POSTELECTION PERIOD

Results Reporting Process

Although most observer organizations applauded the apparent operational success of the electoral process on election day, Democracy International released a preliminary statement two days after the election emphasizing that it was still too early to tell if the process had been a successful step forward for Afghanistan’s democracy. Despite the fact that a large portion of polling centers opened and operated successfully and a significant number of Afghans had successfully cast ballots in an extremely challenging security environment, no winner had yet been declared. In our preliminary statement, we urged the IEC to provide partial results and to be fully transparent about their decisions and procedures. Withholding information could undermine the operational successes of election day and ultimately threaten the legitimacy of the entire process.

Unfortunately, in the weeks following the election it seemed that, rather than releasing results as they became available and fully explaining its processes, the IEC was managing the announcement of results. Preliminary results were tallied and released at a languid pace. Between the election on August 20 and the beginning of the runoff campaign on October 24, 27 days were spent tallying and announcing the preliminary results, and 34 days were spent auditing and reviewing the preliminary results. The delays in tallying and counting did not serve the process well. The delays raised suspicions that the government was improperly interfering in the process of announcing the results. Some in the Afghan press also charged that foreign interference in the election process was responsible for the delays.^[82]

This apparently managed process undermined the IEC’s credibility and therefore the credibility of the entire election process. On September 6, the IEC determined that it would invalidate votes from 447 polling stations it had quarantined, only to reverse itself a day later on the grounds that it did not have the legal authority to invalidate votes.^[83]Some in the press and international community viewed this decision as an effort by the IEC to bring Hamid Karzai’s vote total above 50 percent in the first round, regardless of whether quarantined or suspect ballots were included.^[84]The IEC’s decision to reverse itself also coincided with the ECC’s decision (described below) to issue an audit and recount order for a substantial number of polling stations.

Because of the complexity of the tabulation and audit and the general lack of transparency, most stakeholders, including both leading political campaigns, judged these processes on the basis of their outcomes rather than their integrity. In other words, most stakeholders appeared concerned only about whether or not these processes would result in a runoff. This outcome-oriented approach substantially raised the risks that at least one competing political faction would not accept the results of the process.

National Tally Center

From DI’s observations at the National Tally Center, the tabulation process was insufficiently transparent to ameliorate significant concerns about improper interference. DI observers were present regularly at the National Tally Center, where results in Tamper Evident Bags from the provinces were being received, opened, sorted, and entered into the national database through a standard double-blind entry process. Although a viewing area was established for observers and candidate agents, the process remained confusing and opaque. Tabulation center staff members were generally unhelpful in answering key questions regarding the tally process, such as whether ballot boxes had arrived from certain provinces or if a protocol was in place for tallying and announcing results. On one occasion, officials told DI observers that the IEC was prioritizing ballot boxes received from provinces with a small percentage of results reported, but this was never publically substantiated. The delays and changes throughout the nearly month-long process of tabulation raised concerns, particularly about the order in which districts and provinces were being processed.

During the tabulation process, some of the delays of announcements of the preliminary presidential election results were attributed to the need for the national tally center to simultaneously process results from the various provincial council elections. The IEC, however, did not announce provincial council election results for 30 provinces until 10 days after the preliminary, uncertified result for the presidential election and long after the national tally center had ceased operations. This claim that delays were a result of provincial council elections raised further concerns about the transparency of the results tabulation and reporting process.

The Preliminary Results

The IEC began releasing preliminary results for the presidential election on August 25 and continued releasing results until September 16, when it released full preliminary results that did not include 646 IEC-quarantined polling stations. The preliminary uncertified results released on September 16 comprised 23,345 polling stations and 5,662,758 valid votes, of which Hamid Karzai won 3,093,256 votes (54.6 percent), 261,878 votes above the 50 percent threshold required for victory in the first round of the election. Abdullah Abdullah earned 1,571,581 valid votes (27.8 percent), finishing second.^[85]

DI observers and analysts were present at each press conference where preliminary, uncertified results were released, and DI tracked and analyzed the results as they were made public. Based on this work, we believe that the apparent managed release of results likely contributed to perceptions of fraud in the first-round election. Areas in which Hamid Karzai was likely to perform well were not reported until late in the results-reporting process, and thus many in the international media and the opposing campaigns perceived the large increase in Karzai’s vote share in the last few days of results reporting as illegitimate. The lack of transparency in the results reporting process compounded perceptions of fraud.

The September 8 Audit Order and the Sample Audit

Mounting concerns about electoral fraud prompted the ECC to order the IEC on September 8 “to conduct an audit and recount of ballot boxes in the presidential election in polling stations nationwide for which the preliminary results” met either (or both) of the following two criteria:

1. The total number of votes cast in a polling station for the presidential election was equal to or greater than 600 (since polling stations were issued six ballot books with 100 ballots each);
2. Any one presidential candidate received 95 percent or greater of the total valid votes cast in a polling station.^[86]

The intent of the ECC to investigate potential fraud in a systematic manner should be applauded. The ensuing audit process, however, was flawed. Throughout the process, the IEC and ECC issued multiple methodological and mathematical corrections. More important, no plan was developed for interpretation of the audit results, and the sampling procedure and statistical analysis suffered from significant methodological flaws. 

Before issuing the audit order, the ECC issued a number of orders addressing specific Category 1 complaints (600 or more votes per polling station), which invalidated the results from entire polling stations and made the audit order appear to be a response to widespread fraud. At the time the ECC issued the order, eight days before the preliminary uncertified result was announced, the ECC did not know how many polling stations would meet either criterion or how many votes would need to be included in such an audit. Announcements about the number of polling stations that met the two criteria varied because the IEC committed several technical errors in the implementation of the order. Ultimately, 3,376 polling stations were identified that met the audit criteria, which together comprised roughly a quarter of all valid votes in the preliminary uncertified results.

When the full preliminary results were issued in mid-September, the ECC and IEC concluded that an “audit and recount” of the polling stations covered by the September 8 order could not be completed in a reasonable timeframe that would allow a runoff election, if one were to be required, to be held in 2009. Moreover, there were significant concerns about conducting audits in provincial centers as initially planned. After more than a week of negotiations among the IEC, ECC, and various diplomatic stakeholders, the IEC and ECC agreed to employ a sampling procedure to determine the findings of the ordered audit. The sample audit would examine 358 ballot boxes, which would be brought to an audit center in Kabul. No recount was planned or conducted. Two UN-appointed consultants developed the sampling procedure. Democracy International has not identified any precedent for using sampling to determine an election result in a developing country context.

At the time the sampling procedure was agreed upon, the IEC the ECC had no plan for interpreting the sample results and applying them to the overall set of polling stations covered by the audit. Despite having drawn a random sample of polling stations on September 24, the ECC did not decide on a formula for interpreting the findings until October 4, less than 24 hours before the opening of ballot boxes began in Kabul. The initial announcement of the formula was met with confusion regarding how exactly it would be interpreted mathematically, which led to a second announcement by the ECC outlining the mathematics of the procedure. This second announcement contained mathematical errors that subsequently were determined to make the initial formula inappropriate for the sampling method in question, and thus the formula was re-issued on October 11. This series of decisions and clarifications damaged both the credibility of the ECC and the audit process.

The population of suspicious polling stations was originally stratified into three categories based on the two aforementioned criteria for suspicion: (1) polling stations with 600 or more valid votes; (2) polling stations with 100 or more valid votes in which one candidate received 95 percent or more of the vote; and (3) polling stations in which both 600 or more valid votes were cast and one candidate received 95 percent of the total vote. A random sample of polling stations was drawn from each of the three categories. After the discovery that the sample had not been drawn correctly, apparently as a result of confusion regarding the translation of the order, a decision was made to include three additional categories to cover polling stations that had been unintentionally excluded.

Ultimately, six separate random samples were drawn, one for each category. Next, a “coefficient of fraud” was determined by dividing the number of invalidated votes in each sample category by the number of pre-audit valid votes in that category. Invalidation rates (or “coefficients of fraud”) ranged from 53 percent to 96 percent.

The invalidation rates were then applied to the votes of each candidate in the six strata, which came to be known as the “collective punishment” approach. The approach was best described to one DI observer as “using one stick to beat both the guilty and the innocent.”

Democracy International calculated that the total number of votes invalidated by the audit process was approximately 1.26 million votes, about 1 million of which were cast for Hamid Karzai.

The Audit as a Flawed Approach

The statistical approach employed in the sample audit process did not meet the basic requirements necessary to employ statistical sampling. For a sample-based audit process to produce a result that could instill confidence, the assumption of a normal distribution would need to hold true. There was no evidence to suggest the incidence of fraud was normally distributed across a population of polling stations in Afghanistan. Rather, where fraud exists, a normal distribution cannot be assumed. In short, sampling the entire population based on the assumption of the normal distribution of fraud is simply an invalid statistical approach.

Beyond the lack of foundational integrity, a number of other issues existed with the methodology employed for the sample audit.

First, units of analysis were inconsistent in the sampling and application of coefficients. The audit process investigated the likelihood of the incidence of fraud at the polling-station level, but claimed the ability to produce a coefficient that could be applied at the individual voter level to discount individual votes, not polling stations. By changing from one unit of analysis (polling stations) to another (votes), the process infers findings for a population that has in effect not been sampled or investigated at all.

Second, the ECC inaccurately asserted the Margins of Error (MoEs) for its calculations were close to half a percentage point. This assertion was incorrect because the audit process investigated the proportion of fraudulent ballot boxes in a sample, not the proportion of fraudulent votes. As such, the sample sizes were simply too small to produce such low MoEs and high levels of confidence.

The ECC also implied that based on these “low” MoEs, the application of its “collective punishment” coefficient (invalidating the votes of all candidates at the same percentage) was accurate within half a percentage point. Even if the ECC’s MoEs were accurate, they make no statement about the confidence of applying a coefficient to a population, only in the confidence that if another random sample was drawn, the same coefficient would be calculated.

Third, the stratification of the sample was fundamentally flawed. The inability of DI analysts to recreate the audit category samples, as well as the inclusion of three additional categories late in the audit process, call into question the stratification of the sample. Sample stratification should be based on homogeneous characteristics that are thought to have some varying effect on the question under investigation. Stratifying the sample by provinces, for example, would have merit on the basis that the incidence of fraud might have been more prevalent in some provinces than in others. Unfortunately, there is no statistically sound reason why type of suspicion – such as polling stations with 600 or more ballots or polling stations with greater than 95 percent votes cast for one candidate – can serve as a basis for sample stratification.

Fourth, one of the ECC’s guiding principles was logically inconsistent. As explained above, it was not possible to make judgments on individual votes given the audit procedures, only on ballot boxes at polling stations. The ECC’s audit process, therefore, imposed a double standard. The ECC invalidated entire polling stations based on the principle that where fraud existed the voting process was compromised at that location. But it abandoned this principle by applying a coefficient to a candidate’s vote total, in essence arguing that some votes can in fact be salvaged from compromised polling stations.

Last, the “collective punishment” approach contributed to the creation of a flawed incentive structure for the future. By invalidating the votes of all candidates by the same percentage, all those who tempted to commit fraud in future elections could take from this the following lesson: for every percentage point by which you fraudulently increase your candidate’s vote total, your competitors will be stripped of one percentage point of their votes, fraudulent or not. Although authorities will not allow the guilty party to keep its fraudulent votes, it will reward that party by taking votes away from its competitors.

While the audit process may have provided the means by which to reach a political solution regarding the next step in the election process, the audit was based on a series of flawed approaches.

IEC-Quarantined Polling Stations

Separate from the ECC audit process, the IEC decided on its own to withhold the results from certain suspicious polling stations, pending further review. In a press release on September 12, the IEC laid out three criteria for quarantining stations:

1. The number of the votes cast and recorded was more than the number of the ballot papers handed over to a center on election day;
2. The number of votes cast and recorded in a polling station was more than 1000;
3. The votes were cast at locations that were not scheduled to open or, based on IEC information, did not open as a result of security conditions.

The IEC initially announced these criteria on September 16 to cover 579 stations, but that number was later increased to 646 stations. In early September, the IEC announced it would invalidate the quarantined polling stations, then believed to number 447. But the IEC then reversed its own decision a day later, concluding that it did not itself have the legal authority under the electoral law to invalidate results. The IEC then referred all quarantined stations to the ECC for a determination of their validity. After the ECC review of these stations, 18 were ruled acceptable, having no evidence of fraud. The ECC ruled that for 344 polling stations (not covered in other complaints) there was clear and convincing evidence of fraud. The ECC also ruled that no other polling stations from the quarantined stations could be included in the final results by the IEC without first applying audit and recount “coefficients of fraud,” if applicable. It is not clear that the IEC did in fact apply the “coefficients of fraud” to the remaining quarantined stations.  

The Post-Election Complaints Process

In total, the ECC received 2,854 election complaints on and after election day. Of these complaints, there were 604 complaints ruled to be “Priority A” complaints regarding the presidential election, meaning that the complaint could affect the vote outcome at one or more polling stations. The total number of polling stations covered in these Priority A complaints is not known, but the ECC decided to invalidate 210 polling stations on these grounds. Comparing the total number of votes invalidated by the ECC including the audit, the Priority A complaints process accounted for less than 5 percent of the total votes invalidated.

Many in Afghanistan sharply criticized the ECC process as being led by foreigners and used by the international community to influence the election process. These criticisms came to the fore when one of the ECC Commissioners, Mawlawi Mustafa Barakzai, briefly resigned from the ECC on October 12; he charged that the three international representatives on the ECC were “making all decisions on their own.”^[87]President Karzai rejected Barakzai’s resignation on October 14 on the grounds that Barakzai needed to continue to serve “in the larger interest of the Afghan nation.”^[88]It is not clear what formal authority Karzai had to reject Barakzai’s public resignation. A supreme court appointee, Barakzai never filed a formal written withdrawal, which prevented the supreme court from appointing a replacement. The episode, which occurred just days before the ECC’s announcement of its findings from the sample audit, further reinforced an image of the ECC as embattled and politicized.

Overall, the ECC process, including the audit, began as a challenging procedural, investigative, and legal endeavor. The Commission then compounded those difficulties by failing to clearly explain its work to the Afghan public. Even though the ECC faced enormous political pressure, it was responsible for its own decision-making, particularly regarding the audit and sampling procedure, and thus bears responsibility for the flawed process.

IEC Decision to Order a Runoff

After the ECC issued its decisions publicly on October 19, the IEC faced the question of how to interpret the ECC’s rulings. Some, including the ECC, believed that the interpretation of the decisions was clear-cut: the IEC should implement the announced formulas and announce a result accordingly.

On October 19, DI issued a public statement explaining the significance of the ECC’s decisions. Under Afghanistan’s electoral law, these decisions were binding. As the DI press release explained, after applying the ECC’s decisions to the preliminary results, Hamid Karzai’s percentage of the vote fell below the 50 percent threshold needed to avoid a second-round runoff election. (See Table 2.) As a result, a runoff between Hamid Karzai and Abdullah Abdullah was necessary under the law.

The IEC argued that it had the right to reject the ECC’s findings in the case of the audit because the audit fell outside the scope of the regular ECC complaints process. Only a small fraction of the polling stations in the audit had a separate Priority A complaint against them. Supporters of Hamid Karzai also applied political pressure in an attempt to force the IEC to reject some or all of the ECC’s findings regarding the audit.

After much well-publicized political wrangling, the IEC accepted the ECC’s decisions and announced that they had been implemented, resulting in Hamid Karzai’s vote share being reduced to 49.67 percent. Two days later, the IEC revealed that Abdullah Abdullah had received 30.59 percent of the remaining post-audit valid vote. It is still unclear exactly how the IEC calculated the final certified results, which were inconsistent with DI’s calculations from publicly available numbers. DI requested, but never received, clarification on the calculations used by the IEC.

The Politics of the Runoff Campaign

Official campaigning for the runoff began on October 24 and was scheduled to end on November 5. Campaigning was almost entirely process-oriented, with little public debate about any substantive policy issues. Neither candidate held major rallies with their supporters or made major trips to the provinces. The central public arguments in the aborted runoff campaign concerned whether any IEC Commissioners, most notably IEC President Azizullah Ludin, would be replaced before the runoff election.

At the outset of the runoff process, Abdullah’s campaign issued 17 written demands it considered necessary to ensure that the second round of voting would be conducted fairly. Abdullah’s conditions called mostly for a wide range of officials to be replaced, including three cabinet ministers. President Karzai, who appointed the IEC Commissioners and cabinet ministers, refused to make any changes to either body before the second round of voting. The electoral law and the Afghan constitution were ambiguous as to whether the president had the formal legal power to remove the IEC chairperson. Nevertheless, Karzai would have very likely had the political capacity to force a resignation if he had wanted to do so. The IEC did make some minor changes to its provincial and district election staff, most notably by not re-hiring six provincial election officers and rotating eight more to new provinces.

Abdullah subsequently issued an ultimatum that his demands for personnel changes be met by October 31 or he would withdraw from the race. On November 1, after none of the major personnel changes that he had called for had been made, Abdullah announced in a speech at the Loya Jirga tent in Kabul that he “would not participate” in the runoff election.^[89]

Runoff Preparations

Under international pressure, the IEC began preparations for the runoff in mid-September, well before the runoff was formally announced in mid-October. By mid-September, polling kits were already in the provincial centers. The two remaining items – indelible ink, which has a limited shelf life, and ballots, which could not be printed until there was a formal determination on the results – were produced and shipped in early October, arriving in Kabul a week before the runoff was announced by the IEC. As a result, the major challenges in preparing for the runoff were determining which polling centers would open and whether certain election officials would be remain in place.

The IEC released a new manual of procedures for the runoff election. As a result of the truncated timeline for the runoff, the IEC did not conduct a comprehensive, on-the-ground civic education campaign. Instead, it relied on public service announcements using print, broadcast television, and radio media. The IEC also made a number of other procedural changes, including reducing the size of the polling station staff from five to three and introducing the use of a star shaped hole punch for voter registration cards. (In the first round, circular shaped hole punches failed to penetrate the lamination of voter registration cards.)

In late October, the IEC announced it would attempt to open 6,322 polling centers on the runoff election day, more than were open on August 20, but slightly less than 17,000 polling stations, which was approximately 6,000 fewer stations than were open in the first round. This decision came despite the UN’s recommendation that that the number of polling centers be reduced to 5,817, and it raised concerns that the runoff process would be no fairer than the first round.^[90]Nonetheless, the IEC argued that Afghan and international security forces would be able to secure the increased number of polling centers across the country.

An analysis of the locations of the polling centers also raised concerns about the quality of a runoff election. For example, Paktika, a province that was widely recognized to have significant problems with the quality of its first-round vote, was scheduled to have more than 160 additional polling stations than were reported in the preliminary uncertified results from September 16. Within some provinces there appeared to be evidence of a shift toward a higher density of polling stations outside of major cities. In Kandahar, for example, the province planned less than 100 fewer polling stations than were reported in the uncertified pre-audit results, but Kandahar City had almost 250 new polling stations planned; this meant there were nearly 150 more stations planned outside of Kandahar City than were reported in the preliminary uncertified result. The unexpected increase in the number of polling centers in rural areas in provinces that had proved to be problematic in the first round raised concerns.

Cancellation of the Runoff Election

On November 2, after Abdullah’s withdrawal from the race, the IEC declared the runoff election cancelled and Hamid Karzai elected as president of Afghanistan. The IEC announced that the runoff election was cancelled because only the two leading candidates from the first round could participate and one of those two candidates had announced his nonparticipation.^[91]The legal and procedural basis for this conclusion was debatable, since Abdullah had not filed any formal written withdrawal notice with the IEC, as this had been the procedure for withdrawal of a presidential candidate in the first round. Although the practical value of holding the runoff election with only one candidate participating would have been questionable, the Afghan constitution and the electoral law were ambiguous about whether the IEC had the authority to cancel the runoff.

Back to Top

Provincial Council Results

The IEC released the first results for the provincial council elections, which took place concurrent with the presidential elections on August 20, only on September 26. The preliminary results covered 30 of Afghanistan’s 34 provinces; they did not include Ghazni, Kandahar, Nangarhar, or Paktika. Throughout November and December, the IEC released final certified provincial council results province by province. By December 26, the IEC had released final certified results for all provinces, with Kandahar, Herat, and Takhar the last provinces to have final results released.

Quantitative Review of Preliminary Provincial Council Results

To determine whether there were unusual patterns in the provincial council results, DI conducted a province-by-province assessment of the preliminary results, which highlighted some potential concerns. Suspicious patterns – where multiples of 10, 50, and 100 valid votes – were present in the results. Although far from conclusive, these patterns may have indicated collusion among candidates. The results of this analysis are listed in Table 3.

These tests provided only rough indications of possible fraud, however. The analysis identified polling stations that were the most likely to contain suspicious results. Several other criteria could potentially indicate fraud—such as polling stations with more than 540 total votes and ones with more than 90 percent of votes cast for one candidate—that could suggest more questionable results in each province. Because of the significant problems with the polling-station-level results for the presidential election, DI did not compare the provincial council results with polling station results from the presidential election. Although not a definitive indication of fraud, some of the reported results differed so significantly from what would normally be observed that they raise serious questions about the validity of the provincial council results.

ECC Process of Provincial Council Review

For the provincial council elections, approximately 630 Priority A complaints were filed with the ECC.^[92]The ECC delayed responding to all provincial council complaints until it completed its work on the presidential election on October 18. Analyzing the provincial council results for fraud and taking appropriate corrective action was much more difficult with the provincial council results than for the presidential results because the margins between the candidates were much smaller. As a result, the elimination of even a single station could indeed affect the results of a particular race.

DI analyzed a sample of provinces to assess the ECC review process. This limited analysis indicated that there was not necessarily a strong correlation between suspicious patterns of vote totals in the uncertified results from the IEC and the number of votes invalidated by the ECC. Consistent with its mandate, the ECC appears to have invalidated results in the provincial council elections primarily in response to complaints that presented verifiable evidence. As a result, invalidations appeared to be much more common for reasonably accessible provinces close to Kabul (such as Panjshir and Kapisa), while those provinces that were less accessible (Helmand and Farah) had far fewer ECC-invalidated votes than one might expect based on the vote patterns. These results are displayed in Table 4.

The ECC faced serious methodological challenges identifying and developing sufficient evidence of suspicious voting patterns in provincial council elections. It also faced real financial constraints, which limited its ability to conduct comprehensive investigations in all provinces. While a broad review of the provincial council elections, as occurred for the presidential elections, would not necessarily have been practical, with additional resources and more time the ECC could have perhaps more effectively scrutinized the provincial council results.

Back to Top

RECOMMENDATIONS

Afghanistan’s 2009 elections revealed numerous flaws both in the country’s election system and in the administrative processes that support it.

The Afghanistan and Pakistan Regional Stabilization Strategy released in January 2010 by the U.S. Office of the Special Representative for Afghanistan and Pakistan highlights the importance of supporting Afghan efforts to reform the electoral system. This document put forward two milestones for success: (1) “political agreement on the way forward for elections scheduled for 2010 and beyond,” and (2) “implementation of agreed reforms to elections and related laws.”^[93]

Relevant stakeholders should pursue both a short-term tactical and a long-term strategic approach toward strengthening Afghanistan’s ability to hold credible and acceptable elections. With legislative elections scheduled for September 2010, concrete steps can be taken in the short term that will improve the quality of the election process. At the same time, however, larger reforms that cannot be enacted before an election in 2010 remain crucial for the long-term strengthening of Afghanistan’s democratic system.

Based on its extensive observation of the election process, and as articulated in the findings and analysis of this report, DI urges the government of Afghanistan and the international community to consider the following recommendations for needed reforms.

Back to Top

Strengthening Elections Scheduled for 2010

Maintain the impartiality of the IEC. The perceived partiality of the IEC in 2009 was one major contributing factor to the ultimate failure of the electoral process to win the confidence of the people of Afghanistan and the international community. If the IEC continues to be perceived as biased, the people of Afghanistan may forever question the legitimacy of their election commission and the efficacy of elections as a means of choosing leadership. The IEC must avoid even the appearance of partiality and should make every effort to instill confidence in candidates, voters, and international stakeholders that the IEC is in fact the independent institution Afghanistan needs to run successful elections. If the IEC becomes tainted through its own actions, the government of Afghanistan should act to ensure that impartiality is restored.   

Establish swift and effective methods for prosecuting electoral fraud. The government of Afghanistan should grant an appropriate government body the jurisdiction to swiftly prosecute and punish those implicated in electoral fraud. While the ECC played a vital role in providing an avenue for complaints to be heard and investigated, no existing institution appears to have both the authority and the political will to prosecute individuals guilty of electoral offenses with the speed necessary to create incentives that could deter the fraud committed in 2009.

Eliminate the entry points of fraud. The IEC should examine possible entry points for fraud and develop systems by which to mitigate the risk. Specific examples of entry points for fraud include the distribution of sensitive election materials, the certification of tally sheets, and the counting and tallying processes. By identifying these vulnerabilities, Afghan and international actors can focus their efforts on deterring the fraud that occurred in 2009.

The prevention of widespread fraud cannot and should not depend on the honesty and integrity of each and every election official. Although all effort should be taken to recruit and train an election workforce of the highest integrity, it is even more important that systems that provide avenues for corruption should be reformed. Checks and balances must be better distributed throughout the process of administering elections so that one corrupt district field coordinator is not be able to undermine the entire election in his or her district.

Abandon the use of sample-based audits. Sample-based audits should not be used to correct for fraud. Random sampling should be used to determine whether the incidence of fraud was widespread enough to warrant further concern, not to make calculations about the extent of fraud within potentially fraudulent polling stations. Using random sampling to invalidated election results is both unprecedented and fundamentally flawed.

The specific methodology used in the 2009 sample audit applied the same invalidation rates to all candidates. If adopted as a response to potential fraud in the future, this method could actually create incentives for fraud. The approach also sets a precedent for simplistic expediency rather than thorough investigation; instead of pursuing fraud diligently and comprehensively, it employs a shortcut that compromises the basic democratic principle of counting every vote.

Increase transparency of election administration and improve coordination between relevant election stakeholders. Throughout the 2009 election process the individuals and organizations collectively responsible for the success of the election effort too often came to operate as competing rivals rather than as complementary actors. The lack of transparency, effective communication, and coordination among election stakeholders contributed to an environment of uncertainty and mistrust. Better communication and coordination among electoral authorities, implementers, and international assistance organizations would help strengthen the administration and integrity of the election process by creating an environment of collective oversight that did not exist in 2009.

Develop and present for comment a plan for the tracking and distribution of sensitive elections materials. The IEC should develop a clear and comprehensive plan for the distribution of sensitive elections materials that includes a clear strategy for tracking the transit of materials to and from Kabul. The IEC should maintain a temporary operations room with the specific task of overseeing the distribution of sensitive election materials. Election authorities should brief international and domestic observers before and at the conclusion of major operations for the transport of sensitive election materials.

Establish effective coordination between the IEC and civil society. The IEC and Afghanistan’s civil society organizations should work together to support the election process. Civil society organizations should mobilize their networks to help fill the gaps in the IEC’s staffing capacity. The Afghan Women’s Network, for example, could provide female security screeners at polling stations throughout the country. Such partnerships would help fill critical needs and also strengthen civil society’s confidence in the IEC by establishing CSOs as true stakeholders in the election process. The international community should provide funding for such partnerships.

Strengthen the procedures at the national tally center. The IEC should revisit the design of and procedural rules employed at the national tally center in Kabul. Safeguards in the data entry process should be strengthened. Specifically, the IEC should erect walls between the three data-entry stations and ensure no communication takes place between line managers. The IEC staff should be better trained for communicating the intake and tallying procedures to observers and other stakeholders and explaining specific problems with visibly quarantined ballot boxes.

Back to Top

Reforming Afghanistan’s Electoral System and Improving Long-Term Assistance for Afghan Elections

Replace the SNTV system.The use of the SNTV electoral system stifles the emergence and growth of political parties, limits fair representation for the Afghan people, and provides greater incentives for fraud and intimidation on election day. The international community should support Afghan civil society groups and political actors in their study of alternative systems to SNTV and encourage the Afghan government to change the election system for future legislative and provincial council elections.

Increase assistance to domestic observation groups. As with most fledgling CSOs in Afghanistan, the Free and Fair Elections Foundation of Afghanistan needs to strengthen its capacity. It would benefit from the incorporation of long-term capacity-building and technical assistance into its core plan, particularly when it comes to analyzing data gathered by its observers.

FEFA must also work on identifying potential CSOs for its coalition in all of the provinces so it can coordinate and train these local CSOs well in advance of elections. These CSOs in turn can identify appropriate observers that can receive more in-depth training on the observer process before the election process. Observer reporting forms should also be simplified.

The international community should increase funding and technical assistance to domestic observation groups and ensure that this support extends beyond a specific election cycle. Although  FEFA provided a much-needed domestic observation of the 2009 elections, increased assistance should be provided to help support the development of additional domestic observation capabilities or new organizations. Domestic election monitoring can help build a culture of accountability in the administration of elections and can provide broader election-day coverage than can international election observation missions. Support should be provided between election cycles to ensure the continuity and professionalization of observation groups.

Consider providing election assistance through multiple implementers. The international community should explore whether election assistance can be provided through multiple coordinated efforts, rather than through one implementer, as was largely the case for the 2009 elections. Support to domestic election observation groups should no longer be administered through UNDP-Elect, the organization primarily responsible for providing technical assistance to the election commission. This creates a conflict of interest and places an additional burden on UNDP’s administrative, management, and technical capabilities. By decentralizing assistance, the international community would also increase the likelihood that problems would be detected earlier and should make implementers more accountable.

Reform the selection process for IEC and ECC commissioners. The new leadership of Afghanistan’s electoral authorities has already been appointed for 2010. This should not, however, prohibit long-term reform of the selection process from taking place. President Karzai has an opportunity to show the Afghan people and the international community that he is serious about strengthening his country’s democracy by supporting the parliamentary approval of IEC and ECC appointments.

Create a civil and voter registry. Significant improvements to the voter-registration system cannot be made before parliamentary elections and should not be pursued in that timeframe. The goal of developing a civil registry, however, should not be abandoned. The London Conference on Afghanistan of 2006 culminated in the Afghan Compact, which recommended the creation of a civil and voter registry by the end of 2009. Although this deadline was not met, the government of Afghanistan and the international community should refocus on this goal after the next round of elections.

Re-evaluate electoral capacity-building strategy. The October 2009 quarterly report of the Special Inspector General for Afghanistan Reconstruction (SIGAR) highlights the need for a long-term strategy on building and sustaining Afghan electoral capacity.^[94]An additional SIGAR audit report recommended that UNDP hire a capacity-development advisor who would be tasked with implementing a strategic plan for the transfer of skills and knowledge to Afghan election officials.^[95]The international community must move forward with developing a long-term, sustainable capacity-building strategy for the IEC to contribute to the consistent administration of credible and acceptable elections in Afghanistan. A first step toward this goal would be to implement the SIGAR’s recommendations as soon as possible.

Back to Top

Consensus Recommendations

In addition to the recommendations listed above, DI has compiled and synthesized consensus recommendations from a wide range of Afghan civil society organizations, international observer missions, assistance organizations, and independent election experts.DI drew these recommendations from a comprehensive analysis of 437 recommendations presented by this broad group of stakeholders in Afghanistan’s election process. DI has compiled and made public all of these recommendations in a separate document titled “Recommendations for Electoral Reform in Afghanistan,” which is available for download at www.democracyinternational.com.

From this analysis, Democracy International has identified the following 34 recommendations, which reflect the opinions of a wide range of stakeholders, as the points of major consensus for electoral reform in Afghanistan. DI did not developed these recommendations independently but rather synthesized them from more than 50 source documents. These recommendations should form the basis for an all inclusive discussion of comprehensive electoral reform in Afghanistan.

The key consensus recommendations for electoral reform in Afghanistan are as follows:

1. The use of the SNTV system should be reconsidered. There is broad agreement that the Single Non-Transferable Vote system impedes the development of political parties in Afghanistan and prevents fair and accurate representation of Afghanistan’s diverse population. A public consultation process should take place to solicit the opinions of relevant Afghan actors and international election experts to determine the best alternative system for Afghanistan. One alternative that has been presented is a mixed SNTV-proportional system.
2. Rationalize the electoral calendar: The calendar by which Afghanistan holds elections must be rationalized. Under the current design, Afghanistan will need to hold elections nearly every year for the foreseeable future. This puts unnecessary strain on stakeholders to the election process and prevents proper legislative bodies from undertaking necessary changes to the election system. Some observers have proposed the adoption of a two-tiered election cycle with presidential and parliamentary elections in one tier and provincial council and municipal elections in another, staggered by two years.
3. The IEC and ECC should be vested with more sanctioning authority. Both institutions should be further empowered to impose sanctions in a swift and immediate manner on those found guilty of electoral offenses, including public officials.
4. The Electoral Law should clarify the role the Provincial Council plays in Meshrano Jirga elections. The law should clarify the IEC’s role in such elections, the quorum necessary for elections to take place, and the procedures for replacement of provincial council members elected to the Meshrano Jirga.
5. Appointment process for IEC commissioners should be consultative. Nearly all stakeholders agree that the process for appointment of IEC commissioners, including the chairman, should be changed to allow for a check on executive authority. Most agree the National Assembly should play a role, either by submitting the list of candidates to the president or by approving the president’s selections through the legislative process. Some argue that civil society should be responsible for presenting the list of potential commissioners.
6. IEC must be more proactive in ensuring staff impartiality. The IEC must establish a more sophisticated vetting process to ensure the political impartiality of its staff. Clear penalties should be established for abuse of power and safeguards, such as assigning staff to work in provinces away from their homes, should be put in place.
7. IEC must act more transparently. The IEC should immediately develop and implement a broad civic engagement strategy that builds public confidence in the election process. As part of this strategy the IEC should invite a broad group of stakeholders including members of civil society, candidates, and political parties to attend plenary sessions.
8. IEC must develop a better recruiting program for marginalized populations. The IEC should develop systems to ensure the adequate recruitment of staff to serve marginalized populations, particularly in hard-to-reach areas. The IEC should seriously consider a partnership with civil society in this endeavor.
9. The list of polling locations must be finalized early in the process. Although the changing nature of the security environment may prevent a list of polling locations from being entirely static, the IEC and relevant security actors should strive to identify the final list of polling locations as early as feasible. Ample time should be provided to communicate the list of polling locations to observer organizations and civic educators, and any changes should be announced publicly and with proper explanation.
10. Technical assistance to electoral institutions should be provided through different mechanisms. Assistance to the IEC, ECC, and domestic observer groups should be managed through different mechanisms. By no means should the same advisors have overlapping responsibilities with different institutions. Advisors should be embedded with organizations, and the organizations should have a voice in the nature of the assistance they receive.
11. Support women’s political participation. Increased assistance should be provided to female candidates and legislators. Programs should also aim to increase female participation at all levels of the political process. Efforts should be made to educate male candidates and elected officials about the important role of women in the process.
12. Registration cards should have photos. To prevent proxy voting, particularly by men for women, registration cards should include photos for both men and women.
13. Conduct a national census and consider national ID cards. The international community and the government of Afghanistan should prioritize conducting a national census to create a civil registry. The use of national ID cards should be seriously considered.
14. The voter registry must be addressed. If the voter registry is to be used for elections in 2010 it must be updated and duplicates removed. A consultative process should be conducted with Afghan actors and international experts to assess Afghanistan’s voter registration needs and to present a strategy to solve this problem. The use of a civil registry as a foundation for a voter list should be considered.
15. Constituency delimitation must be prioritized. The MoI and the National Assembly should work in coordination with the Central Statistics Office to prioritize district, municipal, village, and local administrative-unit delimitation. Consideration should be given to creating electoral precincts.
16. Establish a broad-ranging, continuous civic education program. A civic education program should be developed and implemented that covers a broad range of civic responsibility topics and targets all Afghans. The program should be continuous and broadly supported and should specifically target the marginalized and hard to reach areas.
17. The role of the media should be clarified and their independence ensured. The law on mass media should be clarified to ensure reasonable access and to encourage a fair and balanced role for the media in the election process. The independence of outlets should be guaranteed, particularly of Radio Television Afghanistan. Further training should be provided to journalists and other outlets on the media’s role in a democracy.
18. Sustained support should be provided to domestic monitoring organizations. The international community should provide long-term sustained support to existing domestic monitoring organizations and additional CSOs. Technical assistance should focus on expanding their capacity to monitor other aspects of governance and to conduct effective advocacy campaigns.
19. Candidate registration and eligibility should be stricter. To prevent irrational numbers of candidates from contesting elections, the IEC should develop a more robust set of procedures to verify the legitimacy of candidate registration materials. Candidate registration should perhaps require the endorsement of more voters.
20. Additional support for political parties should be provided. The international community should increase its support for political party programs and consider incentives for parties that engage marginalized populations such as youth, women, and minority ethnic groups.
21. Overhaul the process for preventing candidates with links to IAGs. The IEC should not be responsible for determining links to Illegally Armed Groups. The IEC should determine candidate eligibility based on a list provided by another organization. Which organization is responsible for determining ties should be clarified. The international community should provide any intelligence that could improve the list.
22. Guidelines on the use of public resources in campaigning should be established. Access to public resources must be equitable and strict guidelines with clearly defined sanctions should be adopted. A transparent system should be developed and authorities should be more proactive in pursuing violators.
23. Stronger campaign finance regulations should be adopted. All candidates should be required to release public financial disclosures on personal assets as well as periodic reports on campaign fundraising.
24. Develop a long-term electoral capacity development strategy. The international community should work with the IEC and other electoral stakeholders to develop a long-term electoral capacity-development strategy. [Assistance to electoral institutions must be tailored to build self-reliance. International advisors should not be developing plans but assisting in their development.]
25. Investigate the entry points of fraud. A full investigation of fraud in the 2009 elections should be conducted in order to make recommendations to strengthen the process and develop detailed procedures for the detection and mitigation of fraud.
26. Track and secure election materials more effectively. More effort should be taken to secure the chain of custody of sensitive election materials including clear and binding procedures for tracking all election materials.
27. Reassess where counting takes place. The counting and reconciliation process should begin as soon as is practicable after the polls close. Given the security challenges that exist with a count at the polling-station level, serious consideration should be given to conducting the count at the provincial level. If counting is conducted at the provincial level, procedural plans must take into account access requirements for observers and candidate agents at the polling station and provincial level.
28. Redesign the results forms. The results forms should be assessed for improvements, such as including the number of entries on the list of voters at each polling station.
29. IEC should establish clear procedures for investigation, quarantine, and annulment. The IEC should establish specific criteria and transparent procedures for investigation, quarantine, and annulment. This process should be well publicized, and observers and candidate agents should make sure to fully understand it.
30. Constitute the ECC early and provide it with necessary funding. The ECC must be constituted early so as to provide it with the time necessary to build effective and transparent national and provincial-level complaints adjudication processes. It should receive all necessary funding. 
31. The ECC should adopt more open complaints processes. The ECC should adopt more transparent investigation and adjudication procedures and properly educate observers and agents in their use.
32. The ECC should be a permanent body and its authority and independence from the executive strengthened. The ECC should be permanent and should be empowered to impose sanctions. Its independence from executive influence should be strengthened.
33. The authorities of each level of government should be clarified. One institution should be the ultimate constitutional arbiter. The authorities of each level of government should be clarified and checks and balances established. The role of the supreme court in the election process should be clarified.
34. Strict consequences should exist for electoral offenses. Strict consequences should be established for public officials found interfering in the election process and swift and immediate action should be taken. Any citizen implicated in fraud should be referred to and prosecuted in the judicial system swiftly.

Back to Top

[5]International Crisis Group, “Afghanistan’s Election Challenges”, June 24, 2009 http://www.crisisgroup.org/home/index.cfm?id=6176&l=1

[6]The Joint Electoral Management Body (JEMB) originally had nine commissioners and was reduced to seven under the IEC.

[7]Electoral Law, Chapter II, Article 8

[8]The sorting process for provincial council ballots was also slightly different in these four areas. IEC Polling and Counting Procedures Summary.

[9]Electoral Law, Chapter V, Article 18

[10]Unless otherwise noted, the information in this section on logistics and polling staff is from the August 3 logistics update by the IEC.

[11]Figure on sample ballots is from http://afghanelections.org/pdf/2009%20Weekly%20Report%20No27.%2028%20July-3%20August%202009.pdf.

[12]Articles 62 of the Constitution and 14, 15 of the Electoral Law

[13]Article 4.1 of the Rules of Procedure of the Electoral Complaints Commission

[14]Article 4.3 of the Rules of Procedure of the Electoral Complaints Commission

[15]Article 4.5 of the Rules of Procedure of the Electoral Complaints Commission

[16]Articles 9.1 and 9.4 of the Rules of Procedure of the Electoral Complaints Commission

[17]2009 Voter Registration Observation Report, FEFA, p.24

[18]2009 Voter Registration Observation Report, FEFA, p. 36-38

[19]2009 Voter Registration Observation Report, FEFA, p.41

[20]United Nations DSS SIOC Daily Security Reports, Kabul, Afghanistan, 2009

[21]2009 Voter Registration Observation Report, FEFA, p. 43

[22]2009 Voter Registration Observation Report, FEFA, p.44

[23]ECC Press Release, June 9, 2009

[24]ECC Press Release, June 9, 2009

[25]ECC Rules of Procedure, article 7.2a

[26]ECC Press Release, June 9, 2009

[27]ECC Press Release, June 9, 2009

[28]IEC Press Release, June 13, 2009

[29]ECC Press Release, August 18, 2009

[30]“Decree of the President of the Islamic Republic of Afghanistan on Noninterference in the Electoral Affairs of Elections 1388 (2009).” http://www.iec.org.af/assets/PDF/decree/english/presidential_decree_non-interference_in_election_process.pdf

[31]Interview with Commissioner Mohammed Farid Hamidi, Afghanistan Independent Human Rights Commission (AIHRC), July 27, 2009.

[32]Interview with Shahrzad Akbar, Free & Fair Elections Foundation of Afghanistan (FEFA), July 27, 2009 and http://www.fefa.org.af/download/Campaign_Report%20For%20Website.pdf.

[33]http://www.fefa.org.af/download/PressReleaseEN.pdf

[34]DI observer Leslie Knott interview with Sidiqullah Tawhidi, August 15, 2009.

[35]U.S. Department of State, Bureau of Democracy, Human Rights, and Labor. 2009 Human Rights Report: Afghanistan.

[36]OSCE “Election Support Team to Afghanistan Recommendations on Elections 2005,” October 6, 2005, www.osce.org/item/html

[37]“Towards a Successful 2009 Afghanistan Presidential Election,” Workshop report, by Bob Sharp, December 9, 2009.

[38]Jan Malekzade, UNAMA, Interview by Adam Sugar and Tim Fairbank, 6 August 2009, Jalalabad, Afghanistan.

[39]Naim Nazari, Civil Society & Human Rights Network (CSHRN), Interview by Tim Fairbank, Laurie Knop, and Chris Jackson, 26 July 2009, Kabul, Afghanistan.

[40]Mr. Shaharzad Akbar, FEFA, Interview with Adam Sugar and Tim Fairbank, July 26, 2009, Kabul, Afghanistan.

[41]www.fefa.org.af/pressrelease, Election Day Press Statement, August 20, 2009.

[42]Mr. Jandad Spinghar, Executive Director FEFA, Interview with Adam Sugar, August 30, 2009, Kabul, Afghanistan.

[43]Sarwar Hussaini, Executive Director CCA, DI Interview with Tim Fairbank and Laurie Knop, July 28, 2009, Kabul, Afghanistan.

[44]Naim Nazari, Executive Director CSHRN, DI interview with Tim Fairbank, Laurie Knop, and Chris Jackson, July 27, 2009, Kabul, Afghanistan.

[45]Fahim Sadiq, Director ANPO, DI Interview with Laurie Knop and Tim Fairbank, July 30, 2009, Kabul, Afghanistan.

[46]Afghanistan Civil Society Assessment; How Afghans View Civil Society,” by Anika Ayrapetyants, Counterpart International, 2006.

[47]Afghanistan Civil Society Assessment: How Afghans View Civil Society,” by Counterpart International (I-PACS), 2006, Afghanistan.

[48]“Advocacy & Networking: Five Million Women Campaign, www.afghanwomensnetwork.org/advocacy.

[49]Nilab Mobarez, “Provincial Council Election,” United Nations Mission to Afghanistan, May 7, 2009.

[50]Freedom House, Freedom in the World 2009 - Afghanistan, July 16, 2009, available at: http://www.unhcr.org/refworld/docid/4a6452d71c.html.

[51]United Nations Development Fund for Women (UNIFEM), “UNIFEM Afghanistan Fact Sheet 2007,” www.afghanistan.unifem.org/docs/pubs/07/factsheet.html. In addition, in 2004, 87 percent of Afghans believed that women need a male relative's authorization to vote; 35 percent of women believed [they would not have permission/ be eligible] to vote; and 18 percent of men admitted they would not allow their wives to vote.

[52]“Afghanistan’s Election Challenges, Asia Report No. 171, June 24, 2009, International Crisis Group

[53]Nilab Mobarez, “Provincial Council Election,” United Nations Mission to Afghanistan, May 7, 2009.

[54]Najla Ayubi, DI Interview by Laurie Knop, Tim Fairbank, and Chris Jackson, July 27, 2009, The Asia Foundation, Kabul, Afghanistan.

[55]“Afghan Women Candidates Campaign in Burqas,” July 30, 2009 by Rukmini Callimachi, Shirak Pirak, Associated Press.

[56]Female FEFA representative, DI Interview by Laurie Knop and David Aasen, August, 2009, Parwan, Afghanistan.

[57]Vic Getz, IFES, DI Interview by Laurie Knop, August 25, 2009, Kabul, Afghanistan.

[58]Najla Ayubi, Interview by Laurie Knop, Tim Fairbank, and Chris Jackson, July 27, 2009, The Asia Foundation, Kabul, Afghanistan.

[59]Marina Nawabi and Vic Getz, “Barriers and Opportunities for Women’s Participation in the Election Process in Afghanistan,” May 13, 2009, IFES, www.ifes.org/features.      

[60]Najla Ayubi, DI Interview by Laurie Knop, Tim Fairbank, and Chris Jackson, July 27, 2009, The Asia Foundation, Kabul, Afghanistan.

[61]Afghanistan Human Development Report 2007, Center for Policy and Human Development, p. 164.

[62]“FEFA’s report on Voter Registration, Third Phase, p.3, www.fefa.org.af/eng_pages/reports/VRP_Rep/VRP_3.pdf.

[63]International Crisis Group, Asia Report No 171, Afghanistan’s Election Challenges. June 24, 2009, p.1.

[64]Project Athena, Trend Overview. April – June 2009.

[65]“The Situation in Afghanistan and its Implications for International Peace and Security” , UN document A/63/751-S/2009/135, March 10, 2009, p.5.

[66]IEC Stakeholders Meeting, Zekria Barakzai. August 30, 2009.

[67]www.walesonline.co.uk/.../brown-hails-success-of-panther-s-claw-in-afghanistan-91466-24254530/

[68]UNDSS SIOC Daily Security Report. August 1, 2009

[69]“Progress toward security and stability in Afghanistan” Report to Congress in Accordance with the 2008 National Defense Authorization Act, US Department of Defense, January 2009, p.9. In April 2007 the Joint Coordination Monitoring Board (JCMB) agreed to a temporary increase to 82,000 from 62,000. Interview with UNDP Elect Security, August 10, 2009.

[70]WSJ. Afghanistan Enlists Tribal Militia Forces, August 12, 2009. www.online.wsj.com/article/SB124994313594220571.html

[71]USIP Special Report 227, “Afghanistan’s Police – The weak link in Security Sector Reform” August 2009.

[72]United Nations DSS SIOC Daily Security Reports, Kabul, Afghanistan, 2009.

[73]United Nations DSS SIOC Daily Security Reports, Kabul, Afghanistan, 2009.

[74]Overseas Security Advisory Council (OSAC), Warden Message: Kabul Suicide Bombing, August 15, 2009. www.osac.gov

[75]These provinces were Badghis, Balkh, Faryab, Helmand, Jowzjan, Kandahar, Khost, Kunar, Logar, Laghman, Nangahar, Paktya, and Wardak. United Nations DSS SIOC Daily Security Reports, Kabul, Afghanistan, 2009, United Nations DSS SIOC Daily Security Reports, Kabul, Afghanistan, 2009.

[76]United Nations DSS SIOC Daily Security Reports, Kabul, Afghanistan, 2009.

[77]2009 Presidential and Provincial Council Elections: Final Observation Report. Free and Fair Elections Foundation of Afghanistan (FEFA). January, 2009.