(p.130) Appendix 2 The causes of surplus seats at Bundestag elections
(p.130) Appendix 2 The causes of surplus seats at Bundestag elections
The phenomenon of surplus seats (Überhangmandate) received little attention prior to the reunification of Germany. At no election before 1990 were there more than five such seats (in 1961). In the elections between 1965 and 1976 there were no such seats. In no election did surplus seats come close to playing a role in deciding which party or coalition of parties would possess a majority in the Bundestag.
In 1990, things changed. At that election there were six surplus seats. In 1994 there were sixteen and in 1998 thirteen such seats. In 2002 the number of surplus seats fell to five. In 1994 surplus seats almost decided whether the Kohl government (a coalition of the Christian Democrats and the liberal FDP) could continue to govern, since twelve surplus seats went to the CDU and only four to the SPD, a net gain of eight seats for the CDU. These extra eight seats gave the governing coalition 341 seats, the other parties (SPD, Greens and PDS) 331: a majority of ten seats which, without surplus seats, would otherwise have been a majority of just two seats. In 2002 the net gain of three surplus seats (four for the SPD, one for the CDU) enlarged the majority of the SPD–Green coalition from six seats to nine seats. So in both 1994 and 2002 small shifts in voting behaviour, or the addition or subtraction of a small number of surplus seats, could have prevented the incumbent coalition government from continuing in office.
Surplus seats distort the proportionality of the German electoral system since, unlike electoral systems in the Länder (see chapter 6), there is no provision for ‘equalisation’ seats to be allocated to other parties. If a party in a Land election in, for example, Thuringia or Baden-Württemberg, is allocated surplus seats, then further additional seats are awarded to other parties until proportionality of the overall allocation of seats is restored. In Bundestag elections, that does not occur. Apart from the 5 per cent requirement for qualification for allotments of list seats, this is the only (p.131) significant distortion of what otherwise is a system of accurate proportional representation of parties. In 2002 the SPD (the principal beneficiary of surplus seats) was awarded one seat for every 73,644 votes. The CSU needed 74,336 votes and the CDU 74,548 for each seat. The Greens needed 74,697 and the FDP 75,265 votes for each seat.
The causes of surplus seats are varied. One cause is the different basis for calculating how many constituency seats each Land possesses (according to its population), contrasted with the number of its list seats (based on voting turnout). So, even ignoring differences in turnout, discrepancies in these two factors of total population and population of voting age can result in some Länder having more, and others fewer, constituency seats compared to list seats (see Appendix 1). When differential turnout is added into the equation, that discrepancy can increase considerably. In 2002, Saxony-Anhalt had ten constituency seats, but only eight list seats. The SPD in that Land won all ten constituencies (56 per cent of seats) but obtained only 43 per cent of list votes: so was entitled to no more than eight seats. The two additional seats became surplus seats in the calculation of total seats won by the SPD.
Vote-splitting is another factor. Since list seats (and the total number of seats for a party) are awarded on the basis of second votes, but constituency seats are won by first votes, discrepancies in a party's share of first and second votes can give it more constituency victories than its share of second votes would justify. The biggest parties (CDU and SPD) tend to have more first votes than second votes, since ‘vote-splitters’ (see chapter 5) are either supporters of those larger parties who ‘lend’ their second vote to the smaller coalition partner party (usually now the FDP for the CDU, the Greens for the SPD), or supporters of those smaller parties voting ‘rationally’ with their first vote for a constituency candidate of the larger coalition partner likely to win the seat. So even if a party wins all ten constituency seats in a Land with 50 per cent of first votes, and assuming a total of twenty seats in that Land, then if it only has 45 per cent of list votes (an entitlement of nine seats in total), one of its constituency victories would constitute a surplus seat. Indeed, a party can win all the constituency seats (generally 45–55 per cent of seats available in a Land) on a relatively small share of the list votes in that Land. In Saxony in 2002 there were seventeen constituency seats and twelve list seats, making a total of twenty-nine seats. Though winning only 33.6 per cent of list votes, the CDU in Saxony in 2002 won thirteen of the seventeen constituencies in that Land. Thirteen as a percentage of twenty-nine is 45 per cent. So one of the constituency seats was a surplus seat for the CDU.
(p.132) Saxony was also affected by a third causal factor: the strength of ‘third parties’. So in several constituencies the PDS obtained between 18 and 20 per cent of constituency votes, without winning a seat. The SPD in seats won by the CDU sometimes obtained between 30 per cent and 34 per cent. So CDU constituency seats were often won with only between 33 and 40 per cent of first votes.
Finally, in Länder with very small numbers of seats, such as Bremen (2002: four seats in total), a surplus seat can arise because the method of calculating seats results in large arithmetic remainders, and one of these can give rise to a surplus seat, as happened in Bremen in 1994. Indeed, it has been calculated that, in certain circumstances, a paradox can arise. If a party which, having won most of the constituency seats in a Land, had polled fewer list votes in that Land than it actually did, it could end up with extra seats in total (assuming the same national total of votes in each case). This is because it would retain all the constituency seats, more of which would become surplus seats, but win additional seats in other Länder because of the method of allocating a party's seats among the Land lists.
There is an almost metaphysical dispute as to which kind of seats are surplus seats. Some claim that they are constituency seats (after all, the party wins more such seats than its total allocation in that Land, so it would seem that the constituency seats are ‘surplus’). Others point out that the number of constituency seats remains constant with or without surplus seats, so that the extra seats are additional list seats. Therefore surplus seats must be list seats won inequitably by the benefiting party (though which list seats these are is difficult to establish for, by definition, they are not in the Land where the surplus seats seem to have been won). The Constitutional Court seemed to favour, implicitly, the first interpretation. In 1998 it ruled unanimously that in a Land where a party has won surplus seats, should a vacancy occur – by, for example, the death of an MdB from that party – no successor can fill the vacancy until the number of seats for that party in that Land returns to normal. This applies whether the vacancy is in a constituency seat or on the party Land list. So in 1998 the Bundestag began with 669 MdBs (the standard 656 plus thirteen surplus seat MdBs). In 2002 it had been reduced to 665, because four vacancies had remained unfilled following the 1998 Constitutional Court ruling.