publisher colophon

On 18 September 2005, nearly six and a half million Afghans voted in the freest and most competitive legislative elections they had ever experienced. The election vividly demonstrated that vibrant and meaningful elections can be conducted even under the most unpromising circumstances. But the Afghan elections also showed that designing appropriate institutions tailored to the needs of an emerging democratic society remains the greatest challenge. The single nontransferable vote (SNTV) electoral system came about by a path of muddled missteps, and it was a disservice to the millions of Afghans who deserved a clear and transparent tool to craft their first democratic parliament.

On 18 September 2005, nearly six and a half million Afghans voted in the freest and most competitive legislative elections they had ever experienced. Almost half of them were women, which had a significant impact on the power dynamics of the new legislature. Elections were held for the 249-seat Wolesi Jirga (Council of the People) and the indirectly elected 102-seat Meshrano Jirga (Council of the Elders), as well as for local councils. With the convening of the new parliament on 18 December 2005, the international community proclaimed that a substantial part of the 2001 Bonn Agreement had been fulfilled, namely "the establishment of a broad-based, gender sensitive, multiethnic and fully representative government through national elections."1

This election, as well as Afghanistan's October 2004 presidential polling, vividly demonstrated that vibrant and meaningful elections can be conducted even under the most unpromising circumstances. Moreover, the 2004 and 2005 ballotings showed that, when given the opportunity, women can make significant political progress even in highly patriarchal, conservative societies. Regardless of how the country began its transformation from dictatorship to democracy, its two elections have proven that the yearning to choose leaders freely and hold them accountable is the human condition. Indeed, the "ordinary citizen"—the man on the Kabul bus, the woman in the Herat library, the nomad shepherd in the mountains—will vote with confidence when given the chance.

But the Afghan legislative elections also illuminate some far less celebratory truisms about the transition to democracy. First, designing appropriate institutions tailored to the needs of an emerging democratic [End Page 104] society remains the greatest challenge. Second, the electoral system is the most important of these democratic institutions; an inappropriate or flawed electoral system can retard democracy's progress as much as warlords, religious fundamentalists, and corrupt business leaders taken together. Third, if the choice of an electoral system is based on mistaken theory or a poor grasp of what election and party-system mechanics will mean "on the ground," the results are likely to be far from desirable. Indeed, if an electoral system is less chosen than haphazardly assembled by means of drift and accident, an ineffective legislature will be the upshot. Last, Afghanistan exemplifies the difficulty of promoting multiparty politics in an inchoate democracy, where leadership is heavily swayed by the need to incorporate and assuage corrupt and deeply undemocratic sectional interests.

The electoral system ultimately used in the September 2005 Wolesi Jirga elections featured the rarely used single nontransferable vote (SNTV), a system employed in Japan from 1948 to 1993, but today used only in Jordan, Vanuatu, the Pitcairn Islands, and partially in Taiwan (although there change is underway). Under the SNTV system, Afghan voters cast ballots for individual candidates rather than political parties. Each province elects a number of members, some of whom have to be women, but each voter can vote for one and only one candidate. Candidates are elected by simply winning the most votes: If a district has been allotted four seats, then the top four vote-getters are elected. Thus, if collectively a party wins a majority of the votes, it does not necessarily win a majority of the seats—the number of seats won depends on whether individual candidates whom the party has fielded have performed adequately. "Adequate" candidates need not be especially popular, however: In a notional four-seat district, for instance, one candidate could be elected with 90 percent of the votes while three others could be elected with 3 percent each.

The advantages of the SNTV system are that it is simple, both for voters and for those who administer the election; that it promotes representation of independents in a nascent party system; and that it boosts representation and accountability by giving the individual voter a sense of being "invested" in a known individual candidate rather than some faceless party-determined list. But the system is believed to lose its efficacy when the districts are too large in size, because the vote then becomes too fragmented. For that reason, the average district in Japan, Jordan, and Vanuatu had four seats. In Afghanistan, however, the districts ended up being based on provincial boundaries and ranged from 2 to 33 seats—a third of the districts had more than nine seats.

How Afghanistan Ended Up With SNTV

The process for designing Afghanistan's new constitution was laid out by the December 2001 Bonn Agreement. It was an efficient but [End Page 105] closed process, whose product was ultimately rooted in the transitional government's own interests and presented to the public largely as a fait accompli. A nine-member committee appointed by interim president Hamid Karzai came up with an initial draft between October 2002 and March 2003. One of the most important constitutional issues with which these drafters had to grapple was the choice of an electoral system for the new legislature: Should they return to the first-past-the-post (FPTP) system inherited from British colonial rule and used in the brief period of competitive multiparty politics during the 1960s? Or should they adhere to the trend of list-based proportional representation (list PR), which has been the system of choice in the vast majority of postconflict situations since 1989? Or was there another system that would better fit Afghan circumstances?

The drafting committee received expert advice from, among others, the International Foundation for Election Systems, Princeton University's Liechtenstein Institute for Self Determination (LISD), and from the constitutional experts who authored "Afghanistan's Political and Constitutional Development: Summary and Key Recommendations."2 All these advisors sought to steer the drafters away from the old FPTP system and toward a form of proportional representation that had some geographic basis and allowed space for independent candidates. According to the abovementioned report:

The electoral system [needs] to allow for the representation of Afghanistan's diversity, and give all contenders for power enough of a stake in the system that they remain bound to democratic politics. Given the factionalized nature of Afghan politics, the primary goal should be to produce reasonable proportionality. In addition, most Afghans want an opportunity to vote for candidates from their own area, and the distrust of political parties means that voters should be given the chance to vote for individuals, rather than only parties. Voting procedure will also need to be simple and transparent; illiteracy and innumeracy limit the complexity of possible voting systems, and inexperience with voting also means that results must be easily explicable.3

Between April and November 2003, a 35-member all-Afghan constitutional commission selected by Karzai refined the earlier committee's draft, and produced a final document that was presented for ratification to the constitutional Loya Jirga in December 2003.4 The new constitution, which provoked significant unhappiness and was ratified only with the help of strong political pressure,5 did not make explicit the electoral system to be used for the legislative elections. A decision had been reached, however, that some form of list PR was to be used: This was spelled out in an appendix to the constitution. The Transitional Government assumed the task of working out the details of the system in cooperation with the Joint Election Management Body (JEMB) and the UN Assistance Mission in Afghanistan (UNAMA). [End Page 106]

By early 2004, they had designed what they thought was the best alternative: a closed-list PR system using multimember districts based on Afghanistan's 34 historic provinces. Less complicated than open-list PR, such a system would allow party leaders to determine which candidates would appear on the ticket and in what order, meaning that voters would cast their ballots for a party, not a specific candidate. Enayat Qasimi, a young Afghan-émigré lawyer who had recently returned to act as legal advisor to President Karzai, was selected to present this system to the cabinet of the transitional government. By the accounts of some of those present—including cabinet ministers as well as UNAMA and JEMB representatives—Qasimi made an utter hash of presenting the system, demonstrating that he himself was confused about its workings.6 This gave an opening to critics: If the president's own legal advisor could not make the system intelligible to the cabinet, the ministers argued, then how were ordinary Afghans supposed to understand the system?

In such a fragile and distrustful environment—the cabinet being a loose patchwork of feuding technocrats, returnees, jihadi leaders, and unreconstructed warlords—Qasimi's inept presentation of the system opened the door for some ministers to complain that it was a bad system for Afghanistan. This sentiment was fueled by the distrust of political parties common among Afghans due to the chaotic nature of multiparty politics in the 1960s and the subsequent Communist Party rule and Soviet occupation (1978–89). Leading the charge against PR was Minister for Rural Development Mohammad Haneef Atmar, with the backing of several other ministers from the Pushtun southeast.

In the face of this onslaught of objections, President Karzai decided to ask about alternatives to closed-list PR. He asked the international members of the JEMB to draw up a memorandum, detailing which electoral systems would allow Afghans to vote for individual candidates rather than parties, while retaining the province as the base constituency (he understood that for logistical and political reasons single-member districts were not an option). The SNTV system ultimately chosen was the "least bad" of the alternatives that fit these criteria. Thus Afghanistan ended up with SNTV not as a result of extensive deliberation and careful evaluation of its pros and cons, but rather by a fairly random process of elimination. SNTV was simply better than the other systems combining a single vote for a candidate with provincial multimember constituencies. It is important to note that Karzai did not choose SNTV with any understanding of its consequences or history.

How Was It Supposed to Work?

The received wisdom on SNTV—based on forty years of use in Japan and a decade in Jordan—is that the system can be manageable under specific circumstances, but that it is not generally desirable as a means [End Page 107] of translating votes into seats in a democracy. In the case of Japan, the system had been part of the institutions "bestowed" upon the country after the Second World War, and in Jordan it was the product of the late King Hussein's 1993 manipulation of the former bloc-voting system.7 The chief flaw of SNTV is the difficulty of strategic coordination that it creates for parties: How many candidates should a party field within a given constituency, and how can a party induce voters to share their votes equally across these candidates? The system's most undesirable consequences include a high disproportionality between votes and seats won, a tendency to exclude minority parties, the encouraging of clientelism and corruption among those elected, and ultimately the fragmentation of the ruling party.

In September 2004, I joined Andrew Wilder, the director of the Kabul-based Afghan Research and Evaluation Unit (AREU), in speculating as to how SNTV might work were it to be used under Afghan conditions.8 We foresaw five negative consequences involving 1) the translation of votes into seats; 2) the party system; 3) the vote itself; 4) the ability of the elected bodies to govern; and 5) female representation.

1) The relationship between votes and seats won. Founding elections in fledgling postconflict democracies need to do a particularly good job of fairly translating votes cast into seats won for majorities, minorities, and independents. Election results are particularly susceptible to challenge if the losing candidates feel that the electoral system has discriminated against them or their core constituencies. Candidates should also be afforded the reasonable assumption that if they do relatively well in the vote they will get elected. We worried that in Afghanistan SNTV would be unable to live up to these expectations. Because it creates a capricious relationship between votes and seats, we reasoned that seat shares would depend on how many candidates stood in a province and how voters distributed their votes across those candidates. We predicted that in the larger districts STNV would create a lottery effect: It would be entirely random as to who among independents and minority candidates got elected.

2) The establishment of a stable party system. Experience shows that in postconflict environments democracy and stability require the promotion and encouragement of a stable party system. Even in systems that discourage or do not recognize parties, likeminded interests gravitate together to form party-like movements, blocs, lists, and alliances. [End Page 108] Manipulating electoral systems in an effort to eliminate parties merely makes such blocs unaccountable, less democratic, and less able to respond to voters' interests.

The SNTV electoral system weakens the role of political parties, thereby working against the rise of a stable, dynamic, and accountable party system. While Afghanistan's electoral law did not bar political parties from fielding candidates, party affiliations were left off the ballot. We predicted that this would result in a fragmented legislature comprising a multitude of independents and small political factions—making government formation and legislative politics hard to manage—and that any parliamentary factions or alliances would likely be disjointed and personality-driven, beholden to regional bases or strongmen rather than national interests.

3) Ensuring a clear and effective vote. We argued that it was crucially important for the voting itself to be easy, so that Afghans could use their ballots to make their political preferences heard "loud and clear." Should the ballot itself prove too complex, voters would be alienated and unable to see the link between their votes and the newly formed government. We speculated that SNTV, particularly in the larger districts, would encourage a great number of candidates, making ballots long and confusing and causing illiterate voters to suffer most. There could be hundreds of names and symbols on the ballot (one for each candidate regardless of party affiliation), making it difficult for candidates to publicize their unique symbol. Furthermore, the vote-seat anomalies and vagaries of the SNTV system would result in a fragmented parliament and a lack of transparency in government formation, leaving many voters wondering what their ballot had had to do with the whole murky process.

4) Allowing the executive and legislature to govern. In emerging democracies where power is balanced between the legislature and a directly elected executive, it is important that the former is designed to work in harness with the president and not block his or her will at every turn. Considering the pressing need for effective legislation and policy making in democratizing countries, gridlock in government is particularly dangerous. We reasoned that if legislative elections had been held concurrently with the October 2004 presidential election, the SNTV system might have been more likely to produce a solid pro-Karzai bloc in the Wolesi Jirga. Instead, the year-long gap between the two elections was bound to weaken his legislative base. In a memorandum I wrote for President Karzai in January 2005, I argued:

The system [SNTV] will advantage those parties/movements most able to mobilize and manipulate votes. While the President has broad multi-ethnic support he does not have the level of party "machines" that commanders and provincial power brokers have in the North, East and West. Thus if anti-Karzai forces are attuned to the winning strategy of SNTV they are likely to win many more seats than their vote share would suggest.9 [End Page 109]

5) Promoting dynamic women in parliament. Affirmative-action mechanisms, or electoral quotas, generally boost the advancement of women's interests when a) voters do not consider them overtly manipulative; b) they facilitate the election of women who are less dependent on traditional power structures; c) the women elected have an electoral base and enjoy some degree of legitimacy; and d) male-dominated parties find it in their interests to field progressive women candidates who will appeal to both male and female voters.

The Afghan electoral law set aside an average of two seats per province—a national total of 68 seats—to be filled by women candidates. We speculated that, in combination with the SNTV system, these quotas increased the likelihood of women winning such reserved seats with dramatically fewer votes than losing male candidates throughout the country. In an environment not known to be particularly open to women's involvement in political leadership, or even to women holding visible social or professional positions, we feared that this might breed concentrated resentment against women legislators.

Many international organizations, foreign diplomats, UNAMA, and emerging progressive Afghan political and civil society movements in Kabul shared some or all of our concerns regarding the malign consequences of SNTV. Though Karzai had signed the SNTV electoral law on 25 May 2004, intensive efforts to persuade him and his closest confidants to revert to list PR continued until early 2005. UN special representative Jean Arnault and EU representative Francisco Vendrell led the diplomatic lobbying effort, which notably did not enjoy the support of U.S. ambassador Zalmay Khalilizad—who was close to Karzai but was said to doubt the salience of the electoral-system issue. The AREU and the International Crisis Group both prepared reports on the potential problems of SNTV in an emerging democracy, while senior international members of the JEMB wrote confidential memoranda outlining how disastrous SNTV could prove to be for the nascent Afghan democracy.10

On 17 January 2005, 35 of Afghanistan's 40 registered political parties released a joint statement saying that they supported a list-PR system and wanted the electoral law revised along the lines of a draft law produced by UNAMA. To get over the general distaste for political parties, the UNAMA draft law talked of seats being awarded to "lists" rather than parties. In January 2005, the three main challengers to Karzai in the 2004 presidential election—opposition leaders Abdul Rashid Dostum (an Uzbek warlord), Yunous Qanooni (a Tajik from the Northern Alliance), and Haji Mohammad Mohaqiq (a Hazara)—stated that they supported list PR and not SNTV.

The momentum for change seemed overwhelming, and members of the JEMB believed that Karzai had decided to revert to a province-based, closed-list–PR system as had initially been envisaged.11 By early February 2005, however, a new objection to closed lists had surfaced [End Page 110] among Pushtun cabinet members close to Karzai, a Pushtun. They feared that the list system could benefit the charismatic non-Pushtuns Dostum, Qanooni, and Mohaqiq, as well as the opposition candidates riding their coattails. This would injure the strategy of Karzai and his confidants, who envisioned a legislature divided between Pushtun MPs—most of whom, if push came to shove, would support the president—and a fragmented non-Pushtun opposition. Thus, without much theoretical basis, the president and his advisers determined that SNTV was the system that would best serve their interests.

In response to this decision, I drafted a memorandum suggesting an open-list version of PR that would allow voters to cast their ballots in support of individual candidates on the party lists. Solicited by Karzai's national security advisor Zalmay Rassoul (who later presented it to Karzai), this proposal was presented to the cabinet by Interior Minister Ali Jalali and garnered the support of some non-Pushtun ministers. But before the debate could be reenergized, the cabinet on February 15 confirmed the SNTV system—with only three cabinet members voting against it.

What Actually Happened

The final results of the September 2005 legislative elections, delayed by accusations of fraud, were eventually announced on November 12, and the new legislature convened on December 18. Considering the controversy that had surrounded Afghanistan's choice of electoral system, what kind of legislature did the SNTV system eventually produce?

1) The relationship between votes and seats won. Whether SNTV produced a legislature broadly representative of Afghan society's political cleavages is difficult to judge. The party system is so embryonic that one cannot simply compare political movements' vote shares with their seat shares in parliament. There are approximately 33 identifiable parties, factions, and alliances in the Wolesi Jirga, but few of them campaigned on any form of coherent ideological platform. Rather, these groups merely consist of independent MPs allied with regional and national strongmen.

One measure of balance is the degree of ethnic diversity in the new legislature. Indeed, there were concerns that partisan politics would produce a Tajik-Uzbek-Hazara opposition bloc pitted against a collection of Pushtun representatives presumed to support President Karzai. Andrew Wilder's recent analysis shows that the ethnic proportions in the new Wolesi Jirga largely adhere to the estimated ethnic proportions in the country as a whole (though these estimates are controversial, as there has not been a nationwide census for more than three decades).12

Though the new legislature is divided along ethnic lines, ethnicity is by no means the only cleavage in evidence. Only 40 percent of the Pushtun MPs can be categorized as pro-Karzai, while 14 percent are in [End Page 111] the opposition camp, and nearly half are nonaligned—including many of the female representatives. Tajiks are split almost equally among pro-Karzai, opposition, and nonaligned camps; only the Hazara and Uzbek MPs sit overwhelmingly on the opposition benches.

Table 1. Ethnicity and Elected Representation
Click for larger view
View full resolution
Table 1.

Ethnicity and Elected Representation

Both pro-Karzai and opposition factions are likely to push hard for legislation rooted in a conservative interpretation of Islam. Approximately 65 MPs are fundamentalist Muslims, with Abd al-Rabb al-Rasul Sayyaf's and Burhanuddin Rabbani's pro-Karzai factions and Qanooni's opposition faction in the vanguard. These groups may also come to enjoy the support of many of the 47 MPs in the more moderate traditionalist camp led by the Hazara Shi'ites. The smallest faction in the Wolesi Jirga comprises 43 more-progressive MPs, including the 13 liberal democrats and leftists as well as the 20 MPs affiliated with Dostum's secular National Islamic Movement.

Some remarkable facts about the members of the Wolesi Jirga, identified by a Kabul-based human rights group, indicate that they are not entirely representative of the greater population. Among the 249 legislators, there are 40 commanders still linked to militias, 24 who belong to criminal gangs, 17 drug traffickers, and 19 against whom there are serious war-crimes allegations. Moreover, nearly half of all MPs were mujahideen (holy warriors) against Soviet occupation during the 1980s.13

When judging the relationship between votes and seats in the Wolesi Jirga, it is important to stress that the candidates elected are not those for which a majority of Afghan voters cast their ballots. Just over two million of the more than six million votes (32 percent) were cast for winning candidates, and thus four million votes (68 percent) were cast for candidates who lost. The extent of this "wasted vote" is remarkably high; in comparison, only 5.3 percent of votes were wasted in the January 2005 Iraqi elections, and less than 1 percent in South Africa's first democratic election in 1994.

The reason for the high ratio of wasted votes was the abundance of competing candidates, which in turn was an expected effect of the SNTV system. With so many candidates in each province, the votes were spread very thin: The first seat in each region was won with an average of 11.5 [End Page 112] percent of the vote, and the last seat was taken with an average of just 5.7 percent. In Kabul province, more than 400 candidates competed for 33 seats, and the last seat went to a candidate who received only 0.5 percent of the vote.

SNTV indeed caused the lottery effect that we had predicted, especially in the larger districts. On average, there were only 864 votes between the lowest-polling elected candidate and the highest-polling (male) runner up. Such tiny margins brought into dispute the results in areas tainted by vote fraud and campaign manipulation. They also make wild swings of legislative power likely from election to election: Despite incumbency—or perhaps because of it—most members of this Wolesi Jirga could very easily be ousted next time around.

2) The establishment of a stable party system. In the runup to the parliamentary elections, SNTV was expected to retard the development of a stable party system by causing political fragmentation, thereby making national legislation the business not of ideologically coherent political parties but of regional warlords and religious fundamentalists. The election outcome gave credence to each of these concerns. Only 16 percent of the more than 2,700 candidates were from registered political parties; these candidates won less than a third of the seats in the Wolesi Jirga.

As noted earlier, there are as many as 33 identifiable parties, factions, and alliances in the 249-seat legislature, the largest group being Qanooni's New Afghanistan party with 25 seats (10 percent of the total). Members of the National Democratic Front, a new alliance comprising 14 liberal-democratic parties, won only 7 seats, and candidates representing the old leftist parties won just 6. Supporting Karzai is a motley collection of small bands led by powerful individuals whose interests will need to be assuaged to guarantee a voting bloc for the president's legislative agenda.

3) Ensuring a clear and effective vote. There is substantial evidence that many voters found the SNTV system and the poster-sized ballots confusing. Because 49 percent of Afghan males and 79 percent of Afghan females are illiterate, the electoral commission had assigned each candidate an icon that was included on the ballot next to the candidate's name and picture. Because there were not enough different icons, some candidates had multiple icons as their symbol, which further added to the confusion.14 Many candidates, especially those new on the political scene, found it immensely difficult to communicate to illiterate voters either their face or their assigned symbol.

According to an October 2005 opinion poll carried out by Charney Research, the main reason respondents gave for not voting was that they did not find a candidate whom they could support.15 Craig Charney argued that this was compounded by the ballot itself, associating "the low turnout in Kabul [with] the electoral system there, where people, often of low literacy, were confronted with pages and pages of ballots."16 Moreover, [End Page 113] the extreme fragmentation of the vote also indicates that in the absence of parties—affiliation was not listed on the ballot even for those candidates backed by a party—it was difficult for voters to fathom which blocs were likely to be influential in the new parliament.

Further evidence regarding the level of confusion among voters comes from the high percentage of invalid or "spoiled" ballots: 2.9 percent of all ballots were rejected because they were marked in error or for disqualified candidates, and 2.1 percent because they were blank. The total of 5 percent compares to less than 1 percent in South Africa's 1994 elections, 1.1 percent in the January 2005 Iraqi elections, and 2.4 percent in Liberia's November 2005 election. Voter confusion was possibly one of the main reasons behind the low overall turnout, which dropped from 69 percent in the October 2004 presidential election to 50 percent in the September 2005 Wolesi Jirga elections (going as low as 29 percent in the south). In Iraq, by contrast, turnout increased from 58 percent in January 2005 to 70 percent in December 2005.

4) Allowing the executive and legislature to govern. The new legislature is likely to obstruct the passage of President Karzai's reform agenda. The plethora of vested interests and the extreme fragmentation caused by the SNTV system mean that Karzai will have to cobble together a majority for every executive bill by way of piecemeal promises and logrolling.

Karzai won 55 percent of the vote in the 2004 presidential election—more than three times the vote share of his closest rival—but the pro-Karzai bloc now makes up less than a third of the Wolesi Jirga (and is by no means monolithic). The opposition makes up slightly more than a third, and so do nonaligned legislators. The strength of the opposition to Karzai became clear when, as soon as the new legislature convened, Qanooni won the coveted Wolesi Jirga chairmanship by 122 to 117 votes over Sayyaf, Karzai's favored candidate. Qanooni subsequently resigned from his de facto position as opposition leader in favor of Rabbani, who previously had been seen as a member of the pro-Karzai camp.

5) Promoting dynamic women in parliament. Female representation was the only area in which SNTV actually proved to have a positive effect. The quota mechanism, which ensured that a total of 68 women were elected (on average two per province), remained largely unchallenged. The fragmenting effect of SNTV helped 19 women—8 percent of all MPs—get elected in their own right without the aid of the affirmative-action mechanisms. In the large Western province of Herat, for example, female candidate Fauzia Gailani outpolled all male candidates, including those backed by local warlords. In the province of Farah, female candidate Malalai Joya came in second; she had bravely denounced the warlords at the constitutional Loya Jirga, and her election may have been the result of protest votes cast by those alienated from traditional, corrupt, and warmongering male candidates.17 It is also [End Page 114] worth noting that in December, the new Wolesi Jirga chose as its second deputy chairperson Fawzia Kofi, who had been elected without the help of quotas in the eastern province of Badakhshan.

Table 2. Afghan Election Results
Click for larger view
View full resolution
Table 2.

Afghan Election Results

Taken together, the 68 women MPs form a highly significant voting bloc, one that is for the most part unaligned with traditional interests. Yet their future influence in the legislature remains unclear. While our fear that many women would get elected with dramatically lower vote shares than unsuccessful male candidates was not borne out, 49 of the 68 female MPs still owe their election to the quota mechanism, having leaped over 422 male candidates who outpolled them (there was an average of twelve higher-polling males per district).

Even the 19 women elected without the aid of quotas received an average of only 3 percent of the vote; their election was as much a consequence of the lottery aspect of SNTV as of their popularity. Altogether, the 68 women elected polled an average of just 2.3 percent each—significantly lower than their male counterparts. The highest vote share for a woman was 9.2 percent in the Panshjir Valley—the heartland of the Northern Alliance—and the highest number of votes was 9,092 for a woman who came in third in Nangarhar province. In Zabul province, the woman who claimed the reserved seat had polled a total of only 751 votes, while the highest-polling defeated male candidate had received 1,816.

The advantages that women received from the quota system, however, should not diminish the amazing progress that they have made in the Afghan political arena. Their advancement is particularly remarkable considering the oppression that they endured under the Taliban just a few years ago.

What Next?

By and large, the SNTV electoral system based on large districts did just what experts had predicted that it would do: It decreased turnout by confusing voters, it created a fragmented legislature largely unrepresentative of the votes cast, and it diminished the prospects for legislative-executive [End Page 115] cooperation. And it did not work in favor of President Karzai, as he and his advisors had hoped. Karzai and his clique had envisioned the emergence of a loyal Pushtun-majority bloc upon which the executive could rely for legislative support. While the SNTV system did to a degree serve to fragment the opposition and retard the emergence of new parties, it also fragmented the president's largely Pushtun base. Indeed, it transformed the absolute majority that he had won in the presidential election into a disjointed bloc comprising no more than a third of the Wolesi Jirga.

As predicted, the new legislature has already shown itself to be a place of wheeling and dealing, of clientelism and shifting alliances, where men with tainted pasts hold significant sway over the future of Afghan democracy. As the liberal-democratic and progressive parties faced high hurdles in getting their messages across and their candidates elected, most hopes for moderation and nonviolent reform are now pinned to the 68 women MPs, most of whom are nonaligned and independent of traditional power structures. If the reserved women's seats are ever abolished, the prospects for reform and democratic progress will be much bleaker than they are today.

The SNTV electoral system came about by a path of muddled missteps, and it was a disservice to the millions of Afghans who deserved a clear and transparent tool to craft their first democratic parliament. Not only did the system fail to provide such a tool, but it is destined to work in favor of those who seek to strongarm and bribe their way into office. If SNTV is used in subsequent elections, the fragmentation and parochialism of the legislature will increase, and politics in general will remain detached from the masses.

No electoral system can transform an illiberal polity into a representative democracy without a raft of supporting social, economic, and institutional transformations. But an appropriately crafted PR system—one that both is proportional and allows Afghans to vote for individual candidates—can do much to encourage the emergence of a stable party system, better translating votes cast into seats won, simplifying the vote, and promoting cooperation between the executive and the legislature. Only by adopting such a system will Afghanistan be able to avoid the great anomalies that were so apparent in the 2005 Wolesi Jirga elections.

Andrew Reynolds

Andrew Reynolds is associate professor of political science at the University of North Carolina at Chapel Hill. He has consulted on issues of electoral and constitutional design for a wide array of countries, including Afghanistan. His latest book is (2002).

The Architecture of Democracy: Constitutional Design, Conflict Management, and Democracy


1. Preamble to the Bonn Agreement, officially called the Agreement on Provisional Arrangements in Afghanistan Pending the Re-establishment of Permanent Government Institutions, signed 5 December 2001. Available at dh/latest/afghan/afghan-agree.htm.

2. Chris Johnson, William Maley, Alexander Thier, and Ali Wardak, “Afghanistan’s Political and Constitutional Development,” report by the Overseas Development Institute and the UK Department for International Development, [End Page 116] London, January 2003. Available at In the interests of full disclosure, I was a consultant to this study.

3. Chris Johnson, et al., “Afghanistan’s Political and Constitutional Development,” 7.

4. See Barnett Rubin, “Crafting the Afghan Constitution,” Journal of Democracy 15 (July 2004): 5–19.

5. See Larry Goodson, “Afghanistan’s Long Road to Reconstruction,” Journal of Democracy 14 (January 2003): 82–99.

6. Author’s private conversations with international members of the JEMB, an Afghan cabinet minister, and others close to the process.

7. King Hussein believed that limiting a voter to a single vote in a multimember constituency would reduce the capacity of the Muslim Brotherhood to win seats. Subsequent elections gave some credence to this notion. See Andrew Reynolds and Jorgen Elklit, “Jordan: Electoral System Design in the Arab World,” in Andrew Reynolds and Ben Reilly, eds., International IDEA Handbook of Electoral System Design (Stockholm: International IDEA, 1997), 53–54.

8. Andrew Reynolds and Andrew Wilder, Free, Fair or Flawed: Challenges for Legitimate Elections in Afghanistan (Kabul: Afghanistan Research and Evaluation Unit, 2004).

9. The memo was solicited by Karzai’s National Security Advisor, Zalmay Rassoul, and was discussed with Minister of Reconstruction Amin Farhang and Finance Minister Ashraf Ghani before reaching President Karzai.

10. See Andrew Reynolds, Lucy Jones, and Andrew Wilder, “A Guide to Parliamentary Elections in Afghanistan,” Afghanistan Research and Evaluation Unit, August 2005, available at; and International Crisis Group, “Afghanistan: From Presidential to Parliamentary Elections,” Asia Report 88, 23 November 2004.

11. Author’s private conversations with international members of the JEMB, January 2005.

12. Andrew Wilder, “A House Divided? Analysing the 2005 Afghan Elections” Afghanistan Research and Evaluation Unit, December 2005, 8. Available at

13. See “Let’s Make A Deal: A Democracy Arrives, Afghan Style,” New York Times, 4 December 2005, 14.

14. Candidates were given a choice of three symbols drawn at random. These included representations of animals, forms of transport, cell phones, oil rigs, and weather patterns. Thirty-five MPs were elected with animal symbols (with birds leading the way with 16); 32 won with planes, trains, or automobiles (plus four boats in this land-locked nation); and there were 14 fruit-and-vegetable MPs.

15. ABC News poll, “Despite Deep Challenges in Daily Life, Afghans Express a Positive Outlook,” 7 December 2005. Available at

16. Comment by Craig Charney at “Afghanistan at a Crossroads,” a roundtable discussion hosted by the Century Foundation, 7 December 2005.

17. See Andrew Wilder, “A House Divided?” [End Page 117]

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.