Here are links to the most frequently viewed posts on this blog.
Visitors might also be interested in the following series of posts. Please just follow the thread on the archives.
Here are links to the most frequently viewed posts on this blog.
Visitors might also be interested in the following series of posts. Please just follow the thread on the archives.
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 (firstname.lastname@example.org). 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 http://presidential-power.com/?p=7869.
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.
Here is my list of presidential, parliamentary and other regimes as of 1 August 2018. A list of countries with semi-presidential constitutions is posted separately.
Afghanistan, Angola, Argentina
Benin, Bolivia, Brazil, Burundi
Chile, Colombia, Comoros, Costa Rica, Côte d’Ivoire, Cyprus
Djibouti, Dominican Rep.
Ecuador, El Salvador, Equatorial Guinea
Ghana, Guatemala, Guinea, Guyana
Kazakhstan, Kenya, Rep. of Korea
Malawi, Maldives, Mexico
Palau, Panama, Paraguay, Philippines
Seychelles, Sierra Leone, Singapore, Sudan
Tajikistan, Turkmenistan, Turkey
Uganda, USA, Uruguay, Uzbekistan
Parliamentary (M = Monarchy, R = Republic)
Albania (R), Andorra (M), Antigua & Barbuda (M), Australia (M)
Bahamas (M), Bahrain (M), Bangladesh (R), Barbados (M), Belgium (M), Belize (M), Bhutan (M)
Cambodia (M), Canada (M)
Denmark (M), Dominica (R)
Estonia (R), Ethiopia (R)
Germany (R), Greece (R), Grenada (M)
India (R), Iraq (R), Israel (R), Italy (R)
Jamaica (M), Japan (M), Jordan (M)
Lao PDR (R), Latvia (R), Lebanon (R), Lesotho (M), Liechtenstein (M), Luxembourg (M)
Malaysia (M), Malta (R), Mauritius (R), Monaco (M), Morocco (M)
Nepal (R), Netherlands (M), New Zealand (M), Norway (M)
Pakistan (R), Papua New Guinea (M)
St Kitts & Nevis (M), St Lucia (M), St Vincent & the Grenadines (M), Samoa (M), Solomon Islands (M), Somalia (R), Spain (M), Swaziland (M), Sweden (M)
Thailand (M), Trinidad & Tobago (R), Tuvalu (M)
Presidentialism (i.e. popular presidential election) with no PM but cabinet accountability
Presidentialism (i.e. popular presidential election) but president accountability to legislature but not cabinet
Parliamentarism (i.e election of the president by the legislature) with no PM and no head of state/govt accountability and no cabinet accountability
Eritrea, Micronesia, Suriname, Switzerland
Parliamentarism (i.e election of the president by the legislature) with no PM but head of state/govt accountability and cabinet accountability
Botswana, Cuba, Kiribati, Marshall Islands, Nauru, South Africa
Parliamentarism (i.e election of the president by the legislature) with head of state, PM and cabinet accountability
Brunei, Monaco, Oman, Qatar, Saudi Arabia, Tonga, UAE
Bosnia & Herz., China, Iran, North Korea, Myanmar, San Marino
Libya, South Sudan
A broader statement of my philosophy about how regimes should be defined and the definitions themselves can be found at the link below.
The two basic definitional criteria are the method of selection of the head of state and whether or not the cabinet is responsible to the legislature. Obviously, other criteria could be used to refine the list of regimes further. So, for example, I use Shugart and Carey’s distinction between premier-presidential and president-parliamentary regimes to identify sub-types of semi-presidentialism. Here, though, I stick to the two criteria above, which is scientific standard.
The countries are classed on the basis of their current constitution. Bear in mind that the list does not take into account whether a country is democratic. For example, in Bahrain there is a form of collective cabinet responsibility. That is to say, there is a constitutional process by which the legislature could bring down the government. The politics of Bahrain, not least the monarch-appointed chamber that forms one house of the legislature, means that this is very unlikely ever to happen. However, constitutionally, Bahrain meets the requirements for parliamentarism, even if it is unequivocally autocratic. To put it another way, some presidential countries are, in practice, more presidential than others. Some parliamentary countries are, in practice, more prime ministerial than others, etc. The definitions used here are not designed to capture political practice. They are designed to capture constitutional rules.
I welcome corrections and comments.
I am not sure how this case missed my attention. However, I have ‘discovered’ another case of historic semi-presidentialism. The Federal Republic of Yugoslavia was semi-presidential from July 2000 until the dissolution of the state in 2003 and the creation of the State Union of Serbia and Montenegro.
Following the collapse of communist Yugoslavia, a new constitution was adopted in 1992. The text is available here. At that time, the Federal Republic of Yugoslavia was parliamentary. Arts. 103 and 104 make it clear that the PM and cabinet are collectively responsible to the legislature. However, Art. 97 indicates that the president is elected by the legislature.
In July 2000 the 1992 constitution was amended. (I have a copy of the text and can supply it, but the source is gated). There was a new Art. 97, allowing for the direct election of the president. Indeed, in September 2000 the first election was held, with Vojislav Koštunica beating Slobodan Milošević. There are details here as well as a more general article in Journal of Democracy , Volume 15, Number 4, October 2004.
In February 2003 the Federal Republic of Yugoslavia was dissolved and the State Union of Serbia and Montenegro was created. The constitution is available here. This constitution is parliamentary. Art. 21 states that the president is elected by the legislature.
To bring the story up to date, when Montenegro voted for independence in 2006 the State Union of Serbia and Montenegro was dissolved. The 2007 Montenegrin constitution and the 2006 Serbian constitutions are both semi-presidential. Therefore, the dissolution of the State Union of Serbia and Montenegro led to the creation of two new semi-presidential countries.
To complete the story, from 1992-2006 both Montenegro and Serbia were Republics within the Federal Republic of Yugoslavia/State Union of Serbia and Montenegro. Both had their own constitutions and both were semi-presidential. The constitutions are available here and here. So, even though at the national level Yugoslavia was semi-presidential only from 2000-2003, at the sub-national level both Montenegro and Serbia were semi-presidential prior to statehood in 2006.
In a previous post, a question was raised about Madagascar’s constitutional history from 1991 onwards and how it should be classified. This post sets out how I see it. With the exception of the 1995 reforms, the documents are all available in French at the Digithèque MJP.
In October 1991, there was a constitutional convention that governed the transition to democracy. There was a president and a prime minister, but no legislature. This was clearly an interim document that was not semi-presidential (or either presidential or parliamentary).
In August 1992, a new constitution was adopted. ICL provides an English version here. This constitution created a premier-presidential version of semi-presidentialism. Arts. 53 and 61 made no mention of prime ministerial responsibility to the president, only to the National Assembly. Responsibility to the National Assembly took various forms. According to Art. 90, it was on the basis of an investiture vote on the government’s programme, whereby a majority of the members of the National Assembly was required to approve the PM. Art. 91 also stated that if the government decided to change its programme during the course of a legislature, then it could ask for a vote of confidence. If two-thirds of the National Assembly voted against the government, then it had to resign. Art. 94 also allowed for motions of no-confidence. Such a motion was adopted if it gained the support of a majority of the members of the National Assembly. The wording of this constitution is unproblematic.
In October 1995, a constitutional law was passed. The details are in J. du Bois de Gaudusson, G. Conac, C. Desouches (eds.), Les Constitutions africaines publiées en langue française, Paris, La Documentation Française, 1998. This revision did change various articles that are potentially relevant to the classification of the country. There was still no mention in Arts. 53 or 61 of PM responsibility to the president, only to the National Assembly. However, Art. 90 was reworded. A new clause was added at the end stating that the president ends the PM’s functions (met fin aux fonctions du PM) either as a result of a vote of no-confidence (vote de défiance), or as a result of a motion of censure, or for any other determining cause (toutes autres causes déterminantes). This article is now very confusing. It seems to deal explicitly with the terms of the government’s investiture, but the new clause at the end seems more general. Presumably, it put the onus on the opposition to lodge a motion of no-confidence or censure rather than requiring the government to obtain a majority. Even so, the phrase ‘for any other determining cause ‘ seems to apply far more generally than just the investiture. What is more, it is not entirely clear how a vote de défiance is different from a motion of censure. Art. 91 was also reworded to allow the government to ask for a vote of confidence at any time. Again, a two-thirds majority was required to bring down the government. In addition, Art. 94 was amended such that a vote of no-confidence now required a two-thirds majority to bring down the government.
In April 1998 there were further amendments. Art. 53 was amended so that it now read: ‘The president ends the PM’s functions for any determining cause’ (pour toute cause déterminante), though Art. 61 still only makes reference to PM responsibility to the legislature. Art. 90 was amended to remove any need for an investiture vote and the clause that was added in 1995 was deleted. Art. 91 was also amended. A motion of confidence now required only a majority of the National Assembly to bring the government down. Art. 94 remained the same.
There were further amendments in 2007, but the wording of the above articles was not changed, though they were renumbered. The constitution was suspended in February 2009.
So, how do we classify Madagascar from August 1992-February 2009. It seems to depend on two things. Firstly, there is the super-majority requirement in Art. 94 from 1995 onwards. For me, this does not make a difference to the classification of a country as semi-presidential. I have already posted about this issue here. Secondly, there is the phrase “pour toute(s) cause(s) déterminante(s)”. This is a very unusual phrase that I cannot remember seeing in another constitution. I interpret it as meaning that the president can dismiss the PM on the basis of a given reason. To me, this allows the president to dismiss the PM.
Overall, I class Madagascar as semi-presidential from 1992-2009, and as premier-presidential from 1992-1995 and president-parliamentary from 1995-2009.
Here is the list of president-parliamentary and premier-presidential regimes that I am currently working with (as of 28 June 2018).
These terms were first defined by Matthew Shugart and John Carey in Presidents and Assemblies. Constitutional Design and Electoral Dynamics (Cambridge: Cambridge University Press, 1992).
The distinction between the two terms can be expressed as follows:
President-parliamentarism is a form of semi-presidentialism where the prime minister and cabinet are collectively responsible to both the legislature and the president.
Premier-presidentialism is a form of semi-presidentialism where the prime minister and cabinet are collectively responsible solely to the legislature.
A time-series, cross-sectional dataset in country years can be found here.
If there are any classification errors, then do please let me know:
Austria, Azerbaijan, Belarus, Burkina Faso, Cameroon, Central African Republic (2015-), Congo (Republic of), Gabon, Iceland, Madagascar, Mozambique, Namibia, Peru, Russia, Rwanda, Senegal, Sri Lanka, Syria, Taiwan, Tanzania, Togo (2003-)
Algeria, Bulgaria, Cabo Verde, Croatia (2001-), Czech Republic, Dem. Rep. of Congo (DRC), Egypt (2014-), Finland, France, Georgia (2013-), Haiti, Ireland, Kyrgyzstan, Lithuania, Macedonia, Mali, Mauritania, Moldova (2016-), Mongolia, Montenegro, Niger, Poland, Portugal (1983-), Romania, São Tomé e Príncipe (2003-), Serbia, Slovakia, Slovenia, Timor-Leste, Tunisia, Ukraine
There are countries that are no longer semi-presidential or which are currently semi-presidential but which have changed their form of semi-presidentialism over time. These cases are captured in the following lists.
Angola, Armenia (1995-2005), Austria (1929-1933), Burkina Faso (1978-1980), Burundi, Central African Republic (1981, 1993-2003, 2005-2013), Comoros, Croatia (1991-2000), Cuba, Egypt (2007-2011), Georgia (2004-2013), Germany (Weimar Republic), Guinea-Bissau, Kyrgyzstan (1996-2007), Madagascar (1996-2009), Portugal (1976-1982), Rwanda, São Tomé e Príncipe (1990-2002), Senegal (1970-1983), South Vietnam, Tunisia, Ukraine (1996-2006), Yemen
Armenia (2005-2018), Burkina Faso (1970-1974), Chad, Congo (Republic of), DRC (1991-1993), Kenya, Kyrgyzstan (1993-1996), Madagascar (1992-1995), Moldova (1994-2001), Niger, Togo (1993-2002), Turkey, Yugoslavia (2001-2002)
Is Uzbekistan semi-presidential? The answer seems obvious. No. However, a closer look suggests that the situation has changed over time and that Uzbekistan may be (or have been) semi-presidential.
The first constitution was adopted in 1992. It is widely available online in English. There are two key clauses:
Art. 78 (16) [The exclusive powers of the Oliy Majlis of the Republic of Uzbekistan shall include:] ratification of the decrees of the President of the Republic of Uzbekistan on the appointment and removal of the Prime Minister, the First Deputy Prime Ministers, the Deputy Prime Ministers and the members of the Cabinet of Ministers;
Art. 98, which in this version mentions nothing about collective cabinet responsibility or even individual prime ministerial responsibility.
In 2003 the constitution was amended. The 2003 version is also available online in English. Here, the wording of the relevant clauses changes:
Art. 78 (15) [Joint authority of the Legislative chamber and the Senate of the Oliy Majlis of the Republic of Uzbekistan shall include:] consideration and approval of a candidature of the Prime Minister of the Republic of Uzbekistan upon the nomination of the President of the Republic of Uzbekistan;
Art. 98 The Cabinet of Ministers, in its activity shall be responsible to the President of the Republic of Uzbekistan and the Oliy Majlis of the Republic of Uzbekistan.
So, while the legislature has the power (in theory!) to reject the appointment of the prime minister, it cannot reject the appointment of the cabinet as a whole. However, Art. 98 indicates that there is collective cabinet responsibility. Therefore, it seems as if Uzbekistan should be considered semi-presidential on the basis of the definition used in this blog. That said, there are no provisions in the constitution as to how the legislature will hold the government accountable. Therefore, I am reluctant to include it in my list of semi-presidential countries.
Interestingly, there now seems to be a further amendment. RFE/RL reports that the role of the president is being reduced, the procedures for appointing the prime minister are changing and that explicit reference to a vote of no-confidence is going to be included. Now, at this time, there is some confusion. In a separate report on the amendments, the wording of the no-confidence motion is actually provided. It is supposed to be worded as follows:
“The vote of non-confidence in the PM shall be deemed adopted if it receives a vote of at least two-thirds of the lower chamber and the upper house of the Uzbek parliament, respectively. In this case the president decides on the release of the PM from office. The entire composition of the Cabinet of Ministers resigns together with the PM”.
On the basis of this blog, this would not make Uzbekistan semi-presidential because the president still has the final say over the dismissal of the government.
Given the final text does not yet seem to be available, I will reserve judgement. However, clearly Uzbekistan is almost semi-presidential at least constitutionally.
This is series of posts that identifies countries that almost comply with the definition of semi-presidentialism that is used in this blog, but which fail to do so on the basis of a certain, sometimes unusual provision, or where the date when semi-presidentialism started can be contested.
This post examines the case of Cameroon in the 1960s. In 1961 Cameroon adopted a semi-presidential-like constitution. The text is impossible to track down on the internet, but it is available in the Revue juridique et politique d’Outre-Mer at that time.
At that time, the Constitution declared Cameroon to be a Federal Republic, combining the former French and British-administered territories. The French element was politically dominant. However, the federal nature of the system meant that the constitution was organised somewhat differently from its French counterpart, which was the model for most of the countries in the region at the time.
Anyway, this constitution established a directly elected federal president (Art. 9). At the federal level, the president was also the head of government and the cabinet was not responsible to the legislature. So, constitutionally, the Federal Republic of Cameroon was presidential.
However, the constitution also made provision for prime ministers to be appointed in each of the two federated states. The federal president appointed the prime minister in each case (Art. 39) and the PM was subject to an investiture vote that required a simple majority. The state cabinet was also subject to confidence and no-confidence votes (Art. 44) and had to resign if, for example, it was defeated on a no-confidence motion by an absolute majority of the legislature.
So, here is a country that was not semi-presidential, but which had a clear semi-presidential element to it. Indeed, it was the only Africa case to have anything like this type of presidential/prime ministerial arrangement until 1970.
The 1961 Cameroon constitution was repealed in 1972 when the country officially became the United Republic of Cameroon and any reference to the federated states was abolished.
This is an updated version of a previous post.
This is a series of posts on semi-presidentialism in areas other than internationally recognised states. The focus is on areas with, or that have had, full constitutions, but ones that are not recognised as independent states. They may be territories that have declared independence but whose status has not been internationally recognised, or they may simply be self-governing units within or under the protection of another state.
Zanzibar is part of the United Republic of Tanzania. Art. 103 of the 1977 Tanzanian Constitution (2005) states that Zanzibar shall have a president, that there shall be a Chairman of the Revolutionary Council, and that it shall have a House of Representatives. The relations between these institutions are regulated by the 1984 Constitution of Zanzibar. The Zanzibar constitution, as amended up to 2002, is available here.
Art. 26 establishes the President of Zanzibar as the head of government and the Chairman of the Revolutionary Council. In addition, Art. 39 establishes a Chief Minister, who is the “principal adviser” of the president and who has “authority over the control, supervision and execution of the day-today-function and affairs of the Revolutionary Government of Zanzibar”. So, the Chief Minister is the equivalent of the Prime Minister. Art. 41 provides for the individual responsibility of the Chief Minister. So, on the basis of the definition used in this blog, Zanzibar would not be semi-presidential. However, Art. 42 establishes the Revolutionary Government of Zanzibar so-called the Revolutionary Government of Zanzibar, which includes a Revolutionary Council (and which can be taken as the equivalent of the Cabinet). The Revolutionary Council comprises the President of Zanzibar, the Chief Minister and Ministers. This article then states that the Revolutionary Council is collectively responsible to the House of Representatives. There are no details about the procedures for holding the Revolutionary Council responsible and no confirmation that the loss of responsibility requires resignation, but this can reasonably be assumed.
So, is Zanzibar semi-presidential according to this constitution? There is probably enough for it to be classed as such. There is a directly elected president, a prime minister, and explicit mention of cabinet responsibility. The wording of the constitution is somewhat more ambiguous that other texts and the procedures for responsibility are unspecified. However, overall, Zanzibar can be considered semi-presidential on the basis of the this constitution.
As reported in a previous post, there was a significant constitutional amendment earlier this year. I do not have the text of this amendment. (The result of the referendum on the reform is available here.) However, EISA are reporting that the position of Chief Minister has been abolished and that two vice-presidencies have been created. Ministries are now shared proportionally between the two main parties in the system. Given this reform seems to have abolished the Chief Minister (or prime minister), then Zanzibar would no longer appear to be semi-presidential. Tanzania remains so, of course.