In December 1995, after four years of warfare, the three major ethnic parties in Bosnia and Herzegovina convened in Dayton, Ohio, and agreed on a peace settlement. Called the Dayton Accords, they provided the framework and timeline for elections and to establish post-conflict national, entity and local governments.
The Dayton Accords called for elections to be conducted within six to nine months of the signing in December. In April 1996, I was appointed Director General of Elections and within five months, on Sept. 14 to be exact, the first cycle of voting was to be conducted.
Under the watchful eyes of heavily armed NATO troops, Serbs, Muslims and Croats made their way to polling places across Bosnia Saturday to vote in that country’s first post-war national elections.
The elections, mandated by the Dayton peace accords, could determine the shape of Bosnia for the future — whether the country remains reunified under the terms of the accords, or whether the de facto partition that has divided the country since the accords were signed is reinforced.
About 3 million people are eligible to vote for a collective three-member presidency — one Serb, one Croat and one Muslim. Voters will also elect a joint parliament, separate assemblies for the Muslim-Croat federation and the Bosnian Serb republic and regional governments.
According to the Dayton Accords, the Organization for Security and Cooperation in Europe (OSCE) was mandated to “supervise the conduct” of the election.
The September 1996 elections were complicated by a number of factors, including the conflict legacies, operating in a post-communist environment, translating voting messages into three languages (Bosnian, Serb and Croat) and safely moving candidates, voters, supervisors, journalists and observers in a country plagued with some 2 million landmines.
After four years of war, segments of the population were displaced. To determine citizenship and residency for the voter rolls, the 1991 census figures were used because they was considered to be the most accurate since it was conducted before the war.
In order to use the elections as a means to reverse the effects of ethnic cleansing, voters were given a variety of enfranchisement options: To cast their ballots from the current locations of residence; to vote absentee for the former residence; or to declare a “future intended residence” as the voting location.
The Provisional Election Commission was established and election rules and regulations were adopted. It administered the election in partnership with about 140 Local Election Commissions, which were responsible for recruiting poll workers and conducting activities at the polling stations. These Local Election Commissions assembled 300 registration committees, 144 counting center committees and 32,000 polling station committee members at 4,500 polling station sites.
In fulfilling its mandate, the OSCE deployed more than 2,000 supervisors to provide long-term and short-term election supervision at the local level. International election supervisors were afforded rights of intervention if they observed problems with the process.
The North Atlantic Treaty Organization’s Implementation Force was used to provide security, as well as a variety of mission critical services including logistical support for transporting election materials and facilities for voter education training.
There were concerns that there were insufficient international supervisors to genuinely oversee the election process. Unfortunately, the election calendar did not allow sufficient time to train and deploy 4,500 supervisors (one per polling station), as well as the country’s inability to absorb such a large deployment. Instead, the supervisors were assigned two or three stations to supervise on a rotating basis.
Finally, the International Crisis Group asserted that there was a 103.9 percent turnout of voters on Election Day.
The preliminary 1991 census, used for the voter registry, contained 3.5 million names of individuals, some of whom were no longer living. The enfranchisement options permitted people to cast their ballots in one of three locations. Therefore, the exact number of eligible voters could only be estimated since no separate registration exercise was conducted.
