German electoral system
To debate whether electoral systems are an independent or dependent variable is pointless. For we are not dealing with an intrinsic independence; we are simply asking different questions. Electoral systems are assumed to be an independent variable when the question is, what do they do? If the question is, instead, how electoral systems come about and why are they chosen, then the electoral system is treated as a dependent variable. And that is all there is to it. Germany and its false witnessing aside, when is a mixed system veritably mixed? And what are the merits of plurality-PR mixes? Frequently held opinion is that mixed systems are ‘the most attractive forms of solution to meet otherwise contradictory imperatives’. It agreed that we are faced with contrary (though not contradictory) imperatives, namely, (1) the function of mirroring (exact representation) and (2) the function of functioning (efficient government). But are mixed systems a solution that combines as the claim goes — the best of both worlds?
The single national district is a very ‘photographic’ element of the system. The Hare-Niemeyer is one of the less distort PR formulae, or even the least distort, depending on how the proportionality of a formula is calculated. Thus, the manipulative impact of this ‘half’ of the system, if any, is to be found in the five per cent threshold. 
Dating from Duverger’s seminal studies, two effects of electoral systems on party systems have been distinguished: the mechanical and the psychological effect. The mechanical effect of electoral systems is composed by two distinct phenomena, distort and the reductive effects. The distort effect is the induced disproportion between the seat and the vote quotas of each party. The reductive effect consists in the reduction of the number of parties that obtain seats in respect of those that participate into the electoral competition: all parties that do not cross a certain threshold of votes are excluded from parliamentary representation. The psychological effect consists in the pressure exercised on electors not to ‘waste’ their votes, but to cast them for parties for which the votes would ‘count’ in determining the victory of candidates, or in contributing to the quota of seats. This electoral behaviour, called in the literature ‘strategic’ or ‘sophisticated’ voting, aims to prevent the (reduce the chances of) success of those parties or candidates which are less preferred by the elector. 
After the introduction of the five per cent threshold at the national level in 1953, the index displays relatively high values in the elections until 1961, then falls to very low values between 1972 and 1983, and rises again in the most recent elections. These values show that the process of concentration of the German party system that took place between 1953 and 1961 did not only occur at the parliamentary level (where the parties represented in parliament decreased to three) but also in the electorate. Moreover, the index shows that the long ‘three-party phase’ of the years 1961-80 was not ‘manufactured’ by the electoral system (with the exception of the 1969 elections), rather it was based on the concentration of the preferences of the electorate on the three main parties. The distort effect of the five per cent threshold can also be captured by the indices of electoral and parliamentary fractionalisation the difference between the two values in each election measures as distort–that is, de-fractionalising–effect of the electoral system. the values of the parliamentary and electoral fractionalisation over time, displays a U-shaped pattern, with a noticeable de-fractionalising effect in the first few and last few federal elections. The electoral bases of the concentration process are observable in particular after 1953, shown by the rather modest difference between 1972 and 1987 the outcome of the system was of almost perfect proportionality between votes and seats. Finally, the figure shows that after national reunification the five per cent threshold has again had a ‘filter’ effect, turning a higher electoral fragmentation into a relatively less fragmented parliamentary party system.
More specifically, the series of shows that in the first post-reunification elections of 1990 the de-fractionalising effect of the electoral system reached its peak, being even stronger than in the 1950s and determining, among other things, the highest level ever of overrepresentation of the two bigger parties. This result was mainly due to the negative result of the ‘western’ Greens in those elections, which failed to pass the five per cent threshold by 100,000 votes. (11) Obviously, the application of the five per cent threshold on a single territory would have produced even higher levels of disproportional. In 1994, however, when the old system was restored, disproportional was lower than it would have been, had the PDS not brought its 4.4 per cent of second votes in parliament via the alternative threshold. 
The Reductive Effect
A general measure of the reductive effect of an electoral system is the difference between the number of parties competing into elections and the number of parties that obtain parliamentary seats. From the figures reported in Table 3, two characteristics of the German case are immediately evident: first, the number of the parties represented in the Bundestag decreased rapidly between 1949 and 1961, from ten to three, and then rose to four after 1983 and to five after 1990. Second, a high number of lists have constantly been excluded from representation. This is, however, a characteristic of virtually all electoral systems. What is of interest here is to evaluate the specific reductive effect of the five per cent threshold, distinguishing it, on the one hand, from the reductive effect of the ‘natural’ thresholds of the system, and, on the other, from the moderating effect on the five per cent hurdle typical of the alternative threshold. For the 1990 elections the effect of a further characteristic of the system m should be considered: the division of the territory in the application of the five per cent threshold. 
The current fractionalisation and territorial differentiation of the vote are for the most part the consequence of the strong electoral showing of the PDS, a party that has obtained between 15 and 25 per cent of the vote in the Lander of the ex-GDR, whilst almost absent in the western regions of the country. The electoral system has not been able to exclude the party from the Bundestag (actually the alternative threshold gave it the possibility of consolidating its position), and therefore the increase in the electoral fractionalisation has translated into a higher parliamentary fragmentation. Although it is difficult to forecast how the present situation of the German party system will evolve, it cannot be excluded that in the next few elections the consolidation of the PDS in the Federal Assembly might seriously influence the bipolar functioning of the German party system. In fact, apart from sporadic cases of external support by the PDS to SPD-led governments in two eastern regions, the other parties do not consider the PDS to be Regierungsfahige, that is, a possible alliance partner. Moreover, and somewhat paradoxically, the diversification of the electoral panorama between eastern and western regions might instead considerably damage the other two small parties present in the Bundestag, the FDP and the Greens. Problems for the bipolar mechanics of the system might emerge even if all three small parties obtain seats, though. The capability of the German electoral system to manufacture majorities is indeed limited if compared with that of majority systems, and essentially depends on the quota of votes that are excluded from representation. With the PDS being constantly excluded from coalitions, the electoral system might not be able to manufacture a majority if all three small parties (FDP, Greens and PDS) obtain seats–unless the largely unpredictable occurrence of surplus seats comes to help. In fact, as surplus seats have reinforced narrow majority coalitions in the last two elections, it cannot be excluded that they might even manufacture a majority in the future. It is reasonable to think that this eventuality will be kept in mind in a reform of the surplus seats mechanism, now advocated by many. However, a scenario in which resort to a Grand Coalition or to a minority government–the latter would be an entirely new option at the federal level–might prove necessary is not unlikely. And in that case, as already occurred in the 1960s, voices in favour of a sweeping electoral reform might be heard again. 
Massicotte, Louis. German Politics: 2005, Vol. 12 Issue 1, p1-22, 22p.
Capoccia, Giovanni. West European Politics: 2006, Vol. 25 Issue 3, p171, 32p.
Saalfeld, Thomas. German Politics: 2006, Vol. 11 Issue 3, p99-130, 32p.
Rensmann, Lars. German Politics ; Society: 2006, Vol. 24 Issue 1, p67-92, 26p.
Olsen, Jonathan. German Policy Studies/Politikfeld analyse: 2005, Vol. 2 Issue 2, p197, 24p.
Arzheimer, Kai.. European Journal of Political Research: 2006, Vol. 45 Issue 3, p419-443, 25p.
 (Saalfeld, Thomas. German Politics: 2006, Vol. 11 Issue 3, p99-130, 32p)
 (Massicotte, Louis. German Politics: 2005, Vol. 12 Issue 1, p1-22, 22p)
 (Capoccia, Giovanni. West European Politics: 2006, Vol. 25 Issue 3, p171, 32p)
 (Rensmann, Lars. German Politics ; Society: 2006, Vol. 24 Issue 1, p67-92, 26p)
 (Olsen, Jonathan. German Policy Studies/Politikfeld analyse: 2005, Vol. 2 Issue 2, p197, 24p)
 (Arzheimer, Kai.. European Journal of Political Research: 2006, Vol. 45 Issue 3, p419-443, 25p)