How much do electoral systems matter, and does it matter?

It is not for nothing that the design of an electoral system is seen as one of the most consequential decisions for a new political system, nor is it surprising that decisions about electoral reform are taken very seriously. From Duverger and Rae onwards, countless studies have explored the consequences of electoral system design for how a political system operates. That electoral systems matter is generally treated as a given. But how much do they matter? There are undoubtedly some trends that meet theoretical expectations, such as that PR systems tend to have more political parties. But equally there are other trends that appear counter-intuitive – for instance, PR systems often produce stable governments; high levels of right-wing extremist party support are found both in PR and non-PR systems; female representation can be just as high in PR as non-PR electoral systems. And, research on electoral reform outcomes finds little evidence of the reform making much difference (Bowler and Donovan 2013; Renwick and Pilet 2016). This paper makes use of new cross-national data on electoral system consequences in 78 of the world’s democracies to, first, provide fresh evidence on the systemic and strategic consequences of electoral systems. In some respects our findings concur with previous studies, notably relating to the mechanical consequences of electoral systems on proportionality of electoral outcomes and the shape of the party system, but in a number of other respects – for instance, relating to ideological polarization, government stability, the rise of extremist parties, descriptive representation, levels of satisfaction with democracy, or income inequality – the picture is more mixed or nuanced. In short, electoral systems clearly matter in terms of a small number of (albeit significant) outcomes, but appear to matter a lot less in regard to a large number of other areas. Having provided evidence for why electoral systems do not matter quite as much as assumed, the paper then offers a number of explanations for this finding, paying particular attention to the agency of the key actors involved, and the question of wider institutional embeddedness. We conclude with a discussion of why, even if electoral system consequences are more limited than might have been assumed, understanding the increasing complexities of how, why and in what ways electoral systems matter continue to be significant questions for the research community.