A Vote that Counts

Currently, only the votes of a few citizens decide the outcome of general and local elections. The outcome of these elections bears little relation to the choices expressed through the ballot box. Our citizens are effectively disenfranchised on a vast scale, and the power that one citizen has to affect the outcome of an election varies widely from another. Even where proportional representation is used, it delivers inconsistent and disproportionate results. 

Therefore, if we are to enshrine the right to vote in a new constitution, we should ensure that it is not merely the right to the act of voting, but the right to a vote that counts: and each citizen’s vote should count the same as each other’s.

The basic constitutional right could be:

“Every Citizen shall have the equal right to elect the members of the representative bodies.”

This means that the electoral systems used should:

1) Maximise the number of votes that count – i.e. a citizen should be able to vote for the party they want, and it contribute to the standing of that party in the representative body; and

2) Maximise the equality of the votes – i.e. every citizen’s vote should have the same power.

Under some systems, the degree to which a Citizen’s right is upheld varies widely depending on where they happen to live. The outcome differs depending on various factors, such as: where a constituency’s boundaries are drawn; whether it has more or fewer seats; whether a Citizen happens to vote for a winning candidate or not; how many parties are standing; and how many candidates each party puts forward.

To overcome these obstacles, and ensure that every Citizen’s right is upheld, the electoral system chosen should consistently and uniformly deliver 1) and 2).

The optimal result can only be achieved if the allocation of seats in the representative body is calculated using all the votes from across the area electing it – a “global allocation” of seats.

In addition to this global calculation of seats, two other important factors are the divisor method used for the calculation, and the threshold that parties need to pass in order to be included in it. On the one hand these should allow a multi-party system to develop, to help prevent one or two parties from dominating – or going unchallenged – and to encourage the development of more consensual politics; whilst on the other the should allow parties to coalesce into effective units, and not constantly splinter

The global allocation of seats should be calculated using Sainte-Laguë, as it most proportional divisor method. And the legal threshold should be no more than 2% of the nationwide vote. Thresholds higher than this can make the transition from being in an assembly and not being in it too sharp. E.g. with a 5% threshold for an assembly of 600 seats, a party could go from having 30 seats on 5.1%, to having 0 seats on 4.9%. Higher thresholds also mean more votes are wasted, and can prevent new parties from emerging and establishing themselves.

We should also choose an election system that: uses open party lists, to allow Citizens to affect which candidates are elected; and uses multi-seat electoral districts, to reflect the diversity of the local vote.

Finally, it doesn’t matter if the internal mechanism for converting votes into seats is complex, as long as the result can clearly be seen to reflect the way people voted. The closer the proportion of votes matches the proportion of seats, the more simple and transparent the result. The more the proportions diverge, the harder it is to tell if the outcome matches voters’ intentions, or to know what these are.

 

Example Electoral Systems 

The systems below would ensure every Citizen had the equal right to elect the members of their representative bodies. They are based on existing systems that meet them. Example A is based on the Danish system; and Examples B and C are based on the systems used in a number of cantons is Switzerland, including Zurich.

The “front end” of the systems (i.e. what a voter does on the ballot paper) can vary, but my recommendation would be the “Danish ballot”, where a voter can either: i) choose their preferred candidate from a party list (which counts as a vote for that party – unless it is an independent candidate); OR ii) simply choose a party, if they have no preferred candidate.

Also from Demark is the idea that parties can nominate candidates in “nomination districts”, which are subdivisions of the electoral districts. Candidates nominated in a particular nomination district appear at the top of their party’s list on the ballot papers in that district.

Example A (for national elections)

a. The country is divided into multi-seat electoral districts. 75% of seats are allocated to these districts, in proportion to their population size using Sainte-Laguë. Most should have between 7 and 13 seats.
b. Citizens cast their votes using the “Danish ballot” described above.
c. Votes are counted within each district, and seats are allocated to parties (and independent candidates) using the D’Hondt method.
d. All the party votes from across the country are added together.
e. All the seats allocated to parties in (c) are added together from across the country.
f. Any party getting at least 2% of the national vote, or that is allocated at least one seat from (c), passes the threshold, and qualifies for the next stage.
g. The qualifying parties’ votes from across the country are used to calculate what the proportional result for all 100% of the seats should be - using Sainte-Laguë.
h. The difference in seats between what a party is allocated in (c), and what is calculated in (g), is made up from the 25% of seats that are still left to allocate: these are the adjustment seats.
i. The process can stop here. It would give two types of MPs, similar to the systems used in Scotland and Germany. But the Danish system goes further: it allocates the adjustment seats down to the electoral districts – so that there is only one type of MP.
j. The final stage is the assigning of candidates to seats. A candidate’s total votes consist of the personal votes (the ones where voters have identified a preferred candidate) from across the whole electoral district; plus a proportion of the party votes (the ones where voters haven’t identified a preferred candidate) from their nomination district. The candidate with the most votes gets the first of their party’s seats; the one with the next most votes gets the next seat; etc.  

Example B (for national elections) 

a. The country is divided into multi-seat electoral districts. 90% of seats are allocated to these districts, in proportion to their population size using Sainte-Laguë. Most should have between 7 and 13 seats. Unlike system A, there can be more smaller-sized districts.
b. The electoral districts are grouped into regions, based on the standard UK nations and regions.
c. Citizens cast their votes using the “Danish ballot” described above.
d. Votes are counted within each district to calculate the first threshold. A mock allocation of seats is carried out using the D’Hondt method to see which parties (and independent candidates) would get a seat under those conditions.
e. All the party votes from across the country are added together.
f. Any party that gets at least 2% of the national vote, or at least one seat from the mock allocation stage (d), passes the threshold, and qualifies for the next stage.
g. The votes of qualifying parties are added up across each region.
h. These regional votes are used to allocate seats to each electoral district and to each party, such that each district gets the number of seats it needs, and each party gets the right proportion of seats. The method for doing this is called bi-proportional apportionment, which is used in a number of cantonal and municipal elections in Switzerland.
i. All the seats apportioned to parties in (h) are added together from across the country.
j. The qualifying parties’ votes from across the country are used to calculate what the proportional result for all 100% of the seats should be using Sainte-Laguë.
k. The difference between what a party is allocated in (h), and what is calculated in (j), is made up from the 10% of seats that are still left to allocate: these are the adjustment seats.

From here on, it is exactly the same as Example A. I.e. there can either be two types of MP, or seats can be allocated down to electoral districts. And candidates are allocated to seats in the same way. 

Example C (for local elections)

A variant of Example B can be used for local elections.

a. The council area is divided into multi-seat electoral districts. 100% of the seats are allocated to these districts, in proportion to their population size, using Sainte-Laguë. Districts can vary in size from as few as three or four seats, to as many as 12 or 13, depending on local circumstances.
b. Citizens cast their votes using the “Danish ballot” described above.
c. Votes are counted within each district to calculate the first threshold. A mock allocation of seats is carried out using D’Hondt to see which parties or independents would get a seat under those conditions.
d. All the party votes from across the council area are added together.
e. Any party that gets at least 2% of the council-wide vote, or at least one seat from the mock allocation stage (c), passes the threshold, and qualifies for the next stage.
f. The council-wide votes of the qualifying parties are used to allocate seats to each electoral district and to each party, such that each district gets the number of seats it needs, and each party gets the right proportion of seats. The method is the same bi-proportional apportionment used in Example B.
g. That’s it. That’s the final result, in terms of seat allocation, and it is completely proportional because it is based on the council-wide vote.
h. The final stage, once parties have all their seats, is the assigning of candidates to seats. Each candidate’s votes consist of all their personal votes from across the electoral district. The candidate with the most votes gets the first of their party’s seats; the one with the next most votes gets the next seat; etc. 

All these systems ensure that as many votes as possible count, and count equally, because of their global apportionment elements, the highly proportional seat-apportionment methods, and the reasonable thresholds. But they also ensure diverse local representation; and that voters can affect which candidates are elected.

 

 

Further information about bi-proportional apportionment:

http://en.wikipedia.org/wiki/Biproportional_apportionment

http://www.math.uni-augsburg.de/stochastik/pu...im/2008e-en.pdf 

 

Further information about the Danish electoral system: http://goo.gl/IwNNa9

 

edited on Apr 18, 2015 by James Grindrod

Nicholas Charalambides Apr 5, 2015

Now we are into phase two, and so we can consolidate discussion on this topic, could we keep all commentary on this idea in the Electoral Reform Society's post: https://constitutionuk.com/category/2844#/post/79624

Many Thanks!

BananaPlant Apr 10, 2015

Interesting

Share