In one of his many excellent pieces in the run-up to the German election, Alex mentioned the phenomenon of ‘overhang mandates’. These are extra parliamentary seats that a party gains by winning more seats via one of German’s two electoral methods than by the other. This might seem odd enough. What’s even odder is that a party could lose a seat if too many people vote for it.
German electoral law is complex. In a comment to one of Tobias’s posts, Florian recommended the wahlrecht.de website as a good primer on how it works. He also mentioned examples of some of the electoral weirdnesses explained by wahlrecht.de. For example, did you know (asks Florian) that, under certain circumstances, a vote can have ‘negative weight’ — can reduce the parliamentary representation of the party for which it is cast?
Well, it can. And this conundrum is worth looking at closely, because right now it is more than a mere electoral curiosity. There is one electoral district in Germany, Dresden I, that has not yet voted. (Those who’ve been paying a perhaps unhealthy level of attention to the German elections will know that the death of a neonazi candidate has forced the delay of the election.) And in Dresden I, there is a very real chance that a local triumph of the CDU could cause the party to lose a seat in the national parliament. The reason? It’s those overhang mandates that Alex kept mentioning.
Excellent as wahlrecht.de is, it’s in German. Below the fold, then, is a summary explanation of how the CDU could lose a seat by gaining votes. For those who read German and are interested in that sort of thing, there are links to the relevant passages of the BWahlG (German Federal Electoral Act).
In the mean time, we should note that the possible ‘negative weight’ of CDU votes in Dresden I, though perverse and undemocratic, would not affect the overall results in Germany. Even if the CDU are ‘catastrophically successful’ in Dresden I, the Union will still have more seats than the SPD, albeit with a lead of only 2 rather than 3 MPs. The really perverse thing that could come out of the Dresden special election is this: CDU and SPD wind up with an equal number of seats. As the Spiegel explains, however, this is mathematically a possibility, but in real-world terms exceedingly unlikely. To achieve this result, the SPD would need to poll 91% of voters in the district, and every single eligible voter would have to vote.
But back to the ‘negative weight’ problem.
Sadly, No! takes the Wall Street Journal to task for saying that the elections failed to produce a clear result because Germany’s parliamentary system is ‘peculiar’. As S,N! points out, the ‘peculiarity’ of this system, i.e., its use of proportional representation, is far from peculiar. Most other democracies use it too.
There is a bit of truth to the WSJ‘s comment, though. It’s not peculiar at all that, in a proportional system, no one party has a clear majority. But Germany’s electoral law actually mixes elements of proportional representation with UK/US-style ‘First Past The Post’. And it is the inelegant interface of these two systems that, ultimately, means a CDU voter in Dresden I might end up hurting her party by voting for it.
Here’s how we get to this perverse result:
The very first section of the Federal Electoral Act states that the number of MPs is 598… subject to adjustment. Any adjustments will result from the outcome of the second of Germany’s two ways to elect an MP. For in a German election, voters cast two votes: one for a specific candidate, one for a party list.
The first way to elect an MP will be familiar to UK and US readers. This is the ‘direct mandate’. Each party can nominate a candidate in a given district. The candidate who wins that district on a FPTP basis receives the district’s direct mandate. Whatever the number of MPs ultimately turns out to be, direct mandates will account for 299 of them.
The second way to elect an MP is through the party list. In each Land (federal state), each party draws up its own list. Seats in the Bundestag are then apportioned in accordance with the ‘second’, party-list votes each party received.
Two things now come into play to muddle everything up.
First, you’ll recall those candidates elected with direct mandates. In totting up the number of seats each party is to receive based on proportional representation, that party’s number of direct mandates is subtracted. If my party is proportionately entitled to 50 seats, and also won 49 direct mandates, then those 49 are subtracted from my 50, and only one candidate from the list takes a seat. But what if my direct mandates exceed my list mandates? What if proportional representation entitles me to 50 seats, but I won 51 direct mandates? In that case, I get to keep the ‘extra’ Bundestag seat. That seat is an ‘overhang mandate’; and these overhang mandates are what can cause the total number of MPs to vary from the base figure of 598.
Second, though each party draws up its list on a state-by-state basis, the state lists of a given party are ‘linked’ (verbunden). In apportioning seats, the number of seats each party gets is measured by the number of second votes the party received nationwide. But the actual apportionment of those seats goes by the statewide lists. When it comes to putting MP arses on Bundestag chairs, each party is, in a way, competing with the same party in other states.
Overhang mandates, remember, arise when you win more seats directly in your state than the second votes in that state would entitle you to. What could happen in Dresdner I is that the CDU get sufficient second votes that the overhang shrinks. This wouldn’t hurt the Saxon CDU; they’d still have the same number of mandates. But that last mandate would no longer be ‘extra’, and this would hurt the party nationwide. The number of Bundestag mandates would decrease by one, and thanks to ‘linked listing’, the CDU would receive one less proportional seat in the Bundestag; a party-list candidate from a state other than Saxony would lose his or her place.
According to wahlrecht.de, it wouldn’t take an improbably huge number of CDU votes in Dresden I to reduce the CDU’s total number of MPs — about 41,000 would do. And it’s not as though this would be the first time German votes had ‘negative weight’ — it’s the Union in Saxony just now, but according to wahlrecht.de, enthusiastic SPD voters in the city-state of Bremen have more than once reduced the SPD’s overall representation in the Bundestag. This time round, though, the legally-required delay means that the Dresden I election is taking place in isolation, with all political eyes in Germany focused on it. wahlrecht.de predicts, no doubt correctly, that the election will be a carnival of bizarre tactics, with the CDU urging its followers to ease up, while the SPD turn out in droves for the Union. The Dresden I special election won’t make enough difference to determine the next German government. But it should be highly entertaining.