Democracy’s succinct definition, and perhaps its best attribute, is majority rule. But it is unclear that majority rule equates to democracy in places like Lebanon, Iraq and other Middle Eastern countries that are contending with past and present religious or ethnic conflict. Clearly, democracy in such diverse societies would minimally require that citizens of all ethnic and religious backgrounds enjoy the same civil and human rights; it would also require that the government refrain from religious or ethnic persecution. A democracy should also allow its citizens to practice their faith or express their cultural traditions, provided such practices do not contradict other fundamental values the state is bound to uphold. But beyond these minimum guarantees, what constitutes a “just” distribution of political power among majority and minority groups in a religiously or ethnically divided society? And how can that distribution be reconciled with the principle of majority rule?
Who’s Afraid of Democracy?
All democratic political systems have elected parliaments; some also have directly elected presidents with substantial legislative influence. In both parliamentary and presidential systems, it is generally the case that laws must be passed by the parliament. With some exceptions, such as amending the constitution, which may require a “super-majority,” it is also the case that laws are passed by the majority of the members of parliament. As long as the basic human and civil rights of minorities are protected, this kind of “majority rule” is not inherently discriminatory against the minority — if the composition of the minority and the majority can change over time. In a parliament made up of several parties, the combination of parties that comes together to pass one law may be different from the combination that passes the next, so that “the majority” may actually change from day to day. In a parliament made up of two parties, the majority party may vote as a bloc so consistently that that party actually makes almost all the laws by itself. But, in theory, this majority can be overturned in the next elections.
Democracies without permanent majorities may have parties based on ethnic or religious groups that are likely to remain permanently in the minority. Such parties, however, can wield great power when the population is so divided that there is no single majority party. In a parliamentary system in which no party has 50 percent of the seats, a government cannot be formed until some coalition of parties is put together that controls 51 percent of the votes in parliament. Creating such a coalition often requires one of the larger parties to woo one or more minority parties, and even a small minority party can extract important concessions for joining — or not defecting afterwards. Minority parties can also flex their muscles if the parliament is so divided on a major issue that the minority party’s votes are enough to decide it.
But what if one majority ethnic or religious group consistently, election after election, votes to put members of its group into office? If the country has a judiciary committed to individual rights, then the basic human rights of minority groups might still be protected, but minorities will be unable to pass laws differing from the agenda of the majority. This scenario is not inevitable, since the majority might be quite divided along ideological lines and not share a common agenda. But it is also clear why minorities might find democracy threatening — as a permanent majority might abuse its power to their detriment.
For that reason, countries with significant heterogeneity, and particularly those with a history of religious or ethnic conflict, often create democracies with prominent anti-majoritarian features. A country might explicitly mandate direct representation of each group in the political system so that members of each group can look out for their own group’s interests. Or, framers might give as much power as possible to local or regional governments, rather than the national government, so that religious and ethnic groups are largely left to run their own affairs in the particular areas of the country where they are concentrated. Finally, countries might seek to protect various groups’ interests by ensuring that majority groups cannot take and exercise power without gaining the support of some minorities.
Lebanon, whose system of “consociationalism” mediates between 18 ethno-confessional groups, is the classic example of the first approach. In 1932, when Lebanon’s last census was held, the ratio of Christians (Maronites, Greek Orthodox, Greek Catholics, Armenian Catholics and other denominations) to Muslims (Sunnis, Shi‘a, Druze and Alawis) was six to five. Maronites were the single largest group in the population at 30 percent and Sunnis the second largest at 20 percent. The 1943 National Pact, a verbal agreement concluded between Maronite and Sunni leaders, established unwritten rules whereby the president would always be a Maronite, the prime minister a Sunni and the speaker of parliament a Shi‘i, with lesser positions being guaranteed to other, smaller groups. Seats in parliament would be divided in a six to five ratio between Christians and Muslims. Although Christians, who had historically been close to European powers such as France, were a majority at the time of the National Pact, they feared becoming a minority due to the aspiration of many Lebanese Muslims to dissolve Lebanon into a larger (majority-Muslim) Arab state. The National Pact was designed to protect the interests of the Christian majority as much as any minority group, as it was based upon the principle that “Lebanon’s Muslims should cease to strive for unification with Syria and the Christians should renounce Western tutelage.” 
Over time, the relative size of each group in the Lebanese population shifted, so that by 1998 it was estimated that Lebanon was about 70 percent Muslim and 30 percent Christian, and that the Shi‘a had probably become the single largest group. The Ta’if agreement of 1989 that paved the way for the end of the 1975–1990 civil war modestly adjusted the distribution of political power in the direction of this reality. Ta’if stipulated that Christians and Muslims would split seats in parliament 50-50. The agreement also modified a formula specifying the exact distribution of seats among confessions in each electoral district, so that, for example, one particular constituency in the 2005 elections elected five Shi‘is, one Sunni, one Druze, one Greek Orthodox, one Maronite and two Greek Catholics. While keeping the presidency for the Maronite, Ta’if substantively limited presidential powers in favor of the Sunni prime minister and the Shi‘i speaker of parliament, as well as the Cabinet.
In the early stages of the US-led occupation of Iraq, the Coali- tion Provisional Authority (CPA) adopted a Lebanese-style idea of ensuring direct representation in positions of political power for Iraq’s myriad ethnic and religious groups. Larry Diamond, the Journal of Democracy editor who served as a CPA adviser on democratization, notes that when US viceroy L. Paul Bremer created the Iraqi Governing Council in July 2003, “a numerical balance was agreed upon that would be replicated in each batch of appointments, beginning with the ministers. The Shiites, who constituted a majority of the population, would have a bare majority on the Governing Council — 13 of the 25 members — and there would be five Kurds, five Arab Sunnis, one Assyrian Christian and one Turkoman.”  A version of this approach continued from May 2004, one month before the CPA’s dissolution, until the January 2005 elections of a transitional national assembly. In the interim government headed by Iyad Allawi, ministers were selected according to the same strict religious-ethnic calculus.
The idea of guaranteeing such direct representation for ethnic and religious groups is intuitively appealing: how better to protect any group’s rights — majority or minority — than to promise hands-on participation in lawmaking to every group? In practice, however, this type of system tends to freeze existing power distributions in place, to decrease the weight of majority opinion and to hinder the development of an overarching national identity. Although Ta’if changed the equation slightly, Lebanese Christians are still guaranteed more political positions than any other group even though they are no longer a majority. By contrast, Lebanese Shi‘a, who probably make up a larger percentage of the population than all the Christian denominations together, are only allocated 22 percent of the seats in parliament. A new census could produce a more proportional allocation of positions, but political systems which tie power directly to the size of a group’s population turn censuses into sources of potentially violent contention, with strong temptations to engage in rigging. Even if positions were doled out in strict and updated proportions, what are the long-term consequences of inculcating the idea that citizens of a particular religious or ethnic identity automatically have distinct interests that only fellow members of the group can be expected to protect? One obvious corollary is that fellow citizens from other religious or ethnic groups cannot be trusted to have one’s interests at heart.
The Lebanese system is a curious way of institutionalizing this idea, because while each group is guaranteed a certain number of seats in parliament, members of many religious groups vote on which members of each group will take those seats. In a district in which one seat is to be filled by a Sunni, all registered voters in the district choose which Sunni will fill the position, so if Shi‘a were the majority of the district, the Sunni member will probably be chosen by the Shi‘i voters. This system has the advantage of encouraging cooperation across communities, as politicians of various faiths join to create one slate that includes the number of members of each religious group required to win in that district. Nevertheless, this sys- tem entrenches the idea that there are specific and enduring Maronite or Sunni concerns that are different from, and perhaps even require special protection from, other communities. The lack of a strong Lebanese national identity facilitated the outbreak of the civil war, and it is not surprising that Ta’if, while reaffirming the principle of guaranteed positions for each community, also stipulated future review with an eye toward the “deconfessionalization” of Lebanese politics.
An Extreme Form of Federalism
Since the promulgation of a Transitional Administrative Law (TAL) in March 2004, the post-Saddam Iraqi political order has been moving toward a form of federalism that devolves considerable power to regional governments. Federalism was initially advocated by the Kurds, but has since garnered the support of many Shi‘a. Since 1991, with the establishment of the northern no-fly zone and the withdrawal of the Iraqi army from the three majority-Kurdish provinces in northern Iraq, the two major Kurdish parties have enjoyed a great deal of autonomy from Baghdad. Kurdish militias (peshmerga) policed the region, and, though there was much internecine conflict, the two parties formed institutions of a quasi-unified Kurdistan Regional Government (KRG) in 1992 and 1996. The Kurdish parties entered post-Saddam political negotiations, then, with the demands that the peshmerga remain intact and that the KRG retain its existing autonomy. The Kurds pressed for a clause in the TAL saying that a permanent Iraqi constitution, even if approved by a simple majority in a referendum, could be defeated by the opposition of two thirds of the voters in any three of Iraq’s 18 provinces, a requirement crafted to allow the Kurds alone to veto the constitution. On October 15, 2005, however, it was Sunni Arabs and other opponents of the same federalism provisions who came close to defeating Iraq’s new constitution with a two thirds “no” vote in two provinces and a 55 percent “no” vote in a third.
The wide margins of victory for the constitution in majority-Kurdish provinces came largely from the document’s explicit recognition that Iraqi Kurdistan already is an autonomous region, as well as its promise to uphold the TAL’s provision for a future referendum on whether the oil-rich Kirkuk area will be annexed to the Kurdish provinces. Many Shi‘a, however, also came to appreciate the decentralizing thrust of the constitution, which makes it very easy for provinces to form “regions” that enjoy significant autonomy from Baghdad. This autonomy includes the right to amend laws passed by the national government except in such areas as foreign policy, the national budget and national security. Even diplomatic representation would in fact no longer be the exclusive domain of the national government, as Article 116 of the constitution authorizes regions and provinces to “establish offices in [Iraqi] embassies…to pursue cultural and social affairs, and issues of local development.”
Many democracies with significant diversity adopt federal- ism to allow local groups significant power over many issues that directly affect them. But these systems generally tip the balance of powers much more definitively in favor of the national government than does the current Iraqi constitution. The Iraqi version of extreme decentralization could give rise to extremely different social policies from one area of the country to another. To date it is not clear, for example, if there would be any way to prevent some future regions from implementing strict shari‘a law provisions substantially limiting women’s rights in marriage, divorce and child custody, while Kurdistan and other regions will certainly have much more liberal laws. Over time, it appears that the current constitution could give rise to regions living side by side with such different laws that it would be hard to imagine them being, in any meaningful sense, part of the same country. It is also unclear whether all Iraqis would share equally in the benefits of Iraq’s natural resources, particularly oil; the fear that occupants of oil-rich regions would be able to keep most of “their” own profits terrifies the Sunni Arabs, who could be impoverished by such an arrangement. While the kind of consociationalism seen in Lebanon can hurt national unity by enshrining the idea that only members of one’s own group can protect one’s own interests, at least there politicians of different confessions work together to create laws that will apply to all Lebanese.
A third way to deal with diversity is demonstrated by Nigeria, where aspects of consociationalism and federalism are combined with requirements that political parties cannot function, and presidential candidates cannot win, unless they draw support from many different states and ethnic groups. Nigeria’s three main ethnic groups are the Hausa-Fulani with 29 percent of the population, the Yoruba with 21 percent, and the Ibo with 18 percent; over 200 other ethnic groups comprise the remaining 32 percent. In 1954, the British made Nigeria into a tripartite federation in which the Hausa-Fulani constituted the majority in the north, the Yoruba in the west and the Igbo in the east. In addition, the north has always had just over half the population of the country  and has largely dominated politics. After independence in 1960, the three regions were changed to four. These in turn were replaced by 12 states in 1967 to decrease northern dominance as the Igbo east pressed for secession. This failed to prevent Igbo secession, leading to three years of civil war.
Three important changes were subsequently introduced in an attempt to defuse ethnic and regional conflict. The number of states was again increased to form ethnic minority states that, together, had more states than any of the three major groups had, decreasing the importance of the Hausa-Fulani, Yoruba and Igbo identities by forcing politicians to appeal to the smaller minorities. A presidential system was created in which, among other requirements, successful candidates had to win at least 25 percent of the votes in at least two thirds of the states. New conditions on parties included the stipulation that they have “functional branches in, or a governing body that includes members from at least, two thirds of the states.” 
These changes have not made Nigeria into a successful democracy. After they were passed, the military resumed power until 1999. Members of ethnic minority groups in the oil-rich Niger Delta were persecuted and even executed, the number of states continued to balloon as more and more groups pressed for their own states, and corruption was further entrenched. But presidents began to be elected with support far beyond their own regional or ethnic group — with the resumption of democratic rule in 1999 both candidates for president were Yoruban southerners, and Olusegun Obasanjo won the presidency, despite little support from his fellow Yorubans, by winning votes in the north and the east. Nigeria has long had consociational elements in its governments, including a requirement that at least one member of each state receive a cabinet post, but at least since the 1980s, political parties have begun to ensure that their candidates for key government and party posts are distributed across members from many different regions. These practices, too, have hardly been foolproof; Obasanjo received little support from Yorubans in the 1999 election but has subsequently been seen as favoring their interests. 
Nonetheless, the level of inter-regional and inter-ethnic cooperation at the national political level now in Nigeria is something that the new Iraqi constitution is unlikely to produce. If Iraq’s new constitution had created a directly elected and powerful president, it is hard to imagine now an outcome analogous to Obasanjo’s victory — for example, a Kurd who lost the Kurdish vote but was carried to victory by Shi‘i and Sunni Arab voters. The Nigerian experience, however, is also notable for the extent to which the political system in place at independence changed over the next several decades as different recipes for reducing conflict were adopted or discarded. Iraq, if it remains a single nation-state, may well embark on such an experimental trajectory of its own.
 Dieter Nohlen, Florian Grotz and Christof Hartmann, eds. Elections in Asia and the Pacific: A Data Handbook, vol. 1 (Oxford: Oxford University Press, 2004), pp. 170–171.
 Larry Diamond, Squandered Victory: The American Occupation and the Bungled Effort to Bring Democracy to Iraq (New York: Times Books, 2005), pp. 42–43.
 Rotimi Suberu and Larry Diamond, “Institutional Design, Ethnic Conflict Management and Democracy in Nigeria” in Andrew Reynolds, ed. The Architecture of Democracy: Constitutional Design, Conflict Management and Democracy (Oxford: Oxford University Press, 2002), p. 406. The Nigeria section here draws heavily on this chapter and on Larry Diamond, “Nigeria: Pluralism, Statism and the Struggle for Democracy” in Diamond, Juan Linz and Seymour Martin Lipset, eds. Democracy in Developing Countries: Africa (Boulder, CO: Lynne Rienner, 1988).
 Suberu and Diamond, p. 416.
 Ibid., p. 414.