See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228703581

The State and the citizen

Article · April 1998

CITATIONS READS

8 20,419

3 authors, including:

Adam Przeworski
New York University
229 PUBLICATIONS 29,904 CITATIONS

SEE PROFILE

All content following this page was uploaded by Adam Przeworski on 27 February 2015.

The user has requested enhancement of the downloaded file.

 The State and the Citizen

Adam Przeworski
Department of Politics
New York University
przewrsk@is2.nyu.edu

March 6, 1998

Paper prepared for the International Seminar on “Society and the Reform of the State,” Sao Paulo, Brazil,
March 26-28, 1998. I appreciate comments from Nuria Cunill Grau, Fernando Limongi, and Guillermo
O’Donnell.
Abstract

Why in some democracies, in Latin America but also beyond, the state unequally enforces the law?
The paper examines the institutional mechanisms through which rights of all citizens can be protected in
democracies. “Horizontal” mechanisms consist of mutual checks and balances among different branches
of government. “Vertical” mechanisms are those through which citizens exercise their control over the
government. One vertical mechanism are the elections. But elections are a blunt instrument and
additional mechanisms are necessary to enable popular control. Yet these mechanisms may accentuate
political inequality, rather than reduce it.
The alternative hypothesis is that unequal protection of rights is due not to the internal structure of the
state but to the social and economic conditions which it confronts. Perhaps in a highly unequal society, no
state institutions can enforce the law universally, regardless how well the horizontal and the vertical
mechanisms are designed. Thus reform of the state, while necessary, may be insufficient to overcome
political inequality in the face of economic and social inequality.
Introduction

Why in some democracies, in Latin America but also beyond, the state unequally enforces the law? The
inspiration for asking this question comes fromO’Donnell (1993, 1997b), who painted a map of Latin America
in colors. In green areas peace and order prevail, but as the color shades into brown, the state disappears
and lawlessness reigns.1 The favelas of Brazil, rural areas of Colombia, “inner cities” of the United States,
shanty towns of South Africa are areas of widespread private violence, separated by walls, physical not only
economic, from the “gated communities” inhabited by the rich. The state often just patrols the walls.
Note that this is a map of the degree of social order, not economic or social inequality per se. The state can
do only so much to reduce economic inequality. But the essence of the democratic state is political equality
(O’Donnell 1997b). And this equality entails not only the positive rights to vote or to voice opinions, but also a
right to an effective government, what Bresser Pereira (1997a: 10) calls the “republican rights”: “the right every
citizen has that the public patrimony be effectively public, i.e., of everybody and for everybody.” Why is it, then,
that in some democracies the state does not generate political equality, failing to provide even a modicum of
social order for some?
To put the problem in its context, it is useful to begin ab ovo. When people live together, they face the risk that
conflicts over divergent values or scarce resources would lead them to kill one another. The solution that
emerged historically to prevent this danger was the state. I say “emerged historically” to free this formulation
from its Hobbesian framework: the conflicts that threaten individual existence need not be among individuals
but between pre-existing groups, even in the absence of the state conflict is need not break out, and, finally,
the emergence of the state need not result from any contract but merely from a conquest of monopoly over
force. Nevertheless, the Hobbesian framework usefully highlights the two sources of danger to individual
existence: the emergence of the state constitutes a transposition of the “horizontal” danger of people killing
one another into the “vertical” danger of being killed by the state (Dunn, in press). The Chief of Anti-
Kidnapping Police Unit of the state of Morelos was just arrested for organizing several kidnappings: vertical
danger personified.
Whatever form it assumes, dictatorial or democratic, the state rules. Even in a direct democracy, decisions of
a majority are binding on everyone, including the minority which finds them against their opinions or interests
(Condorcet, 1986). In a representative democracy-- our form of government—these decisions are made and
implemented by elected representatives and by appointed officials to whom the representatives delegate
some of the tasks of governing. The representatives decide what citizens must and cannot do, and they
coerce citizens to comply with their decisions. They decide how long children must go to school, how much
individuals should pay in taxes, with which countries must men war, what agreements must private parties
adhere to, as well as what citizens can know about the actions of governments. And they, or the
bureaucracies to whom they delegate these tasks, enforce such rules, even against the wishes of the
individuals concerned. In this sense, the state “rules”: it dictates and coerces. To citeDunn (1996: 29),
“democracy is one (very broadly defined) form of being ruled.... It is not, and cannot be, an alternative to being
ruled.”

1
A similar map of some well established democracies, notably the United States, would also exhibit a broad color spectrum.
The United States has the highest rate of incarceration in the world, more private than public police, discriminatory conviction rates,
and large areas that used to be called "ghettos" and are now euphemized as "inner cities."

1
Once the state emerges, so does liberalism. To put it in crudely utilitarian terms, the question it poses is when
this transposition of the horizontal into the vertical danger is a good deal for individuals. And the answer must
be that it is not enough that the state does not kill, deprive people of their liberty, or act in otherwise arbitrary
ways. It is a good deal only when the state also effectively prevents people from killing one another and
perhaps, more broadly, prevents them from abusing and exploiting one another in some specific ways. As
Holmes (1995: 6) put it, “Constitutions restrict the discretion of power-wielders because rulers, too, need to be
ruled. But constitutions not only limit power and prevent tyranny, they also construct power, guide it toward
socially desirable ends, and prevent chaos and private oppression.” The state must obey the law and enforce
it, both obey and enforce.
One may object against this distinction by arguing that governments obey the law not just when they do not
violate it but only if they also enforce it universalistically. Yet, as we see below, in practical terms there is an
asymmetry between not violating and enforcing. A system of institutions may be effective in preventing
governments from acting illegally, but at the same time it may make it more difficult for them to act in any way,
including doing things they should do. While order may restrict liberty, protection of liberty may reduce the
ability of the state to maintain order. This is an old issue, and it is old because solutions are not obvious and
the balance is always delicate.
The paper examines the institutional mechanisms through which rights of all citizens can be enforced in
democracies. “Horizontal” mechanisms consist of mutual checks and balances among different branches of
government: the hypothesis here is that if the structure of government is well designed, the particular organs
of government will check and balance each other in such a way that citizens’ rights will be enforced. “Vertical”
mechanisms are those in which citizens exercise their control over the government.2 One vertical mechanism
are elections: the hypothesis is that if elections are freely contested, if participation is widespread, and if
citizens enjoy political liberties, then governments will act in the best interest of the people. But there are also
non-electoral mechanisms through which citizens can influence and control governments. Indeed, since
elections are a blunt instrument, additional mechanisms may be necessary to enable popular control.
Yet political institutions function in a society. The failure to enforce the lawuniversalistically may be due not to
the institutional structure of the state but to the social and economic conditions which it confronts. Perhaps in
a highly unequal society, no state institutions can enforce rights universally, regardless how well the horizontal
and the vertical mechanisms are designed. Thus reform of the state, even if it conceived, as it is in Brazil, in
broadly political not only administrative terms, may be insufficient to overcome political inequality in the face of
economic and social inequality.
The paper is structured along these lines. In the first part, I examine the thesis according to which a
government that is divided will be limited. In the second part, I study the ways in which citizens can induce
governments to pursue their interests, distinguishing between electoral and non-electoral mechanisms. In the
last part, I reformulate the question as a causal one, asking why it is that in many democracies the state
enforces the law unequally and sporadically.

2
The terminology of "horizontal" versus "vertical" follows O'Donnell (1994).

2
Horizontal Mechanism: Checks and Balances

The thesis that a if a government is divided, in the sense that different organs have different functions, then it
will be limited in what it can do to citizens, already present in Locke, has been put forth most vigorously by
Kavka (1986). The idea is that if a government is divided, any action requires cooperation of several
independent organs. To put someone in jail, the legislature must pass a law making a particular class of acts
so punishable, a court must find the individual guilty, and a bureaucracy must incarcerate him. Without a court
verdict, the action by the bureaucracy would be arbitrary; without the law, the verdict of the court would be
moot.3 Yet even if a divided government will be limited in what it can do to citizens, it may be also limited in
what it can do for them. For note the converse: to put someone in jail, it is not enough that the legislature
passes a law, or even that it passes the law and the case is brought to justice and adjudicated: the
appropriate bureaucracy must also cooperate. Hence the perennial tension between protection of rights and
effective governance.
Before analyzing this argument, it is necessary to introduce a distinction, offered byManin (1994), between
separation of powers versus checks-and-balances. The difference is the following. In the pure separation
model, each of the organs of government is functionally restricted in what it does: the legislature and only the
legislature legislates, the executive and only executes, the judiciary and only adjudicates. And, since
Montesquieu’s list of functions is outdated, the central bank and only the bank regulates money supply, the
regulatory agencies and only they set prices for natural monopolies, etc.
In the system of separation of powers, even if each organ is restricted to the exercise of particular functions, it
is not limited in this exercise. Yet this is not a system of equals: the legislature is supreme to all other powers,
since no action can be legitimately taken by any part of the government unless it was enabled by legislation.
The parliament is the unchecked power in this system. Moreover, the parliament is not just an expression of
popular will; it defines it. As Kelsen (1988: 23 [1929]) argued, in a representative democracy “popular
sovereignty” means parliamentary sovereignty:volonté generale is formulated by the legislatures, it is volonté
étatique. In a system of separation of powers, the legislature can decide whatever it wishes, with no limits.
This model of separation is contrasted by Manin with one including checks and balances, in which each
function is performed by more than one organ of government. Checks and balances are a device to limit
legislative sovereignty, and at least in the United States were introduced just for that purpose. The legislature
passes a law but this law must be signed and can be vetoed by the president and must not be found ex ante
or ex post unconstitutional by some other power.4 Hence, the legislature is checked by the president and by
other appropriate organs. The actions of the executive are, in turn, enabled by the legislation and checked by
fiscalizing agencies, by courts, and by the legislature, in the extreme case through the threat of censure. The
courts are empowered by the legislation and controlled through the appointment and the censure procedures
by the legislature. The actions of the central bank are checked by the threat of legislation and by the
appointment and censure procedures. Etc.
One immediate question is whether this system as well does not include “unchecked checkers.” In some
systems, Brazil and Spain among them, the independence of the judiciary is almost absolute by constitutional

3
There are gray areas here. The Soviet system used the reasoning by "analogy": even though the specific act was not in the
class defined as punishable by the law, the argument went, it was analogous to those that were.
4
If I remember correctly, in France, Conseil d'Etat exerts prior control and Court Constitutionel ex post control.

3
design. In Italy, the judicial system became practically independent as a result of political conflicts: while in
principle controllable by the parliament, the courts wield the threat of inculpating the very legislators who are
supposed to check it. The office of Special Prosecutor in the United States seems to be rapidly acquiring the
same status. Thus, while in the system of separation of powers, the legislature is the unchecked checker by
the constitutional design, the possibility that some organ acquires this status is present under the system of
checks and balances as well.
With these preliminaries, we can finally approach the substantive issue, namely, the question why
governments obey and enforce the law. The answer cannot be that it is “because it is the law”: the favorite
response of lawyers, constitutional and otherwise. The notion of the rule of law entailed here is one to which
Calvert (1994) refers as “institution-as-constraint.” For example, the law constrains governments not to take
partisan advantage from the incumbency. But the question of regress is obvious: why would governments
obey the law? As Calvert (1994: 3) observes, the institutions-as-constraints approach “takes as given the
institution’s effectiveness in channelling behavior. ‘Rules of the game’ are either (a) impossible to violate, due
to the specification of the game or the guaranteed actions of other players; or (b) backed by some form of
external functions. Either form of rule simply pushes the problem of institutional effectiveness and persistence
back to another level.” The question why the government would observe the law cannot be resolved by a
tautology—“because it is binding.” Clearly, one can imagine politicians who submit, out of a sense of duty, to
the “rules that equally bind the government itself with its citizens” (Grillo 1997). But then these politicians will
not want to exploit partisan advantage: they will have different preferences (seeKreps 1990: 116-120 for a
discussion of this issue). That is, unless politicians internalize constitutional norms, the question why
governments would obey the law remains open.
The rule of law must be understood as an equilibrium, in which “individuals’ actions are dependent upon the
past actions of many others, or upon expectations about the future reactions of many others, to one’s present
actions.” (Calvert 1994: 4). Law rules if no part of the government would want to violate it (or not enforce it) for
the fear of sanctions by other parts of the government and, eventually, by the electorate. A obeys the law
because it fears the sanction by B if it does not. But why would B sanction A for violating and only for violating
the law? The answer must be that B fears sanctions by C, etc. But in a democracy, there is no ultimate
enforcer other than the electorate. Hence, if the rule of law is to result from checks and balances alone, they
must be circular, that is, if A is sanctioned by B, and B by C, then C must be sanctioned by A. The powers of
government must all mutually check and balance one another.5
As always in such a framework, the question concerns instruments and incentives. Instruments empower,
incentives motivate. Both depend on the institutional design of the government. Yet beyond this truism, not
much more can be said. We do know something about the effects of institutional design of government if we
consider it feature by feature: parliamentarism versus presidentialism, the scope of decree powers, the
degree of constitutionalism, the forms of delegation to independent agencies, forms of oversight over
bureaucracy. But we know less about emergent effects: say combination ofpresidentialism versus
parliamentarism with different electoral systems. Finally, even if one believes that some institutional features
are superior to their alternatives—I firmly shareLinz’s (1990 a and b) view that parliamentarism is better than
presidentialism ( Alvarez and Przeworski 1995, Cheibub and Przeworski, in press) -- some reforms are simply
unfeasible. Indeed, in the entire post-war history, there were only three wholesale changes of political

5
For a fascinating example of how such a system might work, see Persson, Roland, and Tabelini (1996).

4
institutions without a breakdown of democratic continuity, two of them in Brazil and the third in France in
1958. Change is costly and not always desirable, even if it were to be for the better.
With this caveat, let me raise just two issues that appear particularly controversial in the current discussions
about the reform of the state. The first one concerns control over bureaucracies; the second the role of
independent agencies, including those engaged in oversight.
In a democracy, the authority of the state to regulate coercively the life of the society is derived from elections.
Yet many of the functions of the state and all of the services that the state supplies to citizens are delegated
by the elected representatives to someone else, specifically to public bureaucracies. Delegation is inevitable.
As Kiewiet and McCubbins (1991: 3) observe, “desired outcomes can be achieved only by delegating
authority to others.”
The relation between elected politicians and appointed bureaucrats is a thorny problem. Democratic
institutions contain few, in any, mechanisms that would allow citizens to sanction directly legal actions of
bureaucrats. As Dunn and Uhr (1993: 2) suggest, we do not even seem to know how to think about
principal-agent relations involved in controlling bureaucracies: “it is by no means clear what place executive
officials are meant to play as representatives of the people. Are they agents of the government or of the
people? If of the former, are they primarily responsible to the executive which employs them, or the legislature
which funds them?” While the bureaucracy is supposed to serve citizens, it is accountable at most to
politicians (or to other bodies appointed by politicians, such as courts or administrative oversight agencies).
Delegation raises the standard principal-agent problems. Since it is impossible to write legislation that would
fully specify the actions of agents under all contingencies, the executive and the administrative agencies are
left with a significant degree of discretion. But the objectives of bureaucrats need not be the same as those of
citizens or of the elected politicians who represent them. Bureaucrats may want to maximize their autonomy
or the security of their employment, renderclientilistic favors to friends and allies, shirk in office, aggrandize
their budgets (Niskanen 1971), or simply get rich—all at the expense of the public. They have private
information concerning the benefits and the costs of their actions and they undertake actions that cannot be
observed directly but only inferred from outcomes or monitored at a cost. Hence, delegation must inevitably
give rise to agency costs. Indeed, given the discretion bureaucrats must enjoy, the question is how to avoid a
regime of “policy without law,” as Lowi (1979: 92) described the U.S. political system.
What can be done to alleviate these agency problems? Since these issues have been recently spelled out by
Bresser Pereira (1997b, see also Przeworski 1996), I will not discuss specific measures. Let me just highlight
two general principles.
The first principle is that an adversarial procedures should be built into the governmental decision making, by
creating multiple principals or multiple agents with dissonant objectives.Tirole (1994) observes that most
governments are divided in such a way that it is not the task of any particular position or agency to maximize
general welfare, yet their interaction is supposed to generate this effect. His example of multiple principals is
the division between “spending ministries” which are supposed to promote substantive goals and the finance
ministry which is supposed to control spending. His example of multiple agents is that separate agencies
should be charged with collecting information for and against the desirability of a particular project or policy
rather than one agent collecting the information and reporting the net conclusion.
The second principle is that horizontal institutional checks should be built into the delegation process.Kiewiet
and McCubbins (1991: 33) point out that “agents are often in a position to do more harm to the principal than

5
to simply withdraw effort: embezzlement, insider trading, official corruption, abuse of authority, and coups
d’etat are all testaments to this fact. Whenever an agent can take actions that might seriously jeopardize the
principal’s interests, the principal needs to thwart the agent’s ability to pursue such courses of action
unilaterally.” Their solution are “institutional checks [that] require that when authority has been delegated to an
agent, there is at least one other agent with the authority to veto or to block the actions of that agent.”
The controversial issue is whether the oversight should be exercised by agencies independent from control by
the elected politicians. Most public agencies are independent from the direct control of citizensqua voters but
they are subordinate to politicians, who are in turn subject to periodic elections. Hence voters can at least
indirectly induce such public bureaucracies to act in their best interest by threatening politicians not to reelect
them unless bureaucracies well serve citizens. To choose an extreme example, generals are responsible for
the conduct of war, but it is the elected politicians who decide whether to go to war, who appoint them and
can fire them.
Yet there are good arguments to the effect that oversight agencies should be independent from the control of
the elected politicians, mainly that if the oversight agencies are controlled by politicians, they tend to collude
with them, hiding information unfavorable to the current government from the public. Examples of
independent oversight agencies include the office of the Promotor in Brazil, Canadian Human Rights
Commission, the Chilean Controlaria, or the French Cour des Comptes. Indeed, according to Sutherland
(1993: 24), at one time there were 650 such independent review bodies in Canada.
These bodies are independent in the sense that they are not subject to partisan control of the current
majority. Independence is accomplished by appointment and removal procedures: members of such bodies
either are appointed for life, or for long, fixed terms, not coinciding with electoral periods, or are removable
only with non-partisan consent. Such agencies differ greatly in their power. In the extreme case, such as of
the Canadian Human Rights Commission, they have direct enforcement power. Weaker powers, such as
those of the Canadian Information Commissioner, include investigation leading to judicial proceedings. Still
weaker are powers to investigate and report to the legislature: this is the case of the Canadian Security
Intelligence Review Committee. Finally, the weakest are those of agencies that can only hear complaints and
suggest remedial actions, such as the Canadian Police Public Complaints Commission. (Based onD’Arcy
Finn 1993.)
These independent oversight bodies raise two issues. One stems from their very independence.Sutherland
(1993: 24) claims, for example, that “’independent’ always means funded by the public and accountable only
to itself.” She lambastes the Canadian Office of Auditor General for pursuing its own objectives without any
control, claiming that in Canada there are “no effective legislative provisions subjugating control
bureaucracies to the House of Commons” (1993: 32). Hence the danger of independence is that agencies
become “unchecked checkers,” free from control by anyone, either other organs of government or citizens.
The second issue is whether such organs facilitate public control over the government. Again,Sutherland
claims that “independent review is not a democratic doctrine, but rather a management doctrine based on a
formula of checks and balances between appointed officials.” She contrasts such “closed” review with a
“popular” review, although she never spells out what she means by the latter term. The fact is that
independent oversight bodies are part of a system of checks and balances internal to the government, rather
than an instrument of popular participation in governance. The danger here is that even such independent,
but bureaucratic, agencies would collude with other bodies, hiding information from the public. Various kinds
of committees designed to oversee security or police agencies best illustrate this danger.

6
The issue of independent agencies, moreover, is not limited to oversight. Recently we have seen the
emergence of public agencies that are independent of the elected politicians, primarily central banks, but also
constitutional courts6 or semi-independent regulatory agencies. These agencies are empowered to pursue
some specific objectives, whether monetary stability, the conformity of the ordinary legislation with the
constitution, or the regulation of natural monopolies. Their personnel is appointed by the elected politicians.
But these agencies are supposed to act independently of the policies pursued by the elected politicians and
their personnel cannot be removed for policy reasons.7
The justification for the existence of such bodies is technical. The argument is that there are some
combinations of policies that should not be made by the same decision maker, whoever it might be. No single
body should pass laws and adjudicate particular cases, no single body should make fiscal and monetary
policy, no single body should fix prices of monopolies and tax them. The technical justification is different in
each case and, at least with regard to independent central banks, it is based on a rather elaborate theory
which claims that governments would pursue monetary policies that are time-inconsistent. Hence, the
argument goes, monetary policy should be made independently of other policies; thus, by a body
independent of the government. But then what guarantees that agencies that are independent of the public
act in the public interest?
This is not a simple question, for the answer depends in part on the theory one believes. If there exists a
uniquely optimal policy in a particular realm—say that money supply should increase at the rate of growth of
output or that natural monopolies should earn the competitive rate of profit—then the only problem is
implementing a rule. The issue is then one of “rule versus discretion,” and the public interest is uniquely given
by the optimal rule. But if no such unique rule exists—say that the public has preferences over the rate of
inflation8—then the issue is no longer of rule versus discretion but of discretion of independent agencies
versus discretion of elected ones. And, asMinford (1995) observed, nothing does guarantee that
independent agencies will pursue policies preferred by citizens. Directors of the central bank may pursue
policy objectives that are preferred by citizens but they may also have objectives that differ from those of the
public, and citizens have no instruments to induce central banks to act according to their interests.9
While specific solutions are not obvious, the issues are clear. Citizens are protected from the state when the
structure of government includes a well designed system of checks and balances. These checks and
balances must be horizontal, not only vertical: not just superiors checking subordinates but principals and
their delegates checking one another. This system can function effectively only when the particular agencies
have correct incentives to exercise their institutional prerogatives. In particular, to avoid collusion, some
agencies must be independent from the partisan interests of politicians. Yet no agency should be free from
control by some other agency or directly by voters: there should be no “unchecked checkers.”

6
On constitutional courts as independent agencies, see Pasquino (1997); on the similarities and differences between courts
and central banks, see Manin (1997).
7
For the evidence that it is the actual turnover of the directors of central banks, rather than the legislation authorizing their
independence that matters, see Cukierman (1992).
8
It is ironic that Kydland and Prescott (1977) in their classical article first (p. 478) claim that "A change in administration ...
reflects a change in the relative costs society assigns to unemployment and inflation..." and then go on to assume that there exists
some uniquely optimal rate of inflation (p.480).
9
Indeed, Minford compared the welfare of citizens under central banks that depend on the elected government, under central
banks that are independent of the government and the public, and central banks that are independent of the government but
elected.
And, at least within his assumptions, the last design is the one that best serves the interest of the public.

7
 While such a system of checks and balances may be effective in protecting citizens from the state, this
argument is vulnerable to accusations of a liberal bias. After all, citizens do not just want to be protected by
the government, they need and want to be served by it. Such a system may prevent governments from
violating the law, but at the same time makes it more difficult to enforce it. In the extreme, with checks and
balances at every step, the government can do almost nothing. The liberalcounterargument, put forth most
eloquently by Holmes (1995: 6), is that “liberalism is a necessary, though not sufficient condition for some
measure of democracy in any modern state.” Liberty, protection from arbitrary encroachments by the state, is
necessary for citizens to be able to exercise their positive political rights, to express their opinions, formulate
the collective will, and to choose good governments; in the end, to make governments do what citizens want
and need. Yet liberty may not suffice for citizens to control the governments unless political institutions provide
effective instruments of control and unless all people enjoy some minimal economic and social conditions
necessary for the exercise of their political rights. These are the two topics that follow.

Vertical Mechanisms: Elections

Under democracy, people exercise their control over governments through elections. Indeed, the central claim
in favor of democracy is that if elections are freely contested, if participation is widespread, and if citizens
enjoy political liberties, then governments will act in the best interest of the people (Dahl 1971). Yet this claim
is nothing more than a hypothesis, to be examined as such.
There are two ways in which elections may cause governments to act in the best interest of the people. In
one—the “mandate”—view, elections serve to select good policies or policy-bearing politicians. Parties or
candidates make policy proposals during campaigns and explain how these policies would affect citizens’
welfare; citizens decide which of these proposals they want implemented and which politicians to charge with
their implementation, and governments do implement them. Thus, elections emulate a direct assembly and
the winning platform becomes the “mandate” which the government pursues. In a second—“accountability”—
view, elections serve to hold governments responsible for the results of their past actions. Because they
anticipate the judgement of voters, governments are induced to choose policies which in theirjudgement will
be positively evaluated by citizens at the time of the next election.
The mandate conception of elections is widespread: scholars, journalists, and ordinary citizens rely on it as if
it were axiomatic. Keeler (1993), for example, explains the major policy reforms introduced byReagan,
Thatcher, and Mitterand as follows: their respective countries faced economic crises, voters wanted change
and expressed this desire at the polls, the respective governments implemented their mandates.
The logical structure of the mandate conception is the following. In elections, parties or candidates present
themselves to voters, informing them about their policy intentions. Specifically, they tell voters which policies
they intend to pursue, for what purposes, and with which consequences. Voters evaluate these proposals, as
well as the personal or partisan characteristics of the candidates, and vote for those which they most like.
Once elected, the victorious candidates pursue the policies which they proposed.
This argument holds when what politicians and voters want coincides or when politicians care only about
winning elections, and to win they must promise and implement policies that are best for the public. But short
of this happy coincidence, politicians may have incentives either to deviate from the mandate in the best
interest of the public or to stick to it at the cost of the electorate. First, conditions may change in such a way

8
that the implementation of the electoral promises, the “mandate,” is no longer best for voters. Secondly, to be
elected, a candidate must promise what voters most want. Yet if the incumbent believes that a less popular
policy is sufficiently more effective than the one voters prefer, he or she anticipates that, having observed its
effects, voters will become persuaded that the correct policy was chosen and will vote to reelect him in spite of
the betrayal of electoral promises. Thus, Stokes (in press) reports that in 45 elections that took place in Latin
America after 1982, 13 newly elected governments immediately took a course diametrically opposed to their
campaign platforms (and four were so vague that they could not be classified). Finally, incumbents may stick
to their promises even if they know that implementing them is not in the best interest of the public, fearing that
if they propose one platform and pursue another, they will not be reelected (Harrington 1993).
To summarize, incumbents may either pursue policies that enhance the welfare of voters by deviating from
the mandate or they may adhere to the mandate even if they think that implementing it is not best for voters. If
voters are not sure which policies best serve their interests, they cannot be certain whether the government is
acting in their best interest when it implements its promises or when it betrays them. And since governments
know that voters do not know, they have room for doing things they, rather than voters, want.
A striking feature of democratic institutions, highlighted byManin (1995), is that politicians are not legally
compelled to abide by their platform in any democratic system. In no existing democracy are representatives
subject to binding instructions. Citizens’ suits against governments that betrayed specific campaign promises
have been rejected by courts in several countries, most recently in Poland. No national level democratic
constitution allows for recall, and, except for the U.S. House of Representatives, electoral terms tend to be
long: on the average 3.7 years for legislatures and 3.9 years for presidents C( heibub and Przeworski, in
press). While provisions for impeachment and procedures for withdrawing confidence are common, they are
never targeted at the betrayal of promises.10 Binding national referenda based on citizens’ initiative are found
only in Switzerland and, in more restrictive forms, in Italy and Argentina. Hence, once citizens elect
representatives, they have no institutional devices to force them to adhere to promises. Voters can sanction
deviations from mandates only after their effects have been experienced.11
Why then are there no institutional mechanisms to force office-holders to be faithful to their platforms?
Historically, the main argument was that legislatures should be allowed to deliberate. People want their
representatives to learn one from another. Moreover, when people are uncertain about theirjudgements, they
may want representatives to consult experts.
Another historical argument was that voters may not trust their ownjudgements. Not only people may be
afraid of their own passions but, if they are rationally ignorant, they must know that they do not know.
Presumably, elections establish the calendar for when the accounts are to be taken. Hence, citizens may
want to give the government some latitude to govern and evaluate government’s actions at election times.
Finally, institutions must allow for changing conditions. No electoral platform can specify ex ante what the
government should do in every contingent state of nature: governments must have some flexibility in coping

10
Occasionally a deviation from mandates provides part of the impetus for impeachment, even though deviation is not the
formal justification. Two recent presidents who abandoned their campaign promises, in Venezuela and Ecuador, were impeached,
one immediately, with no time allowed for the outcomes to materialize.
11
Cunill (1997) lists several instances in which recall procedures and imperative mandates were recently introduced in Latin
America. All of them, however, are at subnational levels.

9
with changing circumstances. If citizens expect that conditions may change and governments are likely to be
representative, they will not want to bind governments by their instructions.12
Hence, there are good reasons why democratic institutions contain no mechanisms enforcing adherence to
mandates. We choose policies that represent our interests or candidates who represent us as persons but we
want governments to be able to govern. As a result, while we would prefer governments to stick to their
promises, democracy contains no institutional mechanisms that insure that our choices would be respected.
Even if citizens are unable to control governments by obliging them to follow mandates, citizens may be able
to do so if they can induce the incumbents to anticipate that they will have to render accounts for their past
actions. Governments are “accountable” if voters can discern whether governments are acting in their interest
and sanction them appropriately, so that those incumbents who act in the best interest of citizens win
reelection and those who do not lose them.
The standard view of how the accountability mechanism operates relies on “retrospective voting.” In this view,
citizens set some standard of performance to evaluate governments, such as “My income must increase by at
least 4% during the term,” “Streets must be safe,” or even “The national team must qualify for the World Cup.”
They vote against the incumbent unless these criteria are fulfilled. In turn, the government, wanting to be
reelected and anticipating the citizens’ decision rule, does whatever possible to satisfy these criteria.
Yet suppose that voters do not know some of the conditions facing the country. Politicians know these
conditions but voters may be unable to observe them at all or they may be able to monitor them only at a cost.
Such conditions may include the negotiating posture of foreign governments or international financial
institutions—something citizens cannot observe—or the level of demand in the major recipients of the
country’s exports—something voters can observe only if they turn into economists. Then voters are in a
quandary. If they set the standard the incumbent must meet too high, the incumbent cannot be reelected
whatever it does and it will not have incentives to act in the public interest. In turn, if voters set the standard
too low, the incumbent will be able to do little for the public and still be reelected. Whatever voters decide to
do, politicians will sometimes escape from their control.
One aspect of incomplete information merits particular attention. If voters are fully rational, they should also
care at the end of the term about the present value of their future welfare: the legacy the incumbent leaves for
the future. If the economy grows because the government cuts all the trees in the country, the voter will live
on champagne during the term, but there will be no trees left to cut. In turn, if the economy declined because
it underwent structural reforms, voters will have suffered economic deprivation, but may have improved their
life-chances for the future. Yet all voters observe is the change of welfare during the term and they have to
make inferences about the future on this basis. Say that their current welfare declines: should they infer that
the government is investing in their future or pursuing some chimeras of its own or just robbing them blind?
Accountability is not sufficient to induce representation when voters have incomplete information.
Note that in a pure mandate model voters use the vote only to choose the better candidate. In a pure
accountability model, voters use the vote only to sanction the incumbent. Yet, for all we know, voters do not
meditate whether to use the one instrument they have, the vote, to choose a better government or to structure
incentives for incumbents. Fearon (in press) offers persuasive stories to the effect that voters want to select
good policies and politicians. Yet voting “to keep them honest” seems also ingrained in the repertoire of the

12
Minford (1995: 105) observes in the context of monetary policy that "if voters have little information, they may prefer to let
governments have complete discretion, regardless of the lack of credibility, rather than tie their hands."

10
democratic culture. The fact remains that voters have only one instrument to reach two goals: to select better
policies and politicians and to induce them to behave well while in office. Madison (Federalist No. 57) thought
that “The aim of every political constitution is, or ought to be, first to obtain for rulers men who possess most
wisdom to discern, and most virtue to pursue, the common good of the society; and in the next place, to take
the most effectual precautions for keeping them virtuous whilst they continue to hold their public trust.” Using
the vote for both purposes—to obtain best rulers and to keep them virtuous—is not irrational: while voters lose
some control over the incumbent (Fearon, in press), in exchange they elect a better government. Yet the
system Madison and his colleagues designed makes it possible to strive for one goal only at the expense of
the other.
These conclusions may appear unduly skeptical. It may well be that either through the mandate and the
accountability mechanisms, voters induce governments to act in the best interest of the public. Empirical
evidence shows that at least in Western Europe the elected government do adhere to their promises
(Klingeman, Hofferbert, and Budge 1994) and that policy is responsive to public opinion. Corruption is
frequent, but most of it is motivated by the need to finance parties and elections—the great unresolved
institutional problem of democracies—rather than by personal enrichment. The class bias of many
governments is obvious, but whether this bias is caused by political or economic structure is not. Thus, one
may adopt a sanguine view that, by and large, elections do cause governments to behave in the best interest
of the public.
Yet my view is that elections are at best a blunt instrument of control over governments. AsBobbio (1989:
157) once remarked, “to pass a judgement today on the development of democracy in a given country the
question must be asked, not ‘Who votes? but ‘On what issues can one vote?’.” An average European
parliament makes about 3,000 decisions during its term, bureaucracy makes thousands more. Voters have
only one, in presidential systems two, decisions to make with regard to the entire package of government
policies. It is not possible to control thousands of targets with one instrument. Hence, elections must
inevitably leave large areas of policy out of citizens control.
Moreover, control through elections requires information. The main difficulty both in instructing governments
what we want them to do and in judging what they had done is that we, citizens, just do not know enough.
True, we will never know all governments know, nor would we want to. After all, governing is a part of division
of labor, and even if we were to select our rulers at random, not on the basis of the privileged knowledge they
already have, they would inevitably acquire expertise. Indeed, we want them to develop such expertise. Yet
we do not know enough, and the reason is not just volitional but structural. Citizens may be better or less well
educated, the media may be more or less aggressive in digging up information, voters may be more or less
attentive. But the issue of information goes deeper.
The peculiarity of the principal-agent relation entailed in the relation of political representation is that our
agents are our rulers: we designate them as agents so that they would tell us what to do and we even give
them the authority to coerce us to do it. And the rules which our agents impose on us include access to
information. The British government, for example, barred independent researchers from access to tissues
extracted from cows suffering from bovinespongiform encephalopathy (“mad cow” disease). The United
States Senate recently dropped a provision that would have allowed employees of the CIA and other
agencies to disclose information related to violations of law to members of Congress without the approval of
their bosses.

11
 Yet, to evoke Kant, “All actions affecting the rights of other human beings are wrong if their maxim is not
compatible with their being made public.” Bobbio (from whom this passage is taken, 1989: 84) comments
further that “a precept not susceptible to being publicized can be taken to mean a precept which, if it was ever
made known to the public, would arouse such a public reaction that one could not put it into action.” We do
not want governments to take actions which they would have not taken had we known why they are taking
them. But this means that we have to know what the governments are doing and why independently of what
they want us to know. Our authorization to rule should not include the authority to hide information from us.
To promote representation, democracy requires a “regime of free information” D ( unn, in press).
Thus, even if elections give governments a broad authorization to rule, this authorization should not extend to
informing us. Our information must not depend on what governments want us to know. Hence, to return to
Sutherland, oversight must be “popular,” not only “closed.” The institutional implications are obvious: we need
independent electoral commissions, independent accounting offices, independent statistical agencies. We
need “accountability agencies,” independent of other branches of government and designed to inform the
public, not just its representatives.

Vertical Mechanisms: “Participatory Democracy”

Elections, however, are not the only instrument of “vertical” control, of governments by citizens. Indeed, the
influence citizens exercise through elections may be only of a minor practical importance, compared to
innumerable other channels.
This fact raises a number of difficulties, concerning legitimacy and equality. In normative terms, in a
democracy the authorization to rule is derived from elections, from voting, and only from voting. No “reason of
state” can justify actions of a government unless the authority of this government to rule is derived from
elections. Hence, elections, as ineffective as they may be as an instrument of popular control, have a unique
normative status in any democracy.
Moreover, elections are at least in principle an egalitarian mechanism, ruled by the principle “one-person one-
vote.” They are not uniquely egalitarian: random sample surveys also are. But surveys do not authorize the
rule. Hence, elections are the only egalitarian instrument establishing the legitimacy of a democratic
government.
Yet even in elections, where the norm is universalistic, the influence of money is legally regulated, and the act
of exercising control itself almost costless13, access is distributed highly unevenly. The fact is that just to exist
and to present themselves to voters, parties and candidates must raise funds. When these funds come from
special interests, they must be exchanged for favors. Presumably, if Philip Morris Co. Inc. contributed in 1996
over two and a half million dollars to the Republican National Committee (New York Times, January 28, 1997,
page 3), it must have expected at least two and a half million dollars in favors: otherwise its management
should have been thrown out by stockholders. Indeed, according to the Center for Responsive Policy (New
York Times, January 24, 1997, page 3), a sugar-price subsidy that adds an extra 50 cents to a five-pound bag
of sugar was supported by 61 senators who received on the average $13,473 from the industry political action
committee, while it was opposed by 35 senators who on the average got $1,461. Exchanges of political

13
The rational-choice theory preoccupation with the "costs of voting" is, in my view, much ado about nothing.

12
contributions for policy favors are distorting through their effects on the allocation of resources. And the social
cost of such distortions is likely to be much greater than of outright theft, which is distorting only through its
effect on distorting taxes.
Inequality of access is much greater in all other mechanisms of influence over governments. Since this is one
focus of a remarkable book by Cunill (1997), I will raise only the points that remain open. Note first that some
the inequality of access is legally institutionalized. Indeed, the essence ofcorporatism is a legal recognition of
some specific interests which, asSchmitter (1984) had pointed some time ago, inevitably excludes other
interests. The idea that all interests can be represented throughcorporatist arrangements, underlying the
South African corporatist chamber, the NEDLAC (National Economic Development and Labor Council) is just
a chimera: the “social sector” just does not have the same organizational capacity as capital and labor do (For
contrasting views of NEDLAC, see Lewis 1998 and Nattrass and Seekings 1998). Corporatism may privilege
labor in relation to capital but only at the cost of excluding other interests O
( ffe 1995). Hence, Cunill (1997:
108) concludes, “the deficits of political equality are not corrected by substituting functional for territorial
representation. To the contrary, to pretend that functional representation has primacy over territorial one is to
accept much sharper social exclusion—caused by criteria of socioeconomic power—than that which should
precisely be neutralized by the principle of democratic equality.”
Yet all forms of political influence require resources, whether money, time, organizational skills, knowledge, or
talents of persuasion. And they are all unequally distributed, all highly correlated with the first. The problem,
then, is, how to extent the mechanisms of representation without trasmitting the effects of inequality of
resources to the inequality of access (Cunill 1997: 71).
Cunill distinguishes instruments of direct democracy, such as referendum and recall, from instruments of
control over bureaucracy. She correctly notes that “the spirit that moves us to instruments of direct democracy
is, precisely, the weakness of representative government” (92). And she observes that there is an intense
political demand for public participation in governance, at all levels of government. To insure that the
bureaucracy functions in the public interest, citizens need direct mechanisms of control over public agencies.
Thus, Haggard (1995: 41-42) argued, “The ultimate check on government must come through institutionalized
forms of participation. This may either be ‘corporatist,’ such as building in NGO participation in areas in which
they have expertise, or ‘legislative,’ such as adopting forms of local governance in which citizen participation
is maximized.”
Yet it is far from clear whether and which among such forms of participation would be effective. Participatory
arrangements are vulnerable to capture by the very interests they are supposed to control.Corporatist
arrangements promote special interests of members of corporations. Elected bodies are vulnerable to capture
by political parties. And all such forms of “participatory democracy” are as likely to increase as to decrease the
inequality of access. Indeed, Cunill (1997: 72) warns that “’citizenship participation, even if it is inscribed in the
principles of deepening democracy and, in particular, in democratizing the state, can lead to the weakening of
the public sphere rather than to its strengthening.” After all, the “civil society” heralded by the prophets of
participatory democracy—this I know from personal experience—are often not more than Barbara from ABC
and Tom from CBA, well-meaning in promoting their elevated, yet particularistic, ideals and interests. And one
should not forget that the most powerful NGOs in our societies are lobbies of business: “lobby” does not

13
 sound as lofty as “NGO” but this is what they are.14 Thus, while the need for a more direct and more extensive
public control over governments is evident, no ready-made solutions are in view.

The State in Society

Time has come to put things in perspective. For, as important as institutions may be, can any reform of the
state, even one conceived in broadly political and not only administrative terms, such as in Brazil, compensate
for economic and social inequality? The causal question is why it is that in some democracies the state
enforces the law in a non-universalistic, unequal, way: because of its institutional design or because of some
features of the societies in which it functions? Is it true that, asCunill (1997: 71) puts it, “the asymmetries of
representation are the source of inequities in the distributive tasks of the state, compromising its capacity to
regulate the economy and the centers of private power, as well as toafirm social rights” or that in
economically and socially unequal societies even an egalitarian system of political representation cannot
overcome the inequality?
The explanation offered thus far focused on political institutions. As we have seen, the degree of control that
citizens can exert over governments through elections is weak, while other participatory mechanisms are as
likely to exacerbate as to reduce political inequality. Thus, a democracy in which the only mechanisms of
control over the rulers are “vertical” is likely to be d“ elegative” in O’Donnell’s (1994) sense. The electorate may
be more or less talented or lucky in electing governments but, once power is delegated through elections,
these governments need not fear popular sanctions, at least not from those who are not powerful or
organized. Yet I think that O’Donnell’s (1994, 1997a) claim that Latin American democracies, as distinct from
the well established ones, suffer from the absence of “horizontal” accountability is just factually incorrect. This
is just an impression, but I see in several Latin American countries legislatures, courts,fiscalizing agencies,
Ombudsman institutions (such as the BrazilianPromotor), political parties, independent press that are not any
weaker in imposing checks and balances on the executive and on one another than they are in Great Britain,
Italy, France, or the United States. Moreover, asCunill (1997) evidenced, several Latin American
democracies recently introduced forms of direct democracy (referendum, recall) rarely found in the OECD
countries. So if one were to engage in Tsebelis’s (1995) count of “veto players”—how many actors can block
an action of the government as a whole—the average number would be probably the same in Latin American
and in other democracies. Clearly, to assess whether these “horizontal” mechanisms function effectively, one
would have to do comparative research guided by systematic yardsticks. But even if the Argentine or the
Brazilian presidents seem to be able to get most of what they want, it is doubtful that they are less
institutionally constrained than the British or the Spanish prime ministers. Hence, I conclude that the unequal
enforcement of law in many Latin American democracies cannot be explained in terms of their institutional
structures.
The alternative hypothesis is that in highly unequal societies the state, whatever its institutional structure, is
just too poor to enforce the law universalistically. No reform of the state is sufficient to safeguard the
republican rights for everyone, because the state has no resources to safeguard and promote these rights.
The impediment is fiscal, not institutional. For, as Bresser Pereira (1997a: 12) emphasized, “to guarantee the

14
Indeed, in the United States some corporations now mask their lobbies under labels that sound like NGOs, such as,
"Citizens for a Better Economy."

14
civil rights, a positive action of the state is also necessary, since it implies administrative costs: after all, the
classical (or liberal) state apparatus—the Parliament, the Courts, the Police, and the Army—exists to
guarantee the civil rights, in the same way that the Social State apparatus, expressed in Social Security,
Education, Health, and Culture Departments, exists to guarantee social rights.”
*** Table 1 Here ***
What is different about Latin America—and increasingly one can generalize about the continent—is the
degree of economic and social inequality. Here we have numbers to compare. As of the most recent date for
which information is available, the average ratio of the incomes of the top to the bottom quintiles of the
distribution was 16.95 in 10 Latin American countries (ranging from 8.58 in Bolivia to 30 in Guatemala), 6.58
in seven Asian countries (from 4.30 in India to 10.10 in the Philippines), 5.12 in 17 OECD countries (from 4.11
in Luxembourg to 10.09 in Australia), and 5.05 in 10 Eastern European countries (from 2.64 in the Slovak
Republic to 14.54 in Russia, a distant outlier; next highest is Estonia, with 5.66). The average share of the top
20 percent of income recipients was 56.1 in Latin America (from 48.2 in Bolivia to 65.2 in Brazil), 45.6 in Asia
(from 39.1 in India to 52.7 in the Philippines), 40.5 in the OECD (from 34.0 in Finland to 46.6 in Australia),
and 38.2 in Eastern Europe (from 31.4 in the Slovak Republic to 53.8 in Russia). And all these numbers must
still seriously underestimate the true degree of inequality: at least the income reported in Mexican household
surveys is by about 40 percent lower than in the national accounts, and most of the difference is the income
of the rich (Cortés 1997).15
Why am I citing these numbers, having announced that this is a paper about political, not about economic
inequality? It is because of a finding that links the two in a surprising way.
One would expect that facing a high degree of inequality, democratic governments would quickly act to
reduce it. Indeed, one of the central tenets of democratic theory, dating back to the first half of the eighteenth
century, is that democracy enables the poor to use their power in numbers so as to redistribute incomes via
the state. This assertion has been elevated in the recent period to a status of a theorem. In the light of the
“median voter theorem” (Meltzer and Richards 1981), the greater the inequality of income, specifically the
more the income of the median voter falls below the mean income, the higher will be the degree of
redistribution resulting from the democratic process. Hence, one would expect that the share in the national
income of government revenues or taxes or consumption expenditures should be greater in societies that are
more unequal, thus allowing governments to correct the pre-tax inequality.
Yet inspecting the numbers in Table 1 generates a surprise. It appears that the contrary is true. The total
revenue of central government, the tax revenue of central government, and government consumption
expenditures are all lower in more unequal countries. Statistical analyses confirm this impression. The higher
the ratio of the incomes of the top to the bottom quintile (Q5/Q1), the smaller the revenue, tax, and
consumption expenditures of government.
*** Table 2 Here ***
Thus, for some reason taxing is more difficult in more unequal societies. I can only speculate why it is so.
According to the “structural dependence theory” (Przeworski and Wallerstein 1988), governments are limited
in their ability to raise tax revenues by the negative impact of taxation on employment and investment,
decisions about which are made by owners of productive assets in pursuit of profit. Nevertheless, it is not
15
Another piece of evidence is that inequality of assets is typically much higher than inequality of incomes, which would make
sense only if the rich saved less or at lower rates of return.

15
apparent that the deadweight costs of taxes are higher in more unequal societies. To the contrary, it seems
that redistribution should be easier under such circumstances. Just calculate: If the top quintile of income
recipients receives 60 percent of all incomes and the bottom quintile receives three percent (roughly Chile),
taxing the top quintile at five percent rate would double the income of the bottom, while still leaving 57 percent
in the hands of the rich.
So it appears that the reasons must be political, not economic: either the rich have more power over the
government in more unequal societies or the poor are less able to use their democratic rights. The data are
consistent with both explanations: The higher the share of the top quintile (Q5), the lower the total
government revenue, the tax revenue, and government consumption expenditures. In turn, the higher the
income share of the poorest twenty percent (Q1), the higher are government revenues and consumption
expenditures.16
To understand why governments would tax and spend less in countries in which the poor receive a smaller
share of income, one can return to J.S. Mill’s observation that “without decent wages and universal reading,
no government of public opinion is possible.” Democracy is a system of positive political rights. But, of itself,
democracy does not create the social and economic conditions necessary for these rights to be effectively
exercised. The nineteenth century way to avoid this problem was to restrict citizenship only to those who
enjoyed these condition. Today, political rights are universal in all democracies. But in many of them masses
of people are not in condition to exercise them. Hence, we witness a new monster, democracy without an
effective citizenship.17
This kind of an argument would imply, however, that what matters for the effective exercise of rights is not
inequality per se but the absolute income of the poor.18 Yet Reis (1996: 121) may have been right that “if
social inequality is rampant, power will be distributed unevenly, thus inevitably undermining the perspectives
of full enjoyment of civil and political rights by all.” The statistical results are ambivalent. When per capita
income (and its square19) is controlled for, the degree of inequality (Q5/Q1) still has a strong negative effect
on government expenditures, the share of the bottom quintile has a positive effect, and the share of the top
quintile a negative effect; all statistically significant. But with regard to taxes and general revenues, while all
the variables have the same signs, they are no longer statistically significant. Hence, it appears that inequality
per se somehow reduces government expenditures, while absolute levels of income dominate government
revenues.

16
Note that the median voter theorem fails as well. The theorem asserts that (1) tax revenues should be high (equality is
tempered only by the marginal deadweight losses) whenever the income of the median is lower than the mean and (2) that the
lower the median in relation to the mean, the higher should be the tax rate. The income of the median person (household in some
data) is Q3 and it would equal the mean if it were 20 percent. Hence, in all countries, the median is below the mean and the tax
should be high. Moreover, regressions based on Q3 show that it has a wrong sign: the closer to the mean, the higher the total
revenues, tax
revenues, and government consumption expenditures. Yet, since not everyone votes, this evidence is not conclusive.
17
O'Donnell (1993) refered to "low density citizenship."
18
Take two countries with average per capita income of $5,000 (1985 PPP) and compare the incomes of different groups with
the distribution in Brazil and Poland. In Brazil, the average person in Q5 will have an income of $16,295, in Q3 of 2,288, and in Q1
of just $620. In Poland, the incomes will be $9,150 in Q5, $4,425 in Q3, and $2,325 in Q1.
19
Government expenditures first decline and then slightly increase in income; government revenues first rise with income and
to decline.

16
Whatever the reason, it appears that the ability of the state to reduce economic and social inequality is lower
in the societies in which this inequality is high. This is a low-level trap: high inequality causes the state to be
poor, poor state cannot reduce inequality.20
None of the above implies that institutions do not matter or that a reform of the state is not necessary. A well
designed system of government, with effective mechanisms of popular control, will be more effective in
transforming tax-payers’ money into services for citizens. But the government must have the enough money
to transform.
So my answer is that the state is just too poor in Latin American democracies, as well as in United States, to
enforce the law in a universalistic fashion. And it is too poor because it is incapable of taxing the rich. The rich
find it more effective to privately purchase various services, whether education, medical care, access to the
judicial system, or police, rather than to pay taxes that would support such services for everyone. The middle
class pays taxes—indeed, it seems that tax burden on the incomes from employment is quite high in some
Latin American countries—and receives the bulk of state services.21 The poor are just outside the purview of
the state: they have no incomes to tax and they receive few services. Hence, a vicious circle: the poor are
politically ineffective because they do not enjoy the conditions for the effective exercise of their political rights;
because they are politically ineffective, they remain poor. The result is a poor state in an unjust society.

20
It is a trap in the sense that if the level of inequality exceeds some level, than it will never decrease given the low level of
government revenue resulting from this level of inequality.

For those who do not mind symbols, this is how this situation appears, judging from the statistical results: (1) For each initial level
of inequality, call it I = Q5/Q1, there exists a maximal level of government reveneues, call it G* = G(I). Say G*(5) = 32, G*(15) = 24,
I
G*(26) = 14.4, etc. (2) For each initial level of I, there also exists a level of government revenues, call it G (I), such that inequality
I I I I
remains at the current level if G = G , it declines if G > G , and increases otherwise. Again, say that G (5) = 2, G (15) = 11, and
I I I I
G (26) = 52. (3) Now compare G* and G for each level of I. If G* > G , then inequality will be reduced; if G* = G , then it will remain
I I
at the same level; if G* < G , it will increase. Say that G* = G when I = 18. Then any country that has inequality larger than 18 is in
the trap.

How does a country get out of the trap? Obviously, by collecting more revenues than in the equilibrium described by the regression
results. Brazil, for example, is the outlier: the total revenues predicted by its inequality of 26.28 are 14.4 but the observed value is
27.8, which represents considerable effort. But is this enough to get out of the trap? In this purely illustrative example, it is not: you
I
see above that G (26) = 52. Hence the actual level, even if above the equilibrium, is too low to get out of the trap.
21
In Brazil, and I am certain in several other countries, the net effect of social expenditures is regressive once education is
included.

17
References

Alvarez, Mike, and Adam Przeworski. 1995. “Parliamentarism and Presidentialism: Which Works? Which
Lasts?” In Antoni Sulek and Jozef Styk (eds.), Ludzie i Instytucje: Stawianie sie ladu spolecznego. Pages 67-
90. Lublin: Wydawnictwo Uniwersytetu Marii Curie-Sklodowskiej.
Bobbio, Norberto. 1989. Democracy and Dictatorship. Minneapolis: University of Minnesota Press.
Bresser Pereira, Luiz Carlos. 1997a. “Citizenship and Res Publica: The Emergence of Republican Rights.”
Manuscript.
Bresser Pereira, Luiz Carlos. 1997b. “State Reform in the 1990s: Logic and Control Mechanisms.”Cadernos
MARE da Reforma do Estado. Brasilia: MARE.
Calvert, Randall. 1994. “Rational Actors, Equilibrium, and Social Institutions.” Manuscript. University of
Rochester. Forthcoming in J. Knight and I. Sened (eds.), Explaining Social Institutions. Ann Arbor: University
of Michigan Press.
Calvert, Randall. 1995. “The Rational Choice Theory of Social Institutions: Cooperation, Coordination, and
Communication.” In Jeffrey S. Banks and Eric A. Hanushek (eds.), Modern Political Economy (pp. 216-268).
New York: Cambridge University Press.
Cheibub, José Antonio, and Adam Przeworski. In press. “Democracy, elections, and accountability for
economic outcomes.” In Bernard Manin, Adam Przeworski, and Susan C. Stokes (eds.),Democracy,
Accountability, and Representation. New York: Cambridge University Press.
Condorcet. 1986 (1785). “Essai sur l’application de l’analyse a la probabilité des décisions rendues a la
pluralité des voix.” In Sur les élections et autres textes. Textes choisis et revus par Olivier de Bernon. Paris:
Fayard.
Cunill Grau. Nuria. 1997. Repensando lo público a través de la sociedad: Nuevas formas de gestión pública y
representación social. Caracas:Nueva Sociedad.
Dahl, Robert. 1971. Polyarchy: Participation and Opposition. New Haven: Yale University Press.
D’Arcy Finn, T. 1993. “Independent review agencies and accountability: snapping at the heels of
government?” Optimum: The Journal of Public Sector Management 24: 9-22.
Deininger, Klaus, and Lyn Squire. 1996. “A New Data Set Measuring Income Inequality.” The World Bank
Economic Review 10: 565-592.
Dunn, John. 1996. The History of Political Theory and other essays.Cambridge: Cambridge University Press.
Dunn, John. In press. “Situating Democratic Political Accountability.” In AdamPrzeworski, Bernard Manin, and
Susan C. Stokes (eds.), Democracy, Accountability, and Representation.New York: Cambridge University
Press.
Dunn, Delmer D., and John Uhr. 1993. “Accountability and Responsibility in Modern Democratic
Governments.” Paper presented at the Annual Meeting of the American Political Science Association,
Washington, D.C., September 2-5.

18
Grillo, Michele. 1997. “Democracy, competition, and the principle ofIsonomia: An economic analysis of the
political exchange as an incomplete contract.” In AlbertBreton, Gianluigi Galeotti, Pierre Salmon, and Ronald
Wintrobe (eds.), Understanding democracy: Economic and political perspectives.Pages 47-63. Cambridge:
Cambridge University Press.
Haggard, Stephan. 1995. “The Reform of the State in Latin America.” Paper presented at the World Bank’s
Annual Conference on Development in Latin Amrica, Rio de Janeiro, June 12-13.
Harrington, Joseph E. Jr. 1993. “The Impact of Reelection Pressures on the Fulfillment of Campaign
Promises.” Games and Economic Behavior 5: 71-97.
Kavka, Gregory S. 1986. Hobbesian Moral and Political Theory. Princeton: Princeton Unversity Press.
Keeler, John T.S. 1993. “Opening the Window for Reform: Mandates, Crises, and Extraordinary Decision-
Making.” Comparative Political Studies 25: 433-486.
Kelsen, Hans. 1929. Vom Wesen und Wert der Demokratie. Aalen: Scientia Verlag.
Kiewiet, D. Roderick, and Matthew D. McCubbins. 1991. The Logic of Delegation: Congressional Parties and
the Appropriation Process. Chicago: University of Chicago Press.
Klingeman, Hans-Dieter, Richard I. Hofferbert and Ian Budge. 1994. Parties, Policies, and Democracy.
Boulder: Westview Press.
Kreps, David M. 1990. Game Theory and Economic Modelling. Oxford: Clarendon Press.
Kydland, Finn E., and Edward C. Prescott. 1977. “Rules Rather than Discretion: The Inconsistency of Optimal
Plans.” Journal of Political Economy 85: 473-491.
Lewis, David. 1998. “Pacts and the Consolidation of South African Democacy.” Paper prepared for
Conference on the Politics of Economic Reform, University of Cape Town, 16-18 January.
Linz, Juan J. 1990a. “The Perils of Presidentialism.” Journal of Democracy 1: 51-69.
Linz, Juan J. 1990b. “The Virtues of Parliamentarism.” Journal of Democracy 1: 84-91.
Lowi, Theodore J. 1979. The End of Liberalism: The Second Republic of the United States. New York:W.W.
Norton.
Manin, Bernard. 1994. “Checks, balances, and boundaries: the separation of powers in the constitutional
debate of 1787.” In Biancamaria Fontana (ed.), The Invention of the Moden Republic. Cambridge: Cambridge
University Press.
Manin, Bernard. 1995. Principes du gouvernement représentatif. Paris: Calmann-Lévy.
Meltzer, Allan H. and Scott F. Richard. 1981. “A Rational Theory of the Size of Government.” Journal of
Political Economy 89: 914-927.
Minford, Patrick. 1995. “Time-Inconsistency, Democracy, and Optimal Contingent Rules.”Oxford Economic
Papers 47: 195-210.
Nattrass, Nicoli, and Jeremy Seekings. 1998. “Democratic Institutions and Development in Post-Apartheid
South Africa.” Manuscript.

19
Niskanen, William A. 1971. Bureaucracy and Representative Government.Chicago: University of Chicago
Press.
O’Donnell, Guillermo. 1993. “On the State, Democratization and Some Conceptual Problems: A Latin
American View with Glances at Some Postcommunist Countries.” World Development 21: 1355-1369.
O’Donnell, Guillermo. 1994. “Delegative Democracy?” Journal of Democracy 7: 34-51.
O’Donnell, Guillermo. 1995. “Partial Institutionalization: Latin America and Elsewhere.” Paper presented at
the International Conference on Consolidating the Third Wave Democracies. Institute for National Policy
Research, Taipei, August 27-30.
O’Donnell, Guillermo. 1997a. “Polyarchies and the (Un)Rule of Law in Latin America.” Paper presented at the
Annual Meeting of the American Political Science Association, Washington,D.C. August 28-31.
O’Donnell, Guillermo. 1997b. “Horizontal Accountability and New Polyarchies.” Paper presented to the Vienna
Dialogue on Democracy and Institutionalizing Horizontal Accountability. Vienna.
Offe, Claus. 1995. “Some Skeptical Considerations on the Malleability of Representative Institutions,” in Erik
Olin Wright (ed.), Associations and Democacracy. London: Verso.
Persson, Torsten, Gerard Roland, and Guido Tabelini. 1996. “Separation of Powers and Accountability:
Towards a Formal Approach to Comparative Politics.” Discussion Paper No. 1475. London:Centre for
Economic Policy Research.
Przeworski, Adam and Michael Wallerstein. 1988. “Structural Dependence of the State on Capital.” American
Political Science Review 82: 12-29.
Reis, Fabio Wanderley. 1996. “The State, the Market, and Democratic Citizenship.” In ElisabethJelin and Eric
Herschberg (eds.). Constructing Democracy. Boulder, Co.: Westview Press.
Schmitter, Philippe C. 1984. “Patti e transizioni: Messi non-democratici a fini democratici?” Rivista Italiana di
Scienza Politica 14: 363-82.
Shepsle, Kenneth A. “Institutional Equilibrium and Equilibrium Institutions.” In Herbert F. Weisberg (ed.),
Political Science: The Science of Politics. New York: Agathon Press.
Sutherland, S.L. 1993. “Independent review and political accountability: should democracy be onautopilot?”
Optimum: The Journal of Public Sector Management 24: 23-41.
Tirole, Jean. 1994. “The Internal Organization of Government.” Oxford Economic Papers 46: 1-29.

20
Table 1: Income Distribution, Taxes, and Government Consumption, by decreasing ratio of top to bottom
quintiles (Q5/Q1)
COUNTRY YEAR Q1 Q3 Q5 Q5/Q1 GOV TAX NONTAX TOTAL

GUATEMALA 1989 2.10 10.50 63.00 30.00 11.7 8.0 1.7 9.7
BRAZIL 1989 2.48 9.15 65.18 26.28 14.5 26.4 1.4 27.8
BARBADOS 1979 2.25 14.75 51.00 22.67 13.1

S AFRICA* 1993 3.3 9.8 63.3 19.18 21 24.7 2.0 26.7

CHILE* 1994 3.50 10.90 61.00 17.43 9 19.1 2.6 21.7
VENEZUELA 1990 3.61 11.65 58.41 16.18 15.4 19.2 5.1 24.3
BAHAMAS 1986 3.05 14.31 48.86 16.02 6.9
COLOMBIA 1988 3.70 12.50 55.90 15.11 11.2 12.4 1.0 13.4

RUSSIA* 1993 3.7 13.5 53.8 14.54 21 19.1 1.4 20.5

DOMINICAN R 1989 4.20 12.50 55.70 13.26 9.9 11.4 1.3 12.7
T&T 1981 3.43 16.09 44.86 13.08 10.0
COSTA RICA 1989 4.00 14.30 50.70 12.68 19.6 21.2 3.4 24.6
PERU* 1994 4.9 14.1 50.4 10.28 10 13.7 1.3 15.0
PHILIPPINES 1988 5.20 13.30 52.50 10.10 15.4 14.5 2.5 17.0
AUSTRALIA 1990 4.60 15.50 46.40 10.09 12.9 25.0 2.8 27.8

NZEALAND 1990 4.58 16.31 44.73 9.77 14.5 34.5 4.6 39.1
ECUADOR* 1994 5.4 13.2 52.6 9.74 7 14.9 2.0 16.9
USA 1990 4.60 16.60 44.20 9.61 13.1 18.2 1.6 19.8
TURKEY 1987 5.24 14.06 49.94 9.53 11.3 16.2 2.3 18.5
THAILAND* 1992 5.6 13.0 52.7 9.41 9 17.0 1.8 18.8
IRELAND 1987 4.93 15.84 44.60 9.05 14.5 43.4 4.9 48.3

BOLIVIA 1990 5.62 14.53 48.23 8.58 18.5 10.3 6.3 16.6
JAMAICA 1990 5.98 14.45 48.37 8.09 14.9

NORWAY 1986 5.10 18.09 40.20 7.88 15.9 40.2 8.2 48.4
MAURITIUS 1986 5.90 14.90 45.70 7.75 16.6 20.3 3.0 23.3
PORTUGAL 1990 5.70 16.90 42.40 7.44 22.3 33.8 2.8 36.6
DENMARK 1987 5.21 19.47 37.78 7.25 21.0 38.1 5.4 43.5
SRILANKA 1976 6.42 15.34 46.42 7.23 20.2

GREECE 1988 6.19 17.04 41.18 6.65 14.3 32.2 3.6 35.8
BANGLADESH 1986 6.99 15.07 46.03 6.59 38.9 7.9 1.6 9.5
FRANCE 1984 6.58 16.74 41.97 6.38 15.5 38.1 3.4 42.5

W.GERMANY 1984 6.59 17.97 38.88 5.90 15.6 27.4 1.8 29.2
SKOREA 1988 7.39 16.27 42.24 5.72 8.6 15.6 1.8 17.4
ESTONIA* 1993 6.6 15.1 37.4 5.66 24 29.1 1.5 30.6
UK 1990 7.78 15.89 40.99 5.27 16.3 34.1 3.3 34.1
SWEDEN 1990 7.40 16.70 38.20 5.16 22.3 37.7 6.7 44.4
NETHERLAND 1986 7.21 18.30 36.83 5.11 12.3 45.3 5.2 51.5

BULGARIA* 1992 8.3 17.0 39.3 4.74 15 29.3 8.9 38.2

PAKISTAN 1988 8.61 16.60 40.39 4.69 19.3 12.3 4.6 16.9
ITALY 1989 8.35 17.32 38.10 4.56 11.7 38.8 1.1 39.9
CANADA 1990 7.54 25.46 33.85 4.49 12.2 18.8 2.5 21.3
FINLAND 1987 7.88 19.53 33.97 4.31 16.2 28.7 2.9 31.6
INDIA 1990 9.10 16.90 39.10 4.30 30.2 11.5 2.8 14.3
SPAIN 1989 8.39 18.68 35.28 4.21 12.7 30.2 1.8 32.0
BELGIUM 1988 8.48 18.67 35.08 4.14 11.7 41.9 1.6 43.5
LUXEMBOURG 1985 8.75 18.05 36.00 4.11 10.2
SLOVENIA* 1993 9.5 17.1 37.9 4.00 21

POLAND* 1992 9.3 17.7 36.6 3.93 19 37.9 4.2 42.1

HUNGARY* 1993 9.5 17.6 36.6 3.85 13
LATVIA* 1993 9.6 17.5 36.7 3.82 22 25.3 1.6 26.9

21
 ROMANIA* 1992 9.2 18.4 34.8 3.78 13 26.5 3.5 30.0
CZECH R* 1993 10.5 16.9 37.4 3.56 22 38.0 3.0 41.0

SLOVAK R* 1992 11.9 18.8 31.4 2.64 24

Notes: * Data are from World Bank Development Reports, 1994. Otherwise, GOV is from PWT5.6, income
distribution
data from Deininger and Squire (1996), tax data from Development Reports, closest years. Data are for the
last
year available. Blanks indicate missing data.
Q1 the income (or expenditure or consumption) share of the bottom quintile
Q3 the income (or expenditure or consumption) share of the median quintile
Q5 the income (or expenditure or consumption) share of the top quintile
Q5/Q1 ratio of top to bottom quintiles
GOV government consumption expenditure
INCOME per capita income in 1985 PPP USD
TAX central government tax revenue as a proportion of GNP (one or two years later)
NONTAX non-tax central government revenue as a proportion of GNP
TOTAL total revenue of central government as a proportion of GNP

22
 Table 2: The Impact of Income Distribution on Taxes, Government Revenues, and
Government Consumption, based on data in Table 1

Dependent Variable: Tax Revenues of Central Government as a Percent of GNP, N =

44

Variable Coefficient Standard Error t-ratio P[³T³≥t] Mean of X

ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ

Q5/Q1 -0.72593 0.22314 -3.253 0.00226 8.957

Q1 1.7500 0.65022 2.691 0.01017 6.162
Q5 -0.65736 0.15410 -4.266 0.00011 45.08
Q3 1.6085 0.66582 2.416 0.02013 15.69

Dependent Variable: Total Revenues of Central Government as a Percent of GNP, N

= 44

Variable Coefficient Standard Error t-ratio P[³T³≥t] Mean of X

ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ

Q5/Q1 -0.80083 0.22465 -3.565 0.00092 8.957

Q1 1.8540 0.67993 2.727 0.00929 6.162
Q5 -0.71561 0.16117 -4.440 0.00006 45.08
Q3 1.7780 0.72374 2.457 0.01824 15.69

Dependent Variables: Government Consumption Expenditures as a Percent of GNP,

N=53

Variable Coefficient Standard Error t-ratio P[³T³≥t] Mean of X

ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ

Q5/Q1 -0.32059 0.10347 -3.098 0.00316 8.870

Q1 0.94546 0.26030 3.632 0.00065 6.278
Q5 -0.20874 0.71448E-01 -2.922 0.00518 44.47
Q3 0.37561 0.25170 1.492 0.14179 15.84

Note: These are results for the income variables, from OLS regressions, with interactiveheteroeskedacity, in
which they were introduced one at a time. The constants are not reported.

23

View publication stats