sector of the economy.

Bureaucratic rents and franchise

Even after elected politicians make taxing and spending decisions, public goods and services do not flow automatically and directly from competitively organized suppliers outside the economy to final consumers within the economy. Governments, when authorized to do so, may purchase inputs from independent private suppliers (individuals and firms) and combine these to produce outputs. Or, alternatively, governments may purchase final outputs after these have been produced by private suppliers and distribute these to beneficiaries. In either case, and much more extensively in the former than in the latter, employees must be hired to implement the complex fiscal transaction between the ultimate taxpayer-purchaser on the one hand and the ultimate beneficiary-consumer on the other, even if, in some net accounting, these may be the same persons. Once elected officials, as representatives of the voters, decide on a quantity and a distribution of taxes, other officials (agents) must be employed to collect the revenues. Accountants must be hired to keep the books; auditors must be added to check the agents and the accountants. Inspectors must be available to search out recalcitrant taxpayers. On the spending side, budget specialists are required to maintain and present details of complex programs and to make comparative evaluations. Purchasing agents must carry out buying tasks in the framework of procedures worked out by still another layer of bureaucratic personnel. And personnel specialists are necessary to get personnel.

All of these would be needed even if no direct production of goods and services takes place within the governmental sector itself. Once direct production is attempted, massive numbers of additional employees are needed. If government produces postal services, mail clerks, postmen, and postmasters must be hired. If government produces education, administrators, teachers, supervisors, and custodians become government employees along with others who must evaluate the credentials of those who produce the services. The list can be extended almost without limit.

If taxpayers-voters, acting through their elected politicians, should be able to secure government employees externally at competitively determined wage and salary scales, the necessary existence of a bureaucratic superstructure need not itself introduce major distortion in the budgetary process. As with the enforcement problem discussed in earlier chapters, however, difficulties arise from the necessity of staffing government with persons drawn from within the political community. The sequence of budgetary outcomes tends to be biased toward overexpansion because of the potential for earning producer rents which government employment offers and because employees hold voting rights in the polity. If bureaucrats could not vote, the existence of producer rents from government employment would increase the costs of public-supplied goods and services, but this alone would not bias the results significantly. On the other hand, even if bureaucrats hold the voting franchise, no problem might arise if governmental wage and salary scales, along with tenure and promotion policies, were competitively determined. In the real world, however, governmental employees have full voting rights, and governmental salaries and working conditions are not settled in competitive markets.

Regardless of his interest as a demander-taxpayer or final beneficiary of a publicly provided good, a person who expects to be or is already employed by the governmental agency that provides this good will tend to favor increases and to oppose reductions in budgeted outlay. (How many medical researchers at NIH would support reductions in federal government outlays on medical research?) If he holds a voting franchise, the prospective or actual employee becomes a built-in supporter of budgetary expansion and a built-in opponent of budgetary reduction, not only for the particular component within his immediate concern but for other components as well. As students of political economy have long recognized, producer interests tend to dominate consumer interests, and the producer interests of government employees are no different from those of any other group in society. Two additional elements accentuate the effects of bureaucratic franchise on budget size. As with elected politicians, those who are attracted to governmental employment are likely to exhibit personal preferences for collective action, at least by comparison with those who are employed in the private sector. More important, because of specific producer interest that a working bureaucrat recognizes, the exercise of ultimate voting privileges is more likely to occur. Empirical evidence supports this inference; the proportion of governmental employees who vote is significantly higher than the proportion of nongovernmental employees. The result is that members of the bureaucracy can exert a disproportionate influence on electoral outcomes.

The franchising of bureaucrats need not involve serious budgetary bias when total government employment remains small. As the public sector continues to grow, however, the voting power, and hence the political power, of franchised bureaucrats cannot be neglected. In modern America, where roughly one in each five employees works for government, bureaucrats have become a major fiscal constituency, and one that politicians seeking elective office recognize and respect.*124

This influence would be present even in an ideally working bureaucracy so long as net rents were earned in government employment. As we must recognize, however, no structure can approach the old-fashioned textbook ideal in which bureaucrats merely carry out or execute policy directives chosen for them by legislative authorities.*125 Bureaucrats, like elected politicians, possess varying degrees of freedom to select among alternatives. A collective decision, as made by a legislative assembly, is never sufficiently definitive to leave no scope for exercise of authority on the part of administrators of the program. Within limits, the nonelected government employee makes final decisions about government actions. Stated in a somewhat converse way, the legislature or elected executive can never exercise full control over the behavior of bureaucrats in the structural hierarchy, and any attempts to gain full control would involve prohibitive costs.*126 Within the constraints that he faces, the bureaucrat tries to maximize his own utility. He is no different from anyone else in this respect. He can hardly be expected to further some vaguely defined "public interest" unless this is consistent with his own, as he defines the latter.

Once this point, again a very simple one, is acknowledged, the influence of the bureaucracy on budgetary results can be predicted to be unidirectional. The individual who finds himself in a bureaucratic hierarchy, who knows that he earns net rents when he compares his situation with his private-sector opportunities, looks directly at the reward and penalty structure within the hierarchy. He knows that his career prospects, his chances for promotion and tenure in employment, are enhanced if the size of the distinct budgetary component with which he is associated increases. He will, therefore, exercise his own choices, whenever possible, to increase rather than to decrease project and agency budgets. There is little or no potential reward to the governmental employee who proposes to reduce or limit his own agency or bureau. Institutionally, the individual bureaucrat is motivated toward aggrandizement of his own agency.*127 And, since the effective alternatives for most governmental employees are other agencies and projects, this motivation for expansion will extend to government generally.

Democracy Unchained

The purpose of the several preceding sections was to demonstrate that even under the most favorable conditions the operation of democratic process may generate budgetary excesses. Democracy may become its own Leviathan unless constitutional limits are imposed and enforced. Historically, government has grown at rates that cannot possibly be long sustained. In this sense alone, modern America confronts a crisis of major proportions in the last decades of the twentieth century. In the seven decades from 1900 to 1970, total government spending in real terms increased forty times over, attaining a share of one-third in national product. These basic facts are familiar and available for all to see. The point of emphasis is that this growth has occurred, almost exclusively, within the predictable workings of orderly democratic procedures.*128

The authors of the United States Constitution, the Founding Fathers, did not foresee the necessity or need of controlling the growth of self-government, at least specifically, nor have these aspects been treated in traditional political discourse. The limits or constraints on governmental arms and agencies have been primarily discussed in terms of maintaining democratic procedures. Rulers have been subjected to laws because of a predicted proclivity to extend their own powers beyond procedural limits, at the presumed expense of the citizenry. But implicit in much of the discussion has been the notion that, to the extent that democratic process works, there is no need for limits. The system of checks and balances, ultimately derivative from Montesquieu, has rarely been interpreted to have as one of its objectives the limiting of the growth of the government. The excesses of the 1960s created widespread public disillusionment about the ability of government, as a process, to accomplish specific social objectives. But, before the 1960s, the checks and balances that were present in the United States constitutional structure were far more likely to be criticized for inhibiting the extent of governmental action than for their inability to accomplish an effective limitation on this action. In this respect, the 1970s and beyond present a new and different challenge. Can modern man, in Western democratic society, invent or capture sufficient control over his own destiny so as to impose constraints on his own government, constraints that will prevent the transformation into the genuine Hobbesian sovereign?

Beyond Constitutional Boundaries

In earlier chapters, we found it useful to make a sharp conceptual distinction between the productive state and the protective state, and the dual functioning of government in these two conceptually different roles was noted. The productive state is, ideally, the embodiment of postconstitutional contract among citizens having as its objective the provision of jointly shared goods and services, as demanded by the citizens. The discussion of Leviathan in this chapter has been wholly concerned with this part or side of government, measured appropriately by the size of the governmental budget. The analysis has shown that budgetary excess will emerge from democratic process, even if overt exploitation is avoided. To the extent that majoritarian democracy uses governmental process to modify the basic structure of individual rights, which are presumably defined in the legal structure, there is an encroachment on the domain of the protective state. Dominant coalitions in legislative bodies may take it on themselves to change "the law," the basic constitutional structure, defined in a real and not a nominal sense. To the extent that the protective state acquiesces in this constitutional excess, the social structure moves toward "constitutional anarchy" in which individual rights are subject to the whims of politicians.

There is, however, an equally if not more significant overstepping of constitutional boundaries when the agencies of government that properly belong to the protective state, and to this state only, begin to act in putative contractual capacities, at both constitutional and postconstitutional stages. A modern treatment of Leviathan would be seriously incomplete if these possible excesses were not discussed. The protective state has as its essential and only role the enforcement of individual rights as defined in constitutional contract. This state is law embodied, and its role is one of enforcing rights to property, to exchanges of property, and of policing the simple and complex exchange processes among contracting free men. In the game analogy that we have used several times before, the protective state is the umpire or referee, and, as such, its task is conceptually limited to enforcing agreed-on rules.

Few who observe the far-flung operation of the executive arm of the United States government along with the ubiquitousness of the federal judiciary could interpret the activities of either of these institutions as falling within meaningful restrictions of the enforcer. Ideally, these institutions may be umpires in the social game; actually, these institutions modify and change the basic structure of rights without consent of citizens. They assume the authority to rewrite the basic constitutional contract, to change "the law" at their own will. At yet other interfaces, these institutions take on legislative roles and effectively displace representative assemblies in making decisions on "public good"—decisions which can in no way be derived from individual evaluations in some quasi-contractual setting. Democracy can generate quite enough of its own excesses even if decision-makers adhere strictly to constitutional norms for behavior. When these norms are themselves subjected to arbitrary and unpredictable change, by decision-makers who are not representative of the citizenry, the omnivorousness of the state becomes much more threatening.

It is more difficult to measure the growth of Leviathan in these dimensions than in the quantifiable budgetary dimensions of the productive state. There is a complementary relationship here, but the two are conceptually independent. An interfering federal judiciary, along with an irresponsible executive, could exist even when budget sizes remain relatively small. Conversely, as noted, relatively large budgets might be administered responsibly with a judicial system that embodies nonarbitrary decision-making. Historically, we observe a conjunction—relatively large and growing budgets along with increasingly irresponsible interpretations of law. Essentially the same philosophical orientation informs both extensions of governmental powers. Burgeoning budgets are an outgrowth of the American liberal tradition which assigns to government the instrumental role in creating the "good society." The arrogance of the administrative and, particularly, the judicial elite in changing basic law by fiat arises from the same source. If the "good society" can first be defined, and, second, produced by governmental action, then men finding themselves in positions of discretionary power, whether in legislative, executive, or judicial roles, are placed under some moral obligation to move society toward the defined ideal.

There is a fundamental philosophical confusion here, one that must be removed if Leviathan is to be contained. A "good society" defined independent of the choices of its members, all members, is contradictory with a social order derived from individual values. In the postconstitutional stage of contract, those outcomes are "good" that emerge from the choices of men, in both the private and the public sector. The "goodness" of an outcome is evaluated on procedural criteria applied to the means of its attainment and not on substantive criteria intrinsic to such outcome. The politician, who represents the citizenry, however crudely and imperfectly, seeks to attain consensus, to find acceptable compromises among conflicting individual and group demands. He is not engaged in a search for some one "true" judgment, and he is not properly behaving if he seeks to further some well-defined ideal drawn from the brains of his academic mentors. The judge is in a distinctly different position. He does seek "truth," not compromise. But he seeks truth only in the limits of constitutional structure. He looks for, and finds, "the law." He does not make new rules. To the extent that he tries deliberately to modify the basic constitutional contract so as to make it conform to his independently defined ideals, he errs in his whole understanding of his social function, even more than the elected politician who seeks the liberal grail.

False philosophical precepts that are so pervasively held cannot be readily overthrown. If our Leviathan is to be controlled, politicians and judges must come to have respect for limits. Their continued efforts to use assigned authority to impose naively formulated constructs of social order must produce a decline in their own standing. If leaders have no sense of limits, what must be expected of those who are limited by their ukases? If judges lose respect for law, why must citizens respect judges? If personal rights are subjected to arbitrary confiscation at the hands of the state, why must individuals refrain from questioning the legitimacy of government?

Leviathan may maintain itself by force; the Hobbesian sovereign may be the only future. But alternative futures may be described and dreamed, and government may not yet be wholly out of hand. From current disillusionment can come constructive consensus on a new structure of checks and balances.