In the implementation of LR one sometimes sees that the so-called quota is rounded up or down to the closest whole number. As will be discussed below, this is an unnecessary and not very logical rule which may give negative side effects. In Zimbabwe the rule is in addition unclear because it is not mentioned in the main rule, but is applied to the examples. The law in its Schedule Eight includes three examples, one for each of the three elections, but the one for the National Assembly and the Senate are equal. One of the examples goes like this: Six seats shall be filled, and 236,634 valid votes are cast for seven parties running. The quota is: Q = 236 634 / 6 = 39 439. The distribution goes as follows: Votes Per cent Integer Remainder Additional seats Total Seats A 89,349 37.76 2 0.26549862 0 2 B C D E F G Total 49,869 40,547 32,064 21,942 1,653 1,210 236,634 21.07 17.13 13.55 9.27 0.70 0.51 100.00 1 1 0 0 0 0 4 0.26445904 0.02809402 0.81300236 0.55635285 0.04191283 0.03068029 2 0 0 1 1 0 0 2 1 1 1 1 0 0 6 Party Seats for Very Small Parties One feature of LR is that the results may seem quite arbitrary in some situations. The method is designed to minimize the deviation from the full proportional result but it does not differentiate between the small and big parties when doing so. This means that very small parties may happen to win a seat with very few votes. By modifying the examples of the act this is illustrated as in tables 3 and 4. Another example which may be realistic in Zimbabwe is the following. There are a total of 217,288 votes for six seats and seven competing parties. Q = 36,214.66… Party Votes Per cent votes A B C D E F G Total 80,123 78,543 39,470 8,134 5,976 3,201 1,841 217,288 36.87 36.15 18.16 3.74 2.75 1.47 0.85 100.00 Or in the terminology of the law which is mathematically equivalent to the table above: Votes A B C D E F G Total 89,349 49,869 40,547 32,064 21,942 1,653 1,210 236,634 Unallocated votes Provisional (Votes – Q*Provisional allocation of seats allocation) 2 1 1 0 0 0 0 4 10,471 10,430 1,108 32,064 21,942 1,653 1,210 78,878 Additional seats Total Seats 0 0 0 1 1 0 0 2 2 1 1 1 1 0 0 6 Table 24: The same example as Table 1 but in the terminology of the Electoral Act 62 2 2 1 0 0 0 0 5 7693.66666… 6113.66666… 3255.33333… 8134.00000… 5976.00000… 3201.00000… 1841.00000… 36214.6666… 0 0 0 1 0 0 0 1 Total Seats 2 2 1 1 0 0 0 6 Table 25: An example where a very small party (Party D) wins a seat Table 23: The National Assembly Example of Schedule Eight of the Electoral Act Party Unallocated votes Provisional Additional (Votes – Q*Provisional allocation of seats seats allocation) The table below shows the “cost” of the seats, or the number of votes behind each seat won by a party, using LR and the more robust division method Sainte-Laguë. The latter method would make the cost more equal, and it would be seen to be much fairer. Party A B C D Votes 80,123 78,543 39,470 8,134 Largest Remainder - Hare Number of Per cent votes Seats votes per seat 2 2 1 1 36.87 36.15 18.16 3.74 40,062 39,272 39,470 8,134 Sainte-Laguë Number of Seats votes per seat 2 3 1 26,708 39,272 39,470 Table 26: The cost of each seat won by the parties in terms of number of votes per seat. Using LR the cost for the seats won by parties A to C is approximately 40,000 and for party D only 8,134, whereas if Sainte-Laguë had been used the cost for all seats would have been between 26,708 and 39,470. 63