14 April 2014

Legal but not Fair: Viktor Orbán’s New Supermajority

with Miklós Bánkuti in Princeton and Zoltán Réti in Budapest*

Viktor Orbán and his Fidesz party coasted to a clear victory in last weekend’s Hungarian election, as expected. The governing party got 45% of the vote, but the new “rules of the game” turned this plurality vote into two thirds of the seats in the parliament. A continuing two-thirds parliamentary majority allows Orbán to govern without constraint because he can change the constitution at will. But this constitution-making majority hangs by a thread.

Orbán’s mandate to govern is clear because his party got more votes than any other single political bloc. What is not legitimate, however, is his two-thirds supermajority. Orbán was certainly  not supported by two-thirds of Hungarians – nowhere close.  In fact, a majority gave their votes to other parties. Orbán’s two-thirds victory was achieved through legal smoke and mirrors. Legal. But smoke and mirrors.

The International Election Observer Mission of the Organization for Security and Cooperation in Europe was extremely critical of the election. The election monitors found that in many different ways “[t]he main governing party enjoyed an undue advantage.” They reported numerous violations of international standards, including a failure to separate party and state, a biased media environment, a partisan Electoral Commission, lack of transparency in determining the electoral districts, and a generally un-level playing field. These, too, contributed to Orbán’s success.

In this post, I will explain why a plurality result in the polls turned into a constitutional supermajority and why that supermajority is due more to Fidesz’s self-dealing than to popular will. I will also show that Fidesz’s supermajority was so close that it depends on every one of the new tricks that the party inserted into the electoral system to benefit itself. One seat less, and the supermajority would be gone.

In saying that Orbán’s supermajority is illegitimate, I am not arguing that Orbán simply stole the election.  No other party came close to Orbán’s 45% of the vote, though when you exclude the new Hungarian citizens from the neighboring countries – people who don’t actually live in Hungary and probably never have – Orban’s support drops to 43.5% among domestic voters.

The left alliance, a collection of five parties with the Socialists in the lead, received only 26% of the vote. It evidently failed to capture the public imagination and capitalize on the fact that about half of the voters wanted a change of government. Jobbik, a far-right party, won a shocking 21% of the vote, up more than 3% from its already large 2010 showing, making it the most successful far-right party in the European Union. The tiny, vaguely Green (but mostly vague) party LMP just barely squeaked over the 5% threshold to enter the parliament.

Orbán’s parliamentary two-thirds hides a complex set of calculations that depends on many factors. The Hungarian election system is fiendishly complicated. The bottom line is that the “two thirds mandate” includes far less popular support than one would guess given its size and far more legal trickery than a democratic government should be permitted to use.  Orbán’s parliamentary majority may have been earned, but the supermajority mandate was not.

That said, Fidesz would probably have won a majority in parliament no matter what sort of electoral system Hungary had. Most mixed electoral systems like Hungary’s top up large pluralities into majorities. Also, many governments – if they can – attempt to shape the rules to aid their own reelection. Some tweaking of the electoral rules to favor the party in power is regrettable but normal.

But the Orbán government went well beyond normal tinkering when it extensively revised the electoral framework during its last four years in office. The new system was designed precisely to give Orbán a vastly disproportionate two-thirds parliamentary majority with less than a majority vote. And it worked.

The international election monitors were not impressed with the new system. They concentrated primarily on evaluating the campaign and the election itself, as we will see later, but they also expressed concern about the election framework and how it had been adopted. As the election monitors noted, the governing party’s “undue advantage” resulted in part from a “legal overhaul” that was “unprecedented” and consisted of laws that were “passed and modified without public consultation or inclusive dialogue with opposition parties.” They found that “[t]he manner in which these laws were adopted and frequently amended, including in the year prior to these elections, led to legal uncertainty and did not provide for effective and inclusive public consultation, contrary to national legislation and good practice.” The election observers concluded that the legislative process used to adopt the new electoral framework “undermined support and confidence in the reform process.”

While written in the diplomatic language characteristic of international reports, this is a serious condemnation.

How did this legal overhaul produce the 2014 election result? It manufactured the appearance of another landslide for Orbán, a landslide that did not occur.

In fact, Orbán’s victory disguised a growing weakness in his appeal to Hungarian voters. Between the 2010 and 2014 elections, Orbán lost a substantial fraction of his supporters. In 2010, Fidesz received more than 2.7 million votes while in 2014 the party won less than 2.3 million. This 16% drop in the total number of Fidesz voters was produced by one of the lowest turnouts in the post-communist period, combined with an 8% decline in the party’s overall support among those who voted. In fact, the three lowest turnout elections since communism ended were the three that Orbán won: 1998, 2010 and 2014.

2014-04-14 Scheppele 1These numbers mask the even more dramatic decline in Orbán’s popularity at home because they include new citizens in the neighboring states to whom the 2012 Fidesz constitution granted citizenship. The new citizens who voted for the first time this year are a formidable force since they voted 95% for Fidesz.

But these new citizens were not eligible to vote in 2010 so we should compare Orbán’s support in 2010 and 2014 without them. Among domestic voters, Orbán received only 2.1 million votes this time, losing 21% of the voters that had supported him in 2010.   His party list vote slipped by 9% in this group as well.

Fidesz lost more than one fifth of its domestic  voters overall in part because of the lower turnout, but the far-right Jobbik party and the left alliance both gained voters, as the chart below indicates.

2014-04-14 Scheppele 3The depressed turnout and yet the increased success of the two other major parties show Orbán’s weakness despite his huge parliamentary majority. Overall, 37% of voters stayed away from the polls, 35% voted for other parties, and only 28% of Hungarians actually cast an affirmative vote for Fidesz (including the “over the borders” voters).

If you consider only those who went to the polls, leaving aside the possibility that the low turnout is itself a signal of Orbán’s weakness, 51% voted for the three other parties that entered the parliament while 45% voted for Fidesz. And yet 51% of voters will get only 33% of the seats. And Orbán will get his two-thirds.

How is it possible for only 45% of the vote to turn into a two-thirds parliamentary mandate? Orbán’s “two-thirds” was enabled by a series of tricks that were legal, but not fair.

(To follow the analysis, you need to know that Hungarians cast two votes at a parliamentary election, one for a representative from their home constituency and one for a party. Fidesz got 45% of both the constituency and the party votes.)

Orbán’s electoral system had the overall effect of creating a huge disparity in how many votes each party needed to gain a parliamentary seat. If you add all votes (party list and constituencies together) and divide them by the number of seats each party actually got in this election, you can see that not all votes were equal in securing representation in parliament. A simple formula describes how many votes for any other party it would take to generate the representation in parliament that one vote for Fidesz did:

1 Fidesz vote = 2.1 left alliance votes = 2.6 Jobbik votes = 3.1 LMP votes

In short, the Fidesz seats were acquired with many fewer votes the others.

In an earlier set of blog posts, I explained the new election system in detail. Now that we have election results, we can see how the different parts of the remodeled election system worked to generate this two-thirds parliamentary mandate.

In my earlier analysis, I alleged that the new electoral districts had been gerrymandered to produce winning results for Fidesz. It’s true that the parliament was reduced in size from 386 seats to only 199, so the districts had to be redrawn. Moreover, the old districts were massively unequal, so the Constitutional Court had required that they be made more equal in size. But neither the smaller parliament nor the more equal districts required gerrymandering.

How can one tell if a district has been gerrymandered? There are many ways to gerrymander by drawing the precise boundaries of districts to achieve a particular result. The full effects are hard to tell without better data than we have.

But the new districts are highly unequal in size, which gives us a way to test for gerrymandering. A gerrymanderer not constrained by the need for strictly equal districts would make the districts of his opponents systematically larger than the districts where his own voters dominate. In larger districts, it takes more votes to elect an MP, so each vote counts less. If the “left” districts are systematically larger than the “right” districts, that is evidence of conservative gerrymandering.

The relationship between district size (here measured by the number of voters who actually voted in each district) and partisanship (measured by the percentage of the vote that went to the left alliance and LMP combined) is substantial. As one would expect in a system gerrymandered to benefit the right wing of the political spectrum, the more left-leaning a district is, the more voters it has. In fact, the explained variation (R-squared) is 37%. In social science, it’s a huge effect when one variable accounts for more than one-third of the variation in another.

2014-04-14 Scheppele 2The system is in fact much more biased against left voters than the one it replaced. Calculated using the 2010 districts and voting data, the same model (votes for the left v. # of votes in a district) explained less than 2% of the variance (R-squared = .019), as the figure below indicates.

2014-04-14 Scheppele 4The old districts were much more unequal in size (which was a problem) but they were not nearly as politically biased, as the nearly flat regression line shows. The new districts are still highly unequal in size. But now they are strongly biased against the left. This is what a gerrymander looks like.

There’s another aspect to the gerrymander that we can also see clearly. In the old parliament, more mandates were determined by the list votes than by individual constituencies. (The pre-2014 system was also disproportionate, but less so.) List votes which are distributed over a set of candidates are more proportional than individual constituencies which are generally more disproportionate because only one candidate can win. In the new parliament, the relative proportions of list and constituency mandates are reversed, so that now 106 of the seats are determined by the individual constituencies and 93 come from the party lists. This makes the effect of the gerrymandered districts bigger in the overall calculation of parliamentary mandates.

Fidesz won 45% of the votes in the individual constituencies in 2014 and yet got 88% of the seats. The effects of the gerrymander, even on the rough analysis above, assisted massively in producing such a disproportionate effect.

Recall that Fidesz unilaterally set the boundaries of all of the districts in a law requiring a two-thirds vote, and did so without allowing the opposition to have any input at all. As the international election monitors noted in their post-election report, “The process of delimitation of constituencies was criticized … for lacking transparency and inclusive consultation.” While the election law required that districts vary by no more than 15% above and below the national average, the election observers noted that districts still fell outside this wide margin. As we can see, that gave more room for the gerrymanderers to play with the district composition.

While gerrymandering is probably the single biggest trick that converted Orbán’s plurality to a supermajority, Orbán’s supermajority is built from many different sleights of vote. In my earlier analysis, I showed how the parties on the left were forced to work together for many different reasons ranging from the elimination of the second round of voting in individual constituencies to the novel system of compensatory mandates on the party-list side.  Because the new “first past the post system” of individual constituencies was designed to reward the largest single party, it was important for the fragmented and fractious left to unite to win.

Five of the parties on the left eventually forged an alliance with each other to overcome the bias against small parties built into the system. The alliance groups squabbled and quite visibly hated working together, which no doubt suppressed the vote they got. Together, they won only 10 of the 106 individual constituencies with their 26% of the constituency vote. But they would surely have lost almost every single constituency if they hadn’t joined together.

As predicted, party fragmentation had huge effects. Had the alliance had managed to attract to its ranks just one more party, the LMP, the number of individual constituencies that a six-party alliance would have gained would have more than doubled – from 10 to 21. Even this small division in the left opposition cut in half the number of individual mandates that the left won. That’s a powerful effect of the system.

LMP had its own reasons to stay outside the alliance, and it guessed correctly that it could (barely) get its own parliamentary fraction by doing so. But if it had joined with the alliance, the left opposition together would have gained an additional 11 district mandates instead of the five that LMP got from its list votes, for a net gain of six.

(The individual constituencies where the Alliance + LMP > Fidesz are Budapest districts #1, #2, #4, #6, #12, #13, #15, #18; Baranya districts # 1 and #2; Szabolcs-Szatmár-Bereg district #1 as you can see here.)

Many mixed electoral systems allow the “unused votes” won by losing candidates to be added to their party-list totals when the list mandates are calculated, so I will call that “normal compensation.” Normal compensation is the tool through which mixed systems become more proportional. But only Hungary has a winner compensation system that throws proportionality to the winds. With winner compensation, the winners of individual constituencies may add difference between the number of votes they received and the number of vote received by the second-place candidate (minus one) to the totals for their party lists for use in calculating party-list mandates. It’s complicated but a simple explanation is here.

With the results in, we can see that winner compensation added six parliamentary mandates to the Fidesz totals, which were essential to the party’s supermajority. We calculated this using the Election Office website data that handily provides the difference in the number of votes between the first- and second-highest vote-getter each district here.  The left alliance would gain three of those mandates, Jobbik would gain two and LMP would gain one.

The table below shows that, if winner compensation were removed from the system, Fidesz would lose six of its parliamentary seats, which by itself would put two-thirds far out of reach.

 Parliamentary Mandates with Winner Compensation

Party

Mandates with winner + normal compensation

Mandates with normal compensation only

Effect of removing winner compensation (in mandates)

Fidesz

37

31

-6

Left alliance

28

31

+3

Jobbik

23

25

+2

LMP

5

6

+1

There was another trick cooked into the legal reforms that we can measure directly from the election results. Orbán’s new constitution gave expedited citizenship, upon application, to ethnic Hungarians who have never lived in Hungary. With citizenship came the right to vote. About 600,000 ethnic Hungarians “over the borders” took out citizenship and nearly 200,000 of them registered to vote. Voter turnout in this group was 81% (much higher than among domestic voters) but 19% of the ballots were disqualified due to errors in the way that the ballots were prepared. Six days after the election, these ballots were finally counted, and Fidesz won 95.49% of the “over the borders” vote.

The “over the borders” votes added 122,588 to the party list votes, enough for 1.4  parliamentary mandates. (Since all party list votes and compensation votes are added together before being divided by the total number of mandates available, all votes contribute to the overall pool and therefore particular sources like “over the borders” voters may produce fractions of mandates.) Given how close Orbán’s two-thirds majority was, the “over the borders” voters were clearly essential to achieving it. Every single mandate was necessary to the two-thirds.

Since Orbán’s supermajority hangs on the votes of the new “over the bo