MARCO TEÓRICO
2.4. SISTEMA SOLAR FOTOVOLTAICO
2.4.1. SUBSISTEMA DE CAPTACIÓN
2.4.1.1. Panel Fotovoltaico
What do voters know about electoral thresholds and how would they behave if the threshold at the federal level were abolished? Only survey data can help to answer these questions. By finding an answer to the first question, one can obtain suggestive evidence on whether behavioral effects due to the reform on the part of voters are likely to be large. The second question relates to external validity of the results at other levels of government. I have asked a random sample of the German population about their voting behavior as part of the
German Internet Panel (GIP)’s wave 30 (Blom et al.,2018). The GIP generates longitudinal survey data on political and economic topics based on a probability sample of the German population (Blom et al., 2015). The current waves’ sample size is varying but approximately includes 3000 participants.33
The topic of electoral thresholds had already been featured in wave 3 of the GIP (Blom et al., 2016), when participants were asked about the five percent threshold in place at federal elections in Germany (“What vote share of second votes does a party require overall in order to be able to obtain a seat in parliament?”).34 The results reveal that a substantial
share of the population is unaware of this electoral institution: 53% of participants answered correctly, 40% answered “Don’t know.” and the remainder answered incorrectly.
Given this rather low number of correct answers, it seems even more questionable that many voters are informed about thresholds at the local level since municipal elections receive much less attention than federal elections. Figure 3.7 shows the fraction of survey respondents by state who believe that there is a threshold in place in their state of residence.35. In all
states, more than 50% of respondents think that a threshold exists while in fact there is no threshold in place in any state (with the exception of city states). The case of North Rhine-Westphalia is quite interesting: the state saw an abolishment in 1999, but a 2.5% threshold was introduced in 2016 (Landtag Nordrhein-Westfalen, 2016). Consistent with this re-introduction, the percentage of the response “yes” is visibly higher than in other states. What matters for the existence of behavioral effects on the part of voters is whether people
learned about the change in thresholds in affected states. Since we do not know what
respondents would have answered before the reforms took place, the only statement that can be made is that less than 50% of respondents could have learned about the change in
thresholds in affected states (because the majority in each state still thinks a threshold is in place).
Can the behavioral effects on the part of voters be expected to vary by municipality size? This would require voter knowledge about implicit thresholds. In order to test whether voters are uninformed about implicit thresholds, a question about the 2014 election in Germany for the European Parliament was implemented. Respondents were asked to guess the implicit threshold for this election, which took place without an explicit threshold in place. Given that there were close to 100 seats available for Germany, the implicit threshold was approximately equal to 0.5% of all votes.36 More precisely, the party “Pirates” obtained a seat with 0.6%
of votes. Figure C.2 reveals that the vast majority did not answer correctly. Therefore, if
33 See
http://reforms.uni-mannheim.de/internet_panel/Methodology/, retrieved August 21, 2018. 34Note that each voter has the opportunity to vote both for a party (the “second vote”) and, independently, for a candidate in the voter’s district at federal elections in Germany.
35 Results for small states are reported in conjunction with another state such that these figures are hard to interpret. This is true in particular for the city states Hamburg (HH), Bremen (HB) and Berlin (BE).
36 Recall the formula from above, 1 2×seats.
Figure 3.7: Is there a threshold at the municipal level in your state of residence? BE/BB BW BY HE MV NI/HB NW RP/SL SH/HH SN ST TH 0 20 40 60 80 % Yes
Source: Wave 30 of the German Internet Panel (Blom et al.,2018)
behavioral effects vary by municipality size, as claimed by Baskaran and da Fonseca(2016b), this must be due to parties behaving differently in municipalities of different size.
The question of external validity is of course crucial since many national parliaments have more than 100 seats which means that implicit thresholds are quite low and therefore the number of parties may increase significantly if explicit thresholds were removed.37 It is quite
likely that salience at the federal level would be higher than at the local level such that behavioral effects could be larger. Stakes are higher at the federal level, such that parties might also react differently. On the other hand, voters might be less willing to experiment at the federal level, knowing that small parties entering a national parliament can have a larger impact than at the local level (e.g. in terms of legislation). Figure C.1 presents answers to the question of whether respondents would have behaved differently in the last federal election if it had taken place without an explicit threshold. Clearly, this is not the case for almost all respondents.
3.7
Conclusions
This paper estimates the causal effect of the removal of an electoral threshold on political and economic outcomes. The findings can be summarized as follows: both the number of parties competing and the number of parties represented on municipal councils significantly increase in response to an abolishment. The effect is significantly more pronounced in large municipalities, which can be explained by differing implicit thresholds. Municipal
37 As for the effects on expenditure, it is quite likely that the common pool problem exists at the federal level, too, since the federal government is typically composed of more than one party.
expenditure increases, consistent with a common pool problem. Survey results reveal limited voter knowledge about thresholds in general and about municipal thresholds in particular, suggesting a limited role of behavioral effects.
Future research may clarify the exact size of behavioral effects, could analyze the role of changing party behavior versus voter behavior due to the reforms and could explore additional policy outcomes. As for the effect on spending, it would be interesting to know which exact channel is at play. In addition, the threshold removals used for identification could serve as an instrument for other political variables, e.g. to explore the impact of council representation of extremist parties on policy outcomes.