What is the Additional Member System?

The Additional Member System (AMS) and Mixed-Member PR (MMP) are voting systems that combine First Past the Post (FPTP) and Party List PR to produce a proportional result.

First used to elect MPs outside Copenhagen in the 1918 Danish election and popularised by its adoption by Germany after the Second World War, AMS is today used to elect the parliaments of Germany and New Zealand, as well as for elections to the Scottish Parliament, Welsh Senedd and London Assembly here in the UK.

How does it work?

AMS systems are characterised by elections to two overlapping categories of seats. The first or ‘constituency’ level consists of a set number of single-member constituencies, usually around half to two-thirds of all seats, elected by First Past the Post. The second proportional or ‘top-up’ level sees the remaining seats elected by List PR on a regional or national basis.

Voters have two votes: one for a candidate in the FPTP constituency, one for a party list.

Seats in the FPTP constituencies are appointed as per a standard FPTP election: the candidate with the most votes wins.

PR seats are then allocated to compensate for the disproportionality of the FPTP results and create a result that is proportional when taking the two types of seats together.

Let’s say you had an AMS region with five FPTP seats and five PR seats, and three parties won 50%, 30% and 20% of the vote. Under a fully proportional allocation of the region, the parties would get five, three and two seats respectively. Accordingly, under AMS, if Party A wins four FPTP seats and Party B one, the compensatory List PR seats go one to Party A, two to Party B and two to Party C.

However, what if the vote in the same region were to split 40%, 36%, 24%, with Party A winning all five FPTP seats? In a ‘straight PR’ system, the parties would be owed four, four and two seats respectively, but with Party A already holding five seats courtesy of the FPTP element of AMS, such an outcome cannot be achieved. Under the versions of AMS practised in the UK, this would lead to Party B being denied a seat relative to a fully proportional allocation. But, under the versions used in Germany and New Zealand, this ‘overhang’ would be dealt with by adding an extra PR seat to ensure that no party was unfairly denied representation. This is sometimes where a distinction is drawn between AMS and MMP, but otherwise they are fundamentally the same.

Why AMS? / What Supporters Say

AMS’s main attraction, particularly for countries with a history of FPTP elections, is its ‘best of both worlds’ approach – retaining the familiar constituency link through its continued use of small single-member constituencies, while still being able to produce proportional outcomes on top of that.

Having two votes also gives voters the opportunity to ‘reward’ a strong local representative, while still being able, should they wish, to vote for a different party whose values align more closely with theirs. This fail safe also guarantees that voters who do not back the winning candidate in the FPTP constituency can still be confident of representation by someone they actually voted for.

And this isn’t the only way that AMS contains a better version of FPTP than FPTP by itself. At present, FPTP constituencies have to be designed so they contain a strict quota of voters, leading to many communities being arbitrarily split to create an illusion of fairness. But the increased proportionality of AMS will enable greater flexibility, allowing constituencies to better represent ‘natural communities’.

Variations

AMS is highly adaptable: it retains the flexibility of List PR systems, including whether to allocate PR seats according to the vote share in a specific region (as in Scotland and Wales) or in the whole nation (as in New Zealand); whether to adopt open or closed lists (most AMS elections used closed lists); whether – and if so where – to set a threshold (the German, London and New Zealand versions require a party to win 5% of the overall vote before they can win seats).

But AMS also introduces its own extra points of adjustability, most notably in the proportion of seats that are elected by PR, and the size of the regions in which they are allocated. Higher proportions of non-PR seats and smaller regions both independently produce less proportional results, benefitting larger – or in some cases more regionally popular – parties. This is why Welsh Senedd elections, where two-thirds of seats are FPTP seats and regions contain an average of 12 members, are less proportional than Scottish Parliament elections, where only 57% of seats are elected by FPTP and the average region elects 16 MSPs.

With thanks to Dylan Difford for compiling this explainer. Dylan is a researcher specialising in elections and voting systems, as well as formerly a regular contributor to the Electoral Reform Society blog.