Elections – Upgrade Holyrood

By Richard Wood

The Baltic states – Estonia, Latvia and Lithuania – all gained their independence from the Soviet Union in the early 1990s after the fall of the Berlin Wall. The famous Baltic Chain of Freedom of 1989 was vital in ending Soviet control of the region, paving the way for three independent republics.

There are striking similarities between the three Baltic democracies and some crucial differences too. All are republics, all use list systems to elect some or all of their MPs who have four-year terms, and all have written, codified constitutions.

Having recently visited Latvia and Lithuania (previously visiting Estonia in 2019), I’ve taken the time to highlight the electoral systems of each of the Baltic state and comparing them to Holyrood and Westminster.

Estonia

Estonia has a population of 1.3 million and is most northern of the three Baltic states, bordering Latvia to the south, as well as Russia to the east.

Since gaining independence in 1990, the Estonian Parliament – the Riigikogu – has had eight elections. Members are elected via a Closed List Proportional Representation system with multiple constituencies and no levelling seats, a model similar to what was used to elect UK Members of the European system. It also resembles the expected new system for the Welsh Parliament, as proposed by Welsh Labour and Plaid Cymru.

A total of 101 members are elected to the Riigikogu across 12 electoral districts ranging from five to fifteen members. This gives the country highly proportional elections while also retaining a reasonable degree of local representation. The three largest districts by representation cover Tallinn (the country’s capital and home of the national parliament) and the surrounding area.

It’s worth highlighting here that unlike at Holyrood and Westminster, when an MP joins the Estonian government they leave they stop being an MP. This gives a more formal separation of the legislature and executive than what we’re used to in the UK. The idea is an intriguing one although it only works due to the country’s list system: when an MP joins the government, the next person on their party’s list replaces them as an MP. This of course has the consequence that when a minister returns to the legislature, their substitute MP vacates their seat.

The country last voted in 2019, following which a coalition was formed led by the centre-right Centre Party. However, the government collapsed less than two years later; liberal Reform Party, led by Kaja Kallas, subsequently formed a cabinet with the Centre Party as the junior coalition partner. That arrangement didn’t last long either, but Kallas remained prime minister, after forming a coalition with the conservative Isamaa and the Social Democratic Party. The next election is expected to take place in March 2023.

As an independent, democratic republic, the country also has a non-executive president as Head of state. Unlike some other European countries with a similar set-up, Estonia’s president is elected by members of its national parliament rather than by a nation-wide vote. Its Baltic neighbours have also have non-executive presidents, but more on that below.

Latvia

Latvia has a population just shy of 2 million, and like its less populous northern neighbour, the country uses a closed list proportional system to elect its members. The Saeima, the country’s parliament, has 100 members, one short of Estonia’s Rijikogu, and is based in the capital city of Riga.

Proportionality is a key principle in Latvian democracy. So much so that Articles 6 and 7 of the country’s constitution enshrine it in law:

6. The Saeima shall be elected in general, equal and direct elections, and by secret ballot based on proportional representation.

7. In the division of Latvia into separate electoral districts, provision for the number of members of the Saeima to be elected from each district shall be proportional to the number of electors in each district.
Latvian Constitution (Article 6)

While Latvia uses a Closed List PR system like Estonia, the system is rather different in design. The Saeima’s 100 members are spread across just five electoral districts. The capital Riga elects 36 members – that’s over a third of representatives. The other four districts elect 26, 13, 13, 12 members respectively.

This voting system leads to highly proportional elections, however, local representation is less than in Estonia, especially in the super-constituency of Riga.

The next Latvian election is scheduled to take place on 1 October 2022. The previous election saw three previously unrepresented parties gain representation in the Saeima. The pre-election governing coalition led by the centre-right, agrarian Union of Greens and Farmers since 2016 lost significant support, leading to a new government. After months of negotiations, a five-party coalition was formed with the centre-right New Unity’s Krišjānis Kariņš becoming prime minister in early 2019. New Unity is the smallest party in parliament, making Kariņš very much a compromise prime minister for the diverse coalition he heads.

Two final things. Like in Estonia, Latvian MPs formally exit the legislature when they join the government, making way for substitute MPs. And finally, Latvia’s non-executive president is elected by members of its national parliament just like in Estonia.

Lithuania

In line with it having the largest population of the three Baltic states (2.8 million people), Lithuania’s parliament (called the Seimas – not to be confused with Latvia’s Saeima) has the largest number of members of the three Baltic states. The country votes for 141 deputies every four years. However, unlike its northern neighbours, Lithuania doesn’t use what can be characterised as a proportional system.

The Seimas’ has 141 members (although the chamber has 142 for the sake of symmetrical design!) are elected in two ways. Just over half of MPs, 71, are elected from single-member districts. Candidates are only elected if they win a majority of votes in their constituency; if no candidate does then a run-off is held two weeks later to determine the elected representative.

The remaining 70 members are elected via proportional lists with a 5% threshold. There is also a degree of openness to the list element where voters can express a preference for a candidate of their choosing.

Unlike similar systems used in the likes of Scotland, Germany and New Zealand there is no link between single-member seats and the proportional element. This results in a Mixed-Member Majoritarian system where overall results are only somewhat proportional. This is also known as parallel voting.

This limited proportionality is best shown via the Gallagher Index, which provides a standardised measure of proportionality to compare different systems at different elections. The closer to 1 an election is, the more proportional it is. The 2019 Estonian and 2018 Latvian elections yielded Gallagher scores of 5.28 and 5.51 respectively. Lithuania’s score was 9.49 overall. Lithuanian elections therefore have proportional elements but are far from fully proportional.

The 2020 election led to the formation of a government headed by popular independent MP Ingrida Šimonytė. This is an interesting situation, which follows the 2019 non-executive presidential election (which unlike Estonia and Latvia was a nationwide election) which Šimonytė lost. However, she remained a popular figure, winning the most preference votes in the 2020 parliamentary election, and has a strong relationship with the centre-right Homeland Union. She now leads a centre-right cabinet consisting of independents, the Homeland Union, Liberal Movement and the Freedom Party.

Lessons for Holyrood and Westminster

When it comes to comparing the electoral systems of Estonia, Latvia and Lithuania with Westminster, the Baltic states come out on top. Even the semi-proportional Lithuanian system is significantly more representative than First Past the Post.

Of the three Baltic states, Estonia has the best balance of proportionality and local representation whereas Latvia’s mega-constituencies reduce local links and Lithuania’s mixed-member majoritarian system limits proportionality. Of course, all three are significantly more representative than First Past the Post, which distorts the link between seats and votes, leads to countless wasted votes and encourages tactical voting.

Were Westminster to adopt any of the three systems from the Baltics, representation would improve significantly.

The situation is somewhat different when making comparisons with the Scottish Parliament, which has used the proportional Additional Member System (AMS) since the advent of devolution in 1999. Unlike the mixed-member majoritarian system of Lithuania, Holyrood’s AMS is mixed-member proportional (MMP) as the constituency election results directly impact the distribution of list seats to compensate for lack of proportionality. The system has some significant flaws – which I’ve written about extensively here – but overall it delivers broadly proportional outcomes. Take a look at the most recent Scottish and Lithuanian elections. The Gallagher score for Scotland was 7.03, making it more proportional than Lithuania’s 9.49.

However, when comparing Scotland’s Gallagher scores to Estonia and Latvia, it is clear that the northern Baltic states have more proportional elections. This highlights the flaw that Holyrood elections are only broadly proportional (due to the balance of MSP types and the fact that votes are only regionally representative).

That said Scotland’s AMS means better local representation that Estonian and Latvia (with all constituents eight MSPs – one constituency MSP and seven regional MSPs). I should say here that I am of course in no position to call for reform in other countries, but the purpose of this to to show comparisons between Holyrood and Westminster with electoral systems around the world.

First Past the Post is significantly flawed. The UK needs to adopt a proportional alternative. Scotland’s AMS is a significant improvement but is in need of reform. When looking to make democratic improvements it is vital to look outward and see what other countries offer. That is a key part of the Upgrade Holyrood mission.

By Richard Wood

Voters in Portugal went to the polls on Sunday 30 January to elect a new Assembly of the Republic, the country’s unicameral parliament. Unlike the UK, MPs are elected via a form of Proportional Representation, meaning that how people vote at the ballot box is fairly reflected in the legislature – in short, seats won broadly match votes cast.

Portugal’s 2022 election – how proportional was it?

The election took place as a consequence of opposition parties voting down the minority Socialist Party government’s budget. In the end, the Socialst Party benefitted from the election, increasing its vote share and winning an unprecedented majority of seats.

Like most European democracies, Portugal uses a form of Proportional Representation to elect MPs. This means that the proportion votes cast per party are reflected by the proportion of seats won by each party in the legislature.

The contrast with UK national elections is stark. Under First Past the Post, MPs elected to Westminster are not reflective of how people vote. In 2019, the Conservatives won 43% of the vote which resulted in them getting over 50% of all seats. Parties like the Liberal Democrats and Greens went significantly underrepresented. The worst modern example of the unrepresentative nature of FPTP was the 2005 election where Tony Blair’s Labour achieved a majority of seats on just 35% of the vote.

Overall, Portugal’s elections are fairly proportional, as seen time and time again at each Portugese election. There is a link between seats and votes unlike in the UK where that link is distorted by single-member constituencies.

However, the 2022 election has exposed a flaw with the particular PR system used by Portugal. The Socialist Party won a majority of seats (117 out of 230) on a minority of votes (41.7%).

Meanwhile the centre-right Social Democrats won 76 seats (33%) on 29.3% of the vote, while the far-right Chega secured 12 seats (5.2%) on 7.2% of the vote. The Liberal Initiative won 8 seats (3.5%) on 5% of the vote, the Communist Party won 6 seats (2.6%) on 4.4% of the vote and the Left Bloc won 5 seats (2.2%) on 4.5% of the vote. People Animals Nature and Livre each won the two remaining seats.

Socialists win unexpected majority in Portugal election https://t.co/Q4mpiyJIgw
— BBC News (UK) (@BBCNews) January 31, 2022

The results are only broadly proportional as the largest party won far more than it should have under a totally fair system. In theory this shouldn’t happen under a PR system but the mechanics of Portugal’s system helped lead to this surprise outcome. It is also worth noting than the country’s smaller parties received less representation than their share of the vote suggests they would be entitled to.

The reasons for Portugal’s disproportionality are explained below.

What type of Proportional Representation does Portugal use? Could it be better?

The country elects 230 MP for provisionally four-year terms (although in practice this is often less than that due to snap elections). The country is split into 20 multi-member constituencies each electing between two and 48 members (the largest being the capital Lisbon), as well as two overseas constituencies (Europe and the rest of the world). This in principle, means there is a strong link between votes cast and seats won, however, in practice that is not always the case.

Crucially, there is no mechanism to ensure national proportionality which in part explains the surprise majority of seats won by the Socialists. In contrast, Scandinavian party list systems have mechanisms to ensure national proportionality rather than just proportionality per multi-member constituency. Furthermore, Portugese seat distributions are determined by the D’hondt system (common across Europe), which on the whole gives a slight advantage to larger parties at the expense of smaller ones. This too can explain how the 2022 Portugese election came to be.

For those interested in how D’Hondt works and how it compares to different methods, the Electoral Reform Society has an excellent summary which can be read here.

But it’s not all about proportionality – how powerful are the voters?

Members are elected on closed party lists meaning that voters vote for parties and have no say in the ordering of lists presented. This means that while voters have a high chance of getting their party into parliament (due to multi-member constituencies), they have no say over who gets elected from their chosen party. Again, this is in contrast with the likes of Denmark and Iceland where voters have a say over the order of candidates elected to the Assembly.

Interestingly, these closed lists and the D’Hondt element make Portugal’s system one that most closely resembles how the UK used to determine members of the European Parliament (although the number of MPs elected per constituency is wildly different).

So how good is Portugal’s voting system overall? When it comes to proportionality, there is a strong link between how people vote and how they are represented in parliament. There are arguably better ways to calculate seat distributions than D’Hondt and the lack of a mechanism to ensure overall proportionality weakens the link between seats and votes on a Portgual-wide scale. This weakening in the link was dramatically exposed in the 2022 election. The system clearly needs reform but overall the system means results are broadly proportional.

In addition, another flaw is the fact that party lists are closed. Unlike in countries such as Norway, voters in Portugal have no say in the ranking of candidates on party lists. Proportionality is extremely important but so to is voter choice. If a country opts for list PR rather than say ranked choice voting in multi-member seats (PR-STV) the lists should be open to empower voters and to avoid giving too much power to party bosses.

What lessons can the UK learn from Portugal?

The first lesson that UK democracy can learn from Portugal is that the principal of Proportional Representation ensures broadly representation outcomes.

However, the second lesson is that not all PR systems are as effective as each other when it comes to ensuring proportional outcomes. Yes, in Portugal there is a correlation of seats and votes but it is far from as accurate as it could be. The UK should take that into account when designing a proportional electoral system.

Lastly, the lack of candidate choice in Portugal’s closed list party system is a significant impediment to voter power. Any future PR system adopted at Westminster should empower voters when it comes to selecting candidates not just parties.

Portugal’s voting system is not the one for the UK, but we should learn from its benefits and its flaws when it comes to designing a representative system at Westminster.

Tag: Elections

How do elections work in the Baltics? Lessons for Holyrood and Westminster