Who Abstained and Why? Voter Turnout for the Referendum on the Treaty to Establish a European Constitution

Who Abstained and Why? Voter Turnout for the Referendum on the Treaty to Establish a European Constitution

Theme: This analysis examines the voter turnout registered in the referendum held on February 20, 2005 to approve the treaty to establish a European Constitution.

Summary: A voter turnout rate of 42.3% was registered in the referendum on the European Constitution. Analysis of the available data leads to the following conclusions. First, territorial distribution patterns were very similar to those of the latest European elections; once again, abstention was greater in the periphery of the country than in the centre. Second, voter turnout revealed a clear sociological profile: abstention was greatest among women, people under 25 years of age, people with the lowest educational levels, the unemployed and housewives. Abstention can thus be interpreted more as a sign of indifference than of rejection. Third, turnout was highest among voters sympathetic to leftist ideology, those who think the government is doing a good job and those who trust Prime Minister Rodríguez Zapatero. Fourth, an unusually high number of blank and spoiled ballots were cast. For lack of more precise post-electoral studies, abstention can be explained as a result of: the fact that the public knew very little about the very complex issues being voted on; the predictability of the result; the attitudes of political parties; and the conflicting pressures and contradictory messages that voters received.

Analysis

Level of Voter Turnout

As soon as this referendum was called, attention focused on voter turnout, which, it was assumed, would be low. Since a ‘Yes’ vote was virtually certain, the number of voters who would turn out to vote became the acid test for the treaty. The result was hardly a surprise, given the pessimistic predictions. On February 20, 14,204,663 voters cast ballots –42.3% of those eligible–. Only a few months before, 15,666,491 voters –44.9% of those eligible– had turned out to vote for their representatives to the European Parliament. Voter turnout was low, as expected, but not much lower than in the European parliamentary elections, to which the referendum must inevitably be compared.

At first sight, this high level of abstention seems to affect all European elections and referenda in a similar way, and not just in Spain. The level of voter turnout in this referendum seems to confirm a well-defined and worrying trend in which turnout for European parliamentary elections has dropped off very sharply. While in 1979 the average turnout in the former European Community as a whole stood at 63%, there has been a clearly downward trend since then, reaching 46% in the last election. The new member countries, with the exception of Cyprus and Malta, have pushed average turnout down even further.

Figure 1. Turnout for the 2005 Referendum and for the 2004 European Elections (Abstention and Voters)

Source: the author, based on Ministry of the Interior figures.

Figure 2. Trends in turnout for European elections in Spain and in the E.U. (1984 = 1986 elections in Spain)

Source: the author, based on European Parliament and Ministry of the Interior figures.

The exceptional mobilization of Spanish voters for the European elections in 1987 (69%) and 1999 (64%) was undoubtedly due to the fact that they coincided with the municipal and regional elections, which were held at the same time. When the ‘voter drag’ effect of this coincidence was no longer present, Spanish turnout, which had been 15% higher than the European average in 1999, dropped to slightly below the European average in 2004.

The territorial distribution of voter turnout

The key question, therefore, is whether this referendum on the European Constitution represents simply one further step in the declining turnout for the European electoral process, or if it is in some way specific or unique. To answer this question, it may be useful to examine some features of the territorial distribution of voter turnout in order to determine whether turnout followed the same territorial patterns as in other elections, or if these were in fact different.

As is usually the case, Spanish voting behaviour revealed significant territorial differences. Voter turnout has tended to be higher in the central regions of the Iberian Peninsula and lower in the surrounding regions. This general trend held true in the referendum on the European Constitution. As Figure 3 shows, turnout was below 30% in Melilla and Ceuta, between 33% and 39% in the two island communities (Balearics and Canaries), the Basque Country and Asturias, and below the 42.3% average in Andalusia, Catalonia, Murcia and Navarre. By contrast, turnout was close to 50% in Extremadura, Castilla-León and La Rioja. Galicia, a community with traditionally higher than average levels of abstention, registered participation higher than the national average, thus continuing its tendency to converge with average turnout levels. The territorial distribution of turnout therefore reveals nothing significantly different from what is seen in Spanish electoral turnout in general.

Figure 3. Turnout for the Referendum by Autonomous Community

Source: the author, based on Ministry of the Interior data.

Table 1. Correlation between Voter Turnout, Blank and Spoiled Ballots, and Votes in Favour

% Blank Ballots% Spoiled Ballots% Votes in Favour
% vote0.400.120.23
% blank ballots0.19-0.22
% spoiled ballots-0.06

N = 52 (provinces).

Source: the author, based on Ministry of the Interior data.

Voter turnout seems to be loosely related to voting in favour of the treaty. Voter turnout was higher in provinces where the result was more homogeneous, that is, with a higher percentage of ‘Yes’ votes (the correlation between turnout and votes in favour is 0.23). Therefore, it seems that abstention was higher where more ‘No’ votes were cast. This may be due to the fact that abstention is usually higher and opposition to the treaty is also greater in the peripheral regions. Another possible explanation is that a more energetic ‘No’ campaign in certain areas discouraged potential voters who were already hesitant to vote in favour, and who ended up abstaining.

Blank Ballots

What does seem clear is that wherever turnout was high, a relatively large number of blank ballots were cast (in this case, the correlation is 0.40). More than 6% of voters cast blank ballots –a figure surpassed only in the referendum on NATO–. It seems there is a type of undecided voter who abstains in some areas and casts blank ballots in others. Catalonia is at the top of this list, with over 7% blank ballots.

Figure 4. Blank Ballots

Source: the author, based on Ministry of the Interior data.

Comparison with European Elections

If we compare the difference between voter turnout in this referendum and in the European elections of 2004, the similarity between the two is confirmed. Province by province, the correlation is 0.90, with certain demobilisation in some communities (Murcia, Madrid, Cantabria, Valencia, Asturias, the Basque Country and Castilla-La Mancha). Compared with the last European elections, there was a very small increase in only three communities: Extremadura, the Canary Islands and Catalonia. There is no obvious explanation for these territorial differences. An analysis of the two most extreme cases, Murcia and Catalonia, illustrates this. Murcia is a community where there was very little political support for the ‘No’ vote, since both the main parties present in the Region’s governmental institutions were in favour of the Treaty. This is also a region where the Popular Party (PP) is clearly dominant and whose efforts to mobilise voters were not comparable to those made in the campaign for the European elections. In Catalonia, the ‘No’ campaign was much more intense, which may have motivated more voters.

Figure 5. Differences in Voter Turnout Compared to the 2004 Elections to the European Parliament

Source: the author, based on Ministry of the Interior data.

Sociological Profile of Abstention: Who Abstained?

The low turnout, which followed a trend very similar to that of last European elections, was largely due to the usual factors: voters’ socio-economic conditions and political attitudes. Although a post-election study would be necessary for a more precise analysis, the pre-election study closest to election day (the CIS barometer in January) shows that, in line with tradition, those with least intention of participating were women, people under age 25, people with low levels of education, the unemployed and housewives.

The relationship between participation and age follows the typical curve: minimal among young people (37%), increasing to age 65 (50%), and dropping off slightly among the oldest segment of the population (47%). The connection with educational level is surprisingly intense: clearly, the groups with the lowest levels of education had the least intention of voting. There is also a classic, but still uncommon urban/rural trend: intention to vote was greatest in the bigger cities (56%) and lowest in the medium-size cities of 100,000-500,000 inhabitants (34%). Small municipalities, meanwhile, fell between the two. In terms of occupation, workers (48%), retired workers (52%), and students (45%) declared the greatest intention to vote, compared to housewives (38%).

Figures 6-10. Sociological Profile of Intention to Vote (by Age, Gender, Educational Level, Population of Municipality of Residence, Occupational Situation)

Data on intention to vote.

Source: the author, based on CIS study 2589.

Indifference or Rejection?

The typical voter who abstained from this referendum clearly belongs to the periphery in socio-economic terms, making it plausible to suggest an attitude of indifference, rather than rejection. Several factors may come to bear on this: voters’ lack of knowledge of the subject and its complexity, combined with the predictability of the result.

All the surveys carried out prior to the referendum emphasised Spaniards’ lack of knowledge about the treaty. In the January CIS barometer, more than 92% of respondents admitted that they knew little, very little, or nothing about the treaty. The campaign, at least to that point, did little to change this situation. Intention to vote is closely related to understanding of the issues: while more than 70% of those who said they were familiar with the Constitution stated their definite intention to vote, only 36% of those who knew nothing about it said they would definitely vote. As is predictable, voters who valued the treaty’s contribution to European integration and who agreed that this referendum should be held, expressed greater desire to participate.

In addition to the inherent complexity of the wording of the treaty, there was the added difficulty of communicating the necessary developmental analysis, that is, situating the treaty within the long and complex process of European integration. This referendum also involved voting on another unknown: the creation of a new political entity. All this resulted in a perception that voters were going to the polls ‘without knowing very much’, which may have favoured abstention, since the perceived cost of obtaining the information was very high. Although party slogans can help in situations like these, they are not always enough to mobilise voters, especially if there are ambiguous messages –Are we voting on the Treaty? On Europe? On the government?– or contradictory pressures.

Figure 11. Trends in Intention to Vote in the Referendum

(Sure to Vote, In Favour, Against, Blank or Spoiled Ballot, Unsure)

Source: the author, based on CIS data.

A second factor that favours abstention is the fact that the referendum had a predictable result, in terms of the percentage of ‘Yes’ votes. All the polls gave a wide margin of victory to the ‘Yes’ campaign, while the ‘No’ side lagged far behind with a percentage similar to that of the blank votes. Voting in favour of the treaty was the option promoted by the two main national political parties (PSOE and PP, though with differing emphasis), by the two most important nationalist parties (CiU and PNV), and by social actors influential in mobilising voters, such as the trade unions. The ‘No’ vote was supported by the left (IU, IC-V) and the other nationalist parties. Only the Catholic Church called for or justified abstention. Such a lack of competitiveness does not generally lead voters to vote en masse. In fact, as we can see in Figure 11, the percentage of voters who said they would definitely vote dropped as the number of decided ‘Yes’ votes rose in the course of the campaign.

The Ideological Profile of Participating Voters

Participation in this referendum also has a relatively clear political profile. In this case, the left seemed somewhat more motivated than the right. Perhaps this was because the left was split between the two possible options, although this effect disappears in a multivariate analysis. People who voted PSOE and IU-ICV in the general elections of March 2004 had the greatest intention to vote in the referendum. The number of available cases is too small to be able to reliably interpret the case of regional political parties, but it does appear that the intention to vote was even lower among those who had previously voted for nationalist parties.

Confidence in the main political leaders and in how they were running the government and opposition, respectively, also significantly affected participation in the referendum. For some, confidence in Zapatero was a strong motivating force: while 67% of those who had great confidence in Zapatero said they would vote, this figure was only 37% for those who had no confidence in him. The effect of confidence in Rajoy is much more tenuous and disappears in a multivariate analysis: those who have the greatest confidence in him declared 57% intention to vote, while this figure was 43% among his detractors. Voters’ impressions of the government’s performance also had a strong impact on participation: over 70% of voters who had a good perception of this said they would vote, compared to only 34% of those with a bad impression. In the case of the PP, if we ignore the top category (which contains few cases), there is a tenuous inverse relationship between positive impressions and intention to vote, which disappears in multivariate analyses (not shown here).

Although there is not enough specific data to analyze the relationship between voter turnout and nationalist attitudes, there seems to be a clear indication that identification with peripheral nationalist ideology reduced the likelihood of voting in this referendum.

Figures 12-15. Political Profile of Intention to Vote (by Votes Cast in the March 2004 Elections, by Ideological Self-definition, by Confidence in Political Leaders, by Impression of Government)

Data on intention to vote.

Source: the author, based on CIS study 2589.

This political profile in which voter participation is associated with leftist ideological stances, a positive impression of the government and confidence in the leader (Zapatero) is quite unusual, according to some studies that affirm that the left has a greater tendency towards abstention. The reason for this unbalanced voter turnout may be found in the parties or in citizens –factors that certainly are related–. In terms of parties, the differences in voter turnout that respond to political orientation may be considered a reflection of the differing amount of effort made by political parties of different ideologies –effort that was more intense or at least clearer on the left than on the right–.

In terms of the voters, differences in turnout may have been motivated by conflicting pressures. Voters who sympathize with the PP may have chosen to abstain before giving a favourable vote to a referendum called by a PSOE government. Nationalist party voters who supported the Treaty (PNV and CiU) may have been sensitive to the argument that the constitutional project did not sufficiently recognise the role of distinct ‘peoples’. Finally, voters for parties that opted for ‘No’ (Iu, Ic-V, Erc, Ea and ChA), with strong pro-European convictions, may also have been subject to contradictory pressures, finally opting to abstain. These contradictions, which have been felt within the parties themselves (many of which went through intense debates before finally determining their positions), may have had an impact on voter turnout. Post-election analyses will confirm or disprove this hypothesis.

Conclusions: Voter turnout for the referendum on the European Constitution was 42.3% –slightly below the turnout for the European elections in 2004–. Territorial distribution trends were very similar in both cases. A positive correlation has been detected between voter turnout and the ‘Yes’ vote, and particularly between turnout and the number of blank ballots cast. Pre-election polls show that participation in this referendum has a clear sociological profile: intention to vote was lowest among women, people under 25 years of age, people with the lowest educational levels, the unemployed, and housewives. This may be partly explained by voters’ lack of knowledge of the subject and its complexity, combined with the predictability of the result. This suggests that abstention was more due to indifference than to a desire to protest. Political stances also had clear effects: voter turnout seems to have been favoured by leftist ideological stances, a positive impression of the government, and confidence in Prime Minister Rodríguez Zapatero. Although the studies done on voter intentions cannot confirm it, abstention may also have responded to conflicting pressures on voters, who had to deal with contradictory arguments when deciding how, or whether, to vote.

Eva Anduiza Perea

UAB