Civil Society and Political Accountability: Propositions for Discussion

Jonathan Fox

A. Introduction1
How do civil society actors contribute to the construction and empowerment of institutional checks
and balances? Civil society clearly matters, but when, how and to what degree? This paper attempts to
identify some of the determinants of the extremely varied patterns of civil society impact on
accountable governance. In the process, this essay explores the challenge of public accountability as it
applies not only to the governance of elected regimes, but also to civil society actors themselves.
This paper draws on the analytical distinction between vertical and horizontal dimensions of political
accountability. The former refers to power relations between the state and its citizens, while the latter
refers to institutional oversight, checks and balances within the state (O’Donnell, 1999). This
distinction locates civil society efforts to encourage accountable governance along the vertical
dimension, as a counterpart to the electoral process. Indeed, electoral competition has been
acknowledged to be insufficient for accountability since the origins of constitutional democracy. Until
recently, however, the role of civil society actors in promoting electoral democracy has received much
more attention than their role in promoting institutional checks and balances (Schedler, 1999b: 340).
Civil society demands for state accountability matter most when they empower the state’s own checks
and balances. 2 By exposing abuses of power, raising standards and public expectations of state
performance, and bringing political pressure to bear, they can encourage oversight institutions to act,
as well as to target and weaken entrenched opponents of accountability. Civil society campaigns may
also drive the creation of certain institutional checks and balances in the first place. However, though
civil society’s contribution to accountable governance has been widely asserted, the causal
mechanisms that determine the patterns of civil society influence on horizontal accountability have not
been well-specified. 3 We still lack analytical frameworks that can account for the conditions under
which civil society actors manage to bolster the institutions of horizontal accountability.
In brief, civil society actors influence horizontal accountability in two main ways: directly, by
encouraging the creation and empowerment of institutional checks and balances, and indirectly, by
strengthening the institutions of vertical accountability that underpin them, such as electoral
democracy and an independent media. The causal arrow also points in the other direction, however.
Weak institutions of horizontal accountability can also undermine vertical accountability, which in turn
weakens civil society actors. For example, inadequate election oversight bodies can permit less-than-
democratic elections, and ineffective official human rights defenders can fail to stop frequent
violations of basic political freedoms. Without adequate checks and balances, the minimum conditions
for political democracy can remain weak or incomplete. Therefore, one most explore the dynamic
interaction between accountability’s horizontal and vertical dimensions to identify civil society
impacts.
This paper begins with a discussion of the some of the challenges involved in developing a working
definition of political accountability. The paper then develops several analytical propositions for
discussion, viewed through the lenses of state and societal accountability, respectively. This approach
complicates the vertical/horizontal framework by taking into account multiple levels of state-society
relations, as well as the interaction between them (subnational, national and international). The
empirical bases for these propositions draw primarily from two long-term research projects. One
focuses on democratization in Mexico, the other on the role of the international actors in state-society
relations, such as multilateral development banks and transnational advocacy coalitions.

B. Defining accountability
This essay is based on the following definition: political accountability limits the use and sanctions the
abuse of political power. Public exposure is necessary but not sufficient to limit or sanction the abuse
of power. Actors and institutions that promote accountability attempt to bind the exercise of power to
specific benchmark standards. Political accountability can be promoted through both state and non-
state institutions, and pro-accountability outcomes often depend on mutually reinforcing interactions
between the two. The following conceptual propositions unpack some of the assumptions underlying
this approach.
1) The relationship between democratic process and accountability is politically contingent.
The concept of accountability is caught in a definitional tension – is it a process or an outcome?
Logically it involves both – as does the concept of democracy. The two concepts do not refer to the
same processes and outcomes, however. Some analysts incorporate accountability into the very
definition of political democracy. 4 Such definitions implicitly suggest that democratic processes
inherently generate accountable outcomes. There are both empirical and conceptual problems with
conflating accountability with democracy.
How do we account for the extraordinary variation in the degree to which pro-accountability
institutions actually manage to limit political power and to sanction its abuse? Empirically, whether or
not democratic processes produce accountable governance outcomes varies widely – across states,
within states and over time. Persistent human rights violations with impunity under elected regimes are
only the most obvious example, as in the cases of repression of agrarian protest in Brazil and Mexico.
5 Many consolidated democracies also experience extended periods and deep pockets of “low
horizontal accountability.” For example, Japan, Italy, the US, and most recently Germany are known
for tolerating quite long-term, systematic political corruption. 6 Moreover, major institutions within
democratic regimes are often insulated from citizen accountability as an integral part of their mission,
such as central banks, intelligence agencies, and certain security forces.
Conceptually, a major collection of essays recently detailed many reasons why the relationships
between electoral democracy, representative governance and accountability are far from consistent or
obvious (Przeworski, Stokes and Manin, 1999). They show that electoral democracy is much more
about voter choices than it is about the inherent vertical accountability of politicians to voters. To
assume that accountability is built in to democracy, by definition, is an example of the common
problem of assuming that all good things go together. If procedurally democratic regimes fail to
produce accountable governance, does that mean that they are not democratic? This would be logically
analogous to the concept of “substantive democracy,” in which regimes that fail to produce socially
equitable policy outcomes are considered, by definition, to be less than democratic. Both accountable
governance and policies that promote socio-economic equity are normatively desirable outcomes of
state actions that may or may not emerge from procedurally democratic processes.
2) Transparency is necessary but not sufficient for accountability.
Some definitions suggest that transparency and “answerability” of authority are part of accountability,
by definition. For example, Schedler proposes a broad definition that includes three dimensions:
“enforcement, monitoring and justification” (1999a: 14). At the same time, he argues that “unless there
is some punishment for demonstrated abuses of authority, there is no rule of law and no
accountability” (1999a: 17). These two approaches are inconsistent. The first one suggests that
transparency is enough, while the second suggests that it requires some kind of sanctions. How are we
to assess the many cases in which transparency does not lead to any kind of sanctions? For example,
most human rights truth commissions come to mind, not to mention many very transparent cases of
corruption or electoral fraud. The World Bank has become significantly more transparent since the
early 1990s, but its public accountability has lagged behind. 7 If transparency is built into the
definition of accountability, then the concept of accountability risks being diluted. To conflate means
(transparency) and ends (accountability) could make it difficult to get at the key analytical problem
here, which is to explain the conditions under which transparency does effectively contribute to
accountability.
3) Accountability is inherently relational.
To operationalize the concept of accountability, one must specify who is accountable to whom.
Accountability refers to power relationships between actors and/or institutions. The accountability
relationships of different state institutions often point in different directions. Even in democratic
regimes, for example, not all state institutions claim to be accountable to the ostensibly sovereign
citizenry. For example, many central banks are designed precisely to be insulated from democratic
institutions. This measure is intended to render monetary policy accountable to other institutions, most
notably private markets and the international financial institutions.
Similarly, supreme courts are also designed to be accountable to non-electoral institutions, specifically
constitutions. When majority will and constitutional standards conflict, supreme courts are mandated
to side with constitutions. They are supposed to be supreme, after all.
Most multilateral economic institutions, such as the development banks, the IMF, the WTO or
NAFTA, are also remote from processes of vertical accountability. Their decisions are formally
accountable to nation-state representatives, via appointed functionaries that represent those nation-state
institutions most remote from national democratic accountability mechanisms: finance and trade
ministries. The European Union is a major exception to the trend, since it includes a regional
parliament and therefore some degree of vertical accountability.
4) Accountability is inherently relative.
To operationalize the concept of accountability, benchmark principles are required. These standards
are inherently socially and politically constructed, and may or may not be enshrined in law. Legal
standards, moreover may or may not be seen as legitimate. Punishments may or may not “fit the
crime.” After all, authoritarian regimes can use laws to hold dissidents “accountable.” Specific
benchmarks are necessary for measuring and explaining varying degrees of accountability. A
“relativist” approach is critical because some partial degree of accountability is usually the most one
can expect, given the power of antiaccountability forces. In case of human rights, for example, even
very limited constraints on the abuse of power involve matters of life and death.
C. Propositions for discussion
The following analytical propositions about the role of civil society in horizontal accountability focus
first on state institutions, then on issues of accountability within civil society itself. These propositions
focus on the interdependence between vertical and horizontal accountability.
I. Political accountability seen from the state:
This section briefly reviews several key state institutions of horizontal accountability in terms of their
relationships with civil society: legislatures, the judiciary, and subnational governments. Then, to
explore the interaction between vertical and horizontal accountability, the concept of “reverse vertical
accountability” is introduced. If vertical accountability refers to citizen power over the state, then
reverse vertical accountability refers to state power over citizens.
1. Legislatures, usually the most important counterweight to executive power, often suffer weak
vertical accountability.
Legislatures are widely seen as more permeable and responsive than executives to organized citizens.
This is sometimes the case, but the pro-accountability potential of each potential counterweight
depends on their specific institutional characteristics. In the case of legislatures in which weak political
parties with shallow roots in society control the lists that determine who is actually elected to congress,
vertical accountability potential is weak. Where access to the mass media depends primarily on private
money, then political parties must manage the tension between being accountable to voters versus
private investors in politics.8 (Efforts to limit the influence of private money on elections by regulating
contributions, the “supply side” are inherently limited. The “demand side” is a more promising arena
of intervention – limiting political parties need for funds by eliminating their major expense, which is
the cost of buying broadcast media time. Brazil’s laws limit the influence of money in politics by
defining campaign access to prime time TV as a public good, removing it from the private market, and
assigning time in proportion to past election results.)

Where parties with shallow roots in society are combined with laws that forbid reelection, then civil
society capacity to hold legislators accountable for performance is further reduced (over and above the
usual high information costs involved in monitoring their actions).
Civil society actors and opposition political parties have somewhat divergent interests in terms of the
empowerment of legislatures as counterweights. Both have a stake in monitoring the government in
power, and empowering legislative oversight is one of the most important means. For most opposition
parties, however, this interest is primarily instrumental. Their leaders do not necessarily want to
institutionalize such checks and balances, in case they should come to power someday. Moreover, they
have an interest in bolstering legislative autonomy not only from the executive, but from voters as
well.9 In contrast to parties, civil society organizations usually do not seek to govern. They therefore
face fewer inherent conflicts of interest and can play critical roles in keeping opposition parties
accountable should they come to power.
2. Judiciaries, one of the most important horizontal accountability counterweights, are rarely
designed to be vertically accountable to civil society.
Judiciaries are designed primarily to be accountable to the rule of law, not to organized citizens (except
in the few systems where some judges are elected). Formal accountability and responsiveness are not
equivalent, however. Legal systems can be responsive to civil society initiatives, which in turn can
encourage efforts to transform them (Méndez, O’Donnell and Pinheiro, 1999; Smulovitz, 1997). This
is the assumption of key international actors, such as the Ford Foundation, which have made
significant investments in public interest law around the world (McClymont and Golub, 2000). Such
investments are critical because the costs of making even moderately responsive legal systems “work”
consistently are inevitably high. Where the courts are relatively autonomous, they can serve to bypass
obstacles to accountability in the legislative or executive branches, as in the case of the US civil rights
movement. Two major questions remain, however. First, what factors permit the courts to gain the
willingness to exercise the autonomy needed to play their role as horizontal counterweight? Second,
what factors determine the judiciary’s capacity to enforce the law? 10
Post-authoritarian conflicts over whether to prosecute human rights violators are among the most
notable cases of politically contingent enforcement of the rule of law. For example, Chile’s judicial
system did little to hold official human rights abusers accountable until recently. This was a dramatic
example of transparency without accountability. However, after a wave of international civil society
initiatives that successfully provoked action by judicial authorities in Europe, Chile’s judicial system
began to defend the rule of law by defining past disappearances as crimes subject to prosecution.
However, Chile’s judiciary was already internally divided, and international pressure appears to have
tipped the internal balance. The judiciary's institutional powers did not change, but its willingness and
capacity to act did, at least for a minority of judges. An approach that focused on formal institutional
powers would not be able to explain this process, since it was a shift in the balance of power between
actors that led to a small but significant change in the correlation of forces, both within and between
institutions. 11
3. Federalism, designed to promote horizontal accountability, can conflict with vertical
accountability.
Relationships between different levels of government look vertical but can be understood as
horizontal. In federal systems, local, provincial and national governments are supposed to checks and
balances on one another. Each level of government usually includes its own combination of horizontal
and vertical dimensions, insofar as provinces have both governors and legislatures, and municipalities
have both mayors and councils. However, in some regimes these relationships are very imbalanced.
Brazil’s state and city governments, for example, combine a high degree of autonomy from the federal
government with weak legislatures. 12
Decentralization is widely assumed to bring government closer to the people, and therefore to
encourage vertical accountability. However, subnational executive authorities often lack their own
checks and balances at the provincial or municipal level. Moreover, some subnational governments are
the size of small or medium-sized countries, so the information costs for monitoring them remain very
high – for central governments as well as for citizens. Central governments share resources and power
with subnational governments based on certain standards for their use, which raises challenges in
terms of monitoring and enforcement of those standards. 13
In addition, some subnational governments may fall short of the minimum conditions for political
democracy. Policies that promote decentralization of authority and resources to subnational
governments that are less than democratic may well strengthen authoritarian elites, at least in the short
to medium term.
Entrenched authoritarian subnational governments challenge national authorities to choose between the
horizontal accountability embodied in the federal system of relatively autonomous subnational
governments, and vertical accountability to citizens deprived of their democratic rights. 14
Even where local governments are not overtly authoritarian, powerful institutional incentives may
encourage exclusion of constituencies they are supposed to represent. 15 The Mexican experience
suggests, for example, that the horizontal accountability of rural local government has a reciprocal
relationship with the nature and density of civil society at the grassroots level. Broad-based
community participation can encourage more transparent and accountable local governance, while top-
down rule by local elites can stifle the potential for vibrant civil society (Fox, 1999).
To develop a more general framework for understanding the relationship between federalism and
horizontal accountability, one needs an interactive approach. First, one can safely assume that degrees
of horizontal accountability of subnational governments vary widely within most national systems
(O’Donnell, 1993, Fox, 1994b). Second, this variation is likely to include some that are more
accountable (in both vertical andhorizontal terms) than the national government, and others that are
less accountable than the national government. Third, the balance of power between levels of
government is both cause and effect of this variation in subnational accountability, but we know
relatively little about the nature of this interaction. For example, under what circumstances do
“advanced” subnational governments induce multiplier or scaling up effects, versus when do they
remain isolated enclaves of accountability? Certainly programmatic political parties, dense civil
societies and independent media are crucial for explaining the sustainability and replication of
innovation, but what else? 16 At the other extreme, under what circumstances do subnational
governments that lag behind progress at the national level catch up? If subnational political elites
manage to exclude part of the electorate from democratic competition, then national political leaders
will have little incentive or capacity to be accountable to those citizens. Electoral systems that
overrepresent the least accountable regions in national politics exacerbates this problem, as in Brazil.
5. Authoritarian clientelism constitutes “reverse vertical accountability” from citizens to state
actors.
Citizenship is supposed to combine a balance of rights and duties. Sometimes, however, state actors
gain the upper hand by holding citizens unduly accountable for certain behaviors, such as political
dissent or culturally proscribed activities. Clientelism, for example, refers to relationships of political
subordination in exchange for material rewards. Specifically authoritarian clientelism enforces such
imblanced bargains with threats of coercion. Other forms of clientelism enforce bargains with threats
of the withdrawal of critical services. In terms of the concept of accountability, authoritarian
clientelism violates fundamental democratic principles in two ways. First, authoritarian vote-buying
renders electoral competition less than democratic, undermining the regime’s potential vertical
accountability. Second, authoritarian clientelism obliges citizens to abstain from participating in
organizations that will be accountable to them, weakening civil society. 17
Furthermore, regimes that use the allocation of public resources systematically to reward and punish
citizens create a form of “reverse vertical accountability,” requiring clients dependent on such
resources for their survival to be accountable to state patrons. These authoritarian relationships are
rarely visible on election day, since the “bargaining” is usually subtle and takes places in advance.
Nevertheless, the regions and social groups most vulnerable to these forms of control are rarely
comprehensively monitored by civic groups even on election day, as the 1994 Mexican and 2000
Peruvian elections show. It is very difficult for even the most consolidated civic movements to monitor
more than a minority sample of polling places.
Widespread confusion also persists regarding the relationship between authoritarian social programs
and electoral politics, as seen in the US media coverage of the Peruvian election and its repeated
references to Fujimori’s strong electoral “support” in remote rural areas. 18 In spite of the fact that the
same reports referred repeatedly to charges of systematic fraud and electoral use of social programs,
the use of the term “support” implies that pro-regime votes from remote rural areas reflect
democratically exercised electoral preferences. Where people vote for the official candidate to retain
access to government programs that feed their children, or out of fear of perceived possible reprisals,
fraud is often not necessary. This problem reveals that the conventional concept of fraud is far too
narrow to encompass the full range of actually existing authoritarian electoral practices (Fox, 1994b;
1996b).
Because the power of vote-buyers depends in part on their capacity to monitor the compliance of
clients, ballot secrecy becomes a critical democratic resource. For one vivid example, one of the
patterns of ballot secrecy violations most widely observed by Alianza Cívica in the 1994 Mexican
presidential election took a puzzling form: voters deliberately revealed their marked ballots to others.
Alianza Cívica concluded that these voters apparently felt pressured to demonstrate to local bosses that
they had kept their part of the vote-buying bargain. In the case of ballot secrecy, “formal” democratic
procedures become most important to the politically weakest members of society. 19
Authoritarian clientelism appears to be especially resistant to NGO efforts. Not only do election
watchdog groups have great difficulties documenting its scope, but they have even greater problems
deterring its practice. When Mexico’s Civic Alliance documented efforts to influence voters
(“coacción”) in more than 25% of polling places observed in the 1994 presidential election, they also
revealed their weak capacity to deter such practices (Fox, 1996b). Transparency was necessary but
insufficient for accountability. Weakening authoritarian clientelism requires on-going monitoring of
state action, as well as some capacity to protect dissidents from reprisals, and therefore depends at least
as much on the empowerment of organized citizens as on external NGO oversight.
The concept of reverse vertical accountability may also be useful for grappling with the unresolved
relationship between the concepts of clientelism and state corporatism. The two ideas are distinct but
overlapping, with clientelism referring to (usually under-specified) practices and state corporatism
referring to certain organizational forms. However, both concepts focus on state-society relations in
which the accountability relations flow more from society upwards than from the state downwards:
reverse vertical accountability. 20
6. Pro-market public policies may reconfigure rather than reduce state intervention, and
therefore
do not necessarily weaken levers of reverse vertical accountability
Pro-market economic reforms are usually associated with the regulatory withdrawal of the state from
the market and social life. One might therefore expect such reforms to remove or to weaken the state’s
levers for inducing reverse vertical accountability. In fact, neoliberal economic reforms in several
Latin American countries weakened state levers for channelling rural politics, contributing to the
politicization of ethnic identities and the rise of broad-based indigenous protest movements (Yashar,
1999).
However, pro-market economic reforms may also be accompanied by social policies that maintain
significant state intervention in economic and social life. In Mexico (and apparently Peru as well), the
central state has withdrawn support for family farming and agrarian reform, but also introduced a
series of highly interventionist programs that re-regulate rural social and economic life (Fox, 1995).
The Mexican federal government now manages three national programs that require it to supervise and
regulate millions of individual rural citizens in three new and unprecedented ways: plot-specific grain
production patterns (PROCAMPO), individual land titles (PROCEDE) and maternal and child
education and health practices (PROGRESA). PROGESA is officially based on reverse vertical
accountability; if state bureaucrat monitors decide that mothers have failed to send their children to
school, to provide labor to local clinics or to submit to mandated medical tests and classes, then the
state will suspend their transfer payments. This is the formal dimension of reverse vertical
accountability, electoral politics often add an informal dimension to the power relationship. Perhaps it
is a coincidence, but before electoral politics was competitive in Mexico, most state regulation of rural
life did not require institutionalized relationships with individuals. 21
7. International actors have contradictory effects on state accountability.
Explanations of the construction of horizontal accountability need to bring transnational actors in to
their frameworks. For example, to refer to international actors as the “third dimension” is conveniently
straight forward (Pastor, 1999), but needs to be further developed to capture the range of patterns of
interaction between domestic and international factors.
International actors have long been active in constraining both vertical and horizontal accountability in
Latin America, most notably the US government and some private enterprises. Since the 1970s,
however, the range of international actors has become considerably more diverse. Local and national
proaccountability civil society actors have found their own sets of transnational allies, ranging from
churches and private foundations to human rights, environmental, women’s rights and indigenous
rights networks. These local/global coalitions are increasingly recognized to be significant political
actors in their own right, as civil society campaigns use “boomerang” strategies to influence nation-
states from both above and below (Brysk, 2000, Keck and Sikkink, 1998; Risse, Ropp and Sikkink,
1999). These coalitions have been very influential in terms of raising public expectations and standards
of horizontal accountability, as well as influencing related public policies. Non-state international
actors usually lack the capacity to sanction, however, other than through their influence on their own
governments. As a result, their impact on actual state behavior has been much more limited and
uneven, which suggests limits of the boomerang strategy. 22
Since the end of the Cold War, the US government itself has invested significant amount of bilateral
aid in institutions of horizontal accountability, as well as to support civil society organizations that
work to improve vertical accountability. So far, these efforts have made a difference mainly at the
margin (Carruthers, 1999). At the same time, however, in some countries the US government
continues to invest in institutions that weaken horizontal accountability. The most notable recent
example involves massive aid to the Colombian armed forces, whose involvement in repression is
quite transparent. 23 Similarly, the multilateral development banks invest simultaneously in both pro
and anti-accountability institutions within states, in part because they themselves are internally divided
over whether to promote transparency, civil society participation and public accountability. Official aid
funding therefore has “net accountability impacts” that vary widely both across and within nation-
states.24
II. Political accountability seen from civil society
If civil society actors are themselves publicly accountable, the may be more likely to contribute to the
reform of the state. What are the institutional factors that encourage civil society actors to be
accountable to public interests, rather than to very particular or private interests? This is one more way
in which vertical and horizontal accountability are interdependent.
To identify mechanisms of vertical and horizontal accountability within civil society requires
distinguishing between two different categories of associations. Membership organizations, whose
main goal is to represent the interests and goals of their members are qualitatively distinct from
nonmembership-based organizaciones civiles (NGOs), entrepreneurial, service and advocacy
organizations that pursue ostensibly broader, society-wide goals. Some civil society organizations
reach across both categories, as in the case of human rights organizations composed of victims’
relatives.
When discussing civil society organizations, public accountability and the reform of the state, it is
important to keep in mind that many civil society actors primarily reinforce institutional arrangements
that limit public accountability and reproduce elitist political cultural legacies. This would characterize,
for example, most broadcast media, as well as elements within some religious hierarchies, traditional
charities and disaster relief organizations. Civil societies also include some movements that oppose the
extension or consolidation of social and political rights sought by other movements, most notably
women’s rights. Looking at civil society in this broad sense of including its powerful pro-status quo
elements reminds us that it includes forces of inertia as well as forces for change, as suggested by
Gramsci’s notion of hegemony. The following propositions focus mainly on how civil society actors
may contribute to horizontal accountability by dealing with their own challenges of accountability as
well.
1. The Iron Law of Oligarchy, though alive and well, is more of a powerful tendency than an all
powerful “law.”
This classic sociological principle contends that the leaders of large membership organizations
necessarilydevelop their own distinct interests, which encourages them to increase their autonomy
from the base. Thispowerful tendency often undermines internal democracy within membership
organizations, but to varying degrees that ebb and flow over time and across organizations. A large-
scale cross-sectional study of 150 organizations found that combinations of horizontal and vertical
linkages are critical (Esman and Uphoff,1986). A longitudinal case study found that the existence of
internal horizontal accountability mechanisms, such as instancias intermedias de participación, are
crucial complements to conventional vertical accountability mechanisms, such as elections (Fox and
Hernández, 1989, Fox, 1992b). In the absence of such participatory subgroups, leaders of large
organizations only need to deal with atomized individuals who usually lack opportunities to share
information and generate alternative opinions, counter-proposals and potential leaders. Participatory
subgroups, in contrast, can increase member power (vertical accountability) by monitoring leader
performance and brokering leader access to member votes and other resources. Members are no longer
atomized. Subgroups may still be insufficient for accountability, however, since they may well
constitute factions that only represent the interests of small minorities. However, as in the case of
national politics more generally, organized factions that oblige those in power to be more transparent
generate a public good even if they are only acting in their own self-interest.
It is important to keep in mind that membership participation is not the only possible means for
influencing leadership accountability. Members may also have exit options, except for extreme cases
of authoritarian state corporatism. Given exit options, members of an organization under oligarchic
pressure will consider putting energy into pro-accountability strategies in so far as the array of allies,
enemies and opportunities make the exercise of voice a plausibly effective approach. This is yet
another example of the mutual interdependence between horizontal and vertical accountability.
2. Societal organizations can encourage state accountability without necessarily being formally
accountable to their base
In Latin America, movements for democracy are defined by their opposition to authoritarian rule
rather than by their internal power relations. However, “democratic movements” are only occasionally
democratic themselves. Contradictory as it may seem, movements for democracy and accountability
do not necessarily need to have formal internal democracy in order to have pro-accountability impact
on the state. For example, in Mexico, repeated waves of protest by civic groups and opposition parties
led to a significant degree of empowerment for the regime’s newly-independent election oversight
body. They thereby contributed to horizontal accountability without necessarily being internally
democratic political parties or civic organizations. Armed protest even helped the process; the
Zapatista uprising bolstered the bargaining power of groups that favored empowering the
“citizenization” of election management, tipped the balance within an internally divided state and led
to a major electoral reform agreement among the political parties in late January, 1994 (Fox, 1994c).
For a Brazilian example, the rural poor are represented primarily by two very different kinds of
organizations, a formally democratic, decentralized, left-led federation of rural workers (CONTAG)
and a direct action-oriented, highly centralized political organization that eschews formal internal
democracy (MST). Observers across the spectrum agree that, in terms of increasing the state’s public
accountability in the area of agrarian reform policy implementation, the MST has been more influential
than CONTAG (Navarro, 2000).
3. The accountability of mass media to civil society is problematic because it is simultaneously a
market and a civil society institution
The mass media often act to fill the gaps left by ineffective state institutions of horizontal
accountability. The media also plays a critical role in encouraging existing institutions of horizontal
accountability to do their jobs.
The mass media are the most notable case of a societal institution that contributes to state
accountability without necessarily being accountable to any other actor. Independence is one of the
media’s great potential strengths, but the media is a market as well as a societal actor. Some media are
accountable primarily to market pressures, and few can ignore them. Only occasionally is there a
convergence between market incentives and journalistic contributions to transparent governance.
Others are accountable to specific ideologies, or to certain partisan political elites.
Broadcast media are especially vulnerable to state control, and often fall short of their potential
contribution to more transparent government (as in Peru and Mexico, among others). In terms of
electoral access, as long as media time is allocated according to purchasing power, then that dimension
of the media will be most accountable to the campaign with the most money to spend on airtime. Since
few media owners or top editors offer incentives to rock the boat, the role of professional norms and
horizontal social capital becomes critical to encourage the media contributions to transparent
governance that are so essential to accountability. This is why the NGOs, universities and membership
associations that defend freedom of the press, individual journalists, and set high standards for
contributions to the public interest are so important.
4. NGOs can be held publicly accountable even though they lack precisely-defined constituencies
NGOs usually lack a clearly-defined constituency, so defining their accountability processes is
inherently problematic. They may be accountable to universal ideals of democracy, justice, or
environmental sustainability, as many claim – or not, as some external critics argue. NGOs tend to
zealously defend their autonomy, so who decides whether they are accountable to their ostensible
goals, and who sets the standards? Since accountability is relational, however, NGOs can be
accountable in many different directions. For example, they can be accountable to grassroots
constituencies, though these power relations vary greatly in practice.25 They can be accountable to
charismatic leaders, political parties, or to certain religious faiths or ideological currents. They can also
be accountable to their funders, whether private or governmental. One of the most promising sources
of NGO accountability is to each other, through coalitions, as discussed below.
One of the most important sources of NGO accountability is the possible distance between the high
standards they publicly set for themselves and their actual practices. 26 If the distance between goals
and practices becomes very large, or very obvious, then the NGO will have a credibility gap. This is a
major potential source of accountability because most NGOs need credibility to survive. They need
credibility with the media to have a public voice, with grassroots partners to have popular legitimacy,
with elites to influence policy, and with funders to gain the material support essential for institutional
survival. 27 Since different NGOs often share these same goals, there is an element of competition that
creates an incentive for mutual oversight, and therefore mutual accountability.28
Many NGOs have also formed networks and coalitions, and these fora can serve to set at least some
“rules of the game” for NGO accountability. When consolidated, networks and coalitions can
constitute “miniregimes” of self-governance within a given region or sector. 29 Coalitions that set
explicit shared goals and carry out joint actions thereby set standards of mutual accountability. NGOs
can sanction those who do not comply with these expectations by excluding them from the coalition,
with an associated loss of credibility. The more general point here is that civil society actors can held
accountable, even if the processes are.not clearly institutionalized.
5. Social movements and NGOs can contribute to political accountability through their influence
on political cultures
According to conventional indicators of mobilization, social movement activity often declines after
transitions to democratic regimes. Many observers conclude that social movement impact on politics
therefore drops off. However, the character of social movement impact on politics is not necessarily
obvious or fixed; its nature can adapt and change. Influence on state behavior has the great advantage
of being tangible, but impact on political cultures, though more difficult to measure, matters greatly as
well. Recall that much of conventional civil society activity in defense of organized interests is largely
instrumental or focused strictly on material issues, and therefore does not necessarily contribute to
changing the culture of politics. In this context, those social movements and NGOs that do focus on
transforming political cultures make an especially important contribution to horizontal accountability.
30
What is the link between political culture and accountability? Many analysts attribute the lack of
accountability to particular sets of values. Others, such as World Bank corruption experts, focus more
on institutions and incentives (Kaufmann, 1999). One does not need to resolve this debate to
acknowledge that political cultures matter for horizontal accountability, however, because even if
dysfunctional state institutions are not caused primarily by cultural factors, the process of reforming
them may well require changing political cultures.
Political culture embodies attitudes, values and behaviors – all of which matter for state accountability
because they influence what citizens expect of the state. In other words, changing expected standards
of behavior is key for promoting accountability, within civil society as well as the state. This
contribution broadens the societal constituency for the state’s horizontal accountability innovations, as
well as reinforcing the legitimacy of the notion that the state should obey standards of accountability in
the first place.
6. Transnational coalitions can change the balance of power among civil society organizations,
with implications for their mutual accountability
When NGOs or social organizations form coalitions across national boundaries, some groups gain
more access to resources, international media exposure, technical assistance and political support than
others. Transnational ties can therefore change the balance of power – not only between civil society
organizations and their targets, but also among different civil society actors. For example, NGOs often
benefit more from transnational networking than grassroots membership organizations, in part because
NGOs are less focused on short-term survival or self-defense and therefore are freer to make the open-
ended investment required. Groups that manage to project their issues in ways that resonate with
internationally influential civil society frames also gain more access to resources and media attention
than others (Bob, 2000). Therefore local groups have incentives to represent their issues in ways that
will resonate internationally. This can raise challenges for accountability within social movements or
NGO networks, once a few interlocutors gain access international circuits and speak in the name of
many who do not.
One study of a wide range of campaigns against World Bank projects explicitly addressed the issue of
the degree to which US and European advocacy groups were accountable to their local coalition
partners in the South, those directly affected by projects. Differences in power, priorities and culture
between Northern and Southern civil society groups inherently risks throwing such coalitions off
balance. Nevertheless, the study found remarkably few cases of so-called “green imperialism” (Fox
and Brown, 1998). Ironically, the World Bank’s frequent attacks on Northern critics’ alleged lack of
concern for the poor increased the pressure on them to be accountable to their Southern partners.
Mexico-US maquila worker support coalitions illustrate some of the relationships between vertical and
horizontal accountability. 31 A decade of maquila worker rights campaigns shows that cross-border
coalitions often help to increase leverage over transnational corporations, sometimes leading to partial
concessions (Williams, forthcoming). However, transnational civil society coalitions, even with access
to mass media, high-level US politicians and trinational labor institutions, have had almost no success
so far at defending workers’ right to freedom of association, which is key for vertical accountability.
32 Specifically, the right to freedom of association depends in part on the functioning of the state’s
labor courts – an institution of horizontal accountability. So far, a decade of local and international
concern, not to mention electoral democracy in northern states, has not made Mexican labor courts
more autonomous from partisan and private sector interests. This case shows how weak vertical and
weak horizontal accountability are mutually reinforcing, even when under international scrutiny.
7. The potential impact of civil society policy monitoring initiatives depends heavily on their
capacity for “vertical integration.”
Most discussions of political accountability stress the importance of civil society monitoring of the
policy process. The information costs required for effective vertical accountability are very high. Once
civil society actors are armed with reliable information about state behavior, however, they can act
strategically to bolster agencies of horizontal accountability within the state. Policy monitoring is
critical to identify not only abuses of power, but also possible opportunities for civil society leverage.
Civil society impact is as much about targeting and weakening the forces of impunity as it is about
bolstering pro-accountability institutions. To be effective, civil society accountability strategies require
reliable information about where precisely to target advocacy campaigns.
The national public policy process is increasingly entangled with multiple levels of authority, both
above and below the national arena. National executive authorities share power not only with other
national horizontal institutions, such as legislatures, but also with international financial institutions,
private investors on the one hand, and with relatively autonomous sub-national governments on the
other. As a result, when national policymakers respond to civic and social organizations that are trying
to hold them accountable, it is very convenient for them to emphasize -- or even to exaggerate --the
increased weight of either international or sub-national policymakers. Decision-makers at different
levels can step aside and point the finger elsewhere, and opaque policy processes make it very difficult
for advocacy groups to assess their claims. 33 For example, the national government can, in the name
of the federalism, sidestep all kinds of anomalies committed by state and municipal governments, as in
the cases of Mexico’s “Deep South” in the 1990s and the US’s Deep South in the 1950s. 34
Turning to international actors, multilateral development banks often respond to criticism by
attributing “problem projects” to national and local governments (as in the recent case of the political
crisis following Bolivia’s water privatization). 35 This is sometimes true, but how can independent
observers know when? Conversely, when a national government makes a socially or environmentally
costly decision, it may be very convenient to have the World Bank look like it forced them to do it, so
that the direct political cost to them is reduced. The World Bank may implicitly accept bearing the
burden of blame, as a political favor to its allies in the borrowing government. In short, public interest
strategists need to know where the key decisions were made, otherwise their efforts will be
mistargeted. This dilemma can be depicted as a process of “squeezing the balloon.” When civil society
organizations they squeeze the balloon over here, it pops out over there.
Since one could describe the multi-level public policy process in terms on vertical integration, civil
society efforts to influence public policies need to integrate vertically as well. “Vertical integration” of
monitoring and advocacy strategies refers to the systematic coordination between diverse levels of
civil society, from local to state, national and international arenas. One can find these vertical linkages
either in specific sectoral issue areas, such as human rights, reproductive rights, the defense of
biodiversity, or in broader multisectoral campaigns that cut across issue areas, such as the coalitions
that campaign to increase the public accountability of the World Bank as an institution.
The most consolidated Latin American effort to vertically integrate civil society accountability
campaigns in this arena is the Brazilian Network on the International Financial Institutions. 36 This
coalition includes both many leading national NGOs and broad-based social organizations. Their main
goal is to work with diverse civil society organizations to bolster the capacity of the federal congress to
exercise its mandate to oversee the nation-state’s relationship with the international financial
institutions. In other words, they use vertical accountability mechanisms to strengthen horizontal
accountability at the national level. At the same time, they submit claims to the World Bank’s own
horizontal accountability mechanism, the Inspection Panel. 37
Another approach to vertical integration starts with intensive, medium-term investments in local
support for grassroots organizations, to increase their capacity for informed participation in and
monitoring of the policy process. The Mexican NGO Trasparencia pursues a two-track strategy,
encouraging greater transparency from above by putting relevant policy information directly into the
hands of representative stakeholders, while also working with those same organizations to involve
them in the process of policy monitoring from below. 38 Because this process is so labor-intensive,
however, it is difficult to focus on more than a small number of regions and policies at the same time.
Both Rede Brasil and Trasparencia’s strategies are based on the principle of “vertical integration” of
policy advocacy.
8. Civil society actors can have a comparative advantage in policy monitoring, but sustained
investments with uncertain payoffs are required.
Since transparency is a precondition for accountability, civil society institutions face the challenge of
investing the resources required to generate reliable, convincing and widely-accessible independent
evaluations of state actions. Where there is a large gap between state commitments and actual
performance, civil society actors have an opportunity to empower those organizations and intellectuals
with the capacity to document that gap.
Some civil society actors have focused their national policy transparency efforts on the multilateral
development banks, in part because most nation-states share much more information with the banks
than with their own civil societies. Therefore, internal development bank documents often contain
important policy information that is not publicly available to a borrowing government’s citizens. The
development banks have increased their degree of public transparency significantly since the mid-
1990s, but it turns out that they often lack their own sources of reliable information about the
performance of borrowing governments, and instead tend to rely on government self-evaluations (Fox,
1997). For example, while development banks have made large investments in social and
environmental initiatives, they rarely bolstered independent oversight mechanisms in either state or
society.
Table One shows the actors and institutions that might do monitoring and evaluation of the actual
results of public policies. This conceptual map shows why independent civil society policy monitors
have a particular comparative advantage: most evaluations of government/multilateral development
bank social and environmental investments are either confidential, produced by interested parties, or
both. Evaluations that remain confidential lose most of their pro-accountability potential. Independent,
public evaluations can fill the information gap and bolster accountability efforts. The information must
be timely, credible, as well as both targeted and accessible to those actors most likely to use it (such as
direct stakeholders in specific policy issues). Information does not generate advocacy or mobilization
by itself, however. For transparency to matter, civil society, political party and state actors need to
have the willingness and capacity to use it politically.
D. Conclusion
This essay’s main theme is that civil society role both frames and is framed by the interaction between
the vertical and horizontal dimensions of accountability. Weaknesses in electoral democracy can
undermine horizontal oversight institutions, and vice versa. Conversely, strong oversight institutions
can empower mechanisms of vertical accountability. This is the dynamic context within which actors
that favor accountability come together in efforts to strengthen checks and balances.
Schedler has identified four main sources of pro-accountability reform: “governments (reform from
above); civil society (reform from below); staff members (reform from within) and international actors
(reform from the outside)” (1999b: 338). In each of these domains, however, pro-accountability actors
are often weak, in relation to the other actors in their respective arenas. Pro-accountability initiatives
most likely encounter resistance both within and between state and society, but the ways in which such
conflicts unfold are not predetermined by a static initial distribution of power resources. The analytical
challenge, then, is to develop a framework that can capture the process of dynamic interaction in which
weak actors gain leverage. After all, pro-accountability initiatives often fail. Sometimes these failures
nevertheless weaken the opposition and therefore constitute steps toward reform, yet at other times the
failure of accountability efforts can actually bolster the forces of impunity. What makes the difference?
Under what conditions can pro-accountability actors set off “virtuous circles” of mutual
empowerment?
An interactive approach is needed to account for how different actors’ capacities to pursue their goals
changes through conflict and convergence. The strength or weakness of pro-reform forces is shaped
through their strategic interaction with each other and with their opponents. An interactive approach
also requires rejecting the still widely-held assumption that state and society are necessarily engaged in
a zerosum balance of power. 39 Pro-accountability actors within both state institutions and civil
society need to find mutually reinforcing coalition strategies that bridge state and society, to make the
whole stronger than the sum of the parts. In the iterative virtuous circles that bolster accountability,
civil society and state actors manage to empower one another and then embed reforms into the state.
These institutional levers, such as transparency and oversight bodies, then further empower pro-
accountability actors within both state and society, contributing to successive rounds of change.
In conclusion, the question of which inter-institutional relationships most effectively promote
horizontal accountability may involve more art than science, with the most promising institutional
configurations depending on particular actors, times and places. Perhaps the challenge can be
understood as a kind of political Feng Shui – the ancient art of placing things in balanced relationships
to one another.