–Shamshad Pasarlay, Herat University School of Law and Political Science. Email: shamshad.bahar[at]yahoo.com
One of the daunting puzzles for scholars interested in constitutional design is how to craft a democratic constitution for a deeply divided society. The challenge is to form a system of government in which all religious, ethnic and linguistic groups of a deeply divided polity are adequately represented such that they accept their concerns are being addressed and their rights protected. Only then, the mutually distrustful ethno-religious communities will cooperate and respect governmental decisions as legitimate.
Among scholars of divided societies, a debate has emerged. On the one side are the advocates of “consociationalism”; and on the other side are the champions of “incentivism.”
Consociationalists argue that divided societies should establish a system of government that assures political representation to all important groups. This type of a government would ideally include the following components:
(1) “grand coalition” executive, meaning that important groups within a divided polity should be allowed to exercise meaningful power in government;
(2) group autonomy, meaning that the state should give power to a region dominated by a particular group, or the state should delegate a certain degree of executive or legislative power to group leaders;
(3) mutual veto rights, giving groups within a state the right to veto government decisions on matters of vital importance to a particular community; and
(4) proportionality in political representation.
By giving all communities of a divided society guarantees of political power, the consociational model of governance tries to form an environment in which rival communities will be willing to work together, leading ultimately to peaceful coexistence and cooperation among rival ethno-religious groups.
Incentivists argue that consociationalism is a poor response to the pathologies of democratic governance in divided polities; it only aggravates the communal conflicts that it is expected to solve, eventually leading to the partition of a divided polity. As an alternative, incentivists suggest that divided societies should form non-consociational unitary governments with powerful presidential systems. They should also adopt electoral systems that reward politicians who are able and willing to reach out to members of rival communities for support in exchange for moderate policies on divisive policy issues. Instead of dividing political power among people based on ethno-religious identities, the system should benefit politicians who are willing to establish policies that promote the interests of people who belong to rival communities. By doing so, incentivism promotes political moderation and compromise on foundationally divisive questions and prevents exacerbating communal conflicts.
Modern Afghanistan provides a typical example of a deeply divided polity. The designers of the country’s 2004 constitution debated both consociationalism and incentivism although they did not use these exact terms. Ultimately, they rejected consociationalism fearing that it might exacerbate communal tensions and push Afghanistan towards partition; instead, they adopted the incentivist approach to governance. The constitution thus establishes a unitary government with a powerful presidency, bans political parties on the basis of ethnicity and religion, rejects proportionality in state bureaucracy and adopts an election system that requires candidates for both parliamentary and presidential elections to reach out to all different communities – a system designed to reward political moderation and compromise across ethnic lines.
In practice, however, the incentivist arrangements did not work effectively. Afghans tended to vote on the basis of ethnicity and there was no cross-ethnic voting although candidates considerably moderated their policies on foundationally divisive questions. In other words, political moderation was not rewarded; ethnic hardliners in fact attracted most of the votes from their communities. As a result, political parties began to be formed on the basis of ethnicity and religion. Furthermore, candidates for all of the past four presidential elections put forward consociational tickets that represented members from at least three ethnic groups (the 2004 constitution requires a presidential candidate to run with two vice-presidents). In the latest 2019 presidential election, candidates moved a step forward towards consociational tickets by opting to include a third de facto vice-president to incorporate representatives from as many ethnic groups as possible in a bid to attract voters from a broad cross-section of Afghanistan’s divided communities.
The highly centralized government and the powerful presidential system adopted in the constitution continued to heighten “the sense of winner-take-all politics in a diverse” and heavily armed polity with powerful local actors. As a result, elections have always begot political crises and threatened Afghanistan’s fragile political stability. This is so because the losers of elections were unwilling to leave the battlefield empty-handed, which would mean that certain communities would have no significant share in political power. Electoral crises were halted only after the leaders of Afghanistan’s divided ethno-religious communities reached a compromise that guaranteed political power for most, if not all, rival community leaders.
For example, Afghanistan’s 2014 presidential election led to a political crisis that nearly divided the country. The crisis was resolved through a consociational power-sharing pact between the two leading candidates, Abdullah Abdullah, an ethnic Tajik with a Pashtun and an Hazara vice-presidents, and Ashraf Ghani, an ethnic Pashtun with an Hazara and an Uzbek vice-presidents. The Agreement provided that Ghani would be declared president while Abdullah will be his Chief Executive Officer (CEO), a position not envisaged in the 2004 constitution. The President and the CEO would share executive power, and together they would form a government that would be inclusive of all ethnic communities. The Agreement further provided that the President and the CEO should arrange for the sitting in the near future of a Loya Jirga – Afghanistan’s constitutional amendment body – to amend the constitution to create a post for a prime minister. The amendment would share executive power between the representatives of at least two most important ethnic groups in Afghanistan. This plan, however, was never realized.
Recently, Afghanistan’s 2019 presidential election once again put the country on the brink of political crisis and civil war. In February 2020, the Independent Election Commission (IEC) declared the incumbent President, Ashraf Ghani, winner of the elections. Ghani’s chief rival, CEO Abdullah, accused the IEC of favoring Ghani, rejected the announcement and “declared himself the winner and said he would form a government of his own.” On March 9, 2020, both rivals swore in as presidents in parallel ceremonies held in Kabul. Both Ghani and Abdullah pledged to form “inclusive and all-encompassing” governments, meaning that they would form consociational governments with representatives from almost all relevant ethno-religious communities. The crisis over the results of the 2019 presidential elections and the inauguration of two rival presidents threaten to paralyze post-election Afghanistan in a critical time – a period when the Afghan government is expected to sit with the Taliban militants to negotiate a political settlement of the Afghan conflict.
Afghanistan’s story of electoral and constitutional crises suggest that in a society as divided as Afghanistan, where decades of civil war and inter-communal violence has deepened distrust and enmity among rival communities, incentivism might not work effectively. Afghanistan’s divided ethno-religious communities do not trust politicians who hail from different groups to promote their interests no matter how moderate a politician’s policies are on divisive foundational issues.
For instance, Uzbek voters would vote for a Tajik candidate only because the Tajik candidate has an Uzbek representative on his election ticket not because the Tajik candidate is moderate. This is true for almost all ethno-religious communities in Afghanistan. Other than direct guarantees of political power, nothing else apparently can incentivize voters in Afghanistan to vote for a candidate who comes from a rival ethnic group. Afghans have not exhibited a willingness to cross-ethnic voting during elections although candidates considerably moderated their policy platforms.
It therefore seems that trust, which could arguably lead to cross-ethnic voting, is one of the most important elements for designing an effective incentivist approach to governance in deeply divided societies. When trust is absent, consociationalism seems to be a much more viable option, at least in the immediate short-term.
In 2014, Afghanistan took a major step towards consociationalism but failed to formally adopt it in the constitution. After five years, Afghanistan is yet again on the brink of political crisis with two rival presidents who are unwilling to see the powerful presidency taken by their rival. The current electoral stalemate, like previous ones, will ultimately be resolved through a consociational power-sharing agreement between the two rival presidents, but elections will continue to breed political crises unless the constitution’s incentivist approach is revisited.
It is thus time for Afghanistan to engage in a more serious effort to rethink the incentivist approach adopted in the constitution and move formally towards consociationalism. Afghanistan is so divided at this point that only consociational power sharing arrangements could possibly create the conditions in which mutually distrustful communities will be willing, however grudgingly, to work together and will hopefully allow these communities to experience the benefits of cooperating rather than fighting. It is on this foundation that trust and more effective cooperation can be built. Over time, ideally, consociational guarantees would no longer be needed.
Suggested Citation: Shamshad Pasarlay, A Country with Two Rival Presidents: Is it Time for Afghanistan to Formally Move to Consociationalism?, Int’l J. Const. L. Blog, Mar. 18, 2020, at: http://www.iconnectblog.com/2020/03/a-country-with-two-rival-presidents:-is-it-time-for-afghanistan-to-formally-move-to-consociationalism?
 A society is deeply divided when two basic conditions are met. First, the society contains a variety of different ethnic, religious, linguistic, or cultural communities. Second, and more importantly, the different ethno-cultural, religious, or other communities in the society are politically mobilized and identification with these communities is the basis of political behavior over a long period of time and a broad variety of policy issues. Ian Lustick, Stability in Deeply Divided Societies: Constitutionalism versus Control, 31 World Politics 325, 325 (1979). In these types of societies, political parties are often organized along ethnic and religious lines, and simple majoritarian democracy risks turning into domination of more populous ethnic or religious faction over all others. Because ethno-political groups do not trust each other, they find it hard to reach consensus on important questions of public policy.
 Constitutional Design for Divided Societies: Integration or Accommodation (Sujit Choudhry ed., 2008).
 Arend Lijphart, Power-sharing in South Africa 25–47 (1985); Arend Lijphart, Consociational Democracy, in The Oxford Companion to Politics of the World 188–89 (Joel Krieger ed., 1993).
 Donald Horowitz, A Democratic South Africa?: Constitutional Engineering in a Divided Society (1991); Donald L. Horowitz, Constitutional Design: An Oxymoron?, in Designing Democratic Institutions (Ian Shapiro & Stephen Macedo eds., 2000).
 From the time Afghanistan emerged as a nation in the late 19th century, its citizens have identified with one of a multitude of distinct ethnic, linguistic and religious communities. Communal identity has always had political salience and translated into political fragmentation. Afghans tended to trust members of their community more than members of other communities. And their trust was rewarded. Leaders tended to promote the interests of the communities from which they hailed. Communal divisions deepened during 25 years of civil war between 1976 and 2001, a period in which communities came to rely on their community’s militias for protection. In 2001, U.S. led forces invaded Afghanistan and in the wake of this invasion the international community sought to help Afghans establish of a democratic government. After years of communal fighting, Afghans tended to look to the leaders of communal militias or figures close to these leaders. They looked to these figures to represent their interests in the new government. From 2001 to 2004, leaders of Afghanistan’s mistrustful communities drafted, debated and ratified a democratic constitution.