Hungarian second round today

Today Hungarians voted in the second round in the 114 (out of 176 total) single-seat districts in which no candidate obtained a majority in the first round on 9 April. Nearly all the 114 contests today were straight one-on-one runoffs between a candidate of the Socialist party of incumbent Prime Minister Ferenc Gyurcsany and a candidate of the main opposition party, the Alliance of Young Democrats (Fidesz) of former Prime Minister Viktor Orban.

On 9 April, in addition to the first round of voting in single-seat districts, voters also cast party-list votes. The Socialists narrowly won a plurality of these votes: 43.2% to 42% for Fidesz.

The Socialists are currently governing in coalition with the Alliance of Free Democrats, and the two parties are cooperating in these runoffs. The Free Democrats won 6.5% of the list votes, giving the incumbent parties 49.7% of the vote. The Socialists and Free Democrats need to win most of the single-seat districts at play today to remain in government. If they do, it will be the first time in five post-communist elections that the incumbent government was returned to power. (If all you care about is the result, scroll to the bottom of the post.)

The Hungarian electoral system is mixed-member, and it is sometimes erroneously placed in the mixed-member proportional (MMP) category. In fact, it is more parallel and majoritarian than it is MMP. However, it is not strictly a parallel system, either. In a parallel system–such as Japan’s–seats are allocated to parties in the nominal tier of single-seat districts (SSDs) and the tier of PR seats independently.* A party adds its proportional share of the list-tier seats to however many SSDs it has won. Under MMP–as in Germany and New Zealand–a party’s aggregate total of seats is based on its party-list votes, and it wins however many list-tier seats it needs to augment its SSDs won in order to equal its total aggregate proportional share. (This aggregate PR share can be determined regionally or nationally, depending on the system.)

If we look at the 2002 Hungarian election, it is easy to see where the assumption that it is MMP comes from: The outcome was close to proportional. The Socialists won 42.0% of the party-list votes, which amounted to 47.4% of the above-threshold list vote. Their seat total was 48.1%. The runner up Fidesz had 41.1% of the party-list votes and 46.3% of the above-threshold vote, and won 47.2% of the seats. This is a nearly proportional result. So, why is it not MMP?

Consider the 1990 election.** Same rules, other than a change in the level of the threshold required to win any list seats (4% then, 5% now). In that election, the leading party was the Hungarian Democratic Forum (MDF), with 24.7% of the party-list vote. Fully 16% of the vote was cast for parties that failed to clear the threshold, so the MDF actually had 29.4% of the effective PR vote. Yet it won 42.5% of the total 386 parliamentary seats. Not very proportional!

In the more fragmented party environment of Hungary’s first post-communist election, the MDF was grossly over-represented in the nominal tier, winning 114, or 64.8%, of the 176 single-seat districts (despite only 23.9% of the votes cast in that tier in the first round). It just so happens that the 114 seats it won is roughly equivalent to its party-list share of the vote (after discarding the below-threshold votes). So, if Hungary had MMP, how many list seats would the MDF have won in 1990? Zero. It already had its full aggregate proportional share on account of doing so well in the nominal tier.

So, we see Hungary’s system is not MMP. So it must be parallel, right? No. Were it parallel, the MDF would have won around 29.4% of the list PR seats and added these to its 114 nominal-tier seats. 29.4% of the 210 available list seats would be 62 (rounding). The MDF would have won 176 seats, or more than 45% of the total. In the actual allocation, it won only fifty list seats (23.8%). In other words, the Hungarian allocation process is neither fully compensatory, like an MMP system, nor completely parallel. It is partially compensatory.

The Hungarian electoral system is mechanically quite complex. However, in its actual workings, it is fairly straightforward. The nominal and list tier seats are allocated in parallel in the sense that the number of seats won in the nominal tier has no bearing on the number of list seats it will win. However, a party’s success in the nominal tier affects the total number of votes that will actually enter into the proportional allocation. The way this works is that any first-round votes that are cast for parties in the first round of nominal-tier district races in which the party is not the ultimate victor (in one round or two) are added to the party’s list total.

In this way, parties that are less successful in the nominal tier will actually win a share of list seats that is greater than their share of list votes. Yet they are not necessarily fully compensated–as they would be under MMP–because the number of seats already won at the nominal tier is not deducted from the aggregate PR total to determine a party’s number of list seats.

By 2002, the party system had mostly aggregated into two large blocs, one centered around the Socialists and the other around the Fidesz. Although parties other than top two are permitted to keep their candidates in the second round (as long as they reach 15% in the district), they usually withdraw in favor of their prospective coalition partner. The end result is that most of the unused votes for one major party are counterbalanced by the unused votes for the other, and thus the overall seat outcome is quite proportional. However, given the mechanics of the electoral system, proportionality is not at all guaranteed. And precisely because proportionality is not guaranteed, the electoral system encouraged the rather fragmented party system at the time of the collapse of communism to aggregate into two blocs–just as any properly functioning majoritarian electoral system will do. (And as the more typical “majoritarian” system–the British and Canadian-style plurality system–often fails to do.)

While it is a very complex system, it has produced its own version of “the best of both worlds”: representation (even quite proportional) for multiple political parties, but two clear blocs permitting stable government between elections and regular alternation in government at elections if the opposition gains sufficient votes at the expense of the incumbents.

At EuroTrib, DoDo is following the counting of the second-round results, and it looks like the Socialists and their neo-liberal partner, the Free Democrats, have won 65 of the 110 districts in which there were runoffs. Added to 34 won by the Socialists in the first round, that would be 56% of the nominal-tier. They should wind up with right around 200 seats*** once the “unused” vote adjustment referred to above is undertaken on the list vote and the national list seats are allocated. They will retain their narrow majority overall, and the result will once again be quite proportional, as after the below-threshold party-list votes are discarded, the two parties had 52.15% of the party-list vote. But once again, the result will be proportional only because the main parties work in two big blocs, and the main parties work in two big blocs only because the system is majoritarian.


*The party-list PR system in Hungary is itself two-tiered, but I will ignore that complication here. After parties’ list votes are adjusted via the procedure explained here, proportional allocation is carried out in that PR “tier” as if the country were a single national district with a magnitude of 210. This portion of the electoral system is a very typical European PR system. It is the relationship between the nominal tier and PR allocation that is atypical even for mixed-member systems–in fact, sui generis.

** As reported in Richard Rose, Neil Munro, and Tom Mackie, Elections in Central and Eastern Europe since 1990. Glasgow: University of Strathclyde Studies in Public Policy, 1998.

*** BBC is reporting:

The governing coalition has taken 209 of the 386 parliamentary seats.

That would be around 54%, on 52.2% of the above-threshold vote for the two parties–a slight majoritarian bonus, but not much. A bit more than in 2002, perhaps because the center-right bloc was less coordinated this time. In fact, the MDF leader was claiming she would join a coalition only if she were the PM (fat chance, given 5% of the party vote!). Such a stance may have resulted in less willingness of their voters to turn out in favor of Fidesz candidates in the runoff. That would be consistent with reports (see the EuroTrib link above) that turnout was down in the runoffs this time (unlike in 2002).

15 thoughts on “Hungarian second round today

  1. I defer to you as to whether Hungary’s partially-compensatory mixed-member system’s glass is half-full or half-empty.

    However, your argument that, in 1990, MDF had 29.4% of the effective PR vote yet it won 42.5% of the seats reminds me that in Scotland, with what looks like a fully-compensatory MMP model, Labour got 38.8% of the seats last time with only 29.3% of the votes. That’s because, in four of Scotland’s eight regions, the seven list MSPs weren’t enough for proportionality, but the other regions weren’t adjusted to correct this, because each region is self-contained, not linked to the others. A fluke? The same thing happened in the same four regions in the previous election.

    When does MMP become MMP-lite? We in Canada very seriously need a definition. Some people think they can add 27 or 23 compensatory seats to a 107-seat Ontario house and call it MMP.


  2. Wilf, Marty and I anticipated that question in that well-thumbed book of ours that you have! As long as the list seats are allocated in a compensatory manner, it is MMP, even if the number of available compensation seats is low. Just as Spain is PR despite not being very proportional in actual outcomes, due to many quite small-magnitude districts, Scotland’s MM system is clearly MMP (or perhaps that should be “MMp”!).

    The reason for low proportionality in (past) Hungarian elections is the absence of fully compensatory allocation, not the magnitude.


  3. “As long as the list seats are allocated in a compensatory manner, it is MMP,” but this is a frustrating definition. In an MMM system with 50% list seats, it is half-proportional. If party A wins 38% of the vote and 60 of the 100 SSD seats, it will end up with 49% of the MPs. Yet if the Jenkins proposal, which you describe with almost British understatement as “not highly proportional,” would give the opposition 15% of the total as compensatory seats, that gives Party A 60 of the 118 seats, 51%. So 50/50 MMM is more proportional than 15% MMP, or should I say MM-not-highly-P. If a rotten apple smells worse than an over-ripe gingko fruit, but I say “it’s still an apple,” will you eat it?


  4. Wilf, if you are asking me, as a political-reformer, would I like the type of “MMp” that Jenkins proposed, I would say no. If you ask me as a political scientist, is that MMP, I would say yes.

    And, as a reformer again, would I prefer Jenkins over what the USA and UK have today? YES!


  5. Just read an interesting paper at the Midwest conference by Susumu Shikano that used agent-based modeling to look at the “contamination” idea in MM systems. The punchlines were many, but one interesting observation was that Germany (but not the East), Japan, and New Zealand were all moving fast toward ENP(s) of 2, primarily because the two largest parties finished 1-2 in nearly every SMD. He compared that to the UK (and could have used Canada or India), where “domination” by the two largest parties is much lower and falling. It seems that the PR tier provides a pressure-relief valve for voters who prefer smaller parties, and clears the district field for big-party domination. In pure SMD, the only route to seats for small parties (and, hence the only basis on which to form viable small parties) is through regional strength and district wins.

    The next question is whether the open field for SMD domination by the two biggest parties allows those parties to be less moderate than they might have to be otherwise, or if it in fact encourages them to obsess about the “median voter” that much more.


  6. As a reformer would I prefer Jenkins over what Ontario and Canada have today? Yes, and I would even prefer Japan’s MMM over FPTP. I just wouldn’t call either one of them MMP, because I want proportional representation, and neither MMM nor MMP-lite are proportional enough to qualify as proportional representation.


  7. neither MMM nor MMP-lite are proportional enough to qualify as proportional representation.

    MMM, a la Japan, is manifestly not proportional representation, by either the institutional definition or the outcome.

    “MMP-lite,” i.e. the Jenkins Commission UK proposal, would not be proportional representation in the literal sense of representing parties proportionally. It would be, however, MMP in the institutional-design sense of allocating the (very small number of) PR seats in compensatory fashion rather than in parallel. They key is in that small number of PR seats, relative to the total of FPTP seats, in any given allocation region.

    Well, it is a dead proposal, anyway. And, from Wilf’s post, it seems that the most Jenkins-like proposal to emerge since Jenkins itself–that of Quebec’s government–may be on life support. Good!


  8. Well, we’re very worried that Jenkins will rise again in Ontario.

    Quebec’s draft bill at least has 40% list seats. It’s MMP, all right, but with the world’s highest threshold. Paul Cliche, Quebec’s leading electoral reformer, was so outraged in December 2004 that he declared the Minister had “transformed the mixed system he had announced into a majoritarian model equipped with a timid compensation. His preliminary draft is so watered down that he finds himself inventing a new voting system that exists nowhere else in the world. . . The German province of Schleswig-Holstein tried to impose a threshold of 7.5% but a court declared it unconstitutional. Faced with this fact, Bavaria gave up a plan to impose a threshold of 10% at the regional level.” In his next blast he called it a new semi-proportional system “unique in the world, which we could describe as majoritarian compensatory.”

    But in Ontario there will be a real temptation to keep the present 107 Single Seat Districts and add 23 or 27 compensatory seats, right down in Jenkins territory: MMP-lite, not real MMP at all.


  9. What would happen if the US came up with a mixed system? We the American voters deserve better than our lousy plurality system. Blah! That’s why I think we should come up with a different type of mixed system. The USA has used and still uses, although in the local level, STV and AV. But I think it would be a mistake if we only had these 2 to make a mixed system. There would be a need to use list PR, hopefully open list or flexible list. Could we probably come up with such a great system? Not saying it’s perfect but at least the results would be more proportional. The MMP could work in the USA but if you’re not careful you could mix the worst of both worlds and have everybody unhappy. Parallel is semi- proportional, since the wasted votes in the district level won’t be compensated by the list seats. But I like the idea of creating a unique mixed system. Think about this: say 50% of the members elected in SMDs using AV, 30% of the members elected in MMDs using STV and 20% of the members elected in regional and national lists (although the nº of national seats would have to be small) and the wasted votes in the district level are compensated. Give it some thinking!!


  10. Pingback: Fruits and Votes

  11. Pingback: Fruits and Votes

  12. Pingback: Fruits and Votes

  13. Pingback: Hungary’s new electoral system | Fruits and Votes

  14. Pingback: Hungary’s new constitution | Fruits and Votes

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.