Hungary’s new electoral system

As long expected, Hungary is about to get a new electoral system, which will include a sharply reduced assembly size. The current system, in use since 1990, has been a mixed-member majoritarian (MMM) system, with partial compensation. Not enough compensation to make it MMP–seats are allocated in parallel, but with unused votes from the nominal tier augmenting the party list votes prior to list-tier allocation. In fact, the evolution of party behavior showed just how much the fundamentally majoritarian nature of the system shaped incentives.

The new system will remain MMM, but perhaps even more so. The following details come from Alan Renwick (of the University of Reading, and one of the leading experts on electoral-system change). Alan has generously agreed to be referenced here.

The assembly size will drop from 386 to 199.* The nominal tier will continue to consist of single-seat districts. However, the formula for these districts will now be plurality, rather than two-round majority-plurality. This tier will amount to 106 seats, or 53.3% of the total.**

The list tier will be one nationwide district, unlike the current system which actually has two tiers of list allocation (regional and national). These 93 seats will continue to be elected from closed lists. As Alan explains:

For the distribution of the list seats, both the list votes and the remainder votes from SMDs are added together, where remainder votes are votes for losing candidates and votes for winning candidates after subtraction of the second-placed candidate’s votes. Seats are then distributed by d’Hondt, subject to a 5% threshold. There are also special provisions for minorities.

For nominations in the single-seat districts, the number of signatures needed is doubled from 750 to 1500,*** even though the increase in the size of these districts is much less than this.

Alan also offers two links, in Hungarian.

This change of the electoral system is in the context of a new constitution, which takes effect on the 1st of January. I previously branded this constitutional reform process as de-facto authoritarian. Alan also recommends a blog entry by Kim Lane Scheppele, who notes that the new constitution also contains a major attack on the independence of the judiciary, the Election Commission, and the electoral district boundary delimitation process.

Thanks, Alan, for this information!

As for Hungary, these are all very troubling developments. In fact, the electoral system change may be the least troubling of this set of changes.

* As the first-linked F&V item above discussed, Hungary has had a significantly “oversized” parliament. The Cube Root Rule would suggest an assembly size of around 200 to 215, or right about where they have now settled.

** Previously the nominal tier was 176 seats out of the total 386, or 45.6%.

*** In the final version of the law, the number was set at 1000, as noted in Alan Renwick’s overview.

11 thoughts on “Hungary’s new electoral system

  1. Maybe the change of electoral system pales before the other constitutional changes in Hungary, but the old electoral system was made this legal coup possible in the first place. It should be a strong indictment against such systems, so why, oh why are states like Egypt embracing them.

  2. That is correct, Harald, and was the main theme of one of my earlier entries that I linked to above.

    For some reason, some experts thought Hungary was MMP. But the super-majority of seats won at the last election on a bare-majority of votes vindicated my coding of the system as MMM (see Shugart and Wattenberg, 2001, for example).

  3. I am so confused by the Hungarian mixed member partially proportionate majoritarian system. It is so complicated. How can surplus votes of loosing candidates be transferred toward the party list votes? Wouldn’t that make the loosing candidate party win more seats?

    I also thought the list-tier at the county level and national level both use different mathematical allocations.

    • The Hungarian system is indeed confusing! Under the current system, once the winners of the single-seat districts are determined (which may not be until after a second round), all votes cast in the first round for the candidates of parties other than the one that won the district are added to the party-list votes in the national pool. Then the PR seats are allocated without regard for how many seats the various parties won in the single-seat districts. This is what makes the allocation of SEATS parallel, but because it is done after the transfer of losing candidates’ VOTES, the result is partially compensatory.

      So the largest party would still have more list votes for allocation in the list tier, assuming that the largest party, as measured by list votes, is also the largest party by nominal-tier (candidate) votes. However, the parties that trailed in the party-list votes will indeed close the gap considerably after the vote transfer from the nominal tier. Of course, the party with the most party-list votes also gets vote transfers, from any districts it did not win.

      I suppose, if you had a close election in party list votes, but the largest party won an overwhelming share of the nominal-tier seats, you could end up with the second largest party winning most of the list seats. But because allocation between the two tiers is in parallel, the party that won the most (list) votes and (nominal) seats would still be far ahead in overall seats. That is, it would win a large (even if not the largest) share of list seats PLUS all those nominal-tier seats. The second party, on the other hand, would have more list seats, but few nominal seats, and still be well behind overall.

      So the second largest party (and other trailing parties) would be better off than if it were “standard” MMM, where no votes are transferred from one tier to the other, but far worse off than if it were MMP, where the list seats are compensatory rather than parallel.

      I hope that did not make it more confusing!

      As for the new system, it should be even more majoritarian, because the party that wins a district also gets some votes added to its list-tier votes. That is not the case currently.

      • The 2006 election is a good demonstration. Taking just the top two parties, it was close in list votes: MSzP 43.2% to Fidesz-MPS 42%.

        Despite how close the two parties were in list votes, the MSzP dominated the nominal-tier races, winning 103 of 176 (58.5%), compared to 69 for Fidesz-MPS.

        In the list seats, Fidesz-MPS won a total of 95, to 89 for MSzP. The Fidesz-MPS pulled ahead in votes with which the list seats were allocated because of those 103 districts where it was the second largest party (and in many, surely a strong second). However, when the MSzP added its 89 list seats to its 103 nominal seats, it still was overrepresented, with 192 of 386, or 49.7% (on 43.2% of the list votes).

        Under MMP, the MSzP would have had only 60-some seats from the list tier. On the other hand, under MMM without vote transfers, it could have expected 91 or more (43% of the total 210 available list seats). You will note that its likely result under standard MMM (i.e. without vote transfers) is much closer to what it actually won than is its likely MMP result. This justifies my classification of Hungary’s (current) system as MMM, albeit with partial compensation, rather than as MMP.

  4. It’s like a Rubik’s cube. I think open party list would be easier.

    On the nominal tier, it is just single member districts, and on the list tier, it is actually two tiers, county, and nationwide. You said that all the list votes are allocated as if it is in one district.

    I meant that I remember reading somewhere that the county tier used the Hare method of calculation, and the national tier used D’hondt to compensate the parties.

    It is good to know that in the nominal tier, that the losing candidate votes is added to the list total, but is it added to the county or regional total or that does not make a difference?

    Perhaps it is smart that they are eliminating the second round as it is not needed under an MMM or an MMP system.

    The allocation of the list seats to the single member districts isn’t calculated until after the second round, right?

    What would the 2010 election result been if it was re-run under a true MMP system?

    The Socialists should have tried to change the electoral system when they ruled for the last eight years, but then they needed a 2/3rds majority to do so.

    Blue Sky exercise; What if NZ has embraced the Hungarian style MMM with partial proportionality? So all the losers in the FPTP nominal tier votes would have transferred to the party list vote. It almost seems as if such a system makes the loser the winner. I wonder what the recent NZ election would have look like under that allocation.

    It almost seems as the Hungarian MMM with with partial proportionality is a system where both votes are equally important especially, when the votes of the losing candidate are transferred and added to the list vote pile, and to add on top of that there are two rounds of the nominal tier if a candidate does not win an outright majority and/or the voting turnout is very low, whereas in most MMP systems, only the party vote is most important, and there is only one round of the nominal tier.

    If the votes of the losing candidate transfered and added to the list votes eliminated, Would that make MMM with partial proportionality more proportionate?

    I wonder does the average Hungarian voter even understand this, and I wonder if they would strategically vote to take advantage of such a system.

    Hungary would have been far better off with a one vote MMM with partial proportionality when it first introduced this system, then it would have made sense for the votes of the losing candidates to brought over as compensatory votes. Would a one vote MMM with partial proportionality have been more proportionate?

    Is the Hungarian MMM with partial proportionality like the Mexican system of MMM parallel, I read elsewhere on your blog that it too is not an MMP system? The Mexican system is a one vote system to boot as well.

    I don’t I would recommend a country to use the current Hungarian MMM with partial proportionality, it is too complicated. Is there a way to keep the system, but to make it more into an MMP system?

    I will quote you below;

    “Just to be clear: as noted in the text above, Mexico’s system is MMM. Allocation is parallel. Exception: no party can have more than 60% of the seats or more than 8-percentage points of over-representation.

    The same system has now been used in 1997, 2000, 2003, and 2006. Only in the first of those elections was either cap triggered (thereby keeping the PRI short of a majority). Otherwise, it has been a pure parallel (MMM not MMP) system, and the results have shown it (as detailed above for the 2006 election).

    The list tier allocation to parties is nationwide; the regional districts matter only in terms of which lists a given party’s seats come from (i.e. the list tier is effectively districted only on the intraparty dimension).”

  5. Alan Renwick looks at the political implications of the new, “more majoritarian” Hungarian system. As Alan notes, there was no objective need for a more majoritarian system, as the current (MMM) system has permitted stable, full-term governments and electoral alternation in power–precisely the supposed advantages of majoritarian electoral systems.

    One fact that I did not previously know about the current system is that there has been no re-districting since 1990. The new law, Alan notes, provides for a new district map that is even more favorable to the ruling Fidesz than the old one, despite the reduction in malapportionment.

    • Given the popular vote share of Fidesz in Sunday’s election, just under 50%, one can calculate a plausible upper limit on the party’s seat total as follows. Suppose they could win all 106 of the district seats, thanks to fragmentation of opposition plus a soupcon of gerrymandering. Suppose further that the average winning margin of their district candidates was 20%. Then the addition of “wasted” votes for district candidates to the votes for party lists would entitle Fidesz to 3/8 of the 93 proportional seats, for a total of 141 seats. (I assume also no net gain/loss from split voting that involves Fidesz.)

      With Fiesz expected to win 133 seats, the system has worked for them pretty much as designed: 66.8% of seats from just under 50% of party votes.

  6. Hungary’s new electoral system seems to have a more than passing resemblance with the mixed system used to elect the Italian Chamber of Deputies between 1994 and 2001. The major differences between the two systems appear to be: 1) the higher proportion of single-member constituency seats in Italy (three-quarters as opposed to just over half in Hungary); 2) the criteria for allocating list seats (in a nutshell, the Italian Chamber formula didn’t take into account votes cast for single-member constituency candidates, although it did feature the subtraction of votes for winning candidates from the party’s vote total in the same terms as Hungary’s new system); and 3) the use of the largest remainder method for allocating Chamber list seats in Italy (however, d’Hondt was used for the proportional component in elections to the Senate, which had a similar but distinct system).

  7. The country with the electoral system with the most similarity to Hungary’s old (1990-2010) electoral system is actually Lithuania, which elects half the seats using two-round system in single-seat districts and the other half by list PR with a 5% threshold. The tiers are not linked, so formally Lithuania has mixed-member majoritarian. However, unlike Hungary, which has generally had majoritarian-type outcomes and often something approaching a two-party or two-block system, Lithuania’s party system is far more fragmented, and on its face virtually indistinguishable from the party system you tend to have in simple PR systems (comparable, for instance, to that of neighbouring Latvia). This divergence in outcome is quite striking, and I wonder what account for it… Rune, perhaps you can shed some light on this?

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 )

Google+ photo

You are commenting using your Google+ 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.