Government organization and power.

Brown, Stephen P.A. ; Saving, Jason L.

I. INTRODUCTION

Economists have systematically examined the economics of government finance and long ago established that the means of financing government creates inefficiencies in the market. For instance, see Baumol and Bradford (1970) and Diamond and Mirrlees (1971a, 1971b). What may be less well examined and understood is that--apart from the distortions that arise from the means of government finance--government officials can have an incentive to provide a combination of services and taxes that are nonoptimal, as in Niskanen (1997). The ability of government to exercise coercive and pricing power creates these incentives.

A growing number of articles explore the effect of government power on its size. For example, Anderson and Tollison (1988) and Rogers and Rogers (1995) examine the implications of pricing power for government size. In his work on bureaucracy, Niskanen (1971) considered the implications of coercive power for government size. Olson (1991), McGuire and Olson (1996), Niskanen (1997), and McGuire (2001) examine the implications of the joint exercise of coercive and pricing powers by an autocratic government. To our knowledge, however, no previous work explicitly distinguishes between the coercive and pricing powers of government, and none systematically examines all the possible combinations of coercive and pricing powers.

In the present analysis, we develop a simple theoretical framework that allows us to examine on an integrated basis the exercise of government power. The framework abstracts from any distortions in the means of government finance and separates government power into two dimensions--pure coercive power and pure pricing power. Pure coercive power is the ability of a government to compel consumers to accept more of the public good than they desire at each tax price through fiat. Coercive power typically arises when citizens are unable to exit a jurisdiction costlessly, and the government is able to push citizens beyond what they would accept in a voluntary relationship. Pure pricing power is the government's ability to restrict output along a given demand curve and earn rents by doing so. Pricing power typically arises when the government has limited competition and faces a downward-sloping demand curve for its services.

We consider four polar combinations of coercive and pricing power in comparison to the social optimum: Lindahl democracy, monopoly, Niskanen bureaucracy, and autocracy. As shown in Figure 1, the government in Lindahl democracy exercises neither coercive nor pricing power. A monopoly government is able to restrict its output and exercise pricing power to earn rents in a traditional monopolistic fashion, but it cannot engage price discrimination or expropriation to obtain any remaining consumer surplus from its citizens. A Niskanen bureaucracy can exact all of the surplus that consumers obtain from the public goods it provides, but it must use the surplus in the production of public goods. It cannot exercise pricing power and restrict output to obtain rents as a monopoly would. An autocracy exercises both coercive and pricing power.

In practice, democratic governments are likely to have both coercive and pricing powers, rather than to conform to the Lindahl ideal. Our analytical framework allows us to clarify and reconcile the differences between some of the previous contributions to the literature on government power and size in democracy. A government can simultaneously use its coercive power to grow too large and its pricing power to restrict its output and earn rents. Furthermore, government size or rents alone may be poor indicators of the extent to which government fails to achieve optimality in its provision of services. Government power may be used to generate rents, provide too much government service, or a combination of both.

II. SOCIAL OPTIMUM

Consider a jurisdiction or economy with n identical individuals and two goods--a purely public good and a purely private good. (1) For each individual, utility ([U.sub.i]) is a function of the amount of the purely public good that is consumed jointly (G) and the amount of the purely private good that is consumed by the individual ([X.sub.i]):

(1) [U.sub.i] = [U.sub.i](G,[X.sub.i]) for i=1,2,...,n.

The production opportunities available to the economy are described by an implicit production function for the two goods

(2) F(G,X) = F

where X = [SIGMA] [X.sub.i].

Maximizing utility for a representative individual subject to production, aggregating over n individuals, and substituting the ratio of marginal costs for the marginal rate of transformation yields the familiar optimality condition for provision of a public good, as in Samuelson (1954),

(3) n([U.sub.Gi]/[U.sub.Xi]) = [MC.sub.G]/[MC.sub.X],

where [MC.sub.G] is the marginal cost of providing the public good, [MC.sub.X] is the marginal cost of providing the private good, and [U.sub.Gi] and [U.sub.Xi] are defined as [partial][U.sub.i]/[partial]G and [partial][U.ub.i]/[partial][X.sub.i], respectively. (2)

III. LINDAHL DEMOCRACY

The Lindahl rule is a particularly appropriate model of democracy when one is considering a society of n identical individuals and wants to abstract from the possibility of the redistributional coalitions that can form under majority rule. Under the Lindahl rule, unanimous consent is used to make decisions about the provision of public goods.

Each individual faces the budget constraint

(4) [Y.sub.i] = tG+[PX.sub.i] for I + 1,2,...,n,

where [Y.sub.i] is individual income endowment, t is the tax price that each individual faces per unit of public good, and P is the price of the private good. For each individual i, maximizing utility (1) subject to the budget constraint (4) yields

(5) [U.sub.Gi]/[U.sub.Xi] = t/P.

Simplifying assumptions about production allow a sharper focus on the organizational factors that shape government size. Under the assumption that the private good is produced by a competitive industry operating at constant costs, the price of the private good will equal the marginal cost of producing it,

(6) P = [MC.sub.X].

Similarly, under the assumptions that production of the public good is characterized by constant costs and the costs of providing the public good are distributed across the n identical individuals at the per unit tax rate t, the tax revenue per unit of output equals marginal cost,

(7) nt = [MC.sub.G].

Aggregating the optimality condition (equation [5]) over n individuals and incorporating the equalities (6) and (7) yields

(8) n([U.sub.Gi]/[U.sub.Xi]) = nt/P = [MC.sub.G]/[MC.sub.X].

As shown by (8), the Lindahl rule satisfies the conditions for optimal provision of the public good under conditions of constant cost.

The Lindahl optimality conditions are illustrated in Figure 2. In the upper panel, the representative individual faces the budget constraint labeled [Y.sub.1] and maximizes utility by selecting the combination [X.sup.*] and [G.sup.*] along the indifference curve [U.sub.3]. For n individuals taken together, the demand curve shown in the lower panel shows the tax and output combinations that satisfy the utility maximization condition. The quantity of the public good provided is [G.sup.*] and the per unit tax revenue, [nt.sup.*], equals the marginal cost of producing the public good, [MC.sub.G].

IV. MONOPOLY

If the government faces no competition in providing the public good, it will not necessarily produce the optimal amount of the public good at the optimal tax price as determined by the Lindahl rule. (3) Instead, the government can obtain monopoly rents by exercising pricing power and restricting output of the public good while raising its tax price, as in Downs (1961) and Anderson and Tollison (1988). Government rents can be represented in this framework as the difference between the tax revenue collected and the cost of producing the public good,

(9) [PI] = nt(G)-[C.sub.G],

where [PI] is the government rent obtained from provision of the public good, and [C.sub.G] is the total cost of its production.

Following Anderson and Tollison (1988), we consider a government that has pricing power but can neither price discriminate because it lacks information to do so and nor exercise coercive power because a fiscal constitution--such as suggested by Brennan and Buchanan (1980)--prevents such activity. Consequently, the monopoly government maximizes the rent from provision of the public good by restricting output of the public good to the point where the marginal tax revenue obtained from provision of the public good equals marginal cost,

(10) nt+[nt.sub.G]G=[MC.sub.G],

where [t.sub.G] is defined as [partial]t/[partial]G. Because the tax price citizens are willipg to pay falls with increased provision of the good, [t.sub.G] is negative, and the aggregate tax price (nt) is greater than the marginal tax revenue.

Assuming government cannot compel its citizens to pay for more of the public good than they would choose to purchase at any particular tax price, taxpaying voters must still agree that the quantity of the public good being provided is the quantity that would maximize individual utility at the tax price each individual faces for an additional unit of the public good. Therefore, the utility-maximizing condition shown in equation (5) would still apply. Aggregating equation (5) over n individuals and incorporating equations (6) and (10) yields

(11) n([U.sub.Gi]/[U.sub.Xi]) = nt/P > (nt + [nt.sub.G]G)/P

=[MC.sub.G]/[MC.sub.X].

As shown by the equalities and inequality (11), a government that exercises pricing power but does not engage in coercion (or price discrimination) will produce less of the public good than is optimal to earn rents. (4)

The monopoly conditions are illustrated in Figure 2. For n individuals taken together, the demand curve shown in the lower panel shows the tax and output combinations that satisfy the utility maximization condition. For the quantity of the public good, [G.sub.LM], the marginal tax revenue equals the marginal cost of producing the public good, [MC.sub.G]. The corresponding monopoly price is [nt.sub.m]. At the tax price [t.sub.m], the representative individual faces the budget constraint labeled [Y.sub.2] in the upper panel and maximizes utility by selecting the combination [X.sub.LM] and [G.sub.LM] along the indifference curve [U.sub.2]. As shown, monopoly provision of the public good yields a lower level of utility for the representative taxpayer, which must be the case because the tax price of the public good is higher than in the case of Lindahl democracy.

V. NISKANEN BUREAUCRACY

Conceiving of a government with pricing power but no coercive power is difficult. In fact, popular writers often describe government monopolies in the language of coercion. (5) Yet the simple monopoly model presented above ignores the exercise of coercive power. In contrast, Niskanen's model of bureaucracy captures the coercive potential of government but assumes the government uses the coercive power to maximize output rather than to earn rents.

According to Niskanen (1971), a government bureaucracy will expand its output beyond the optimal level to the point where consumers are indifferent between receiving and not receiving the public good under a balanced budget constraint. (6) Bureaucracies engage in this behavior because the compensation (either monetary or psychic) is an increasing function of size. In Niskanen's (1971) analysis, a bureaucracy cannot push its output beyond the point at which consumer surplus is exhausted, however, because an exogenous entity from which the bureau obtains its funds knows the maximum it is willing to pay for each level of service and will not permit the bureau to spend any more. (7)

Alternatively, one might argue that taxpaying voters will abolish a government that pushes output beyond the point at which consumer surplus is exhausted because they would prefer no government at all to such an excessive provision of the public good. (8) Because elimination of government strips existing government officials of their ability to choose prices and quantities for the public good (and extract whatever rents are available by so doing), the officials adopt the strategy of producing up to the point where consumer surplus is exhausted but no more. (9)

Therefore, taken to an extreme, a Niskanen bureaucracy has the ability to capture a sizable consumer surplus, but it uses that surplus to produce more of the public good. The surplus is equivalent to that obtained through perfect inframarginal price discrimination. (10) The involuntary nature of direct expropriation is more consistent with the exercise of coercive power, however, and is the mechanism envisioned here.

A Niskanen bureaucracy that exercises coercive power but no pricing power can be represented by the maximization of the public good, G, subject to a utility constraint, [U.sub.i] [less than or equal to] [U.sub.i], and a balanced budget constraint, ntG [greater than or equal to] [C.sub.G]. (11) Solution of the problem yields the following Kuhn-Tucker conditions.

(12) [U.sub.Gi]/[U.sub.Xi]

[less than or equal to] t/P, ([U.sub.Gi]/[U.sub.Xi]-1/P)([U.sub.i]-[U.sup.*.sub.i])

= 0 for each individual i

(13) nt [greaterthan or equal to] [MC.sub.G], (nt-[MC.sub.G])([partial]G/[partial]t) = 0

Aggregating the Kuhn-Tucker conditions given in (12) over n individuals forms a downward-sloping boundary at G ng(t). At points along this boundary, the representative individual is exactly indifferent between no government and the public good and tax price combinations implied by the boundary. In other words, points along the boundary exactly exhaust the consumer surplus that the representative individual obtains from provision of the public good. The boundary might be called the Hicks-Niskanen demand for the public good.

The Hicks-Niskanen demand curve represents the maximal combinations of tax price and quantity of the public good that the representative individual will accept before demanding abolishment of the government. As shown in (12), the representative individual has a lower marginal rate of substitution for the public good along the boundary than the tax price of the public good relative to the price of the private good. This inequality means that the representative individual would be better off with a smaller provision of the public good at each given tax price. In this sense, Niskanen bureaucracy can be seen as coercive.

The Kuhn-Tucker conditions given in (13) form a lower horizontal boundary at nt = [MC.sub.G]. Therefore, a government that seeks to maximize production of the public good must be content to just cover its costs.

The quantity of the public good, G, is maximized along the two boundaries. Combining the boundary conditions obtained from (12) and (13) together with equation (6) yields

(14) n([U.sub.Gi]/[U.sub.xi]) <nt/P = [MC.sub.G]/[MC.sub.X].

As shown by the inequality and equality (14), the Niskanen bureaucracy that exercises its coercive power to maximize output will produce more of the public good than is optimal. (12) Because consumers are indifferent between the Niskanen bureaucratic outcome and no public good at all and prefer Lindahl democracy to monopoly, and monopoly to no public good at all, the outcome under Niskanen bureaucracy is necessarily worse than those obtained under either monopoly or Lindahl democracy.

The Niskanen bureaucratic outcome is illustrated in Figure 2 under an assumption of constant marginal utility for the private good. (13) In the upper panel, the indifference curve U, shows the combinations of the public and private goods that leave the representative individual no better off than zero provision of the public good. For n individuals taken together, the Hicks-Niskanen demand curve shown in the lower panel traces out the tax price and public good combinations that would leave the representative taxpayer on the indifference curve [U.sub.1]. The maximum provision of the public good is [G.sub.CB] with corresponding tax revenues of [nt.sup.*] per unit of public good. The representative individual faces a per unit tax rate of [t.sup.*] which is consistent with the budget constraint [Y.sub.1], shown in the upper panel. Every taxpayer is forced to consume the combination [X.sub.CB] and [G.sub.CB] along the indifference curve [U.sub.1], but every individual would prefer the (optimal) combination [X.sub.*] and [G.sub.*] along the indifference curve [U.sub.3] or the monopoly combination [X.sub.LM] and [G.sub.LM] along the indifference curve [U.sub.2]. Niskanen bureaucracy yields a lower level of utility for the representative taxpayer than can be achieved under either Lindahl democracy or monopoly.

VI. MCGUIRE-OLSON AUTOCRACY

A government's use of coercive power simply to expand output of the public good beyond the optimal level provides no clear benefit to either taxpayers or government officials. Niskanen (1971) resolved this problem to some extent by assuming bureaucratic salaries are tied to output, but it is far from clear (at least for the U.S. government) that bureaucrats in small departments actually receive less monetary and nonmonetary compensation than do bureaucrats in large departments. Moreover, Niskanen's approach limits the rent-earning capabilities of the government by implicitly assuming consumer surplus can be used only for production.

Olson (1991) discusses the coercive and pricing capabilities of an autocracy. Olson describes autocrats as "stationary bandits" who take income from their citizens (as taxes)--an action that is clearly coercive. He also argues that an autocrat has an incentive to provide public goods that increase private production, but he expects the quantity to be less than optimal because the autocrat's tax rate yields less than the full marginal product of the public good.

McGuire and Olson (1996), Niskanen (1997), and McGuire (2001) further develop Olson's concepts of autocracy. In their models, the autocrat expropriates private income through an income tax and provides a public good that is a complement in private production. (14) The extent to which a rational autocracy taxes its citizens is limited by the discouraging effects that income taxation has on the process by which income is generated. Provision of the public good is limited by the extent to which it yields additional tax revenues by raising private income.

VII. AN ALTERNATIVE REPRESENTATION OF AUTOCRACY

In the model presented here, however, the means of government finance are nondistortionary, and the public good enters utility directly. Consequently, the Hicks-Niskanen demand curve for the public good represents the maximal combinations of the public good and the tax prices that the government can obtain through coercion. In this framework, autocracy can be represented by the maximization of government rent, II, subject to the utility constraint [U.sub.i] [less than or equal to] [U.sub.i], which is similar to Brennan and Buchanan's (1980) concept of Leviathan. Solution of the problem yields the Kuhn-Tucker conditions for taxpayer utility (12) and the rent-maximizing condition (10), as shown above.

Once again, aggregating the Kuhn-Tucker conditions given in (12) over n individuals forms a downward-sloping boundary at G = ng(t) that is the Hicks-Niskanen demand for the public good. This demand curve represents the maximal combinations of tax rates and provision of the public good that the representative individual will accept before demanding abolishment of the government. To maximize its rent, the autocratic government selects a tax and output combination along the Hicks-Niskanen demand curve that satisfies equation (10). Combining the boundary condition obtained from (12) with equations (10) and (6) yields

(15) n([U.sub.Gi]/[U.sub.Xi] <nt/P> (nt+[nt.sub.G]G)/P = [MC.sub.G]/[MC.sub.X].

Autocracy yields the same level of utility for the representative taxpayer as Niskanen bureaucracy, which is less than can be achieved under either Lindahl democracy or monopoly. However, government rents are twice those earned by a monopoly government. Interestingly, the joint exercise of coercive and pricing power in autocracy can be said to result in both too much government--in the sense of Niskanen (1971)--and too little government--in the sense of Downs (1961) and Anderson and Tollison (1988).

As shown by the inequalities and equality (15), an autocracy may produce more or less of the public good than is optimal. Unless the public good is an inferior good, however, the autocracy will produce less of the public good than is optimal. (15) In either case, the combination of coercive and pricing power substantially reduces taxpayer welfare and yields rents to the government that are much greater than could be earned through pricing power alone. In exercising its coercive power, an autocracy produces more of the public good than taxpayers would desire at each tax price and, in doing so, substantially reduces welfare. In exercising its pricing power, an autocracy reduces output along the Hicks-Niskanen demand curve. Exercising these powers jointly yields greater rents than can be obtained through pricing power alone. (16)

The autocratic outcome is illustrated in Figure 2. Again assuming constant marginal utility for the private good, the maximal price and output combinations are achieved along the Hicks- Niskanen demand curve shown in the lower panel-which holds the representative taxpayer on the indifference curve [U.sub.1] shown in the upper panel. The associated marginal revenue curve is labeled H-N MR and, in this example, is superimposed on the (normal) demand curve. (17) For the quantity of the public good [G.sub.CM], the marginal tax revenue equals the marginal cost of producing the public good, [MC.sub.G]. The corresponding monopoly tax price is [nt.sub.m]. At the tax price [t.sub.m], the representative individual faces the budget constraint labeled [Y.sub.2] in the upper panel and would maximize utility by selecting the combination [X.sub.LM] and [G.sub.LM] along the indifference curve [U.sub.2], but is forced to consume the combination [X.sub.CM] and [G.sub.CM] along the indifference curve [U.sub.1].

VIII. COERCIVE AND PRICING POWERS IN DEMOCRACY

Above we describe Lindahl democracy as a polar combination in which the government has neither coercive nor pricing power. In practice, however, democracies are created with the coercive power necessary to prevent free riding in the provision of public goods, as in Olson (1965), and many democracies have pricing power because they face limited competition. Any government (whether it is an autocracy or a democracy) that is capable of acting in its own interest (or the interests of a subset of its citizens) has an economic incentive to exercise both its coercive and pricing powers. Numerous authors--such as Downs (1957, 1961), Tullock (1959, 1967), Barzel (1973), Breton (1974), Barzel and Deacon (1975), and Brennan and Buchanan (1980)--have described democratic institutions as using what amounts to coercive power to achieve outcomes that are similar to those found along the Hicks-Niskanen demand curve.

Tullock (1959) developed the concept of logrolling to show how majority rule could result in the overprovision of public goods. Barzel (1973) and Barzel and Deacon (1975) extend the analysis to government provision of private goods. With logrolling, the coercive power of government is used by the majority to obtain revenue from a minority of citizens (who are unable to exit the jurisdiction) that is in excess of the value that the minority places on government goods. In a manner similar to Niskanen bureaucracy, the excess revenue is consumed in the production of additional government goods.

Weingast and Moran (1983) show that a congressional committee that represents a minority of constituents can create a tyranny of the minority. If the committee is capable of controlling which bills affecting the bureaucracy for which it has oversight can be considered by the whole Congress, the committee can thwart majority rule by carefully selecting which bills it puts forward. An outcome similar to Niskanen bureaucracy is possible, but so is underprovision of the good produced by the bureaucracy.

Brennan and Buchanan (1980) argue that rational voter ignorance cedes power to a democratic government which becomes a Leviathan. The Leviathan government uses this power to force citizens to accept more government service than they want at each given tax price, allowing the government to grow. Brennan and Buchanan assume a Leviathan government seeks to maximize its revenue, which is the point of unitary elasticity on the Hicks-Niskanen demand curve (assuming the limit on government expansion is the size at which citizens become indifferent between no government at all and the expanded levels of government services and taxes). Depending on the shape of the demand curve and the cost of providing government goods, revenue maximization could be consistent with a range of outcomes along the Hicks-Niskanen demand curve from autocracy to Niskanen bureaucracy.

Downs (1961) and Anderson and Tollison (1988) recognized that a democratic government has the ability to earn rents through its pricing power. Some of these rents may be dissipated by the ruling party to retain its power, as in Breton (1974), or to generally serve special interests, as in Downs (1957), Bernholz (1966), and Tullock (1967). The use of pricing power to earn rents reduces government size, whether the rents are retained by the government or are dissipated by paying citizen voters to maintain power.

The apparent conflict in the literature about whether government power results in too much or too little government can be resolved by distinguishing between coercive and pricing powers. Those authors emphasizing the government's use of power to grow too large are mainly concerned with its coercive power. Those authors emphasizing the government's use of power to restrict its output and earn rents are primarily concerned with its pricing power. As we have seen above, a government can be too large--in the sense that it exercises its coercive power and provides more government than citizens would want at each tax price-and simultaneously too small, in the sense that it uses its pricing power to restrict output and earn rents.

IX. POLITICAL CORRUPTION

The use of government power to obtain rents for personal gain may be viewed as political corruption, as in Shleifer and Vishny (1993). The distinction between coercive and pricing powers may have a bearing on political corruption. As shown, the larger rents obtained through the joint exercise of coercive and pricing powers provides a greater means for personal gain, but the larger government obtained by emphasizing coercive power over pricing power may provide a greater opportunity to conceal personal gains.

X. LIMITING GOVERNMENT POWER

A number of institutions can limit a democratic government's ability to exercise its coercive and pricing powers. These institutions include constitutional restraint, as in Brennan and Buchanan (1977) and Buchanan and Tullock (1962); interjurisdictional competition, as in Tiebout (1956), Buchanan (1965), and McGuire (1972, 1974); and political competition, as in Breton and Wintrobe (1975). (18) We briefly examine the aspects of restraining government power that are affected by distinguishing between the coercive and pricing powers of government.

With sufficient mobility between a large number of jurisdictions, citizens can exit jurisdictions with allocations they find nonoptimal, and competition could prevent even the stationary bandit rulers of competing principalities from obtaining autocratic rents. (19) If there are many competing jurisdictions and citizens are able to exit costlessly, each government will face a perfectly elastic demand and also be unable to coerce its citizens into paying for more of the government good than they desire at each tax price. Because interjurisdictional competition can be an effective means of restraining government power, the centralization of government power away from the local level toward the state, national, or international level could reduce citizen welfare even with representative democracy, as in McGuire (1998). (20)

In contrast, the effectiveness of constitutional restraint is likely to be less affected by centralization and reduced interjurisdictional competition. Consequently, constitutional restraint is likely to be of greater importance in limiting the behavior of higher levels of government, where there is likely to be less extensive interjurisdictional competition.

Even in a constitutionally constrained democracy with interjurisdictional competition, government managers have an economic incentive to exercise both the government's coercive and pricing powers. To ensure a government's performance, citizens must monitor both rents and output. With government having both coercive and pricing dimensions to its power, neither government size nor rents alone are likely to prove reliable as indicators of the extent to which the exercise of government power has prevented the maximization of taxpayer utility. For instance, the loss in utility associated with shifting from Lindahl democracy to an outcome along the Hicks-Niskanen demand curve is consistent with a range of rent and output combinations--from autocratic behavior that yields maximum rent and little change in output of the public good to a bureaucratic behavior that yields no rent and maximum output of the public good.

If citizens monitor government rents (to ensure they are zero) but not output, a government with coercive power could behave like a Niskanen bureaucracy. If citizens monitor government output (to ensure that the provision of the public good maximizes utility at a given tax rate) but not rents, the government could behave like a monopoly. If citizens follow the common practice of measuring government size as total revenue or expenditures and seek a smaller government, a government with coercive power that has production costs and is operating on the elastic portion of its demand curve can earn at least some autocratic rents by restricting its output and reducing its total revenue and expenditure below the bureaucratic level.

Monitoring of the government is itself a public good and, as such, is likely to be underprovided through voluntary exchange. The complexity of monitoring the government, particularly in a world with large governments providing multiple goods that must be monitored in two dimensions, suggests that monitoring itself may be carried out by a part of the government. That part is typically elected officials who can compete for the job of monitoring government activity by offering to restrain government size and rents. (21)

These elected officials have their own incentives to develop and exercise the government's coercive and pricing powers, as in Weingast and Moran (1983). Citizens, the media, competing candidates, other jursidictions, and other branches and levels of government act to monitor elected officials and/or provide competition, as in Breton and Wintrobe (1975). At the same time, bureaucrats, politicians, and special interest groups act to reduce competition and weaken monitoring. (22) Tactics to insulate government from competition can vary from simple to elaborate. Some U.S. states have enacted statutes that create significant obstacles to minor party candidates for public office. The French government reduces the mobility of its citizen by limiting their exposure to English while asserting the obstacle benefits its citizenry by preserving French culture. The exact degree to which the interplay between competitive and anticompetitive forces acts to restrain or promote government power is beyond the present analysis .

XI. CONCLUDING REMARKS

Government power can be thought of as having two dimensions: coercion and pricing power. The exercise of coercive power allows a government to compel its citizens to pay for more of the government good than they want at a given tax price. In contrast, the exercise of pricing power allows the government to restrict its output along a given demand curve to earn rents.

As we have modeled them, Lindahl democracy, monopoly, Niskanen bureaucracy, and autocracy represent four polar combinations of these two aspects of government power. A Lindahl democracy does not exercise either power and produces the optimal quantity of public service. A monopoly government uses its pricing power to restrict output and earn rents. A bureaucratic government uses its coercive power to produce too much government service for the implied tax prices. An autocratic government exercises both its coercive and pricing power to produce too much government service for the implied tax prices and (at the same time) restrict its output to earn rents.

In practice, democratic governments are likely to have both coercive and pricing powers, and numerous authors suggest democratic governments are likely to use their powers to achieve various nonoptimal results that seem to range from an emphasis on coercive power and too much government to an emphasis on pricing power and too little government. To some extent, these differing conclusions can be reconciled by distinguishing between a government's coercive and pricing powers. Through the simultaneous exercise of these two powers a government can be both too large and restrict its output to earn rents.

Because the combination of coercive and pricing powers provides a government with two means to set its taxes and services at nonoptimal levels, neither size nor rents alone may be good indicators of the extent to which government fails to achieve optimality. Government power can be used to generate rents, provide too much of the public good, or some combination of both. Constitutional restraint, as well as interjurisdictional and political competition reduces government power in both of its dimensions and increase taxpayer welfare.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

(1.) Because a pure public good does not suffer from congestion, the population of the jurisdiction must be exogenously determined. See McGuire (1972, 1974). The assumptions that the population is comprised of n identical individuals and the government provides a pure public good simplifies the analysis but leaves the conclusions unaltered. Substantially similar results to those shown can obtained for a heterogeneous population with the government providing a congestible public good or a purely private good. The exercise of coercive and pricing powers is not dependent on the type of good the government provides.

(2.) The marginal rate of transformation between G and X is also defined as [MC.sub.G]/[MC.sub.X].

(3.) Short of autocracy, there are several reasons why a government may have pricing power. Perhaps taxpayers cannot monitor government activities or fail to overcome the collective action problems associated with doing so. Alternatively, taxpayers might be fully aware of the government's activities but are unable to call government officials to account.

(4.) The rents may be retained by government officials; dissipated to retain power, as in Breton (1974); used to serve special interests as in Downs (1957), Bernholz (1966), and Tullock (1967); or distributed as a combination of all three.

(5.) For an example, see DiLorenzo (1999).

(6.) Brennan and Buchanan (1980) echo this theme when they argue Leviathan will appropriate all surplus for its own use unless constitutional constraints are imposed on it.

(7.) Other models take a somewhat different approach to the bureaucracy. For example, Weingast and Moran (1983) provide evidence that bureaucracies act in accordance to the will of the congressional committees that oversee them. The resulting behavior could be identical to that of a Niskanen bureaucracy or substantially different, depending on the incentives facing the politicians on the committee.

(8.) For analytical simplification, we assume voters can costlessly abolish the government but cannot use the threat of such abolition to foster more efficient behavior by bureaucrats. For a more rigorous analysis of this issue and the types of situations in which these assumptions are appropriate, see Romer and Rosenthal (1979).

(9.) Our reformulation of the bureaucratic model eliminates the need for an exogenous source of funds and more firmly integrates the bureaucratic model into our analysis.

(10.) Consistent with a pure public good, the price discrimination is by unit of output--not by individual.

(11.) Niskanen modeled the government bureaucracy as maximizing revenue. To examine the implications of coercive power in the absence of rent-seeking, we assume the Niskanen bureaucracy maximizes output of the public good. Our assumption will yield somewhat different conclusions than Niskanen reached.

(12.) In fact, a bureaucracy seeking to maximize output under a Lindahl rule will produce the optimal quantity of the public good.

(13.) This assumption simplifies the graphic analysis but does not alter the conclusions. Under this assumption, the total revenue (price times quantity) at a point along the Hicks-Niskanen demand curve is the same as the area under the normal demand curve at the same quantity of the public good. This relationship indicates that Hicks-Niskanen demand represents the price-quantity options that can be obtained under perfect inframarginal price discrimination or direct expropriation that is limited to exhaustion of the consumer surplus.

(14.) In these analyses, coercion and perfect infra-marginal price discrimination are not equivalent.

(15.) In the McGuire-Olson framework, an autocrat also can have an incentive to produce more or less of the public good than is optimal.

(16.) Olson's (1991) autocrat retains the rents, but autocratic rents also may be partially dissipated to retain power, as in Niskanen (1997), or used to serve special interests.

(17.) If the private good were not characterized by constant marginal utility, the Hicks-Niskanen marginal revenue curve would not be superimposed on the (normal) demand curve.

(18.) Taylor's (2000) survey of the empirical literature on governmental competition finds that interjurisdictional and political competition enhance government performance.

(19.) Epple and Zelenitz (1981) have shown that interjurisdictional competition eliminates rents if all factors of production are costlessly mobile across jurisdictions.

(20.) Arguments for centralization typically involve the interjurisdictional externalities that can arise when there are competing governments. In some cases, however, the gains from reducing potential interjurisdictional externalities may he more than offset by the welfare losses that can result from the increased centralization of government power. In other words, market failure can be replaced by an even more severe government failure.

(21.) Some candidates compete by creating and appealing to redistributional coalitions.

(22.) See Olson (1965, 1982).

REFERENCES

Anderson, G. M., and R. D. Tollison. "Legislative Monopoly and the Size of Government." Southern Economic Journal, 54(3), 1988, 529-45.

Barzel, Y. "Private Schools and Public School Finance." Journal of Political Economy, 81(1), 1973, 174-86.

Barzel, Y., and R. T. Deacon. "Voting Behavior, Efficiency and Equity." Public Choice, 21(1), 1975, 51-58.

Baumol, W J., and D. F. Bradford. "Optimal Departures from Marginal Cost Pricing." American Economic Review, 60(1), 1970, 265-83.

Bernholz, P. "Economic Policies in Democracy." Kyklos, 19(1), 1966, 48-80.

Brennan, G., and J. M. Buchanan. "Towards a Tax Constitution for Leviathan." Journal of Public Economics, 8(3), 1977, 255-73.

_____. The Power to Tax: Analytical Foundations of a Fiscal Constitution. Cambridge: Cambridge University Press, 1980.

Breton, A. The Economic Theory of Representative Government. Chicago: Aldine, 1974.

Breton, A., and R. Wintrobe. "The Equilibrium Size of a Budget Maximizing Bureau." Journal of Political Economy, 83(3), 1975, 195-207.

Buchanan, J. M. "An Economic Theory of Clubs." Economica, 32(1), 1965, 1-14.

Buchanan, J. M., and C. Tullock. Calculus of Consent. Ann Arbor: University of Michigan Press, 1962.

Diamond, P. A., and J. A. Mirrlees. "Optimal Taxation and Public Production I: Production Efficiency." American Economic Review, 61(1), 1971a, 8-27.

_____. "Optimal Taxation and Public Production II: Optimal Taxation and Public Production." American Economic Review, 61(2), 1971b, 261-78.

DiLorenzo, T. J. "Monopoly Government." Free Market, Ludwig Von Mises Institute, August 1999, 1-2.

Downs, A. An Economic Theory of Democracy. New York: Harper and Row, 1957.

_____. "In Defense of Majority Voting." Journal of Political Economy, 69(2), 1961, 192-99.

Epple, D., and A. Zelenitz. "The Implications of Competition among Jurisdictions: Does Tiebout Need Politics?" Journal of Political Economy, 89(6), 1981, 1197-217.

McGuire, M. C. "Private Good Clubs and Public Good Clubs: Economic Models of Group Formation." Swedish Journal of Economics, 74(1), 1972, 84-99.

_____. "Group Segregation and Optimal Jurisdictions." Journal of Political Economy, 82(1), 1974, 112-32.

_____. "Mancur Lloyd Olson, Jr. 1932-1998: Personal Recollections." Eastern Economic Journal, 24(3), 1998, 252-63.

_____. "Redistribution, Political Power, and Public Goods." Manuscript, Department of Economics, University of California, Irvine, 2001.

McGuire, M. C., and M. Olson. "The Economics of Autocracy and Majority Rule: The Invisible Hand and the Use of Force." Journal of Economic Literature, 34(1), 1996, 72-96.

Niskanen, W A. Bureaucracy and Representative Government. Chicago: Aldine-Atherton, 1971.

_____. "Autocratic, Democratic, and Optimal Government." Economic Inquiry, 35(3), 1997, 464-79.

Olson, M. The Logic of Collective Action. Cambridge, MA: Harvard University Press, 1965.

_____. The Rise and Decline of Nations. New Haven, CT: Yale University Press, 1982.

_____. "Autocracy, Democracy, and Prosperity," in Strategy and Choice, edited by R. Zeckhauser. Cambridge, MA: MIT Press, 1991, 131-57.

Rogers, D. L., and J. H. Rogers. "Political Competition, Causal Relationships between Taxes and Spending and Their Influence on Government Size: Evidence from State-Level Data." International Finance Discussion Paper no. 500, Federal Reserve Board of Governors, 1995.

Romer, T, and H. Rosenthal. "Bureaucrats versus Voters: On the Political Economy of Resource Allocation by Direct Democracy." Quarterly Journal of Economics, 93(4), 1979, 563-87.

Samuelson, P. A. "The Pure Theory of Public Expenditure." Review of Economics and Statistics, 36(4), 1954, 387-89.

Shieifer, A., and R. W Vishny. "Corruption." Quarterly Journal of Economics, 108(3), 1993, 599-617.

Taylor, L. L. "The Evidence on Government Competition." Economic and Financial Review, Federal Reserve Bank of Dallas, second quarter 2000, 2-10.

Tiebout, C. "A Pure Theory of Local Expenditures." Journal of Political Economy, 64(5), 1956, 416-24.

Tullock, G. "Some Problems in Majority Voting." Journal of Political Economy, 67(6), 1959, 571-79.

_____. Toward a Mathematics of Politics. Ann Arbor: University of Michigan Press, 1967.

Weingast, B. R., and M. J. Moran. "Bureaucratic Discretion or Congressional Control? Regulatory Policy Making by the Federal Trade Commission." Journal of Political Economy, 91(5), 1983, 765-800.

Jason L. Saving *

* We would like to thank Susan August Brown, Joseph Haslag, Martin C. McGuire, Lori L. Taylor, two anonymous referees, and participants in seminars at the Federal Reserve Bank of Dallas and the University of Nevada Las Vegas for helpful comments and suggestions. The views expressed are solely those of the authors and should not be attributed to the Federal Reserve Bank of Dallas or the Federal Reserve System.

Brown: Director of Energy Economics and Microeconomic Policy Analysis, Research Department, Federal Reserve Bank of Dallas, P.O. Box 655906, Dallas, TX 75265.5906. Phone 1-214-922-5152, Fax 1-214-922-5194, E-mail [email protected]

Saving: Senior Economist, Research Department, Federal Reserve Bank of Dallas, P.O. Box 655906, Dallas, TX 75265-5906. Phone 1-214-922-5167, Fax 1-214-922-5194, E-mail [email protected]