Tuesday, 28 January 2020
Belgium’s federal elections were held on 26 May 2019. They have been followed by months of laborious negotiations with ten parties involved. In the United Kingdom, by contrast, the national elections were held on 12 December, and the government could start working straight away, with the party in power taking well over half the seats in the House of Commons.
Is this not blatant proof of the superiority of the UK’s “first-past-the-post” electoral system — more technically, plurality with single-member constituencies — over Belgium’s proportional representation, with party lists competing in multi-member constituencies?
It is certainly a good illustration of one significant advantage of a “winner takes all” system, whereby only the party with the most votes in a constituency can turn votes into seats. Compared with systems in which seats are distributed in proportion to votes, any such system is strongly biased in favour of large parties or strongly concentrated ones.
Consequently, barring a proliferation of regional parties, it facilitates the formation of governments: there are just two or three main parties and there is a good chance that one of them gets an absolute majority of the seats.
You said that this is a significant advantage. Do you regard it as decisive?
No. The most obvious downside is an unfair allocation of seats. Consider the results of last December’s election. With nearly 12% of the votes, the Liberal Democrats got less than 2% of the seats. And while the Scottish National Party, with less than 4% of the votes, got forty-eight seats, the Green Party, with nearly 3%, got just one, and the Brexit Party, with 2%, none at all.
The problem of misrepresentation is even more serious than these figures suggest because of the role played by strategic voting: if the party you like the most has no chance of having its candidate elected in your constituency, either you don’t bother to vote or you vote for the electable candidate you dislike the least.
What emerges is the following big picture. With a total population of over 66 million, the UK has a registered electorate of nearly 48 million. The Conservative Party got less than 14 million votes, including a sizable number of strategic ones of the sort I mentioned. Yet this was enough to give the party a comfortable parliamentary majority of 56%.
This is not exactly a parliament that can claim to constitute a fair reflection of the country’s diversity of voices. The ease of government formation is bought at the expense of a distorted representation that supporters of smaller or more dispersed parties can be forgiven for regarding as grossly unfair.
Does this provide a sufficient reason for rejecting “first-past-the-post” as an electoral system for assemblies?
It proved sufficient in Belgium, the first country to abandon “first-past-the-post” in favour of proportionality at the end of the 19th century. Until then, Belgian politics had been essentially bipartisan, with a catholic party and a liberal party competing for power. At some elections, however, a tiny majority of votes enabled one party to pocket an overwhelming majority of seats. Moreover, with industrialization and the extension of suffrage, the socialist party was quickly gaining ground.
If the electoral rules were not modified, it was expected that urban areas would only elect socialist candidates and rural areas only catholic candidates, with the liberal party squeezed out of the parliament, despite the wide support it enjoyed, especially among the more educated, throughout the country.
A reformist association for the adoption of proportional representation was set up in 1881 under the intellectual leadership of Victor D’Hondt, a law professor at the University of Ghent, whose name is still attached to a method for allocating seats under proportional representation used in several countries.
In December 1899, after two decennia of lobbying, arguing and even fighting, list-proportional representation was introduced in Belgium. It later spread to a majority of democracies in the world, though not to France, nor to India and the United States, which imported the British system and are still stuck with it.
Can it then be said that the choice of an electoral system faces a fundamental dilemma between, on the one hand, ensuring that a government enjoying a parliamentary majority can be quickly formed after each election and, on the other, ensuring that the parliament represents fairly the various political forces in the country?
I think this captures the core of the trade-off. But many more arguments have been used. For example, in a book published in 1861, the great British philosopher John Stuart Mill articulated a whole array of arguments against a system that forced British voters to choose “merely from the assortment of two or three perhaps rotten oranges, which may be the only choice offered to them in their local market,” instead of being able to vote for candidates whose opinions are close to theirs whether or not they live in their town.
Mill was confident that the UK would give up such a silly system and move to proportional representation — albeit closer to Ireland’s current “single transferable vote” system than to the more familiar “list-proportional” formula. He refused to believe that the British people “deserved to be stigmatized as being insurmountably prejudiced against anything that can be proved to be good either for themselves or for others”.
Yet, one and a half centuries later, despite innumerable attempts to abolish it, including through a referendum, “first-past-the-post” is still in place — and this makes a big difference. Does anyone believe that the Brexit referendum would have been triggered under proportional representation?
Whatever virtues Mill and yourself may see in proportional representation, should one not admit, when observing Belgium’s impasse, that it has a serious problem?
There may be historical circumstances in which the advantages of quick government formation weigh more, but overall I keep believing that it is better to have a greater variety of voices represented in a country’s deliberative assembly and also that it is better to have governments supported by a majority of the voters, even if that often means painful compromises that need some time to work out.
In Belgium, however, the splitting of all national political parties along the linguistic lines in the 1970s makes the achievement of compromises particularly arduous. If parties compete for separate electorates, they have incentives to make promises at the expense of the part of the country in which they can neither win nor lose votes. This is a pathological situation that needs mending as much as possible.
Moving in the direction of “first-past-the-post” would make things even worse. What would alleviate the problem is the creation – in addition to the eleven provincial constituencies – of a country-wide or “federal” constituency. This was proposed fifteen years ago by the Pavia Group of academics (www.paviagroup.be) and is analogous to the creation of a transnational constituency now proposed by many for the European elections.
Even if only 10% of the seats are allocated to this federal constituency, the dynamics of Belgium’s federal politics will be profoundly affected. If well designed, the leaders of all parties will want to stand as candidates in this constituency and thereby become accountable, not just to their own province or community but to the whole of the country they want to govern.
This will not make government formation as straightforward as under a system that tends to give an absolute majority to one party, but it will diminish an unnecessary source of delay and deadlock.