Final report of the 2018 voters’ roll audit

Executive Summary

The ZESN conducted an audit of the 2018 Biometric Voter Registration (BVR) voters’ rolls. The audit was conducted to provide independent non-partisan information on the quality of the 2018 PVR and the 2018 FVR to all stakeholders, including: the ZEC, political parties, civic organizations, and citizens. The methodology of the voters’ roll audit encompassed conducting computer audits, people-to-list and list-to-people field tests. ZESN examined accuracy, currency and completeness of the voters’ rolls provided.

ZESN conducted the audits for voters roll released in electronic format by the Zimbabwe Election Commission (ZEC) on 16 June 2018 and the 25th of July 2018. ZESN notes that unlike in past elections the ZEC, provided an electronic copy of the 2018 FVR to stakeholders in accordance with the Electoral Act (Chapter 2:13) as provided for in Section 21 (7) which prescribes that where a voters’ roll is provided in electronic form …, its format shall be such as allows its contents to be searched and analyzed.

The 2018 FVR will play a critical role in the conduct of the 2018 harmonized elections. It will be used to determine who is able to vote and who is not on Election Day. Errors on the 2018 FVR regardless of whether they are intentional or not could prevent eligible individuals from voting, or enable illegal voting. Further systematic problems that affect particular groups of individuals can fundamentally undermine the credibility of an election. While no voters’ roll is perfect, if the 2018 FVR is widely perceived as accurate, current, and complete it will enhance confidence in the overall process for the 2018 harmonized elections.

Understanding the quality of the 2018 FVR is particularly important given serious challenges identified with the 2013 preliminary voters roll (PVR). Analysis of the 2013 PVR showed a clear pattern of over registration in rural areas and with older voters and under registration in urban areas and with younger voters. These issues undermined the credibility of the 2013 harmonized elections.

In terms of accuracy, the 2018 PVR has complete data for all registrants and few obviously duplicate registrants, but potential errors were identified related to identifying information (in particular date of birth data for registrants in Mashonaland Central). In terms of currency, the overall number of registrants decreased form -3%. The percentage of new, first time registrants is 43%. Urban registrants increased from 28% to 33%, but registrants for Bulawayo decreased by -14%. Female registrants increased from 52% to 54%. The number of registrants 18 to 22 increased by 372% and those 23 to 27 increased by 83% while the number of registrants 93 to 97 decreased by -91% and those 98 or older decreased -94%.

The analysis also compared the 2018 FVR voters’ roll, with the 2018 PVR and ZimStats 2018 projected census figures. On the whole, the registered voter population of 5,695,706 is lower than both the eligible voters deduced from the2012 census population (6,805,455) as well as those from the 2018 census projections (7,224,128). However, there is a marginal increase of 0.21% in the number of registrants from 5 683 936 reported in the 2018 PVR to 5 695 706 in the FVR. However, the majority (80.40%) of these additions are found in Mashonaland West province. Of the total adult population of registrants contained in the FVR, roughly 54% are women and 46% are men. The majority (68%) of registered voters reside in rural Zimbabwe compared to 32% who stay in urban areas. The analysis managed to flag out a number of records that belong to suspected multiple registered individuals. There are also a few strange cases where the FVR registrants are more than the official 2018 adult population projection.

In addition, ZESN also conducted complementary field tests audit of the 2018 FVR. The field test audit had two components: a list-to-people test which involves randomly identifying individuals from across the country on the 2018 FVR and then deploying enumerators to locate the sampled individuals and verify their registration information; and a people-to-list test in which enumerators randomly select individuals from across the country and request their registration information and then compare this to what is in the 2018 FVR. From the field tests ZESN notes that there is a high level of consistency in entering data of registrants on the voters’ roll. ZESN could not locate 8.3% of respondents, who were unknown at address given in the voters’ roll.

The findings of these audits provide an overall assessment of the 2018 FVR. Overall, ZESN’s audit found that the 2018 FVR is an improvement over the 2013 PVR, but that challenges remain. ZESN’s computer audit found a wide variety of mistakes in the 2018 PVR and 2018 FVR. While every voters’ roll contains inaccuracies, more could have been done to clean the 2018 FVR. ZESN’s computer audit found that while registration rates for urban and young voters on the 2018 FVR have increased significantly from the 2013 PVR, concerns remain about under registration in urban areas and with younger voters relative to registration rates for rural and older voters. While the ZEC allocated additional BVR kits to rural areas that lead to high registration rates for rural and older voters, the ZEC did not, despite requests, make similar efforts to ensure higher registration rates for urban and young voters. This disparity undermines the inclusiveness of the elections.

However, registration rates for urban areas and younger registrants continue to lag behind those of rural areas and older registrants in both the 2018 FVR and 2018 PVR – though not as much as in the 2013 PVR.

Recommendations

The ZEC released the 2018 FVR too late in the process to address any shortcomings identified for the 2018 harmonized elections. Had the 2018 PVR been released in electronic format earlier or had the ZEC made more detailed information available in electronic format at the end of each phase of the BVR process then a computer audit could have been conducted much earlier and recommendations could have been offered in a timely fashion to enhance the inclusiveness of the BVR process and the quality of the resulting 2018 FVR. This could have helped build public confidence not only in the 2018 FVR, but also in the ZEC itself in advance of Election Day.

Based on our analysis of the 2018 FVR, ZESN offers the following recommendations for future voter registration processes:

  • The ZEC should equitably distribute BVR kits to urban and rural areas,
  • The ZEC should ensure that all eligible voters are given the opportunity to register to vote as the current voters’ roll exclude 21% of potential voters.
  • The ZEC and civic organizations should conduct more targeted voter education and voter registration activities targeting young voters between the ages of 18 to 32;
  • The ZEC should provide more timely detailed registration information in electronic format for each phase of the BVR process;
  • The ZEC should compare registration information across voters’ rolls to identify possible errors;
  • The ZEC should also explain what constitutes a material or immaterial error on the voters’ roll, in order to prevent disenfranchisement of registrants whose registration details might have errors;
  • The ZEC should consider posting voters’ rolls outside inspection centres to allow for public scrutiny for removal of ineligible voters by reason of non-residence, or death;
  • The ZEC should further clean the voters’ roll by, investigating all unusual entries and removing duplicate entries;
  • Achieving a credible voters’ roll should be possible if ZEC continues to invest in good data management and analysis skills to deal with complex nature of the voters’ roll data;
  • ZEC needs to check out strange instances where the registered voters exceed projected population figures especially for age groups above 39 years.

Source: Zimbabwe Election Support Network (ZESN)

Download PDF (3.8MB PDF)