Report discusses the findings of the Election Violence Monitoring (EVM) project, which was implemented by the Trust for Democratic Education and Accountability – Free and Fair Election Network (TDEA-FAFEN), to monitor electoral violence across Pakistan during the 2013 General Elections. The monitoring period for this report was February 1 to June 11, 2013. The principle findings discussed in this paper are as follows:

  • During the pre-election period (February 1 to May 10), reports of high-explosive election-related violence far exceeded those of more “traditional” forms of electoral violence, especially in Khyber-Pakhtunkhwa and Balochistan. Sindh was the only “electoral militancy-hit” province where reports of election-related gun violence far surpassed those of high-explosive electoral attacks.
  • In Khyber-Pakhtunkhwa and Balochistan, targets of overall electoral violence were mostly candidates and party leaders. In Sindh, however, far more attacks on party workers rather than on candidates and party leaders took place, particularly in Karachi.
  • Most militant-claimed incidents were high-explosive attacks, because militant organizations tend to claim responsibility for high-profile, high-lethality attacks due to the visibility such attacks buy them. However, four shootings – all of them successful assassinations of candidates – were also claimed by militants.
  • Punjab was entirely safe from election-related high-explosive violence (i.e. violence involving bombs, rockets or other explosive devices) during the entire reporting period. The only two incidents of “suspected militant” violence in Punjab (both incidents of targeted gun violence) occurred in southern districts towards the end of the election campaign period and were both aimed at the PPP.
  • The Awami National Party (ANP) was, by far, the most targeted political party in high-explosive violence – followed by the Pakistan People Party (PPP) and Jamiat Ulema-i-Islam (JUI-F).
  • While it is commonly believed that only selected political parties were targeted in high-explosive attacks, the reality is that candidates and workers of virtually all major political parties were targeted – albeit in varying degrees, different regions and at different stages of the electoral cycle.
  • Out of the constituencies that experienced the most election violence, the majority were those where one or both of the following applied:
    • the margin of victory (MoV) in the previous election and/or the expected MoV in the 2013 was low;
    • races were dominated by “personality politics” i.e. candidates had high incentives to cultivate personal or
      family votes.
  • Out of the constituencies that experienced the most election violence, the majority were those where incumbents ended up being re-elected (in many cases, despite having switched political parties between 2008 and 2013) or where incumbent parties returned.
  • Political parties that were most often involved as perpetrators or participants in incidents of violence in a given province were also the winners of the most National Assembly seats in that province. The only major party this did not hold true for was the Awami National Party (ANP), because the ANP had lost too much popular support for the use of violence to make any advantageous difference.
  • Most major political parties appeared to follow strategies for where to employ violence, based on factors such as party performance in the 2008 General Elections and candidate incentives to cultivate personal/family votes.
  • There were reports of polling day violence in nearly all the constituencies where rejected votes exceeded the MoV; most of these constituencies had reports of major clashes on Election Day – including one constituency where two people were killed in Election Day violence.
  • In constituencies where polled votes were significantly less than in 2008, high intensity of pre-election and/or Election Day violence was unequivocally a factor. However, this does not explain why many highly violent constituencies experienced increases in polled votes compared to the last election. Furthermore, voter turnout as measured by number of votes polled against registered voters is not an accurate measure of voter behavior in elections where vote-counts are often fraudulently inflated.

To download complete report, click here