What is the role of elections in postconflict societies, especially those divided by deep religious, ethnic or racial cleavages? The peacebuilding literature underscores the crucial and instrumental functions elections play in stabilizing postwar societies (Lyons, 2002; Stedman, 2001). Elections mark the official end of the civil war, and the transition from war to peace. By relaunching constitutional processes disrupted for decades and institutionalizing new democratic norms, institutions and procedures, elections also contribute to sustainable peace building. Moreover, the process of “demilitarizing politics” in postconflict societies is intimately connected to the choice of electoral institutions (Lyons, 2002: 223). Partial electoral laws that reward excombatants unequally may generate spoiler problems or a return to full-scale ethnic conflict and the remilitarization of society (Stedman, 1997).
Studies of electoral system designs for deeply divided societies also recognize the role of elections in managing and moderating ethnic conflict (for example, Horowitz, 1985; Horowitz, 1990; Reilly, 2002; Reynolds, 2002). Elections, and more accurately the electoral systems that shape voters' choices, determine the possibility and viability of democratization and conflict management in plural societies, especially those exiting civil wars. The prospects of centripetal and moderate politics in postconflict plural societies, expressed through interethnic accommodation, multiethnic political parties, cross-ethnic electoral alliances and reciprocal vote pooling is in large measure a function of electoral engineering and the institutional arrangements selected (Reilly, 2002).
Designers of electoral systems for deeply divided societies converge on two main approaches: consociationalism and what Andrew Reynolds labels “moderation-focused incentivism” (2005: 57). The former is identified with Arend Lijphart's elite power-sharing prescription for plural societies. Lijphart is categorical that proportional representation (PR) is the optimal electoral system for divided societies, guaranteeing the election of a broadly representative legislative body. To this general prescription Lijphart adds a number of desiderata: a high degree of proportionality of seats to votes, middle-size multimember districts, and closed or almost closed list PR (2004: 100–101). This, according to Lijphart, “encourage[s] the formation and maintenance of strong and cohesive political parties” (2004: 101), a necessary though insufficient condition for “the promotion of effective and sustainable democracy” according to the authors of The International IDEA Handbook of Electoral System Design (Reynolds and Reilly, 2002: 124). But Lijphart's power-sharing recommendations ultimately give rise to the proportional representation of ethnic rather than national parties in the legislative assembly. Democracy, then, or a particular consociational variant of it, is a viable option for deeply divided societies on the condition that societal divisions are institutionalized in the political system through power-sharing mechanisms, such as a PR electoral system, and when interethnic cooperation is bracketed through the consociational principle of segmental autonomy (Lijphart, 1990).
Lijphart's recommendations have been criticized as the kind of “radical therapy” that may harden societal divisions and “bring disaster to a country that does in fact have potential for … accommodation” (Reynolds, 2005: 57). According to this view, power sharing and autonomy, the key attributes of Lijphart's consociational democracy model, may actually contribute to, rather than protect against, state collapse in a deeply divided society (Hudson, 1988; Andeweg, 2000). Consequently, instead of duplicating ethnic divisions in the legislative assembly, critics of Lijphart's model advocate designing strategies that may moderate the power of societal divisions and their hold on political mobilization. Electoral systems play a decisive role in this alternative strategy.
The centripetal function of electoral systems in divided societies is identified chiefly with Donald Horowitz. Horowitz proposes designing electoral mechanisms that promote the election of moderate representatives through interethnic coalitions, vote pooling and bargaining (2003). Where ethnic groups are not regionally concentrated, heterogeneous, single-member constituencies plus electoral incentives that promote interethnic vote pooling may give rise to multiethnic coalitions, moderate politics and interethnic accommodation. Preferential electoral systems, which allow voters to rank their preferences among different parties or candidates on the ballot paper, tend to promote interethnic accommodation because they make politicians reciprocally dependent on the votes of ethnic groups other than their own (Reilly, 2002: 157–159). The “alternative vote” (AV) and the “single transferable vote” (STV) are two such electoral systems advocated by proponents of the preferential option (Reynolds and Reilly, 2002: 37–39 and 83–84). In these systems, candidates compete for both the first-preference votes of their own community and the second-choice votes of other communities (Reilly, 2000/01: 179–182). Electoral systems, when properly designed, thus play a crucial role in engineering ethnic cooperation and accommodation in postwar plural societies.
Lebanon's postwar parliamentary elections confront us with a puzzle. These elections have been held regularly since April 1992, following the promulgation, on September 21, 1990, of the constitutional amendments adopted in the Taif Accord. They are contested by interethnic or, more accurately, interconfessional and intersectarian electoral alliances in multi-confessional electoral districts. Interethnic coalitions, vote pooling and bargaining have structured the results of these elections, as have the electoral laws demarcating the boundaries of the electoral districts. Democratization, peacebuilding and ethnic harmony have been the main victims of these interethnic alliances, however. Postwar elections, far from expanding the parameters of accountability, representation and contestation, have instead restricted citizens' electoral and hence political choices. The elections have also served to consolidate an already-institutionalized confessional and clientelistic system, thus hardening, rather than ameliorating, sectarian cleavages. This is mainly the consequence of unbalanced electoral laws constraining, and at times eliminating, the political opportunities and choices of the Christian communities. Moreover, elections were instrumental in imposing “authoritarianism by defusion” (El Khazen, 2003) in postwar Lebanon, namely by manufacturing a postwar pro-Syrian political elite whose main function has been to institutionalize Syria's hegemony through the state's central legislative, executive and intelligence institutions (El Khazen, 1993; El Khazen, 2000).
This paper unpacks this Lebanese puzzle by looking at the role of cross-ethnic alliances by otherwise antagonistic actors in three consecutive postwar parliamentary elections. The following section situates interethnic alliances in their Lebanese context. Next I undertake an analysis of the determinants of postwar parliamentary elections—namely the nature of the electoral laws and the dynamics of interethnic alliances—in the 1992, 1996 and 2000 elections. The paper closes by evaluating the viability of alternative electoral arrangements to ones hitherto used.
The Varieties of Interethnic Alliances in Postwar Lebanon
Ethnic cooperation in societies divided along deep vertical cleavages is an anomaly to the logic of majoritarian democratic theory (Lijphart, 1977). In plural societies, political organization overlaps with ethnic divisions and loyalties, inviting divisive politics at both the local and national levels, and rendering interethnic alliances difficult, if not impossible.1
The exceptions to this trend include plural societies with either a very high or a very low degree of ethnic fragmentation (Reilly, 2000/01).
Parliamentary elections in postwar Lebanon have produced an array of interethnic electoral alliances (Nasif and Boumonsef, 1996; El Khazen, 2000). These alliances have at times joined same-sect actors divided along ideological affiliations, particularly with respect to the role of religion in society, such as the electoral alliances between the Shi‘a groups Amal and Hizbullah in the South and the Beqa‘. Nor has ideology hampered the formation of electoral alliances between purportedly secular-ideological political parties and sectarian-based ones. Other alliances have brought together sectarian groups divided along ideological choices and foreign policy orientations, namely the role of the state in economic and social development, the kind of relations Lebanon should have with Syria and the West, as well as her posture vis-à-vis Israel, and the status of Hizbullah as an extra-legal military organization despite the postwar demobilization. Even the country's cultural and historic identity remains a deeply divisive subject. Although the Taif Accord declared Lebanon a country with “an Arab identity and belonging” (Mansour, 1993: 249), most Christians insist on Lebanon's cultural distinctiveness from its Arab milieu (Haddad, 2002).
Examining instances of interethnic cooperation also raises a num-ber of theoretical considerations: Does cooperation among actors with varying ideological or sectarian loyalties advance the prospects of democratization, tolerance and ethnic conflict management in deeply divided societies, as the ethnic management literature expects of interethnic vote pooling and electoral alliances (Horowitz, 1990; Reilly, 2002; Reilly, 2000/01)? Or do these alliances serve narrower short-term electoral interests, with negative long-term consequences on democratization, confessional polarization and ethnic moderation? What if interethnic alliances are deployed to protect and consolidate the clientelistic power of ethnic politicians, or to void politics of its contestatory content, thus hardening the influence of ethnic loyalties? And what role do electoral laws play in transforming cross-ethnic alliances from agents of democratization and ethnic moderation and tolerance to instruments of manipulation serving the interests of ethnic politicians? Horowitz notes that ethnic conflict is not rooted solely in cleavages and primordial animosities, but is also a result of “the institutional structure in which conflict and restraint find expression” (1990: 452). An examination of the institutional structure of Lebanon's postwar parliamentary elections, namely the electoral laws shaping ethnic leaders' electoral incentives and strategies, explains how cross-confessional alliances serve to consecrate ethnic conflict rather than ameliorate it in postwar plural societies, and how these alliances negate the institutionalized uncertainty characteristic of democratic elections (Przeworski, 1991: 14).
Institutional Determinants of Postwar Elections
The International IDEA Handbook of Electoral System Design catalogues Lebanon's electoral system under the Party Block Vote (PB) type (Reynolds and Reilly, 2002: 36–37 and 141). This system consists of multimember districts, but with electors casting one vote for predetermined party lists rather than for candidates. The party that wins a simple plurality of the votes in a specific district wins all the seats in that district. Consequently, the entire list of party candidates is elected. PB is thus an atypical electoral system because often when simple plurality is coupled with multimember districts, electors are given as many votes as there are seats in the district, as is the case with the Block Vote (BV) electoral system (Reynolds and Reilly, 2002: 36). PB, when properly used, may contribute to the management of ethnic conflict because it allows for the formation of multiethnic party lists and interethnic vote pooling.
Lebanon's postwar (and prewar) electoral systems deviate from the generic description of PB in significant ways, however. In Lebanon voters choose among individual candidates rather than closed party lists. Even if cross-confessional alliances form unified interethnic electoral lists, voters are allowed to subtract and add names from outside the list—a practice dubbed tashtib (cross-out)—as long as the sectarian proportions of the list are not disturbed. Moreover, none of the postwar electoral systems created genuinely heterogeneous territorial constituencies with incentives for moderation-serving interethnic vote pooling. Instead, the electoral districts in all postwar elections have been purposefully gerrymandered to favour one political leader or another, or to serve the electoral interests of one pro-Syrian group or another. With the use of simple plurality to determine the winners among the different sects in each district, the size of the electoral districts, namely the proportion of Muslim to Christian votes, is the determining factor in the elections, a structural constant shared by both pre- as well as postwar electoral systems (Hudson, 1966). Malapportionment has been a fixed feature of Lebanon's postwar electoral systems, rewarding Syria's allies and punishing those who opposed her hegemony over Lebanon (Atallah, 1996: 18; El Khazen, 2003: 67).
Cross-confessional and cross-ideological electoral, not political, alliances are negotiated within this institutional structure. They may bring together any permutation of same-sect, ideologically opposed or cross-sectarian, ideologically opposed or allied actors as long as the alliance guarantees a substantial bloc of votes on election day. The nature of the alliance varies with the balance of sectarian demography and political cleavages in the district. Districts that enjoy a clear demographic majority by a particular sect but are divided along political lines experience cross-ideological alliances between the main sectarian groups, with allied candidates incorporated into the list to complete the required confessional distribution of seats. This is usually the case in the electoral districts of the South and the Beqa‘, where the Amal-Hizbullah alliance secures an electoral landslide. Where the balance of sectarian demography is roughly equal, as is the case in Zahlé (central Beqa‘), Ba‘abda-‘Alay (Mount Lebanon) and some of the districts in the North, cross-confessional alliances are stretched to the limit in an attempt to mobilize the greatest number of votes for rival electoral lists. Here vote pooling is used to eliminate rival lists rather than as a measure to defeat extremist ethnic leaders. Finally, districts dominated demographically by one sect and a single political leader of the same sect end up with electoral lists formed by the latter, into which are incorporated other sects to complete the confessional distribution of the seats. The Shuf district in Mount Lebanon is such an example, as are the first and second districts in Beirut's 2000 electoral law.
Postwar electoral laws thus provide no incentive for ethnic engineering and accommodation through the negotiation of durable interethnic political alliances. The result is elections by mahadel (rollers), sweeping all or almost all seats in a given electoral district, and divesting elections of their institutionalized uncertainty and margin for competition and contestation. Moreover, the electoral alliances intricately choreographed to achieve maximum electoral gains on election day unravel as soon as the elections end, leaving no trace on the choices and strategies of parliamentary blocs, a regular feature of pre- as well as postwar parliamentary elections (Salibi, 1988: 189). The electoral system, originally designed to promote ethnic accommodation and long-term national unity, is instead deployed for tactical and temporary electoral ends.
Within these institutional structures, the emergence of a permanent multiethnic coalition to occupy the centre of the ethnic spectrum, a prerequisite for successful interethnic accommodation via vote pooling (Horowitz, 1990: 464–466), is almost impossible. Gerrymandered electoral districts privilege local sectarian political agendas over national secular ones. This consecrates the embedded neopatrimonial and clientelistic features of the confessional political system, and produces a chauvinistic sectarian discourse aimed at mobilizing ethnic followers. Consequently, electoral laws in postwar Lebanon have served as a vehicle for the hardening of ethnic loyalties, rather than as a tool for ethnic engineering and moderation (Saghiyé and Saghiyé, 2004: 38–39; Salam, 2004: 14–17). The Taif Accord established the principle of “mutual coexistence” (al-‘aysh al-mushtarak) between Lebanon's different sects as the main objective of postwar parliamentary electoral laws (Mansour, 1993: 260). These laws have fallen far short of this objective.
The instrumental use of electoral laws in postwar Lebanon is also evident in Syria's intimate involvement in the formation of electoral alliances. By grouping otherwise dissonant political actors in an electoral alliance, whether from the same sect or across sectarian divides, Syria ensured that her allies controlled a substantial percentage of parliamentary seats, and concomitantly held control over presidential elections, cabinet formation and legislation. This very visible Syrian hand was present in the 1992, 1996 and 2000 elections. Syria's intelligence chiefs in Lebanon, the late Ghazi Kan‘an (1983–2002) and his successor, Rustum Ghazali (2002–2005), as well as Vice-President ‘Abdul Halim Khaddam and former chief of staff Hikmat al-Shihabi, vetoed the inclusion of anti-Syrian candidates in certain lists, forced the leaders of some lists to take in pro-Syrian candidates, balanced religious candidates with secular ones in some districts, and balanced among their own allies in other districts (el-Huss, 2001: 51–63 and 293–297; Nasif and Boumonsef, 1996; Sabra, 2005). In the Beqa‘, a region of significant strategic importance for Damascus, Syrian officials combed the lists “name by name,” weeding out candidates whose loyalty was suspect, a practice they repeated in the North (Nasif and Boumonsef, 1996: 202). In the South and the Beqa‘, Damascus was chiefly responsible for the making of the Amal-Hizbullah electoral alliance, and participated in negotiations over how to allocate the districts' seats between the two parties. In Beirut and Mount Lebanon, Damascus imposed her candidates on the winning lists to ensure their entry to parliament. The postwar electoral laws were also the work of Syria's intelligence chiefs and their Lebanese lieutenants. Former prime minister Salim el-Huss minced no words in his account of the birth of the 2000 electoral law. El-Huss had attempted to market a new proposal based on the Mixed Member Proportional (MMP) system, whereby 28 of parliament's 128 seats would be elected through PR, with the whole country turned into a single, heterogeneous constituency, and the balance of parliamentary seats elected on the basis of plurality-majority small districts. To his surprise, however, a ready-made law was presented to cabinet for endorsement. “It was common knowledge,” el-Huss wrote, “that Major General Jamil el-Sayyid, Director General of the Suretè Gènèrale, played a role in producing this draft law in coordination with the Syrian authorities at ‘Anjar” (2001: 61), then the headquarters of Syrian intelligence forces in Lebanon. Syria's allies in cabinet endorsed the law automatically, despite el-Huss' protestations.
The Politics of Electoral Laws and Alliances: 1992, 1996 and 2000
In the 1992 parliamentary elections, the country was divided into three large-size electoral districts in the North, the South and Beirut, and nine middle-size districts in Mount Lebanon (six) and in the Beqa‘ (three). Table 1 maps the distribution of sectarian seats in the different electoral districts. These choices were not without purpose, however.
Sectarian Distribution of Parliamentary Seats in 1992 Elections
In the North, the South and Beirut, large electoral districts served the electoral interests of pro-Syrian politicians, while the same objective was served in Mount Lebanon by middle-size districts. In Mount Lebanon, for example, where the Druze population is concentrated in compact communities, middle-size districts worked to the advantage of the Druze leader Walid Jumblatt and pro-Syrian Christian politicians. This is especially true given the low turnout rates among Christian voters alienated from the postwar political system. In the South, the large electoral district served the electoral alliance between Amal and Hizbullah, as did the middle-size district in Ba‘albak-Hermel. This same-sect cross-ideological alliance camouflaged the not-so-subtle tug-of-war between the supporters of the two factions, and a competition over the political representation of the Shi‘a community inside as well as outside public institutions. Similarly, the Hizbullah-Rafiq el-Hariri electoral alliance in the South glossed over what at the time were deep differences between the two groups over the role of the state in society and the appropriate strategy vis-à-vis Israel. Whereas in Mount Lebanon the 1992 law consecrated Jumblatt the undisputed Druze leader, for the Shi‘a community leadership was to be shared between Amal and Hizbullah. In the North, the large electoral district forced an alliance between two erstwhile rivals, Slayman Franjieh and ‘Omar Karami, joined, however, by Syria in a powerful cross-confessional alliance (for the 1992 elections see El Khazen, 1998).
Nor were the 29 parliamentary seats added to the prewar parliament distributed in a manner mirroring the postwar geographic-demographic-sectarian map of the country. The Taif Accord raised the number of parliamentary seats from the prewar total of 99 (54 Christian, 45 Muslim) to 108, divided equally among Muslim and Christian deputies. According to Law 51 of 1991, the nine new Muslim seats were to be allocated to areas with clear Muslim demographic concentrations, thus correcting the sectarian representation of these areas in parliament. Table 2 details the distribution of the nine seats along sect and electoral district.
Allocations of Nine New Seats by Sect and District
Rather than implementing Taif's stipulations, Law 154 of 1992 raised the number of parliamentary seats to 128 instead of 108, thus adding 29 new seats to the prewar parliament. This, in turn, entailed raising the number of Christian deputies by 10 to maintain parity between them and their Muslim counterparts in the postwar parliament. Table 3 details the distribution of the 29 seats along sect and district. However, the distribution of the seats created for Christian deputies was governed by political, rather than objective, reasons. New seats for Christian, but especially Maronite, deputies were created mainly in areas where the balance of votes favoured Muslim rather than Christian voters. For example, the addition of a Maronite seat in Tripoli and another in the Beqa‘ does not correspond to demographic changes in these regions. After all, the number of Maronite voters in each of these districts falls well short of the national average of Maronite voters per deputy. Instead, a Maronite seat should have been added in Beirut, where Maronite voters tend to be underrepresented in some districts. Similarly, the addition of a Druze seat in Beirut served purportedly to represent all sects in Beirut, but does not correspond to a substantial increase in Druze voters in the capital. In other cases, new seats were created to facilitate the entry of pro-Syrian politicians, both Muslim and Christian, to the postwar parliament.
Allocations of Twenty-Nine New Seats by Sect and District
The geographic distribution of the new seats also favoured areas traditionally identified with solid Syrian influence and support (Hanf, 1993: 625). As Table 4 demonstrates, 55 per cent of a total of 29 new seats were allocated to the North and the Beqa‘, where Syria has enjoyed a strong military and intelligence presence since 1976 and until the withdrawal of her troops from Lebanon on April 26, 2005. Mount Lebanon, with its historic sensitivity to Syrian influence in Lebanon, was allocated only 17 per cent of the total seats. Thus both the increase of parliamentary seats to 128 rather than the 108 stipulated in the Taif Accord, and the division of electoral districts in the 1992 electoral law, rewarded pro-Syrian politicians, both Muslim and Christian, and punished the mainly Christian opposition politicians. Indeed, some interpreted the 1992 law as nothing less than a deliberate attempt by Syria and her Lebanese proxies to restrict the political influence of the Maronite community in postwar Lebanon (El Khazen, 2000: 72–82). The elections were also timed to manufacture a new, pro-Syrian political elite, one that would not request a redeployment of Syrian troops in Lebanon as stipulated by Taif (Naser, 2004: 50–51). Little wonder most Christian politicians, and voters, opted to boycott these elections.
Regional Distribution of New Seats
In the 1996 elections, one single amendment was introduced to the 1992 electoral law: the three districts in the Beqa‘ were amalgamated into a single, large electoral district. This neutralized the Christian vote and weakened the Sunni one (Atallah, 1996: 16–18). Middle-size electoral districts in Mount Lebanon remained unchanged, to accommodate the electoral interests of Walid Jumblatt in the Shuf and Ba‘abda-‘Alay, as well as those of pro-Syrian Christian candidates in the Metn. With a clear demographic majority in the Beqa‘, the amalgamation of the three electoral districts allowed the Amal-Hizbullah alliance to dominate the seats of this district. Even though Amal and Hizbullah were allies in the South and the Beqa‘, they joined different camps in Ba‘abda-‘Alay and in Beirut. Electoral politics in these districts entailed alternative electoral alliances. In Ba‘abda-‘Alay, a cross-confessional and cross-ideological alliance joining heavyweight Shi‘a (Amal), Sunni (Rafiq el-Hariri's Future Movement), Druze (Walid Jumblatt's Progressive Socialist Party) and Maronite (Elie Hobeika) political groups was arrayed against Hizbullah's list. The Hariri-Amal cross-confessional alliance ensured the defeat of Hizbullah's candidates in Beirut. In the North, a number of cross-confessional electoral lists collided in a bitter contest for every vote. However, Syria's list, gathering an alliance of Sunni, Maronite and Greek Orthodox politicians, albeit penetrated by opponents from rival lists, was able to secure 17 of the 28 seats reserved for the North (for the 1996 elections see Nasif and Boumonsef, 1996; El Khazen, 2000: 167–212). In 1996, as in 1992, cross-confessional alliances determined to a large extent the results of the elections beforehand. They also succeeded in depoliticizing electoral contests, transforming them from battles over policy to battles over seats.
The 2000 parliamentary elections were atypical by postwar standards. They were held in the shadow of an open contest between two main camps in the country: on the one hand, President Emile Lahoud and his security team, led by el-Sayyid and supported fully by President Bashar el-Asad of Syria and his intelligence apparatus in Lebanon; on the other hand, Hariri and Druze leader Jumblatt (Sabra, 2005). Lahoud's election to the presidency on October 15, 1998, and Hariri's subsequent refusal, on November 30, to form a new cabinet, citing the unconstitutional manner by which Lahoud conducted his deliberations with the parliamentary blocs leading up to Hariri's nomination as prime minister-designate, set the stage for a hard battle in the upcoming parliamentary elections (Salman, 2006). Invariably, the electoral law, and its concomitant electoral alliances, would determine the size of each side's parliamentary bloc and, consequently, the balance of power between them, inside as well as outside state institutions.
The 2000 electoral law re-organized substantially the electoral districts of the 1996 elections in an open maneuver to reward the state's allies and contain her foes. Beirut was divided into three electoral districts, a ruse targeting Hariri in an attempt to reduce the size of his parliamentary bloc.2
Accounts vary as to why Beirut was divided into three electoral districts. Marwan Hemadé insists that the 2000 law targeted Hariri. See his interview on Future TV's Khalik Bel-Bayt, May 3, 2005. Others contend that the division of Beirut into three districts was the idea of Ghazi Kan‘an, then chief of Syrian intelligence in Lebanon, and his Lebanese lieutenant, Jamil el-Sayyid. It was Kan‘an's way of “offering” Hariri Beirut in exchange for the latter's acceptance to include on his lists a number of pro-Syrian candidates (Sabra, 2005).
Sectarian Distribution of Parliamentary Seats in 2000 Elections
The animosity between Lahoud and Hariri, and the political stakes involved on both sides, produced a hard electoral battle in the 2000 elections. One report estimates campaign costs at US$200 million, spent mainly to buy votes, entice voters to cast their ballots for certain candidates, or to secure seats on winning lists.3
See the report prepared by Information International on the 2000 elections, “Al-Intikhabat al-Niyabiyya fi ‘Am 2000 fi Dirasa lil-Duwaliyya lil-Ma‘ loumat,” published in an-Nahar. http://www.annaharonline.com/htd/TAHKIK050512-1.HTM (May 14, 2005).
A number of cross-confessional and cross-ideological alliances shaped the results of the 2000 parliamentary elections (see ‘Assaf and Haddad, 2000; El Khazen, 2000: 213–237). In the South, the cross-sectarian and cross-ideological alliance between Amal, Hizbullah and Hariri, urged and blessed by Syria and Iran, ensured the three parties a sweep of all 23 seats assigned for this electoral district. Hizbullah, though forced by Syria to accept an alliance with Amal rather than contest the elections alone after riding a powerful wave of popular support on the morrow of Israel's withdrawal from south Lebanon on May 24, 2000, nevertheless increased its share of the votes cast as compared to the 1996 elections as a result of the alliance with Bahiya el-Hariri in Sidon. Yet despite the fact that the electoral law had divided the South into two districts, elections in the South were conducted on the basis of a single, large district, under the pretext of the Israeli occupation, even though Israel withdrew her troops from Lebanon three months ahead of the scheduled elections. The single, large electoral district in the South played to the advantage of the numerically superior alliance between Amal, Hizbullah and Hariri, ultimately marginalizing the Christian votes of Jezzine. In fact, Hizbullah, against the wishes of Maronite voters, imposed its own Maronite candidate for one of the Maronite seats in the district; the chosen candidate, George Najm, hails from ‘Ayn el-Mir, a village with only 200 Maronite voters, whereas the favored candidate vetoed by Hizbullah, Slayman Kan‘an, hails from Jezzine, with its 20,000 Maronite voters.4
See the report prepared by Information International on the 2000 elections, “Al-Intikhabat al-Niyabiyya fi ‘Am 2000 fi Dirasa lil-Duwaliyya lil-Ma‘ loumat,” published in an-Nahar. http://www.annaharonline.com/htd/TAHKIK050512-1.HTM (May 14, 2005).
In the Ba‘albak-Hermel electoral district, a cross-confessional alliance gathering political parties corresponding to different, albeit all pro-Syrian, sectarian groups, guaranteed an uncontested electoral victory. The alliance of Hizbullah, Amal, the Syrian Social National Party, the Ba‘th Party and the pro-Syrian wing of the Phalange Party gathered a mix of religious and secular political parties. Moreover, under Syrian urging, and in exchange for an increase in its share of seats in the South, Hizbullah included on its list Nader Sukkar, a former advisor to anti-Syrian Maronite leader Bashir Gemayel and a former member of the Lebanese Forces now turned pro-Syrian. Syria intervened forcefully among local leaders to vote for Sukkar.
In the Ba‘abda-‘Alay district, Jumblatt fought a hard battle to consecrate his local leadership, despite attempts by Lahoud and el-Sayyid to decrease his parliamentary bloc. Given the sectarian demography of the district, the Maronite vote is a substantial and important one. Jumblatt's main challenge was thus to mobilize Maronite voters behind his list. Toward this end, he weaved around himself an intricate—if not impossible—web of cross-confessional alliances. Prior to the elections, he opened up to almost all of the Christian factions in Mount Lebanon, including the supporters of exiled former commander of the Lebanese Army and former interim prime minister General Michel ‘Awn and his Free Patriotic Movement, the Lebanese Forces, Amin Jumayyel's wing of the Phalange Party, and the supporters of Raymond Eddé's National Bloc. In his pre-election speeches, Jumblatt presented a number of themes aimed at mobilizing the Christian vote behind his electoral list. These included a call for the pardoning of jailed Lebanese Forces leader Samir Ja‘ja‘, and the repatriation of the exiled ‘Awn. Moreover, and in a calculated move to bridge the gap between his then pro-Syrian orientation and the anti-Syrian mood of Christian voters in Mount Lebanon, Jumblatt went so far as to indirectly criticize the Syrian presence in Lebanon, calling for a rectification of relations between the two states. Furthermore, in a deeply symbolic gesture designed to mobilize Christian voters behind his electoral list, Jumblatt met in his Mukhtara residence with his one-time nemesis, former president Amine Jumayyel, whom he had once labelled “Somoza Ba‘abda” (Ayoub, 2000). This cross-confessional and cross-ideological alliance, supported by Hariri, succeeded in sweeping all 11 seats in the Ba‘abda-‘Alay district. It instrumentally served Jumblatt's short-term electoral interests despite his previous political alliance with Syria.
Another indication of the instrumental nature of electoral alliances during parliamentary elections may be gleaned from the competition between Amal and Hizbullah in the Ba‘abda-‘Alay district. Though Amal and Hizbullah joined forces both in the South and the Beqa‘-Hermel electoral districts, they did not replicate their alliance in Ba‘abda-‘Alay. In fact, the Amal-Hizbullah competition for one of the Shi‘a seats in this district—waged between Amal's Salah Haraké and Hizbullah's Ali ‘Ammar—was one of the fiercest electoral battles witnessed in this district, proving that electoral alliances serve only political expedience, and are subject to change as political calculations shift (Ayoub, 2000).
In Beirut, the battle was over more than just the capital's 19 seats. Hariri's political stature and future was at stake (Farshakh, 2006: 364–366). He thus marshalled all his political, financial, audiovisual and international resources for the battle. Salim el-Huss complained loudly about the excessive campaign financing by Hariri in Beirut (2001: 6). Given the determining role of the Sunni vote in Beirut's three electoral districts, Hariri also invoked sectarian loyalties to mobilize voters behind his lists. Most interesting, however, was the cross-sectarian and cross-ideological alliance between Hariri and Hizbullah in Beirut's second electoral district. At Syria's urging, and at the eleventh hour of the elections, Hariri withdrew the Shi‘a candidate on his list—Ghazi Yousef—to allow for the victory of the Hizbullah candidate in this district, Muhammad Berjawi. Given his neoliberal socioeconomic policies, which have been detrimental to Hizbullah's urban and rural constituency,5
Including the Elisar Project, which requires the eviction of a strong pro-Hizbullah constituency in the Uza‘i area of Beirut.
Prospects for Interethnic Accommodation
Elections in postwar Lebanon have so far failed to moderate ethnic conflict and generate the institutionalized uncertainty necessary for viable democratization. Rather, interethnic electoral alliances are deployed instrumentally by ethnic politicians to guarantee electoral victory in the context of electoral laws conducive to temporary sectarian coalitions, in contrast to the permanent multiethnic centrist coalitions Horowitz proposes for the successful management and moderation of ethnic conflict in plural societies. Interethnic alliances in postwar Lebanon have instead institutionalized the clientelistic confessional political system, serving the interests of ethnic rather than national politicians, and concomitantly hardening sectarian animosity and robbing the electoral process of its prewar contestatory dynamics. It has also imposed a postwar, pro-Syrian, political elite on society and state institutions, and shackled political life and choices by authoritarian Syrian fiat. All this unravelled under Syrian political and military control. With Syria's withdrawal from Lebanon, what future awaits electoral politics? Will elections continue to serve the same narrow, local political interests of the confessional political elite? Are Lebanese doomed to live in a state where institutionalized sectarianism and clientelism hamper all efforts toward genuine political reform? Or can a new electoral law open up possibilities for durable and moderate interethnic coalitions at the national level? The most recent May-June 2005 parliamentary elections, albeit held after the Syrian withdrawal, generated the same instrumental interethnic coalitions witnessed in previous elections (Sa‘d, 2005). In part, this is because they were contested under the 2000 electoral law. What then are the future prospects for electoral engineering in Lebanon?
The Taif Accord established two objectives for postwar parliamentary elections: mutual, peaceful coexistence between the different confessional groups (al-‘aysh al-mushtarak), and their proper political representation (sihat al-tamthil al-siyasi). This, Taif notes, must be based on a reorganization of prewar electoral districts, thus establishing heterogeneous electoral districts conducive to interethnic coalitions and accommodation, and ultimately national unity (Mansour, 1993: 260).6
What this means exactly is subject to some debate. Those who participated in the Taif deliberations insist that postwar electoral districts should be based on the qada', which in postwar Lebanon refers to reorganized middle-size districts.
Despite the absence of a recent national census, most observers agree that the Muslim-Christian demographic proportion is around 70:30 percent respectively (Kassir: 2005).
According to 2005 statistics, total Muslim voters number 1,772,187 (58.82%), while Christian voters number 1,230,550 (40.89%) (al-Sha‘er, 2005). The concentration of sects in some areas, and their dispersion in others, complicates the formation of equitable heterogeneous constituencies throughout the country. The confessional balance of power is ultimately determined by the size of the electoral district. Large-size districts result in a majority Muslim electorate in the South and the Beqa‘, though less so in Beirut and the North, and a Christian majority in Mount Lebanon. Middle-size districts allow for greater confessional diversity in some districts, but invariably produce a large number of confessionally pure districts—for example, in the Metn, Kiserwan, Bshari, Zgharta, Dennieh, Batroun, Al-Koura, Tyre, Nabatiyé, Jezzine and Bint Jbayl.
One solution is to turn the whole country into a single, heterogeneous constituency and use a PR system with closed, unalterable national lists. The latter caveat ensures that voters vote for the whole list rather than for selected names on the list. Proponents of this option argue that it leads to durable interethnic coalitions rather than instrumental ones, ultimately paving the way for national unity (Salam, 2004). The success of PR in plural societies “presumes some kind of recognized party structure,” however, “since voters are expected to vote for parties rather than individuals or groups of individuals” (Reynolds and Reilly, 2002: 66). Others have called for an MMP system, with the use of simple plurality in middle-size districts to elect 100 parliamentary seats, with the balance of seats elected via PR, on the basis of the country being considered a single heterogeneous constituency. The advantage of this system over the PR one is that, in a small country like Lebanon with its sectarian topography, it ensures equitable confessional and regional representation while at the same time guaranteeing interethnic national groups a share of parliamentary seats. It thus promotes national, rather than local, loyalties, and paves the way for the emergence of interethnic, moderate political parties (el-Huss, 2001: 60; Salem, 1996; Slayman, 1996).
Common to these proposals is the desire to establish an electoral institutional structure conducive to the emergence, in the long run, of permanent and centrist, rather than instrumental and localized, cross-confessional coalitions and parties that foster national, rather than sectarian, loyalties and durable interethnic accommodation. Such proposals, whether PR or MMP, aim either at replacing the simple plurality (FPTP) system of vote counting, or complementing it with proportionality. After all, simple plurality has been a constant of both prewar and postwar electoral systems. By allowing the use of cross-confessional alliances for instrumental electoral objectives, it hardens clientelism, the confessional political system, and consequently ethnic chauvinism; in turn, this works to the advantage of established local ethnic rather than national politicians (Saghiyé and Saghiyé, 2004: 38–39). Other systems recommended for ethnic management in deeply divided societies, namely AV and STV, have their limitations when transposed to the Lebanese context. AV is not congenial to Lebanon's multimember constituencies. The success of STV, on the other hand, presupposes the existence of strong political parties, and its ability to create vote pooling across ethnic lines lacks convincing empirical evidence.8
Although Saghiyé and Saghiyé (2004: 57) expect STV to generate ethnic moderation in Lebanon, in Ireland and Northern Ireland STV has not led to interethnic vote pooling (Reynolds and Reilly, 2002: 86).
Yet others insist that middle-size electoral districts, ranging from three to six seats each, best reflect Lebanon's sectarian sociology and topography, and the primacy of geographic representation in a small, plural society lacking a strong party system (El Khazen, 2000: 233–234). In Lebanon, simple plurality in middle-size districts encourages intraethnic competition in ethnically pure districts, and interethnic competition in ethnically mixed districts. This also hardens the institutionalized clientelism of the confessional system and hence sectarian identities. One study suggests that middle-size districts in a parliament of 108 deputies, as stipulated by the Taif Accord, reduces the number of Christian deputies elected by Muslim votes to eight, with Christian votes determining the electoral outcome of only three Muslim deputies (al-Sha‘er, 2005). Though this may satisfy the demands of the Maronite patriarch, who insists on the equitable political representation of the Christians, à la Taif, it may actually undermine prospects for interethnic coexistence given the emphasis on voting along sectarian lines (Kassir, 2005).
A future electoral system striving to furnish institutional structures that promote centrist interethnic coalitions fostering interethnic accommodation must resolve the structural dilemma Taif bequeathed postwar Lebanon, namely the contradiction between the demands of equitable sectarian political representation and the hard realities of demography and geography. Otherwise, electoral laws and cross-ethnic alliances structuring election results will continue to reproduce the same clientelistic and sectarian dynamics characteristic of Lebanese politics since independence.
Acknowledgments
For helpful comments and input, I would like to thank Hala Kawsarani, Marie-Joëlle Zahar, the participants of the workshop on “Crossing Ideological Divides” at the Rockefeller Foundation's Bellagio Study and Conference Center, August 8–13, 2005, and this Journal's anonymous reviewers. I also thank Mohamad Kassem for his research assistance.
