19
between Spain and Portugal. While in the former case religious voters have adopted more distinctive voting
intentions as time has passed by, in the latter the irrelevance of religion has become even more apparent: in
2002, every religious group split evenly between the two large options. But nothing of the sort has happened in
the march 2004 general elections in Spain; in this case, nuclear Catholics (around 20 per cent of the total
population) opted for the PP, while Non-believers (around 40 per cent) preferred the PSOE.
The tables also confirm that religiosity is not bounded to ideology in Portugal in the same way that it is so
in Spain. For each of the percentages resulting after cross-tabulating religion and party choice, the table
displays a distribution according to ideological categories. For instance, the percentages shown in the fourth
row of Table 9 represent the ideological distribution of those nuclear Catholics that voted for AP in 1982 (28
per cent), and for the PP in 2004 (40 per cent). We are interested to see whether or not, for each different party,
the distribution adopts a different shape according to religion. Should this be the case, we could conclude that
ideology is closely linked with religion. Otherwise, the conclusion will be that ideological identities are formed
regardless religious identities, thus suggesting a definition of ideology free of value orientations. As expected,
the contrasts are striking. While in Portugal ideological distributions for each political party replicate the same
pattern across the different religious groups, in Spain the shape of this distribution seems to depend highly on
the degree of religiosity. An example based on PP voting on 2004 will illustrate the point. In the case of those
nuclear Catholics who voted for the PP, ideological conservatism (right and extreme right) amounted to 60 per
cent of the category. In contrast, only 47 per cent of those non-believers had voted for the PP were
ideologically conservative. Variations of this kind cannot be found in the Portuguese case.
Tables 9 and 10 about here
Concluding remarks
Portugal and Spain present an interesting counterpoint in terms of electoral behaviour. Despite evident
similarities anchored in history, culture, and geography, these two countries have come to develop different
scenarios for the development of religious voting. Indeed, our models seem to challenge the perception of
Iberian countries as split along religious lines. In the context of a recently completed transition period, these
elections took place in an environment were religious conflicts had been deliberately silenced. Moreover, the
identification with strong religious identities were also explicitly excluded in the strategies followed by party
leaders when building ex novo new party systems in both Iberian countries. Consequently, religious identities
were doomed to play only a secondary role in the definition of electoral alignments at the beginning of the
1980s. All that notwithstanding, our data confirm the existence of a religious voting of considerable strength in
Spain. Perhaps without becoming determinant, we have seen that religiosity, particularly in the case of the
diminishing group of nuclear Catholics, still help voters to opt among political parties.
Convention