Presidents and/or Prime Ministers: A Historical Dataset

• In Event: Legislatures and Leaders: New Perspectives on Political Institutions

Abstract

This paper presents new historical data on countries’ systems of government. Although prominent datasets on historical political institutions have recently emerged, long-term historical information on how countries choose their national Head of Government (HoG) is surprisingly lacking. To help fill this gap, we offer a yearly assessment of this for all sovereign countries since the French Revolution (16,910 country-years).

Yet, instead of relying on arbitrary judgment about whether each system is presidential, parliamentary, semi-presidential or semi-parliamentary, and on assuming an arbitrary level of democratization for those concepts to make sense, we propose different routes. First, we do not read countries’ histories in terms of “or”, we read them in terms of “and”; our scheme incorporates the recognition that presidential and parliamentary systems are not mutually exclusive. We identify, for each country-year, whether a President (PR) was elected and whether a Prime Minister (PM) was elected. While there are rare ambiguous cases in which it is hard to decide whether a HoG is PR or PM, most often that proves to be much less ambiguous than classifying a system as being presidential versus parliamentary. Hence, since we are not interested in low level details about each system of government or about the HoG’s specific powers, instead of trying to arbitrarily decide what exactly a mixed system is, we can simply talk about countries that elect PRs, that elect PMs and that elect both. Second, with our minimalist definition of “being elected” - i.e. “put in office by an election where more than one competitor ran” - we do not differentiate between democratic, semi-democratic, semi-authoritarian or authoritarian regimes. On the contrary, it is our interest that our data would help scholars studying any types of regimes and that we can use them to talk about the evolution of methods of choosing HoG regardless of the quality of democracies.

More specifically, we catalog four main binary variables: whether or not a country has an elected PR, an unelected PR, and elected PM, or an unelected PM. We therefore differentiate between unelected PR or PM and having no PR or PM at all. Our notion of presidentialism is quite literal: any sovereign government that includes a figure who is called the "president" is considered to be a presidential system (some secondary variables represent rare edge cases). Similarly, we categorize a country as having a parliamentary government if its HoG is in a parliament.

Some messy examples illustrate the advantages of our categorization scheme: a country that elects its PM holds an election for a previously unelected PR (see the Republic of the Congo in 1961); a country that elects its PR begins to also elect its PM (Sri Lanka 1994); PM elections are suspended while a country continues to elect its PR (Zimbabwe 2014); a country ends PM elections but retains an unelected PM, while also abolishing its PR (Burma 1963); a country with an elected PM simply adds an appointed PR (Greece 1973). Our dataset completely captures these scenarios, but its greatest strength is recognizing the non-mutual exclusivity of these positions. For example, from 1996-2002 scholars often consider the PM of Israel to have effectively been a PR, in addition to Israel’s appointed PR. So we simply assign a “1” to both “elected PR” and “unelected PR”.

Each of these subtleties represents a substantive reality that binary indicators would ignore. Consider Fiji, which has always elected its parliament, but instituted an appointed presidency in 1987. Did Fiji become less parliamentary? Our dataset also captures major upheavals; we can watch the unelected Napoleonic presidency fall to the French 2nd Empire’s unelected legislatures, which turn in 1870 into elected PMs and unelected PRs. By categorizing these four crucial variables across 16,910 country-years, we allow researchers to answer much more subtle questions about the individual consequences of having an elected PM, an unelected PM, an elected PR, or an unelected PR.

Lastly, after presenting the data we use them to describe the evolution of PR and PM elections over time and space. We show, for instance, that from the 18th century until 1900, differently from what is often assumed, elections for HoG around the world were evenly divided between choosing PMs and PRs. It is in the beginning of the 20th century that PM elections took a minor lead, but since the end of First World War, the share of exclusive PM and of exclusive PR elections never ceased to lose space to the increased number of cases where both are elected. Nevertheless, a steep decline in exclusive PM elections since the end of Cold War has left the world at a stage of almost perfect division, where exclusive presidential elections, exclusive PM elections and cases where both are elected currently correspond each to around 30% of all sovereign countries, with 10% not having any elected national HoG.

Authors

• Fabricio Vasselai, University of Michigan

• Samuel Baltz, University of Michigan

• Allen D. Hicken, University of Michigan, Ann Arbor