The health crisis has made it a fairly lean season for elections, but South Korea – which has so far been one of the most successful countries in dealing with Covid-19 – is going to the polls today to elect a new legislature, or National Assembly.
South Korea has a full separation of powers, with an executive president, directly elected for a non-renewable five-year term, and a legislature elected separately every four years. But a lot has happened since the last legislative election, in April 2016.
At the time, the centre-right’s Park Geun-hye was president. Her centre-right party lost its legislative majority, winning just 33.5% of the vote and 122 of the 300 seats. The new legislature duly impeached Park for corruption and she was removed from office in March 2017, almost a year before the expiry of her term. (She was subsequently sent to prison, as was her predecessor.)
An early presidential election then produced a comfortable victory for Moon Jae-in, from the centre-left Democratic Party, at his second attempt. He’s had a fairly exciting three years, playing an improbable double act with Donald Trump in diplomacy with North Korea – which nonetheless seems to have done a good job in averting the threat of war.
Moon’s approval ratings, which were indifferent for most of last year, have been on the improve lately, presumably due to his response to the coronavirus. But his prospects of significant achievements in his remaining two years in office (and therefore on his party’s chances in 2022) will be affected by how sympathetic a legislature he gets.
South Korea has a mixed electoral system: 253 seats are elected by first-past-the-post in single-member constituencies, and the remainder by proportional representation (on a largest-remainder system). Prior to this year, the proportional seats were simply additional, not offsetting the disproportionality of the rest.
So in 2016, centre-right and centre-left, with 58.3% of the vote between them, won more than 80% of the seats. In fact, although the centre-left was 8.7% behind, it finished with one more seat, 123 against 122. The centrist People’s Party (now called the Party for People’s Livelihoods), with 25.8%, outvoted the Democratic Party but only won 38 seats. The leftish Justice Party won 8.2% and six seats, and there are 11 independents.
President Moon, however, prevailed upon the legislature last December to approve a reform to the system, making it more like a mixed-member proportional system (as used in New Zealand). Now, 30 of the 47 proportional seats will be allocated as compensatory seats to parties that are under-represented in the constituency seats.
Unfortunately, the centre-right opposition (now called the United Future Party) devised a scheme to get around this. It has established a satellite party, the Future Korea Party, which will only contest the proportional seats while its parent only contests the constituencies: that way, it will get the full benefit of its votes in both, with no offset.
The Democratic Party was (understandably) outraged, but felt obliged to follow suit so as not to be disadvantaged. So it too now has a satellite, the Together Citizens’ Party. I don’t know if either of the major parties in New Zealand has ever thought of trying this trick, but let’s hope it doesn’t catch on.
Opinion polls show the Democratic Party with a big lead in the constituency seats, so it looks as if the president is likely to have a friendly legislature – which should be proof against suffering the same fate as his predecessor.
6 thoughts on “Election preview: South Korea (legislative)”
One way to prevent this rorting/ gaming of mixed-member proportional would be to authorise the runner-up in each district to nominate which party list that district’s winner should be deducted from. It doesn’t matter how vehemently the local favourite swears she’s an independent; if the runner-up is, say, Labor and signs a declaration that she considers the winner politically aligned with the LNP list, so be it.
You could really finesse it by legislating that if there are multiple runners-up, and they disagree over which partisan column the district winner belongs in, you add up their votes. So if, eg, Liberal (25%) thinks that Independent (30% plurality) should be counted as Labor for compensatory-seats purposes, this would prevail unless, eg, Labor (23%) and Socialist (22%) agree that no, the local MP is actually a Liberal. Note that the actual party labels should be only advisory in this determination.
However, as far as I know, no actually existing MMP system (not Germany, Hungary, NZ, Scotland, Wales, etc) actually uses this rule. Hence the potential for manipulation, deterred only by political pressure and shame. (Which does not pass my personal litmus test, which is “What – if anything – would stop Joh Bjelke-Petersen exploiting this loophole?”)
There’s a very simple solution to this. Scrap the candidate vote/party vote dichotomy. If you vote for a candidate, you vote for their party.
Germany used that for its first election or two with MMP. They switched to two separate votes because otherwise small parties had to run candidates in every one of the 248 constituencies to get any votes at all (shades of the Australian Senate 1984-2016 – run 2+ candidates, get a GVT square). This cluttered the ballot, and meant that local victors were elected with smaller pluralities on more widely-split votes.
There is perhaps a case for basing each party’s overall vote on the sum of votes for both its local candidates and its national or regional party list. This would deter, if not rule out, such tactics as “Oh no, those 50 local districts were all won by the Queensland Liberal Party, which polled 0.02% of the party votes. We’re the Queensland _National_ party, you see, and none of our candidates won at electoral level, so give us the 40 seats we’re owed…”
true, with my system there might be some districts where a party could run two candidates and get first and second place (like the Democrats in California). But this is pretty rare. And they’d need about 67% of the district votes to pull that off, which is almost what an overall quota for a seat would be anyway (eg, in the German Bundestag, 67% in one of 300 districts is near enough to one-600th overall that it wouldn’t distort overall representation that much).
I’m inclined to go with David’s suggestion as the simplest answer – it’s what the new system in Italy uses, for example. But there’s no perfect solution.