Category Archives: Semi-presidentialism

Semi-presidentialism, premier-presidentialism and president-parliamentarism – A new country-years dataset

This new dataset provides time-series, cross-sectional data for the presence of both semi-presidentialism and the two sub-types of semi-presidentialism – premier-presidentialism and president-parliamentarism – since 1900. The dataset uses the same country names, country years, and country ids. as the V-Dem data set, allowing them to be easily merged.

The dataset (v2.0) is available here.

There are two codings of semi-presidentialism in v2.0.

In sp1, semi-presidentialism is defined as the situation where a country’s constitution establishes both a directly (or popularly) elected president and a prime minister and cabinet that are collectively responsible to the legislature (Elgie 2011). This coding includes cases where a constitution requires a super-majority for the dismissal of the prime minister and cabinet by the legislature.

In sp2, semi-presidentialism is defined as the situation where a country’s constitution establishes both a directly (or popularly) elected president and a prime minister and cabinet that are collectively responsible to the legislature by no more than a vote of an absolute majority of one or more houses of the legislature. In other words, this coding excludes cases where the PM and government can be held collectively accountable only through a super-majority vote in the legislature.

In sp1, the following countries are classed as semi-presidential, whereas in sp2 they are not: Algeria (all years), Burkina Faso (1977-80), Burundi (1992-96), Cameroon (all years), Central African Republic (2016), Egypt (2007-11), Kyrgyzstan (1996-2007), Madagascar (all SP years since 1996), Mali (all years), Republic of Congo (2016), Rwanda (all years since 2003), Togo (all years), Tunisia (1989-2001), and Vietnam (all years).

The presence of semi-presidentialism (both sp1 and sp2) is coded as 1, its absence as 0. The start year is the year of the introduction of semi-presidentialism in the constitution if the date is on or before 30 June. If the start date is 1 July or later, then the following year is recorded as the first full year of semi-presidentialism. The end date is recorded for the year that the constitution ceased to be semi-presidential at whatever point in the year it ended. The end of semi-presidentialism is marked by a constitutional change. This can be a constitutional amendment introducing another type of system, or a suspension of the constitution.

This version also codes the premier-presidential and president-parliamentary sub-types of semi-presidentialism. The definitions are:

President-parliamentarism is a sub-type of semi-presidentialism where the prime minister and cabinet are collectively responsible to both the legislature and the president.
Premier-presidentialism is a sub-type of semi-presidentialism where the prime minister and cabinet are collectively responsible solely to the legislature.
These sub-types were first identified by Matthew Shugart and John Carey. The above definitions are consistent with Shugart and Carey (1992).

In the dataset, pp1 and pp2 code premier-presidenetialism as 1 and president-parliamentarism as 2. If a country is not semi-presidential, then the coding is 0. All pp1 codings are based on the definition of semi-presidentialism in sp1. All pp2 codings are based on the definition of semi-presidentialism in sp2.

If there are any mistakes, then please let me know ( If there are any questions, please contact me at the same email.

Please cite the dataset as:

Robert Elgie (2018), Semi-presidentialism, premier-presidentialism and president-parliamentarism – A new country-years dataset [Blog post, 3 April]. Retrieved from


Elgie, R. (2011), Semi-presidentialism: Sub-Types and Democratic Performance, Oxford: Oxford University Press, 2011.

Shugart, M. S. and J. M. Carey (1992), Presidents and Assemblies. Constitutional Design and Electoral Dynamics, Cambridge: Cambridge University Press.

Weaker presidents, better semi-presidentialism?

This post was first published at Presidential Power


Sophia Moestrup and I have just published another edited volume on semi-presidentialism. This time the focus is on Semi-presidentialism in the Caucasus and Central Asia. There are contributions from Alex Baturo on vertical power in the post-Soviet space, Alexander Markarov on Armenia, Jody LaPorte on Azerbaijan, Malkhaz Nakashidze on Georgia, Dmitry Nurumov and Vasil Vashchanka on Kazakhstan, and Matto Fumagalli on Kyrgyzstan. Sophia and I contribute two chapters. The first addresses some misconceptions about the notion of semi-presidentlaism, such as the idea that semi-presidential regimes must have quite powerful presidents but never very powerful or very weak presidents, and also that autocracies cannot be semi-presidential – they can, not least because semi-presidential regimes do not have to comprise only countries with quite powerful presidents. Our second chapter sums up the contributions to to the volume and argues that weaker presidents make for better semi-presidentialism. This is a brief summary of this second chapter.

The main attraction of institutional analysis is that it has the potential to generate better political outcomes. Given the assumption that institutions matter, we may be able to craft them so as to mitigate or even eradicate some of the negative outcomes that would otherwise be caused by the behaviour of political actors. We wish to draw one institutional policy recommendation from this book. All else equal, countries with weaker presidents are likely to experience better outcomes than countries with stronger presidents.

There is evidence from Armenia, Georgia, and Kyrgyzstan that weaker presidents have been associated with better outcomes. In Kyrgyzstan the decline in the president’s constitutional powers has been dramatic. That said, the shift to a weak president is relatively new, dating back to 2010. Kyrgyzstan also has a history of democratic reversals. So, we should avoid any definitive judgement at this early stage. More than that, the shift occurred in the context of the collapse of the previous regime and the desire on the part of the constitution builders to trammel the power of the presidency, which was seen as one of the main obstacles to democratic consolidation under the previous regime. This suggests that any positive effects of the weak presidency may be endogenous to the choice of the new institutional framework. All the same, we note that the early period of the new constitutional framework has been marked by less presidential posturing, less executive/legislative conflict, and, for now at least, less democratic backsliding. These are positive signs.

In Armenia, the decline in presidential power has been less dramatic. The president’s constitutional powers were never as great as the other countries in the region. Moreover, even after the passage of the 2005 reforms, the president still enjoyed some not inconsiderable constitutional powers. What is more, as in Kyrgyzstan, the context in which the president’s powers were reduced in 2005 means that we have to take account of the problem of endogenous institutional choice. Further still, Armenia remains a hybrid democratic regime in which there is plenty of political competition, but where democratic procedures have been manipulated to the advantage of incumbent power holders, although perhaps less so in the most recent elections than previously. In this context, we have to be careful about any lessons that we might we wish to draw from the Armenian case. Even so, we might benefit from thinking counterfactually. What would be the situation if there were now a super-president in Armenia? Would the situation be worse? We cannot know. Yet, we do know that in practice there was a form of super-presidentialism after the passage of the 1995 constitution. We can also confidently claim that this period marked the low point of democratic performance in Armenia to date. Armenia has not experienced a weak presidency, but it has experienced very strong presidents. It is not unreasonable to conclude by comparing the experience of the 1995-2005 super-presidency and the post-2005 period that the latter was less problematic.

By far the strongest evidence, though, comes from Georgia. Here, there were two periods when the problem of endogenous institutional choice was at least partly offset because of a dramatic change in the political context. In the first period there was a very strong president. In the second period, there was a very weak president. In this latter case, it is tempting to think in terms of quasi-experimental conditions. In the same historical, cultural, economic, and social context, there was an institutional treatment, namely the move to a weak presidency. The result has been much better political performance. The period of cohabitation under the previous president-parliamentary form of semi-presidentialism was marked by intense president/prime ministerial conflict as well as conflict between the president and the government and legislature generally. By contrast, the recent period under the weak presidency and a premier-presidential form of government has, to date, been characterized by much calmer relations. Indeed, this latter period is doubly interesting because the president distanced himself from his former political allies immediately after his election. The resulting situation should not be classed as a period of cohabitation, but it is certainly not a period where the president’s loyalty to the ruling party has quashed, perhaps artificially, any political competition within the executive branch. While there have been major disagreements between the president and the government, they have not become regime threatening. Indeed, arguably, post-2013 president/government relations in Georgia resemble those in countries like the Czech Republic or Slovakia where weak but directly elected presidents act as a counterweight to the government, but where there are no serious attempts to assume real presidential power.

If we are right to conclude that weaker presidents are better presidents, then we also wish to assert that the party system is an important intervening variable, as indicated above. It is perhaps no coincidence that in Georgia there has been a solid parliamentary majority since 2013. In other words, the president has not had the opportunity to try to offset his weak constitutional powers by building an alternative and potentially destabilizing pro-presidential coalition within the legislature. We might add that there has also been a relatively stable legislative majority in Kyrgyzstan since the 2010 reforms. Again, the president has not had the incentive to craft a majority that is personally loyal to him and that often requires the distribution of state resources in a geographically skewed and perhaps even corrupt way. In Armenia, by contrast, presidents have not always enjoyed a parliamentary majority and have been forced to forge coalitions in the legislature. This perhaps helps to account for the continued presence of a patronage president in a way that harms the rational functioning of the regime and democratic performance. Indeed, the recent constitutional reform that will introduce a parliamentary system after the next electoral contests might confirm this suspicion. The introduction of a parliamentary system and a weak president should be a positive development on the basis of our logic, but it may merely be a way of maintaining patronage politics in the context of an uninstitutionalized party system.

So, we acknowledge that many economic, social, and political factors affect political performance. We also believe that the party system is a particularly important variable for determining the practice of presidential politics. Even so, we claim that political performance is likely to be better when presidents have fewer powers. This suggests that constitution makers should consider the benefits of reforms that reduce the power of their presidency. We are aware that our conclusion assumes that institutions matter and, therefore, is susceptible to the problem of endogenous institutional choice, but we would like to address the endogeneity problem by arguing that even endogenously chosen weak presidents are better than endogenously chosen strong presidents. In other words, we believe that there are benefits to be gained from the endogenous selection of weak presidents. We should endeavour to create the conditions for decision makers to calculate that their system would benefit from a weak presidency. Fundamentally, if we are right that weak presidents bring benefits, we are unconcerned whether this outcome comes about endogenously or exogenously. That said, even if institutions are chosen endogenously, political actors still have to interpret the institutions with which they are faced. At some point, the economic, social, or political context is likely to change. At that point, if not before, institutions may have at least a partly exogenous impact. In those circumstances, it is better to have a weak presidency in place than a strong one. In other words, we would encourage upstream efforts to create the conditions for a constitutionally weak president. We believe that there are benefits to be gained from a system in which actors are willing to work without the presence of a super-president and that these benefits are likely to be both endogenous to institutional choice and at some point exogenous too.

We wish to make one final point. We promote the idea of a weak presidency, but we also wish to promote a weak presidency in the context of a wider constitutional and political system in which there is a genuine separation of powers and checks and balances. For example, we are not convinced that there are benefits to be gained from replacing a system in which there is a super-president and a weak prime minister by one where there is a weak president and a super-prime minister. This merely shifts the problem. It does not replace it. And it may characterise what is about to happen in Armenia. Let us express this point differently. We are not opposed to weak but directly elected presidents. As we argue in our introduction to the volume, semi-presidential constitutions are consistent with both very strong and very weak presidents. We prefer the latter. Let us make the same point in another way. We do not believe that parliamentarism with a weak but indirectly elected president is necessarily a guarantee of better political performance if there are no checks on the prime minister in the parliamentary system.

To sum up, we are happy to recommend a directly elected president as long as the president’s powers are weak and are exercised in the context of a system in which power is not concentrated in any political actor.

Congo-Brazzaville – Return to semi-presidentialism

This is a cross-post from Presidential Power

In October 2015 the Republic of Congo – Congo-Brazzaville – held a referendum to ratify a new constitution. In a country with a 2016 Freedom House rating of Not Free, it is unsurprising that the new constitution was overwhelmingly ratified. The official figures show that turnout in the referendum was 72.4% and that 92.3% of those voting supported the new constitution.

The immediate motivation for the reform was to secure the legal position in power of the incumbent president, Denis Sassou Nguesso. Sassou came to power in 1997 following a civil war, which brought an end to Congo-Brazzaville’s brief five-year flirtation with a semi-presidential democracy. In January 2002 a new constitution was passed. This constitution reintroduced direct presidential elections and stipulated a two-term presidential limit. In March 2002 President Sassou was elected president with over 89% of the vote. In 2009 he was re-elected with nearly 79%. With a new presidential election due to be held in March 2016, President Sassou needed to find a way to remain in power in a manner that conformed with the law. The solution was to adopt a new constitution that, if passed, would mean that President Sassou had not served any terms under the new regime. With the counter reset to zero, he was free to stand for election in 2016. As Sophia Moestrup reported in a previous post, with the new constitution in place President Sassou was duly elected at the first round of that election, winning 60.4% of the vote.

While the main motivation for the new constitution was to maintain President Sassou in power, the reforms were quite wide ranging. For example, even though the president is still limited to serving two terms, the presidential term was itself reduced from seven to five years. The age limit for presidential candidates, which had been 70 under the old constitution, was also abolished. President Sassou was born in 1943. The text of the new constitution is available here.

One element of the new constitution was a return to semi-presidentialism. Art. 98 re-established the position of prime minister, which had been present from 1992-1997, but was abolished from the constitution when President Sassou assumed power. In addition, Art. 100 states that the prime minister is responsible to the National Assembly. There is no investiture vote. Indeed, Art. 103 makes this explicit. However, Art. 159 states that the prime minister can call for a motion of confidence. If it fails, then the government has to to resign. The National Assembly can also table a vote of no-confidence, which if passed means that the government has to resign. In short, the new constitution is clearly semi-presidential.

The text of the constitution also makes Congo-Brazzaville an example of a president-parliamentary form of semi-presidentialism, though there is some ambiguity. In the original French, Art. 83 reads “Le Président de la République nomme le Premier Ministre et met fin à ses functions.” This is usually translated into English as: “The President of the Republic shall appoint the Prime Minister and shall terminate his term of office”. This is clearly designed to allow the president to sack the PM if necessary. This point becomes clearer if we compare it with the original wording of the 1958 French constitution, which once again is the model for a francophone country. The French constitution states: “Le Président de la République nomme le Premier ministre. Il met fin à ses fonctions sur la présentation par celui-ci de la démission du Gouvernement.” This is translated as: “The President of the Republic shall appoint the Prime Minister. He shall terminate the appointment of the Prime Minister when the latter tenders the resignation of the Government.” In other words, in France the president merely transacts a decision that has been made elsewhere. In Congo-Brazzaville, the president can act independently.

At this point, such constitutional niceties are unlikely to worry President Sassou. The country may have adopted a semi-presidential constitution, but, as yet, no prime minister has even been appointed. More importantly, there has been violence after the election (or re-election, take your pick) of the president. This left a number of people dead. The opposition still contests the result of the election, though the Constitutional Council, unsurprisingly, has validated the election. Since the election, there have been various arrests of opposition figures. However, the opposition fears that violence is being exaggerated and even manipulated by the regime to justify an even harsher crackdown. President Sassou is still in control, but the situation remains volatile.

Countries that have debated introducing semi-presidentialism – Hungary

This is another in a series of occasional posts on countries that have actively debated introducing a semi-presidential constitution, but that eventually decided not to do so.

In the rush to democratisation in Central and Easter Europe in the late 1980s/early 1990s, Hungary very nearly adopted semi-presidentialism. I am no expert on Hungary (so comments are welcome), but there seems to have been a two-stage process.

Hungary was one of the early democratisers in 1989. As I understand it, one of the reforms introduced by the old regime was the direct election of the president. This election was scheduled to take place on 3 December 1989. However, the democratic opposition to the regime was concerned that it would not be able to compete effectively and that Imre Pozsgay would be elected. Pozsgay had played a major role in reforming the Hungarian Socialist Workers Party, but the opposition considered that his victory would have legitimised the old regime in its new form. In November 1989 the opposition was able to force a referendum at which voters were asked to decide whether the presidential election should take place in December or be delayed until after parliamentary elections scheduled for 1990. The vote was very close and the delay was approved by 2,145,023 votes to 2,138,619.

Parliamentary elections took place March 1990. Unlike most other countries, Hungary did not immediately adopt a new constitution. Instead, the old constitution was reformed. As part of this incremental process another referendum was held on July 1990 about whether or not to introduce the direct election of the president. This would have made Hungary semi-presidential. There was an overwhelming majority for the proposal (85.9%), but turnout was only 14% and so the result was invalid. Therefore, Hungary continued to elect its president indirectly and remained parliamentary.

So, Moldova and Hungary, if memory serves me correctly, are the only two countries to have rejected the direct election of the president in a referendum. On both occasions, this was due to low turnout.

The difference between the number and the ‘effective’ number of presidential candidates

Last weekend, Ireland held the first meeting of its Constitutional Convention. One of the topics for debate was whether the president’s term should be reduced to 5 years. However, an issue that generated much more discussion and passion was whether or not citizens should have a role in nominating presidential candidates. In the end, the members of the Convention voted overwhelming (94%-6%) to open up the nomination procedure to citizens. In that context, I thought that I would provide just a little more information about how nomination processes operate comparatively.

Currently, the Irish Constitution provides three ways of nominating presidential candidates: nomination by at least 20 members of the two houses of the legislature (there are 226 at the moment); by 4 elected councils (there are 34, I think); plus incumbent presidents may re-nominate themselves. This has meant that until recently political parties have dominated the nomination process. Since 1937, there have been 7 contested elections and a total of only 24 candidates. That said, in 1997 there were 5 candidates and in 2011 there were 7 candidates. So, there is perhaps some evidence that parties have less of a hold now than previously.

What happens elsewhere? In a recent article in Irish Political Studies, I provided some information about how 29 semi-presidential countries choose their presidential candidates. The take-home point is that probably only Mongolia and Turkey have more restrictive procedures than Ireland and most other countries have some sort of citizen nomination process.

However, two points need to be made.

Firstly, the rules for citizen nomination procedures vary greatly. Some countries simply require a number of signatures from registered electors. Even so, the number is very different from one country to the next. So, Taiwan and Montenegro stipulate 1.5% of the total electorate. By contrast, Poland requires 100,000 signatures. Given there were nearly 31 million registered voters in Poland in 2010, this is a much smaller fraction. In some countries, there is also a geographical requirement i.e. a certain number of signatures from a particular number of provinces or counties.

What all of this means is that even when the nomination system is opened up to citizens, the openness of the system will still vary as a function of the particular rules that are chosen. The up-side of this point is that it does not necessarily mean hundreds or dozens of candidates will inevitably stand. The down-side is that the system is not necessarily much more open. Apart from the cost of a campaign, which will exclude most people, the rules may still serve to restrict the number of candidates.

Secondly, even if the rules are changed in a way that increases the number of candidates, there is a big difference between the number of candidates and the ‘effective’ number of candidates. To put it another way, there may be a larger number of candidates, but how many of them are ‘serious’?

What are the figures? Leaving aside Mali and Peru, the highest figure was 24 candidates at the 2004 presidential election in Ukraine, then 21 in Bulgaria in 1992. Other countries recorded 16, 17, and 18 candidates in some cases. However, the average effective number of candidates was much smaller. For example, whereas Ukraine has had an average of nearly 14 candidates across all its presidential elections, it has had an average of fewer than 4 effective candidates. In fact, France has had the highest average effective number of candidates at just under 5. For most countries the average effective number of candidates is between about 2.5 and 3.5 and the figures are remarkably even across all countries, regardless of how restrictive the procedures are for citizen nominations.

In other words, even if opening up the nomination process to citizens may increase the number of candidates, it does not necessarily make the election more competitive. There may be more candidates from small parties and more candidates from non-partisan backgrounds, but the party candidates are still almost certain to win. This is a function of party identification, previous notoriety, campaign funding resources, and so on.

Overall, the cross-national evidence suggests that opening up the nomination process to citizens is unlikely to change the type of candidate who is actually elected to the presidency. Parties will still dominate. However, opening up the nomination process to citizens may simply be something that people want as a matter of principle.

Lithuania – President rejects ministerial nominations. When have other presidents done the same?

The formation of the new government in Lithuania has been very difficult. The difficulty has arisen not so much from the negotiations between the parties in the coalition, but from the attitude of the president towards the government. The president’s actions have raised a more general issue.

The legislative election in Lithuania was held at the end of October. A three-party agreement on a new government was reached very quickly and a fourth party was soon added. However, from very early on the composition of the prospective government was contested. President Dalia Grybauskaitė refused to accept the nomination of certain ministers from the Labour Party. More specifically, at first she seemed to imply that she would refuse the nomination of any minister from the party. However, the president has now approved the government and, as reports, it does include Labour Party ministers. However, the government is incomplete. President Grybauskaitė has indeed refused the nomination of two Labour Party ministers and their replacements have yet to be found. In addition, the report states that the initial nominee for the Minister of Culture was rejected and another name was found.

This got me thinking. During periods of cohabitation, how often have presidents refused the nomination of government ministers?

Well, first, it should be stressed that Lithuania is not experiencing a period of cohabitation. President Grybauskaitė is an independent. All the same, she has clearly aligned herself with the Conservatives. So, what we have in Lithuania is a ‘difficult case’ of cohabitation. In previous posts, I have recorded certain other examples.

If we leave the specific Lithuania issue aside and concentrate on the general problem, then three examples of presidents refusing to nominate particular ministers during cohabitation have been identified.

In France in 1986, President Mitterrand rejected the appointment of the Foreign Affairs and Defence Ministers. At that time, he intimated that because he had constitutional competences in these areas and also because the president does have to approve the appointment of ministers nominated by the PM, then he was entitled to oppose the proposed nominations. The new PM, Jacques Chirac, did not make a fuss and they agreed on two new appointments. So, the rejections were known, but it was not a stand-off or a crisis, partly because neither the president nor the PM saw it in their interests to create one.

My colleague, Iain McMenamin, reminded me of a similar case in Poland. He recounts the case in his chapter in the book edited by myself and Sophia Moestrup, Semi-presidentialism in Central and Eastern Europe. In the early 1990s, President Walesa insisted that he appoint three ministers in his “special special areas of responsibility”. In a slightly later Cabinet, Walesa refused to accept an SLD nomination for Finance Minister. As McMenamin then writes (p. 130): “Eventually, after a prolonged standoff, Walesa got the SLD to produce a new nomination for finance minister, while he accepted coalition-nominated deputy ministers in the presidential ministries.”

Another colleague and great friend to this blog, Cristina Bucur, also told me of a similar situation in Romania. In December 2007 President Băsescu refused to appoint Norica Nicolai as Minister of Justice. There are some details here.

I can’t think of any other examples, but I would guess that there are some. If anyone has any further examples, then do please comment.

The bottom line is that President Grybauskaitė is following the example of certain other presidents. That said, presidential refusals of ministerial nominations under cohabitation do seem to be pretty rare. If so, then it is perhaps more evidence that cohabitation is not necessarily as problematic as it is sometimes portrayed.

The problems of semi-presidentialism and cohabitation in Africa (updated and corrected)

Yesterday, Octavio Amorim Neto kindly drew my attention to an article in the Jornal de Angola. I posted a link to it on the Facebook site. It is an opinion piece about semi-presidentialism in Africa. It argues that semi-presidentialism is problematic, particularly because of the prevalence of cohabitation. The author, Faustino Henrique, argues in favour of a presidential system and also appears to prefer parliamentarism to semi-presidentialism. If you don’t have Portuguese, then Google Translate does a pretty good job of conveying the argument.

I am responding to the article not because I think that semi-presidentialism is the best of all systems of government and that it needs to be defended from its critics. Indeed, as regular readers of this blog will know, I like promoting the study of semi-presidentialism, but I have severe reservations about recommending semi-presidentialism as a system of government. Instead, I think I was driven to respond partly because doing so helps to promote the study of semi-presidentialism and partly because I was frustrated by the way in which the author defends the argument.

The author’s basic argument is that semi-presidentialism has been the cause of political crisis in Africa. Particularly, the author argues that cohabitation between the executive and the parliamentary majority is the cause of political problems.

What struck me most about the article was the absence of empirical evidence. There was only one example, the case of São Tomé and Príncipe where a representative of the opposition MLSTP-PSD party is quoted as saying that “the country is heading for political instability”. Apart from that one quote, we are simply told that cohabitation has created “many problems”.

São Tomé e Príncipe is indeed one of the most cohabitation-prone countries in the world. Indeed, the most recent period of cohabitation ended in 2011. However, despite the fact that it has had so many periods of cohabitation, it has remained democratic. What about cohabitation in Africa more generally? Well, it was certainly associated with the collapse of democracy in Niger in 1996. However, the only other African country to have experienced cohabitation is Cape Verde, where it began in September 2011. To the best of my knowledge, cohabitation in Cape Verde has worked without incident since this time. So, while cohabitation in Niger was a very difficult experience and is a sign that cohabitation can definitely cause problems, I am not sure that many countries have faced the problems of cohabitation that the author alludes to in the article.

Interestingly, when pointing to the causes of instability, the author identifies not only periods of cohabitation but also periods where an independent president has been at odds with the government or the parliamentary majority. Here is where São Tomé e Príncipe comes back in. The current president, Manuel Pinto da Costa, ran as an independent in 2011, though he was formerly a member of the MLSTP-PSD. The government is headed by Patrice Trovoada from the ADI. It is a minority government. The MLSTP-PSD is the main opposition party in parliament. Currently, the government is having problems passing legislation in parliament. It is also possible that the president is making the government’s life more difficult, even if he is no longer formally a member of the MLSTP-PSD. So, the presence of an independent president may be problematic.

However, the impact of ‘independent’ president is one that needs more research. The term ‘independent’ can cover a lot of situations. It can refer to people who claim to be independent but are really supported by a party/group or block of parties/groups. It can also refer to people who are populist and who do not have a party backing but do have a strong social base. It can refer to people such as technocrats. I don’t know of any studies that have tried to systematically assess the impact, positive or negative, of independent presidents on democracy in Africa. I can certainly think of the problematic situation in Guinea-Bissau following the election of the late João Bernardo Vieira. He had been a member of the PAIGC, but had left the party at the time of his re-election in 2005. Generally, though, I don’t think we know enough about independent presidents to draw any systematic conclusions.

So, this is what I think. Cohabitation is unique to semi-presidentialism. There is evidence that it can be problematic. However, in Africa it has not occurred very often and it has only been associated with the collapse of democracy in one case. Independent presidents are not unique to semi-presidentialism. They can occur under presidentialism too. So, if independent presidents are problematic, and they may be, then presidentialism is not necessarily a solution. Finally, minority governments are not unique to semi-presidentialism. They can occur under presidentialism and parliamentarism. If they are problematic, and as before they may be, then again presidentialism is not necessarily a solution and neither is parliamentarism in that regard.

Overall, I have no particular desire to promote semi-presidentialism as a form of government. I think there are arguments in favour of and against parliamentarism, presidentialism, and semi-presidentialism. However, if we are arguing the merits and demerits of specific regimes, then I think we need to think carefully about how the argument is made. What is the empirical evidence? (And I haven’t even mentioned the need to include all sorts of control variables in order to determine the independent impact of semi-presidentialism and its various forms). What is the alternative? Are the problems of semi-presidentialism likely to occur under other forms of government too?

(This post replaces an earlier version. Thanks to Gerhard Siebert for corrections. Obviously, this post represents only my own opinion).

President or PM? Who attended the NATO summit in Chicago?

Yesterday’s post about the foreign visits of the Lithuanian president made me think about another article I had read. This was an article about the Finnish president and the fact that he was attending the NATO summit in Chicago. Given the Finnish president is now a symbolic figure, though, admittedly, with more legitimacy in foreign affairs than any other area, it was interesting that he was attending such an important summit.

Anyway, I checked the Wikipedia site for the summit and it gives a list of heads of state and government who attended. Obviously, the presidents of presidential countries attended and the prime ministers of parliamentary countries attended, but who attended for semi-presidential countries?

Part of the answer is easy. For semi-presidential countries with a strong presidency, the president attended (Azerbaijan, Armenia, France, Georgia, Romania, Ukraine). For semi-presidential countries with a figurehead presidency the prime minister usually attended (Croatia, Iceland, Ireland, Slovenia). Part of the answer, though, is more complicated. For a couple of semi-presidential countries with a very weak president the president still attended (Finland and Slovakia). For some semi-presidential countries where the president is more than a figurehead but where the Prime Minister is generally the more powerful figure, the president attended (Bulgaria, Lithuania, Poland, Turkey). Finally, for one country in this latter category the prime minister attended (Portugal).

So, there is no obvious pattern. Presumably, the answer should be constitutional. For example, in Finland, even though the president is now very weak, Art. 93 still states that the “foreign policy of Finland is directed by the President of the Republic in co-operation with the Government.” So, whereas the PM is clearly responsible for EU policy, there is a reason why the president should go to a NATO summit. Equally, in Romania, even though there is a period of cohabitation at the moment, the president has clear powers in terms of foreign and defence policy. So, it is natural that he should go. There is some information on constitutional powers in foreign policy topic in Tapio Raunio’s article in Journal of European Public Policy, 19:4, pp. 567-584, 2012.

However, the case of Croatia is interesting. Even though the president of Croatia is a figurehead, Art. 94 of the constitution of Croatia states that the “President of the Republic is responsible for the defense of independence and territorial integrity of the Republic of Croatia”; Art. 99 states among other things that the “President of the Republic and the Government of the Republic of Croatia shall cooperate in the formulation and execution of foreign policy”; Art. 100 states that the “President of the Republic is the Commander-in-Chief of the Armed Forces of the Republic of Croatia”; and Art. 101 also gives more details of the president’s powers in time of war. If you were the President of Croatia, this would probably be enough for you to insist on attending. And yet, the PM attended and, I suspect, without any presidential protest.

Moreover, previously there have been disputes in countries such as Poland and Slovakia as to whether the president should attend equivalent summits.

So, while the constitution holds some of the answer, my guess is that power politics matters as well. If anyone has any more systematic answers, or any country experiences to reveal, then please feel free to comment.

How many semi-presidential countries have a Vice President & what do they do?

Constitutionally, the answer to the ‘how many’ question is five: Bulgaria, Peru, Senegal, Taiwan and Tanzania. In practice, though, the answer is four. In 2009 Senegal passed a constitutional amendment creating the position of vice president. However, President Abdoulaye Wade, who was behind the reform, never appointed anyone. So, the answer to the how many question is fairly straightforward.

The answer to the ‘what do they do?’ question is more difficult. Peru and Tanzania have a prime minister and collective cabinet responsibility, but, in practice, they operate as presidential systems, even if Tanzania has a one-party dominant system as opposed to Peru’s typical Latin American fragmented party system. So, the role of the vice-president in these countries resembles those in equivalent presidential systems.

In Taiwan, as far as I can tell, the role of the vice-president is merely to replace the president if the president dies, resigns, or is incapacitated. My understanding is that President Ma’s first vice-president, Vincent Siew, also met with representatives of mainland China at certain meetings, which was politically sensitive. However, the post seems to be mainly honorific. Any more information would be welcome.

In Bulgaria, the vice-president assists the president (Art. 92 -2). In addition, Art. 104 states that the president “shall be free to devolve to the Vice President the prerogatives established by Art. 98 items 7, 9, 10 and 11.”. They are the power to “7. appoint and remove from office other state officials, established by law”, “9. grant, restore, relieve from and withdraw Bulgarian citizenship;”, “10. grant asylum;”, “11. exercise the right to pardon”. Does the vice president exercise any of these powers? The answer is yes.

In Bulgaria, the vice-president’s use of the right to pardon has become politically controversial. For example, the Sofia News Agency, Novinite, reports that in the ten years from 2002-2012 Vice President Angel Marin signed 49 pardon decrees even when the Pardons Committee had recommended that the pardon should not be used. In total during this period the President and Vice-President together issued 529 pardons, including more than 200 people convicted for murder. Interestingly, since taking office in January 2012 the current Vice President, Margarita Popova, has issued only one pardon. When she did so, she identified very clear grounds for issuing the pardon, trying to differentiate her use of the pardon from her predecessor’s.

So, Vice Presidents are very rare under semi-presidentialism. They are even rarer still in semi-presidential countries where the president is not a strong figure. Bulgaria is the only country where there is a weak president as well as a vice president. However, even in this latter case, the role of the Vice President can be controversial.