D'Hondt Method - Allocation

Allocation

The total votes cast for each party in the electoral district is divided, first by 1, then by 2, then 3, right up to the total number of seats to be allocated for the district/constituency. Say there are parties and seats. Then create a grid of numbers, with rows and columns, where the entry in the -th row and -th column is the number of votes won by the party divided by . For the purposes of this explanation, call the highest entries in the entire grid 'special entries'; each party is given as many seats as there are special entries in its row. The entries are sometimes called 'distribution figures'.

Another, equivalent way to describe the same is that the distribution figure of each candidate is the votes for his party divided by his rank within his party; the candidates with highest distribution figures are elected.

Example: if 8 seats are to be allocated, divide each party's total votes by 1, then by 2, 3, 4, 5, 6, 7, and 8. An example is given in the grid below. The 8 highest distribution figures are highlighted in bold, ranging from 100,000 down to 25,000. For each distribution figure in bold, the corresponding party gets a seat.

Government formation: This is repeated for each district/constituency in the country. Often the party with the highest number of seats will get to select the chief negotiators to form a government and is awarded the Prime Minister's seat. Its seats may be majority, minority, or a coalition, according to whether voters have spread their votes across few or many parties.

Proportionality: D'Hondt does not produce absolute/pure proportionality: in relation to their total vote, the two larger parties are slightly advantaged. Party C gets what it deserves, and Party D received too few votes for a seat. To dispute Party A's fourth seat (*25,000) Party D would have needed a minimum of 25,000 votes. So the seat allocation across these 4 parties is a fair, if not absolutely proportional, reflection of their vote.

Party system produced: The method tends to allow for 3 nationwide parties to be present in parliament, or sometimes four if voters spread their vote more evenly across the four leading parties. However, regionally-based parties with concentrated support in certain districts may well get enough votes there to win a handful of seats in parliament. In other words the D'Hondt method also allows for regional-nationalist parties to be represented.

District magnitude effects: With the D'Hondt seat allocation method, district magnitude or constituency size has quite an effect. If our model constituency, above, had only 5 seats, they would all go to Parties A (*100,000, *50,000, *33,333) and B (*80,000, *40,000) and Party C would not get the (*30,000) seat, being short of 3,334 votes to take it from Party A. And in the opposite case, if the district contained just one more seat, then Party D would have a chance of getting it (contending with A and B for the 20,000 distribution number, the next highest in the table). So the larger the number of seats in a district/constituency, the more likely that a fourth nationwide party will get a seat - as long as voters spread their votes.

From party seat to winning candidate: How the parties' allocate the seats they have won to their candidates: in 'Closed List' PR systems, parties tend to present a list of candidates for all available seats (as if they could win them all), ranked by order of priority. Then the top candidates receive the seats won in order of their rank. In an 'Open List' system, in addition to the party's order, voters can add their own ranking of candidates to show where they disagree with the party's ranking. Or the party may have listed their candidates without any preference, leaving it to the voters to rank them, by number of votes or other method.