Following is a discussion of concrete examples of vote counting and how this is affected by electoral systems.
Which electoral formula is in force has an effect on how vote counting is
accomplished. What has to be counted, and where the counting is to be done is examined.
1. What has to be counted?
Following is a summary of the counting requirements for each electoral formula:
- First-past-the-post (FPTP) in single-member districts - count the votes for each candidate
- FPTP in multi-member districts (the Party Block Vote) - count the votes for each party list
- Limited vote - count the votes for each candidate
- Single Non-Transferable Vote (SNTV) - count the votes for each candidate
- Proportional representation (PR) with closed list, no panachage between list - count the votes for each list
- PR list with panachage allowed - count the votes for each candidate. If voters are allowed to cast a single list vote instead of voting for individual candidates, also count the number of votes cast for each list.
- Two-Round System in single-member districts - count the votes for each candidate. If a second ballot is necessary, count the votes again for each candidate standing at the second ballot.
- Two-Round System in multi-member districts, with closed lists - count the votes for each party list. If a second ballot is necessary, count the votes again for each party list standing at the second ballot.
- FPTP in multi-member districts (the Block Vote), with panachage allowed, there are two possible ballot arrangements:
a. Each seat in the district is numbered in a distinctive way (Seat
'A', Seat 'B', etc.), with one distinct competition for each seat, and each
elector has a vote for each seat.
b. There is no distinct competition for each seat and the voter has as many votes as there are members to be elected in the district.
In both cases, count the votes for each candidate.
Two-Round system in multi-member districts, with panachage
allowed. There are two possible ballot arrangements:
a. Each seat is numbered in a distinctive way (Seat 'A', Seat 'B',
etc.), with one distinct competition for each seat.
b. No distinct competition for each seat, voter has as many votes
as there are members to be elected.
In both cases, count the votes for each candidate. If a second ballot
is necessary, count the votes again for each candidate standing at the
second ballot.
PR list when preference votes for individual candidates are allowed.
- Scenario A (Finland, Netherlands): Electors must vote for a single
individual candidate, and this vote is also deemed to be a vote for the
party sponsoring that individual. In this case, count the votes for each
candidate.
- Scenario B (Belgium, Denmark): Electors may vote either for a party
list or for one individual candidate within a party list. In this case,
count the votes for each party list and count the votes for each candidate.
- Scenario C (Czech Republic): Electors must vote for a party list
and may, in addition, vote for one or many of the candidates within that
list. However, individual preferences cast for the candidates sponsored by
a given party will be taken into account only if 10% of the ballot papers
cast for that party are so marked. In this case, first count the votes for
each party list. Second, count for each party the number of ballot papers
which bear preference votes for candidates. Third, count individual
preference votes for each candidate.
- Scenario D (Latvia): Electors must vote for a party list and they may
vote for one of the candidates within that list. They may, in
addition, cross out the name of a candidate within that list. In this
case, first count the votes for each party list. Second, count the votes
for each candidate. Third, count the number of ballots where the name of a
candidate was crossed out.
- Alternative voting in single-member districts: Count only first
preferences for each candidate. If necessary, count second or other
subsequent preferences expressed on eliminated candidates' ballots (the
latter count must be done at the electoral district level, though Australia
provides for a preliminary unofficial counting of second and subsequent
preferences at the lowest level).
- alternative voting in multi-member districts: Each seat in the district is numbered in a distinctive way (Seat 'A', Seat 'B', etc.), with one distinct competition for each seat. Count the first preferences for each individual candidate for each seat. If necessary, count the second or other subsequent preferences expressed on eliminated candidates' ballots.
- Single Transferable Vote (STV): Count only first preferences for each candidate. If necessary, count second or other subsequent preferences expressed on eliminated candidates' ballots (the latter count must be done at the electoral district level).
- Mixed Member Proportional Systems (see Germany: The Original Mixed Member Proportional System and New Zealand: A Westminster Democracy Switches to PR): Count the votes for each party, and the votes for each candidate. If the two cannot be done simultaneously, count party votes first as these are the most decisive.
- Parallel Systems (Japan): Count votes for each candidate and votes for each party. If the two cannot be done simultaneously, start with the category of members which is the most numerous.
- Mixed Coexistence Systems (French Senate): In each district, use the relevant counting procedure, depending on which formula is used in the district.
2. Where are votes to be counted?
Determining which level will be the counting unit is an ad hoc decision made by national legislation.
The simplest option is to count the vote in individual polling stations. This is the most generally recommended alternative since the officials who perform the operation are already on the spot. There are also reasons of timing: electors and candidates are anxious to know the results, so why not count the votes right after polling is completed? There are also security and public trust reasons: counting the votes at a higher level necessitates transporting the ballot boxes to a different physical location, an operation which involves the risk of tampering or theft, and of course fosters the suspicion that this might happen.
Votes can also be counted at a higher level, be it in counting centres including numerous polling stations, at the electoral district level (United Kingdom) or even nationally. This necessitates safety precautions to overcome the security and public trust considerations mentioned above: ballot boxes must be sealed carefully so that their contents cannot be emptied or tampered with during transportation. Reliable officials must do transportation. A worthwhile precaution is to require inspectors from the various parties to sit in the vehicle transporting the boxes. Eventually, armed protection of the vehicles transporting ballot boxes and party inspectors may be deemed necessary.
While counting the votes at a higher level is often not the best alternative, there may be sound reasons for selecting this option. For example:
1. Protecting voting secrecy: It may happen that all voters in a given polling station vote for the same candidate or party, in which case their vote is known to all. Counting the votes at a higher level, where the contents of numerous polling stations are mixed together, minimizes the likelihood of such an occurrence.
2. Protecting voters: if the elected Member knows which polling stations supported him or her and which did not, there is a risk of retribution during his or her term of office against those parts of the district that voted 'the wrong way'.
3. Protecting national or social cohesion: If electoral districts include communities sharply differentiated by language, social class or otherwise, making the results available for smaller subdivisions within each district may reveal sharp cleavages within the electorate. In hot political situations, this may lead to the fingering, by losing candidates, of a particular sub-group (for example, an ethnic or linguistic minority) as the one that 'tilted the balance'. Counting the votes at a higher level helps to conceal those cleavages and, presumably, to preserve national and social cohesion.
4. Material considerations: Polling stations may be ill-equipped
physically to count the votes. For example, if they are located in open-air
spaces due to the unavailability of appropriate buildings, or are too
poorly lighted for the operation to be done in an efficient way. In such
cases, it may be better to count the votes in larger counting
centres located in public buildings (like schools or administrative offices) that are better equipped and better protected against intruders.
5. Size considerations: In very small countries, it may be deemed
easier to centralize counting in view of the small distances to be traveled.
6. Ensuring that the rules governing the rejection of ballot papers
will be uniformly and fairly applied: While counting the votes for
individual candidates or parties is a simple and straightforward operation,
which does not require high skills, deciding on the validity of each ballot
paper demands higher qualifications. Even trained judges may have
difficulty in deciding some specific cases. Further, the rules governing
that operation may be understood differently by lower officials. Presumably, making such decisions at a higher level ensures they will be made by trained officials and that the law will be applied the same way in all dubious cases.
7. Electoral systems requirements: While national electoral legislation often differs as to whether counting the votes is to be done locally or at higher levels, electoral formulas impose minimal constraints to legislators in this regard. In other words, whichever of the above-mentioned formulas is chosen, it is possible for vote counting, as distinguished from determining the winner(s), to be done on the spot, in voting stations, or to be done at higher levels if necessary.
This evaluation is dictated by empirical observation, which indicates that no electoral formula, even the most complex ones, actually precludes vote counting from being done in polling stations, nationally or somewhere in-between. For example, under the simplest formula, First Past the Post in single-member districts, vote counting is done in polling stations (Canada) or by the returning district for the whole electoral district (UK).
Under the Alternative Vote (AV), counting of first preferences was done in polling stations in the three Canadian provinces which had that system, two of them for about 30 years, while in Australia it is done in larger counting centres. Even under Single Transferable Vote (STV), the counting of first preferences was done in polling stations in the two Canadian provinces which had that system, again for about 30 years, while in Malta counting is done nationally.
AV and STV impose only one important constraint with regard to vote counting: while counting of first preferences can be done in polling stations, the transfer of second and other subsequent preferences must be done at the electoral district level. The reason is that only on the basis of the aggregate figures for the whole district can the decision actually be made as to whether any transfer will be necessary, and as to which candidates, if necessary, will be eliminated and have their second preferences counted and transferred. The crucial point to be ascertained before deciding to count second or subsequent preferences is whether or not a candidate has obtained a majority of first preferences in the electoral district, a decision that obviously can be made only on the basis of district-wide figures. If it is found that such a majority has been reached, counting stops there and second or subsequent preferences will never be counted. If the opposite is found, then it is up to the officials for the electoral district to eliminate the weakest candidates, to count second preferences on the ballots where first-preference votes have been given to such candidates, and transfer them to the remaining candidates, up until one candidate secures the required majority or quota.
This implies that the winner(s) of the election will be known later than election night under both AV and STV. This is one of the drawbacks pointed out by the opponents of these formulas. On election night, one can only speculate, on the basis of the trends of previous elections, to which candidates the second or subsequent preferences of the weakest candidates will go. To overcome that perceived weakness, Australia decided in the early 1990's that vote counting in counting centres (a level, which is between polling stations and the electoral district,) would involve the counting not only of first preferences but also of second or subsequent preferences. This allows election officials to immediately provide the media and the public with a reliable guess as to where the second and subsequent preferences of the weakest candidates will go, and who the likely winner is. However, this is only a preliminary count, to be repeated later at the electoral district level, and only the latter count has any official value.
