Anna Fruhstorfer – Constitutions and Presidents. How formal rules constrain and empower

This is a guest post by Anna Fruhstorfer. It summarizes the main argument and findings of her PhD thesis, Constitutions and Presidents. How formal rules constrain and empower, which was Defended at Humboldt-Universität zu Berlin in July 2015 (PhD committee: Silvia von Steinsdorff, Zachary Elkins, Ellen Imoergut.



Although presidential power is a hot topic in political science research, little attention was paid to presidents in non-presidential systems with a powerful parliament and prime minister. So far, scholars have mostly been focusing on semi-presidential systems, whereas parliamentary systems with an indirectly elected president like in Germany or Estonia are hardly ever discussed (with the exception of Tavits 2009). Thus, in my research, I study presidential power in both parliamentary and semipresidential systems. In addition, little is known about the role of constitutions in presidential power. Treating constitutions as the explanatory variable is unusual. However, some influential studies have sought to explain the central role of constitutions in the conceptionalization of presidential power. From different angles, Amorim Neto and Strøm (2006), Tavits (2009), Schleiter and Morgan-Jones (2010) or more recently Bucur (2013) have significantly contributed to our understanding of presidential power. Nevertheless, because these studies are mostly concerned with explaining why a discrepancy between constitution and reality occurs, they do not pay much attention to how constitutions influence reality. In my thesis I answer the how question. I argue that specific characteristics of constitutional power make presidential institutions more sensitive to outside influences than others. Therefore, presidents that act within these institutions adapt their behavior accordingly. What emerges from these actions and what we can observe are different patterns of presidential leadership.

Empirically, that means that I conduct a comparative case study of 46 countries for a time frame of up to 75 years.[1] First, I compile an original dataset for these countries (CPS dataset)[2], which is based on a new measurement tool of constitutional presidential strength (CPS index). Secondly, using the dataset, a principal axis factor analysis is used to confront the unidimensional perspective of presidential power and form two dimensions of constitutional power. This two-dimensional perspective then lays the ground for the third step; a typology of presidential institutions with four types (CPS typology).

Measurement of constitutional presidential strength

In the course of the development of the stated argument and the conceptualization of power, it became clear that established measurement tools of presidential constitutional power do not always adequately describe the president’s role in parliamentary systems. Nevertheless, these facts were necessary for my research project. Hence, I have developed the index of constitutional presidential strength (CPS) for this study. It advances established tools to better capture the functional logic of parliamentary systems, to facilitate both low-level and high-level constitutional competences, and to enhance methodological and conceptual issues.

Conceptually, the CPS index emphasizes the functional logic of parliamentary systems. This means that it treats the power distribution with the government’s survival placed in the parliament’s hands as its most important element. Methodologically, the CPS index builds on Fortin’s (2013) description of measurement shortcomings of other measurement tools. The CPS dataset, established by implementing this measurement tool, provides a unique data collection of presidential power. From a historical-comparative perspective, the measurement and the data display a picture of great diversity of presidential power within a system of checks and balances. As a result, I observe consistently higher values compared to the normalized overview proposed by Doyle and Elgie (2015), as well as to individual measurement tools. The main reason for these differences is that the CPS index provides a more pronounced portrayal of the presidential role in power-sharing modes. The differences in the constitutional power in Germany and India as parliamentary systems and Austria and Mozambique as semi-presidential systems stand in stark contrast to each other. Some directly elected presidents are therefore not even strong on paper, such as in Austria, Iceland, Ireland, Macedonia or Montenegro, while indirectly elected presidents are already strong on paper, e.g. Bangladesh (1972-1974; 1991-2013), Estonia, India, Latvia, or Slovakia.

From a large-N perspective, empirical evidence shows that the two groups of directly and indirectly elected presidents differ significantly from each other. Hence, statistically, direct election goes hand in hand with a higher level of constitutional power (as well as the president’s role in the cabinet and the command of the armed forces)

T-Test for differences between directly and indirectly elected presidents

Levene’s Test for Equality of Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference
Lower Upper
Constitutional Power Equal variances assumed 129.167 .000 -22.415 3285 .000 -6.12560 .27329 -6.66143 -5.58977

Patterns of presidential de-facto power

Based on this comparative perspective, I further stressed the effect of constitutions on how their power shapes and determines presidential behavior in decision-making. Whereas these effects are diverse, there are also clear patterns. The strong-weak spectrum of constitutional power cannot fully account for these different patterns of de-facto power. An exploratory factor analysis (PA) of the measurement data indicates that constitutional power is indeed a two-factor construct (see also Fortin 2013).

These 2 dimensions are the basis for a typology of presidential institutions with 4 types, which can be characterized by the differences in their level of discrepancy between constitution and reality. Semi-presidentialism research has already shown that countries have alternate patterns of leadership (Elgie 1999, 283). These patterns vary between “dominant pattern(s)” (ibid.), be it either a dominant president, a dominant executive, or “a shift from one dominant pattern of leadership to another” (ibid.). What I argue and provide evidence for on a case-based description, is that these patterns are driven by the constitutional structure.

Presidential institutions labeled the notary (e.g. Austria, Germany, Albania, Czech Republic), i.e. with little power on both dimensions, are not able to change their de facto role. Empirically neither a patrimonial leadership legacy combined with a problematic democratic development, such as in Albania, nor exceptional political situations, such as in 1999/2000 Austria, nor even the newly introduced direct election in Czech Republic allow for an increase of presidential de-facto power.

The same pattern can be observed for presidential institutions like the almighty (e.g. Georgia, Bangladesh, Ukraine). Presidencies with above-average competences on both dimensions are so powerful, that they do not (have to) vary their de facto role. These almighty presidencies are largely insensitive to outside influences. Their dominance in most power-sharing situations offers hardly any room for other political actors to establish an influential position. The strong negative correlation with the democracy level of this type is therefore no surprise.

For presidencies labeled custodian and firefighter (e.g. France, Slovenia, Moldova, Poland, and Ukraine), varying patterns of de facto power can be observed. These cases confirm expectations and established literature regarding the questions of why the de-facto power of presidents varies[3] – surprisingly both for directly and indirectly elected presidents within these two types. The Estonian case (and Moldova after 2000), for example, provides some evidence for different patterns of presidential de-facto power. Presidential involvement or ‘activism’ (Köker 2014) increases in times of cohabitation; Estonian Presidents for example veto legislation more often and with a higher frequency in this situation. In times of cohabitation, indirectly elected presidents do not have roots in the ruling party. In most cases, the constitution does not even stipulate a role in cabinet meetings. Presidents therefore lose any influence on the decision-making process within the cabinet and the parliamentary majority. Thus, it is no surprise that these presidents use their legislative veto power frequently.


Research on presidential power points to multiple, interrelated causes for the varying de-facto power of presidents. I claim that constitutions structure the choices and thereby create a path for the direction of the power distribution. By defining how the game is played and laying the ground for the battle, they frame who gains and who loses power. This argument has been illustrated for several countries, but this can only be the starting point for further research. A large-N test of this may use for example the number of candidates in presidential elections as proposed by Cheibub and Chernykh (2008) as a dependent variable for a comparison of presidential de-facto power in different democratic and regional settings.


Amorim Neto, Octavio, and Kaare Strøm. 2006. “Breaking the Parliamentary Chain of Delegation: Presidents and Non-partisan Cabinet Members in European Democracies.” British Journal of Political Science 36 (4): 619–43.

Bucur, Cristina. 2013. “Who fires ministers? A principal-agent approach to ministerial deselection.” PhD Thesis, School of Law and Government, Dublin City University.

Cheibub, José A., and Svitlana Chernykh. 2008. “Constitutions and Democratic Performance in Semi-Presidential Democracies.” Japanese Journal of Political Science 9 (3): 269–303.

Doyle, David, and Robert Elgie. 2015. “Maximizing the reliability of cross-national measures of presidential power.” British Journal of Political Science.

Elgie, Robert. 1999. “Semi-Presidentialism and Comparative Institutional Engineering.” In Semi-Presidentialism in Europe, edited by Robert Elgie, 281–99. Comparative European Politics. Oxford [u.a.]: Oxford Univ. Press.

———, ed. 1999. Semi-Presidentialism in Europe. Comparative European Politics. Oxford [u.a.]: Oxford Univ. Press.

Fortin, Jessica. 2013. “Measuring presidential powers: Some pitfalls of aggregate measurement.” International Political Science Review 34 (1): 91–112.

Hicken, Allen, and Heather Stoll. 2013. “Are All Presidents Created Equal? Presidential Powers and the Shadow of Presidential Elections.” Comparative Political Studies 46 (3): 291–319. DOI: 10.1177/0010414012453694.

Köker, Philipp. 2014. “Czech Republic – A new government and the evolution of semi-presidentialism.” Accessed October 20, 2014.

Schleiter, Petra, and Edward Morgan-Jones. 2010. “Who’s in charge? Presidents, assemblies, and the political control of semipresidential cabinets.” Comparative Political Studies 43 (11): 1415–41.

Tavits, Margit. 2009. Presidents and Prime Minsters. Do Direct Elections Matter? Oxford [u.a.]: Oxford Univ. Press.


[1] For this I gratefully acknowledge the support of the Comparative Constitutions Project, in particular the University of Texas, Austin.

[2] I label this constitutional presidential strength (CPS) to avoid any possibility of confusion with the different conceptualizations of presidential power presented and with other measurement tools.

[3] This list of arguments explaining why presidential de-facto power varies under certain conditions is infinite. Every country holds its unique set of surprises. One important explanation that proves solid is the role of the president as de facto party leader (Bucur 2013). Additionally party cohesion in combination with cohabitation (Amorim Neto and Strøm 2006), (Elgie 1999), and the proximity of presidential and parliamentary elections (Hicken and Stoll 2013) explain the situation in a variety of countries.

Anna Fruhstorfer is a postdoctoral researcher at Humboldt University Berlin at the Department of Social Sciences. Her main area of research is Comparative Politics, with a regional emphasis in Eastern Europe. Her research concentrates on presidents, parliaments and the relation of law and politics, in particular constitutional politics. She is also affiliated with the University of Texas at Austin via the Comparative Constitutions Project.

6 thoughts on “Anna Fruhstorfer – Constitutions and Presidents. How formal rules constrain and empower

  1. JD

    I think it’s problematic to call heads of state in non-presidential systems “presidents”. The systems are categorised according to their having a popularly elected head of state or not (and their role in appointing the cabinet) so the two types of head of state should be distinguished terminologically accordingly.

  2. Anna Fruhstorfer

    Dear JD,
    thank you very much for your comment. One of the goals of the thesis was to show if there was a difference between directly and non-directly elected presidents, thus of presidents in parliamentary and semi-presidential systems (with the exception of monarchies and microstates), which is recently debated in political science research (in particular since Tavits 2009). Perhaps I should have made this point clearer, as I tried to do in the last part of the contribution.

    1. JD

      Thanks Anna.

      The point was very clear. I’m just using the opportunity to state my objection to the use of the word ‘president’ as anything but popularly-elected head of state/chief executive. Otherwise it has no technical meaning but it just a title. But we shouldn’t let titles dictate what we call things, otherwise we might, for example, think that Germany and Austria have no prime minister, or South Africa for that matter.

      This is not just your use, I see it quite often on this site. This was just a particularly good opportunity to make my argument, considering the fact that the difference is the very subject of your post!

        1. JD

          I always write ‘indirectly-elected heads of state’ and later in the paragraph ‘head of state’. I can imagine, though, that a more compact term would be desired. I have some ideas (republican figurehead; state chairman; first citizen; chief of state); none of them strikes me as being exceptionally well fit for duty, but I think they’re decent.

          1. JD

            Actually, the more I think about it, the more I like ‘First Citizen’, but I don’t mind if something else is chosen.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.