Governments can be structured quite differently in different countries. For example, while some emphasise the role of parliaments, others might have elected presidents. But there are also systems that combine these elements. What effects do different institutional set-ups have on democracy and government performance? This blog post investigates this question with a focus on the two sub-types of semi-presidentialism.
In the last decades, semi-presidentialism has become a widespread choice among constitution makers around the world and the study of semi-presidentialism has emerged as a burgeoning research field in its own right. After the fall of communism, it has become the dominant political system in Central and Eastern Europe and in the post-Soviet countries. In general, semi-presidentialism could be viewed as a mix – or a compromise – between the two main form of constitutional systems; parliamentarism and presidentialism. Leading scholar Robert Elgie defines semi-presidentialism as a political system where the constitution includes both a popularly elected president and a prime minister and cabinet accountable to the legislature. Employing this definition there are more than 50 countries with some kind of semi-presidential constitution.
For quite some time, scholars have discussed the pros and cons of semi-presidentialism. Some scholars have endorsed it as a rather flexible and power sharing system, while others have associated it with institutional conflict and political instability. It has illustratively been described as a constitutional system that is “easy to choose but difficult to operate”[i]. In line with more recent scholarship, we argue that in order to empirically assess the performance of semi-presidential regime, we need to acknowledge the diversity within the family of semi-presidential regimes and also that such assessments should be based on comparison to parliamentarism and presidentialism.
So, do semi-presidential regimes perform worse than other regime types? We argue that it certainly depends on what type of semi-presidentialism we are talking about. In line with Shugart and Carey we distinguish between two different types of semi-presidentialism – premier-presidential and president-parliamentary regimes. Shugart and Carey define premier-presidentialism as a system where (1) the president is elected by a popular vote for a fixed term in office; (2) the president selects the prime minister who heads the cabinet; but (3) authority to dismiss the cabinet rests exclusively with the parliament, and president-parliamentary systems where (1) the president is elected by a popular vote for a fixed term in office; (2) the president appoints and dismisses the prime minister and other cabinet ministers; (3) the prime minister and cabinet ministers are subjected to parliamentary as well as presidential confidence. The key institutional relations of the four regime types are illustrated in the figure below.
Popular elections and cabinet survival under different regime types
A substantial number of studies have empirically assessed the functioning and performance of semi-presidentialism. With the notable exception of Elgie, however, there is a lack of large-N studies where democratic and government performance are measured across the two subtypes of semi-presidentialism (premier-presidential and president-parliamentary regimes). Elgie’s systematic and comprehensive study offers several important findings on the performance of the two types of semi-presidentialism, but it does so in isolation from parliamentary and presidential regimes. Our study is an attempt to address this issue.
By using indicators on regime performance and democracy from a dataset containing 173 countries, we examine the performance records of premier-presidential and president-parliamentary regimes in relation to parliamentarism and presidentialism.
Taking our departure in Linz’s argument on the “perils of presidentialism”,[ii] and by Matthew S. Shugart and John M. Carey’s proposition that president-parliamentary regimes are more perilous to democracy than other regime types, we test three expectations.
- Parliamentarism performs better than other regime types in terms of democracy and government performance.
- Premier-presidentialism performs better than president-parliamentarism and presidentialism in terms of democracy and government performance.
- President-parliamentarism performs on a par with, or worse, than presidentialism in terms of democracy and government performance.
We use four frequently used indicators to measure democratic performance: A combination of Freedom House’s index of civil liberties and political rights and Polity IV; Polity IV on its own; The Economist Intelligence Unit’s Index of Democracy; and the Executive Constraints indicator from Polity IV, which refers to the extent of institutionalized constraints on the decision-making powers of chief executives. For measuring government performance, we use the Government Effectiveness indicator from the Worldwide Governance Indicators, the Corruption Perceptions Index from Transparency International, the Empowerment Rights Index from CIRI Human Rights Data Project, and the Human Development Index from UNDP.
Following a series of descriptive reports, we run analyses with a conventional set of controls including GDP/capita, population size, ethnic fractionalization, proportional representation, and different world regions.
Overall, our findings do not support the proposition that parliamentarism performs better than all other regime types in terms of democracy and government performance (expectation 1). Rather we observed a pattern where premier-presidentialism performs almost as good – and on some measures even better – as parliamentary regimes. Neither the measures of democracy nor the measures of government performance show significantly better records for parliamentary regimes than for premier-presidential ones. This indicates that a parliamentary constitution with an indirectly elected president does not necessarily go along with better political performance than a premier-presidential one with a popularly elected but weak or medium weak president. Thus, to the extent that we think about semi-presidentialism in terms of premier-presidential regimes, we have reasons to question strong propositions about the “perils of semi-presidentialism”.
However, the picture certainly looks different with regard to president-parliamentary regimes. While premier-presidential regimes are closer to parliamentary regimes, president-parliamentary regimes display performance records more similar to pure presidentialism, and it performs even worse on most indicators (expectations 2 and 3). When it comes to the level of democracy, the only regime type to perform significantly worse than the parliamentary one – on four separate measures and with conventional controls – is the president-parliamentary regime type.
Regime types and democratic performance (OLS coefficients with 95% confidence intervals)
The differences in terms of government performance are less pronounced. Although there is a tendency of slightly poorer performance by presidential-parliamentary regimes also in terms of government performance, and significantly so on one indicator, our results demonstrate that the type of constitutional system seems to affect democracy more strongly than government performance.
Regime types and government performance (OLS coefficients with 95% confidence intervals)
Shugart and Carey’s general recommendation to stay away from the president-parliamentary form of government certainly finds support in our data. In our study, we mostly refrain from making claims about causal mechanisms behind the observed pattern. However, we allow some general comments on the importance of presidential powers in relation to the four regime types. We show how variation in presidential powers follow closely the four regime types – weakest among the parliamentary regimes and strongest among the president-parliamentary regimes. We know that case studies on e.g. post-Soviet countries where the system has shifted from president-parliamentary to premier-presidential constitutions provide additional support to the negative impact of president-parliamentarism on democracy. For instance, Elgie and Moestrup show that reduced presidential powers and a shift to a more balanced semi-presidential system have been associated with better democracy records in e.g. Armenia, Georgia, and Kyrgyzstan. A general trend among the post-Soviet countries is that the presidents have used their control over the administration to curb the opposition and thereby directing the trajectory of constitutional developments in their own favor. The outcome has been increased power of already powerful presidents – a straight road to the consolidation of autocracy.
While we recognize some drawbacks of our study and encourage future research to continue researching this issue, we reveal a general pattern with regard to the four regime types on performance. Based on our findings, we claim that democratic performance is likely to be better with a parliamentary or premier-presidential form of government. If the most positive accounts about semi-presidentialism are relevant, such as executive flexibility, power-sharing, and a uniting president, those are most likely to be identified under the premier-presidential form of government. Our data give no support for general recommendations to avoid dual executives or popularly elected president with limited powers.
Finally, and well in line with more recent scholarship, we argue that discussions about the pros and cons of semi-presidentialism should include the distinction between its sub-categories as well as considering dimensions of presidential power.
Kilde: Sedelius, Thomas and Jonas Linde (Forthcoming), «Unravelling semi-presidentialism: democracy and government performance in four distinct regime types«, in Democratization. The full text article is free to download here.