In a recent post, I linked to a new time-series, cross-sectional dataset on semi-presidentialism. The dataset provides an annual, cross-national coding of semi-presidential countries since 1900. V2.0 is available here.
The dataset contains two codings of semi-presidentialism. One conforms to – let’s call it – the standard definition. Here, semi-presidentialism is where the constitution provides for a directly (or popularly) elected president and a prime minister and cabinet that are collectively responsible to the legislature. The other adds another clause. Here, semi-presidentialism is where the constitution provides for a directly (or popularly) elected president and a prime minister and cabinet that are collectively responsible to the legislature other than by a super-majority vote. The second coding was added to V1.0 along with codings for countries that conform to – let’s call them – the standard definitions of premier-presidentialism and president-parliamentarism plus countries that confirm to those definitions with the addition of an equivalent super-majority clause.
Where does the need for a coding that includes a super-majority clause come from? I have been aware for some time that Samuels and Shugart (2010, p. 30, fn 4) excluded countries, such as Madagascar, from their list of semi-presidential regimes because of the introduction in the Constitution at a certain time of a super-majority requirement. In fact, they classed Madagascar as presidential for this reason (e.g. ibid. p. 33 and p. 258). Yet, I don’t remember seeing any definition of semi-presidentialism that explicitly includes this clause. Also, as far as I am aware, it isn’t part of any formal definition of the concept that Samuels and Shugart provide and the equivalent clause isn’t included in their (or Shugart and Carey’s) definition of premier-presidentialism or president-parliamentarism. So, it seems to be post-definitional add-on, or an implicit assumption of the formal definition.
In one sense, I’m indifferent as to whether a super-majority clause should be included as part of the definition of semi-presidentialism, because even if it is included it still allows for the reliable classification of countries. No expert knowledge is needed to determine whether a country should be classed as semi-presidential or not. We just need to apply certain rules to publicly available constitutional information. This reliability is the most important part of the classification process.
Three points, though. First, if it is to be operationalised, then I think the clause should be stated as part of the definition. If it isn’t stated, then for me semi-presidentialism still includes countries with a super-majority requirement. If it is stated, then it obviously excludes them. In other words, we should avoid post-definitional add-ons or implicit definitional assumptions.
Second, I think it is still better to class countries with a super-majority requirement as semi-presidential (or as a sub-category of semi-presidentialism) rather as presidential. After all, the constitution does still allow the legislature to bring down the government, whereas under presidentialism, by definition, it does not. Sure, it might take an extraordinary and almost unimaginable set of circumstances for, say, a two-thirds majority to come together and bring a government down, but constitutionally it could happen. (Think how opposing parties can vote together to end a nominally fixed-term legislature). In other words, whether or not it happens is a matter of politics not the constitution. If we are classing countries on the basis of constitutions, which is the only reliable way of doing so, then surely it is better to think of a country with a super-majority clause as being semi-presidential not presidential? The survival of one part of the executive is still not separate from the legislature.
Third, a super-majority requirement has implications for the classification of parliamentary regimes too. Maybe there are no examples, but what if there was a super-majority clause in a nominally parliamentary regime? For me, this would still make the country with such a clause parliamentary, although we might want to think about classifying the country as a sub-category of parliamentarism. Whatever the choice, I would be wary of classifying that country as presidential.
This is all very nerdy. But why stop there? Next week, I am going to discuss the classificatory implications of introducing other clauses.