of registered voters per day. Available figures indicate a significantly higher proportion of young voters, under-35 years of age, registered nationally in 2018 compared to 2013. Recommendation 7: Ensure enhanced voter registration in “under registered” districts of the country to ensure universal and equal suffrage. In determining a target figure for the estimated eligible voter population, ZEC (and also ZESN) used a purely mathematical population projection (based on a 1% annual growth of population against the census). However, using a more robust projection by ZIMSTATS35 it would suggest that the total number of eligible voters should be some 900,000 more, which would mean that the ZEC capture rate for the 2018 elections would be below 70% nationwide. One of the key elements of the roll is the residence address of a voter, which determines the polling station and constituency allocation. However, in many cases it is hard to verify the real address of a voter, as there is no national household list. This problem is exacerbated by the fact that registrars accepted the same address in various formats and also by the use of pre-commissioned affidavits36 during the registration period which possibly are not entirely accurate. Recommendation 8: In order to ensure universal and equal suffrage as well as the right to vote, there is a need to review procedures for the identification of voter addresses (lessening the overuse of pre-commissioned affidavits), to ensure the correct allocation of polling station / ward / constituency for each voter, and to ensure a consistent format for the capturing and registering of voter addresses. EU EOM research of the June voter roll found 41 duplicate ID numbers, three voters born in the 19th century37, and close to 2,500 duplicates. It is hard to verify how many of these cases are real duplicates and how many rare coincidences38. The June voter roll was examined and audited by various CSOs and think tanks. An audit by ZESN39 showed that 4% of voters could not be located, because they had moved and an additional 9% (which represents over half a million people extrapolated to the entire population) were not found at their given addresses and were not known in the neighbourhood. Most probably these findings reflect a casual approach to the proof of residence during registration, notably through the use of pre-commissioned affidavits. According to ZESN, the roll contained 37,640 addresses with voters registered at the same address but allocated to different wards, affecting 5% of all registrants.40 In a separate report, by Team Pachedu41, more than 9,100 records are assigned to 35 ZIMSTAT Population Projections Thematic Report, August 2015. It is foreseen that a person can sign an affaidavit in lieu of documentation confirming their address. In the Zimbabwe context in some areas people may not possess such documentation or their residenc may not have a formal address as such. Thus it was reported that affadavits were filled out in advance by officials for convenience sake, and other a common point of “address” – such as a church or other identifiable location - was used for a number of people in a given area 37 Most of these mistakes were cleared with issuing the final VR in July. 38 11 cases are exactly the same (with the same address written alternatively). 39 Preliminary unofficial results from the random sample of 1200 voters. 40 In Manicaland 10% of all registrants are listed in the same locations but registered in multiple wards. 41 http://www.teampachedu.org/index.html 36 16