Is direct democracy effective? Yes, if it is citizens who start the process

The EU referendum this week has sparked extensive debates about the pros and cons of direct democracy. But Lucas Leemann writes that landmark votes like this may be the most atypical – and arguably worst – examples of direct democracy. He indicates that in cases where citizens have the ability to launch initiatives and call for referendums, it can play an important role in resolving problems on non-redistributive issues.

Credit: Philippa Mckinlay CC BY-NC-ND 2.0

While landmark votes, like the UK referendum next week, attract a lot of attention and spur debates on direct democracy they might actually be the most atypical forms of direct democracy.

Direct democracy can enrich purely representative systems and have a beneficial impact on various outcomes. For most readers, the most relevant effect of direct democracy will be on policy congruence – i.e. the extent to which policies reflect the wishes of a majority of citizens. The optimistic promise of direct democracy is exactly that it will yield policies that are “more” democratic, in the sense that they are supported by a majority of the citizens.

We can distinguish two important types of direct democracy: one where governments decide to grant their citizens the right to vote on a specific measure and another where the citizens can force the government to hold such a vote. The question of who can start the direct democratic process may appear secondary at first sight but it is actually one of the most important institutional details.

When citizens can, via the collection of a sufficient number of signatures, force a vote on a subject, legislators and the executive live in constant fear of facing defeat at the ballot box. Drafting bills, debating them, and passing final legislation is a time and resource intensive process. If a law is eventually blocked at the ballot box, all legislative efforts are wasted. But when governments decide on whether a referendum takes place or not, that constant threat does not exist and legislative politics may be at times further away from the majority’s preferences.

In a forthcoming research article, co-authored with Fabio Wasserfallen, we look at when and how direct democratic institutions can increase policy congruence. We exploit the fact that all 26 Swiss cantons offer direct democratic institutions to their citizens, but that the costs of using these institutions varies considerably across cantons (e.g. signature requirements, maximal collection period). We focus on ten policies that are relevant issues in cantonal politics and run two surveys among citizens and legislators. Based on these surveys we can measure the preferences of the median voters and median ministers across cantons. The main result is that the effect of direct democracy is especially pronounced when the government and the citizens disagree strongly.

We then checked to see if actual policies matched citizen preferences because they had been changed by a direct democratic vote, but found no evidence of this. Consequently, we infer that it is legislators themselves who craft policies that are more aligned with the wishes of their voters – in fear that to do otherwise would simply see their laws overturned by a referendum. That is the indirect effect of direct democracy. As early as 1970, Leonhardt Neidhart wrote that the referendum is most successful when we do not observe it happening because politicians compromise. This is akin to speeding tickets that are a deterrent – they “work” even when they are not being issued, as long as the expectation of being caught is relatively high.

How does direct democracy otherwise impact on a political system? When citizens can start a signature collection and launch a referendum it is also possible that parties start relying on these institutions. A party’s goal might not always be purely policy-oriented – it is equally possible that parties pursue the direct democratic avenue because it allows them to force a debate on specific issues which might be electorally beneficial for them. In a research article last year, I show that over 100 years there has been a disproportional increase of direct democracy usage by parties and that these votes often concern issues that are politically less salient. If elections are fought mostly over economic questions and redistribution – and the electoral system provides a somewhat unbiased way of mapping votes into seats – there should be little difference between the median legislator and the median voter on economic issues. But on the less salient issues, for which attitudes amongst voters tend to not be correlated with redistribution preferences, one would expect more disagreement and greater opportunity for legislative minorities to push their agenda via the ballot box.

In the Swiss case has seen a disproportionate rise in votes pertaining to the cultural conflict axis – e.g. questions about reproductive medicine, LGTBQ issues, foreign policy and the EU, anti-immigration measures, and environmental politics. While these questions have an economic impact they are first and foremost about culture and identity. They illustrate the cleavage between two poles: the green alternative libertarian visions and traditional authoritarian and nationalistic convictions. While the specific questions might vary from country to country, they do all share a tendency to not be part of the primary dimension of political conflict.

Taken together, this line of reasoning implies that direct democracy is really only efficient if citizens have the ability to launch initiatives and call for referendums. The reason is that if citizens can potentially force a vote, politicians are constantly under that threat and so can be expected to pursue policies closer to the median voter’s preference. Actually having such a system will then mostly lead to votes on non-redistributive issues – at least as long as distributive issues are the main dimension of conflict. In that sense, direct democracy helps to resolve problems on the secondary conflict dimension.

In the context of the looming UK Brexit referendum, one final thought seems relevant for a discussion of referendums. When referendums are very infrequent they might not only be affected by the specific question at hand, but also come to be partly a vote of confidence in the government. This then might be the worst of all possible combinations of representative and direct democracy, conflating issue decisions with government popularity.

Note: This article represents the views of the author and not those of Democratic Audit or the LSE. Please read our comments policy before posting.

Lucas Leemann is lecturer for quantitative political science at University College London. He obtained his PhD from Columbia University where he majored in comparative politics and minored in quantitative methodology. His research in comparative politics focuses on institutional origins and direct democratic institutions. In Quantitative Methodology he is interested in both measurement (IRT, MrP) and modeling (mostly hierarchical). His articles have been published or are forthcoming in the American Political Science Review, the American Journal of Political Science, the Journal of Politics, Political Analysis, Electoral Studies, the Swiss Political Science Review.