In the last years there has been a decrease of voter turnout in most ofthe European countries. Factors influencing voter turnout are many andamong them there are economic factors such as costs related to the act ofvoting; demographic factors such as the individual characteristics of voters;technological and institutional factors and so on. Low turnout has among itsconsequences an unfair representation of the population and therefore, it isusually considered to be a socially unwanted outcome. Consequently, there havebeen numerous studies which aim was to identify the determinant factors ofturnout, in order to encourage participation in the political process. Despitesignificant study into the matter, the causes and consequences for the changesin voting turnout are still not clearly determined. In his work Jackman(1987) finds that political institutions and electoral law have an importantimpact on voter turnout.
Heshows that electoral districts which are more competitive at national level,foster voter turnout. He also finds that Multipartyism decreases turnout,since it weakens the relevance of elections in the formation of a government. Among these causes, of relevantimportance is the impact of the electoral system on voter turnout.Blais and Carty(1990) assess the effect of different electoral formulae on voter turnout inwestern democratic states and find higher turnout rates in states adopting the PRsystem.Several empirical studies have attempted to explainthe consequence of having more freedom on the choice of a candidate on voterturnout. Among them, there are severalpapers about the effect of Open List vs. Closed List, with dissimilar views ontheir impact on voter turnout. Blumenauat al (2014) find that open list systems allow for intra-party competition whichpermit voters to choose among not only parties but also among candidates,fostering competition among candidates and allowing for smaller parties to getmore representative power.
In contrast, Robbins (2010) find that party-centeredclosed list systems compared to candidate-centered open list systems, increasevoter turnout. This is a consequence of the more active lobbying activities ofparties in closed list systems, which aim is to have a higher influence onvoters’ preferences. Ingeneral, the empirical evidence is unsettled and the effect of the pollstructure on turnout is still an unsolved query. Sanz (2015) studiesthe effect of electoral system on voter turnout in Spain. He compares voterturnout among municipalities withdifferent electoral systems as well as different number of citizens. He showsthat open list systems compared to a closed list system enhances voter turnout,as a consequence of the possibility given to voters to choose for individualcandidates from the same or different party list.On the same line, with the present work the aim isto assess the impact of electoral system on voter turnout in Italy.
I will makeuse of the particular institutional framework of Italian local elections wheremunicipalities are divided into two main groups, namely those that apply thesingle ballot plurality system and those that apply instead dual ballotmajority system. I will then assess the impact of these two different rules onvoter turnout by means of regression discontinuity design. Hence, the present work aims to contribute to thealready existing collection of works, hoping to put more light on theconsequences that the freedom of choice of voters has on turnout. Moreover, tomy knowledge there are no other studies which use the RD design to estimate theimpact of the two electoral systems on voters’ casting choice and consequentlyvoter turnout of Italian municipalities. 1.
Electoral system in Italy Theadministrative divisions in Italy are: Regions, Provinces and Municipalities.Municipalities are autonomous and independent administrative divisions. Theyare included in the Province. However, they can entertain direct relations withthe regions and the State due their autonomy. As of February 2017, there were7,981 municipalities in Italy. The political organs of a municipality include municipalcouncil (a legislative body) and the municipal committee (an executive body),both organs assist the mayor. Elections are held every five years. Sincethe implementation of the law n.
81 of 1993, the mayor is directly elected bycitizens which happens simultaneously with the election of the municipalcouncil. Moreover, such law requires that municipalities are subject to twodifferent electoral rules, according to their population size.Municipalitieswith a population of less than 15000 residents apply the single ballot pluralitysystem. Residents in this categoryvote for both the municipal council members and the mayor. Voting for a mayoralcandidate means voting also for the list supporting the candidate. It will beelected mayor the candidate that gets the plurality of votes.
The electiontakes place in a single round, except for the eventuality in which the two mostvoted candidates have obtained the same number of votes. In this case, thesecond ballot will proceed and if a further parity were to be registered, itwill be elected the eldest one among them. The list of the elected mayor willget a two-thirds majority of the seats in the municipal council, whereasone-third of the seats will be proportionally distributed among the defeatedlists.
Onthe other hand, municipalities with a population of more than 15000 residentsapply the dual ballot majority election rule. If no candidate has themajority of votes (i.e. more than half) in the first round, then only the twocandidates with the most votes proceed to a second round (called “closedballot”), which will be held two weeks later.
In the second ballot, becausethere are only two candidates, one candidate will attain with certainty amajority. In this category, each candidate is supported byone or more lists. Besides that, eachvoter is entirely free to change the candidate he votes for, even if hispreferred candidate has not yet been eliminated but he has simply changed hismind (called “panachange vote” or “disjoint vote”). The “panachangevote” allows citizens of municipalities with more than 15000 residents tovote, in the first round, in three possible ways: 1. Electorscan vote for a list. By voting a list they also vote the candidate associatedto that list.2.
Electorscan vote for a list and they can specify two candidates of that list, providedthey are of opposite sex.3. Electorscan vote a candidate. By voting this way, their vote is valid only for thecandidate and not for the list or lists supporting the candidate.Inthe second round, voters can vote only between the two most voted mayoralcandidates and cannot vote for any list.
The two candidates can choose to besupported by a different list than the one they were initially supported by.The winner is the candidate that gets the majority of votes. Once the winnersare announced, seats will be assigned to each list or group of lists with theproportional d’Hondt method.
Themunicipal council is elected in a single round with a complex system that wecan define as mixed in some specific cases, while it can be proportional inothers. In the former case it is called a “mixed system”, since it can bepartly proportional and partly majoritarian. More precisely, we can have the followingsituations:1. The mayoris elected in the first round (by majority), however the list or listssupporting the candidate do not reach the 50% of votes. In this case seats areassigned in a proportional way.2. The mayoris elected in the first ballot by majority, but the list or lists supportingthe mayor gets 40% of votes and no other list gets more than 50% of votes. Inthis case the list supporting the mayor gets 60% of the seats in the counciland the remaining seats are assigned proportionally.
3. The mayoris elected in the second ballot (by plurality), but the list supporting themayor already got in the first ballot 40% of the votes and no other list gotmore 50% of votes. Also in this case, the list supporting the mayor gets 60% ofseats and the remaining seats are assigned proportionally. Lists that have notachieved at least 3% of votes are not assigned any seats.
An important thing to keep in mind, is that thepreference vote was maintained in both municipality groups. Moreover, in bothcases, expressing a unique preference allows the voter to change the order ofthe candidates in the list and to establish their ranking. The main differencebetween larger and smaller municipalities, is therefore, the possibility ofexpressing a disjoint vote in theformer group. 4. Data and Methodology 3.1 Data source In order to estimate the causal effect of the twoelectoral rules on voter turnout, I will make use of a rich data set from theofficial interior ministry site, with results from all elections since theimplementation of the national rule (data available from 1993 to 2017).
Thedata set contains detailed information on candidates and voters in eachelection year, both at the aggregate level and at the municipality-individuallevel (number of electors and of voters, number of invalid votes, sex of votersand candidates, etc). Moreover, I will use publicly available data fromthe National Statistics Institute (ISTAT). ISTAT attributes to each existingmunicipality a six-digit code. The first three digits of the code identify theprovince to which the municipality belongs, while the subsequent three identifythe municipality within the province.
Following each administrative change, thelist of Italian municipalities and their codes is updated. It also takes careof an archive containing the list of the existing municipalities, the list ofdeleted municipalities, the territorial and administrative changes of themunicipalities since 1991 and a table of codes and names of geographical areas,regions and provinces. 3.2 Methodology To assess the impact of the electoral system onvoter turnout I will exploit the regression discontinuity design.The control group will be represented by themunicipalities with a population less than 15000, while the treatment groupwill be represented by the municipalities with a population higher than 15000. Moreover,I will add to the basic equation municipality and year fixed effects.
Hence, we can let the value of the unobservedvariables be different in the treatment and in the control group, as long asthey remain fixed over time and at individual level. Error terms will beclustered at municipality level. Thanks to its peculiarities, the Italianinstitutional context of local elections enables to trustfully identify thedifferences in outcomes between treatment and control group, as the causaleffect of the electoral system.
Meaning, the number of municipalities and theyears of all elections since the adoption of the national law are high (dataabout elections available from 1993 to 2017). Accordingly, the number ofobservations will also be high which will lead to increased precisionin estimates. Moreover, both groupsof municipalities will differ only in their electoral rules at the threshold.
Additionally,I will restrict my analysis to only those municipalities where elections didnot have a second ballot, for several reasons.First, as a consequence of the different nature ofthe two electoral systems, the control group and the treatment group will havedifferent probabilities to proceed to the second turn of elections. Therefore,the estimates will be more precise if we only compare municipalities similar withrespect to the prior probability of having a second turn of elections. But,knowing this, candidates would behave differently if they were in amunicipality with risk of run-off, regardless of the electorate’s inclination.
Forexample, smaller party lists may decide to present themselves with their owncandidate and measure their consent in the first round, significantly fragmentingthe electoral offer. Eventually, in a second moment, they could form acoalition with one of the two competitors for the run-off, in order to alsoenjoy the eventual majority prize. On the other hand, voters themselvescan set up strategic behaviors, casting their vote only for the candidate-mayorwith higher possibility of victory and not for the list supporting thatcandidate (disjoint vote). In both cases, estimates would be biased.Therefore, by considering exclusively municipalities with no second ballotelections, we reduce such biases. 4.
Challenges Itshould not be possible to precisely manipulate the assignment variable,otherwise the RD design will not be the right approach to follow. Lee andLemieux (2010) suggested different methods to test the assumptions of the RDdesign which I will now discuss. Away to test the imprecision of control over the assignment variable is toexamine its density. We should have no jump in the density at the cut-off,otherwise there may be sorting around the threshold, which invalidates thesuitability of the RD design. We could also examine if the treatment and thecontrol group are similar conditional on observable variables, most of whichshould be the same. Finally, we could implement a falsification test and assumeto find continuity in predetermined variables at the treatment cut-off. Sincethese variables were determined before the treatment decision, treatment statusshould have no effect on them.
If the predetermined variables show adiscontinuity at the treatment cut-off or in other points, we should not beconfident about the accuracy of our results and look at the RD design sceptically.Another problem might be that ofendogeneity. However, as already said, being municipalities subject to theimplementation of different electoral systems as a consequence of having adifferent number of inhabitants in their territory, reduces considerably theproblem of endogeneity, moreover since this rule derives from a national law. Aswe know, another key issue when using RD design is the choice of the bandwidth,since both precision and bias increase as the portion of bandwidth consideredincreases. In order to choose the most appropriate bandwidth, we can provideresults for different measures of it and see how sensitive the results are to thesedifferent adoptions (Lee and Lemieux, 2010).The biggest limitation of this approach is thatonly municipalities with single-ballot elections will be considered in theanalysis. This implies a restriction in the number of observations, but moreimportantly, it could mean that we are only considering those municipalitieswhere voters are less indifferent to certain choices than to others.
This together with the possibility of strategicbehavior of candidates and voters, as already discussed, would bias ourestimates.