How electoral systems shape election outcomes

Valgresultat er ikke kun et enkelt uttrykk for velgernes vilje. Valgordning påvirker nemlig måten preferansene utrykkes på. I Nederland opereres det med ren proporsjonal representasjon. Det nederlandske valget våren 2017 førte til et ekstremt fragmentert parlament. Hvordan hadde resultatet sett ut dersom Nederland hadde den britiske eller franske valgordningen?

Av Simon Otjes*

Recent elections in the United Kingdom and France saw unexpected results. The new party of the new French president, Emmanuel Macron, La République en Marche, won a large majority of the seats in the French National Assembly out of nothing. The British prime-minister, Theresa May, lost her majority in the House of Commons. The election results are not just a simple expression of the “will” of the electorate. The electoral system affects the way in which these preferences are expressed.

The Netherlands also held elections in the spring of 2017 which resulted in an extremely fragmented parliament, as citizens voted for many different parties and the Dutch electoral system of pure proportional representation translated this as nearly perfectly into seats on a one-citizen-one-vote basis. But what would the result have looked like if it had adopted another electoral system, like the British or the French?

Dutch Political Parties

Going through the results allows us to introduce the Dutch political parties and compare them to the Norwegian party. The Liberal Party (VVD) of the Dutch prime-minister Mark Rutte, which is similar to the Høyre in Norway, got 21 percent of the vote and 33 of the 150 seats in the Dutch parliament (Tweede Kamer). The Eurosceptic anti-immigration Freedom Party (PVV), which is similar to the Progress Party got 13 percent of the vote and 20 seats. The Christian-Democratic Appeal (CDA), a more moderate pro-European version of the Norwegian Christian People’s Party, got 12% of the vote and 19 seats. D66, a pro-European version of Venstre, also got 12% of the vote and 19 seats. The GreenLeft (GL), a pro-European party with a green and left-wing profile similar to Miljøpartiet De Grønne and Sosialistik Venstreparti, got 9 percent of the vote and 14 seats. The Socialist Party (SP), similar to the SV, got 9 percent of the vote and 14 seats. The Labour Party (PvdA), similar to the Arbeiderpartiet, fell from 25% to 6% of the vote. The ChristianUnion, a smaller, Christian-democratic party with a left-wing orientation on immigration and the environment, got 3 percent of the vote and 5 seats. The Animals Party, a Euroskeptic green party got 3 percent of the vote and 5 seats. 50Plus, a single interest pensioners’ party similar to the Pensjonistpartiet, got 3 percent of the vote and 4 seats. The Political Reformed Party got 2 percent of the vote and 3 seats. DENK, a party formed by two former Labour Party MPs with a Turkish-Dutch backgrounds, that seeks to represents Dutch people with an immigration background (similar to the Norwegian Innvandrerpartiet), got three seats. Finally there is the Forum for Democracy, which like the Freedom Party is a Euroskeptic anti-immigrant party got 2 percent of the vote and two seats.

The British electoral system

In the British electoral system every seat is elected by a single district. The party that wins the most seats in that district gets that seat. So, I have divided the Netherlands into 150 equally-sized districts. I respected provincial borders, attempted to respect municipal borders and within the four largest parties the borders of wards. All districts are contiguous (excluding islands). Some districts are larger than one seat. For the four large cities, I took the results in wards into account, but I did not do this for smaller municipalities and two seats were assigned on basis of the results. On average every district has 113.000 inhabitants. One district is much smaller, the district of Bonaire-Sint Eustatius-Saba, the three Caribbean islands, which form part of the Netherlands. Then I took the results of the Dutch elections in 2017 and identified which party would have been the largest in each district.

8.SOtjes_Uitslag2017

Legend: VVD: blue; CDA: yellow; PVV: orange; D66: pink; SP: red; GL: green; light-blue: DENK; PvdA: purple (not in Figure 1, but in Figure 2).

The results are in

While the actual 2017 elections yielded a very fragmented parliament. The combination of the British electoral system and the voting behavior in the election would have yielded a supermajority of the Liberal Party: it gets 109 out of 150 seats. Figure 1 shows the results per district with a large swath of district colouring VVD-blue. The party is 8 percentage points larger than the second largest party and therefore it is largest in most districts. The Freedom Party would have been the second party with 15 seats. The third party is the CDA (10 seats). This means that centre-right parties win 134 out of 150 seats. The centre-left D66 wins seven seats in the Dutch-Caribbean and university cities. The GreenLeft wins four seats. The immigrant party DENK wins two in a districts with a large share of immigrants. The SP wins one seat.

The French electoral system

For the French electoral system we take the same districts as the British, but instead of assigning them to the largest party in the district, the two largest party per district have a run-off election. We simulate this second round by looking at the positions of parties. The voters of every party out of the first round are assigned to the party than runs in the second round that is closest to it. To define closeness we take the Chapel Hill Expert Survey, which asks political scientists to place political parties on, among others, a left-right dimension. Two parties are not included in the survey: DENK, which is placed between the GL and the SP, and Forum, which is placed between the PVV and the VVD.

8.SOtjes_France2

Compared to the British result, the result is less overwhelming, but still a plurality of the districts (64) colour VVD-blue. It wins when it faces the Freedom Party when it unites all parties to its left. It also wins when it faces D66, because it can unite the Freedom Party, the Christian-Democrats and the Liberals. However, when it faces the CDA, however, the Christian-democrats are able to win the seat with the support of the parties to its left. The CDA wins 43 seats, mainly in Northern provinces.  D66 gets 32 seats, mainly in university cities. The Freedom Party only gets two seats, because it is only able to win when it faces the SP. The GL wins four seats. The PvdA wins one seat. DENK wins one seat in Rotterdam, where the party faces a candidate from the Freedom Party.

Ifs and Buts

There is one obvious concern about this kind of ‘simulation’ of electoral results. It takes the behavior of voters and parties as a given. Voters are likely to change their behavior in order to influence the outcomes. So a left-wing voter under the British electoral system, may instead of voting for the party of their first preference (say the SP) and seeing their district go a centre-right candidate, vote for a more moderate centre-left candidate because she believes that this candidate has a better chance of winning. Parties may also cooperate. The two small Chrisitian-democratic parties SGP and CU may decide not to contest the same districts and ask their voters to support each other’s candidates. A CU-SGP alliance could win six seats. A GL-Labour Party alliance could win 21 seats. Cooperation between the Labour Party-GL-SP could win 82 seats.

The same kinds of patterns may occur in the French system, with voters switch to parties that are more likely to win in the run-off in the first round. Party cooperation is not necessarily positive, because in such cases for instance a Labour Party candidate may replace a D66 candidate and where the D66 candidate could beat the Liberal Party, the Labour Party cannot. Moreover different dimensions may play a role in the minds of voters, such as an EU dimension. Interestingly this would also benefit the Liberal Party as it is more centrist on the EU dimension compared to the left-right dimension.

Denne saken bygger på tidligere saker på den nederlandske bloggen ‘Stuk Rood Vlees’.

*Simon Otjes er forsker ved ‘Documentation Centre Dutch Political Parties’ ved Universitet i Groningen, Nederland.

Legg igjen en kommentar

Fyll inn i feltene under, eller klikk på et ikon for å logge inn:

WordPress.com-logo

Du kommenterer med bruk av din WordPress.com konto. Logg ut /  Endre )

Google+-bilde

Du kommenterer med bruk av din Google+ konto. Logg ut /  Endre )

Twitter-bilde

Du kommenterer med bruk av din Twitter konto. Logg ut /  Endre )

Facebookbilde

Du kommenterer med bruk av din Facebook konto. Logg ut /  Endre )

w

Kobler til %s