Abstract
Passarelli’s book is about variants of list proportional systems which differ from closed list PR in that either the voter can potentially affect the order of candidates on the party list of the party they support; or voters can express preferences that are not restricted to a single party list. Among such systems, there are further variations, such as the number of positions on a list a voter may be able to affect, the degree of importance for outcomes of the preference ordering for the party list initially specified by the party itself, and whether the casting of a candidate preference is compulsory or optional. Although there has been work on which Passarelli draws classifying such systems, and a number of case studies, the consequence of such deviations from the more familiar closed list systems have not been well studied or well understood in a comparative perspective. So, a book length empirical treatment like this one - one that is genuinely comparative and theoretically grounded - is very useful.
Forms of list PR where voters have some influence on candidate choice and not just on party seat share are far more widespread than I had imagined - going well beyond the Finnish case and the pre-1993 use in Italy with which I had previously been familiar - and they are especially common in Europe. Passarelli’s comprehensive study includes Austria, Belgium, Brazil, Croatia, Cyprus, the Czech Republic, Denmark, Estonia, Finland, Greece, Italy, Kosovo, the Netherlands, Peru, Poland, Slovakia, and Sweden. In most of these countries, the study is of elections to the national parliament (usually the lower chamber), but in some countries the relevant elections are (or include) regional elections or elections to the European Union.
While Passerelli identified many possible implications of party list voting where voters may affect candidate choice (see especially: p.48), his main findings have to do with the impact of voter preferences on the ability of voters to affect who is chosen in any given election, the degree of intraparty competition, and incumbent re-election rates (taking into account any change in position on party list). Chapter 4 presents his findings on a country-by-country basis. Because he has national election data for 19 countries over the post-World War II period from 1945–2016, and data on EU elections from 1979–2014, he is able to track time trends.
Summaries and analysis of country-specific findings are in Chapter 5. One of the clearest findings is about systems where the power of voters to affect outcomes is limited (what he refers to as flexible list and latent list systems). In such systems, on average, of the 43% of the cases where an incumbent was not re-elected, retirement is the main reason for incumbent removal, accounting for nearly half the cases. In the 23.4% of the cases which were incumbent defeats, 25% (5.9/23.4) can be attributed to preference voting. The main reasons for incumbent defeat are changes in party support levels and changes in the internal ranking of the candidate on the party list (see Table 5.1). In contrast, in systems where candidate voting might, in principle, be more important (what he refers to as open list and quasi-list systems), retirements are about one-and-a-half times more common and, among incumbent defeats, preference voting explains more than half of the cases (see Table 5.2). Although these average values conceal a great deal of inter-country variation, the flexible list systems are ones where voters tend to have the lowest impact of incumbent turnover, with Croatia, Estonia and Bulgaria, key exceptions, while there are higher rates of preference vote defeat in the open list vote systems with the exception of Brazil, Chile and Poland (see Tables 6.3 and 6.4). Thus, the results go in the expected direction in that the more power voters have in principle, on average the more power they have in practice. (When I ran a t-test for this comparison, I got a p value of 0.015 using a one-tailed test; see also Table 6.1.)
Passarelli also runs regressions comparing the four types of list PR rules that allow for voter input into candidate choice in terms of incumbent retention levels and finds the expected relationship with types other than open list generating higher levels of incumbent retention. He also develops two indices that track the properties of these electoral rules from the standpoint of voters and of parties, respectively. For voters, he offers a composite measure of voter influence which takes into account whether voting is compulsory, whether expressing a candidate preference in compulsory, how many candidate preference votes it takes to changes list order, how many candidate preference votes are allocated. For parties, he provides an index that incorporates whether the candidate preference vote is the sole determinant of outcomes, whether the party list first choice is guaranteed selection if the party is represented in the constituency, whether other top candidates on the original party list are guaranteed to remain at the top of the party list, and the proportion of total votes cast (if any) needed to move a candidate up the preference list. In cross-national regressions with just these two variables, he finds both indices have the expected effects, but with the voter index of much greater importance, both substantively and in terms of statistical significance. Finally, when both the four election types and the two indices are entered into the same regression, three of the four election types retain statistical significance as does the voter index. In addition, he treats the two indices as axes for a two-dimensional representation of the structure of party list systems allowing candidate choice.
In Chapter 6, Passarelli also examines more briefly potential systemic effects of his range of electoral rules on voter turnout, the degree of fragmentation of the party system, and voters’ perceptions of the importance of candidate characteristics. The expected directionality of effects is found for each of these variables, e.g. less fragmentation the more there is candidate choice.
In the preface, Passarelli explains how he became interested in electoral rules in observing that the pre-1993 Italian system of open list proportional representation, as it operated in practice, involved the parties (including the Christian Democrats and the Communists) telling their voters the order of candidates to place on the ballot: ‘I always doubted that preferential voting genuinely conferred any decisive power on voters, at least in Italy. Nevertheless, I needed more sophisticated information than my intuition to answer this question, and I needed to compare more than one case to have a larger picture of the variant ways in which preferential voting can manifest itself. The evidence gathered here confirms my intuition, as well as the importance of investigating how voters’ freedom and power to affect electoral outcomes and parliamentary turnover can vary widely across preferential voting systems’ provisions.’
In sum, this book provides a thoughtful and exhaustively researched study of list PR systems which permit candidate choice, and one that should be on every specialist’s bookshelf. Moreover, more generally the points the author makes about the limited ability of voters to translate their preferences into genuine influence on outcomes are very important for students of empirical democratic theory.
My only real quibble with Passarelli is with his title. The book title is a very general one, preferential voting, but in actuality it is only about preferential voting in list PR systems. While he offers a classification scheme that justifies excluding systems such as the single transferable vote (STV) - see discussion on this point on pp. 26–46 - given standard usage, the title remains mis-leading as to the scope of the study...