Basileus ton Rhomaion – “Emperor of the Romans”: sounds like a position of power and glory. Yet recent studies (by Ralph-Johannes Lilie or Anthony Kaldellis, for instance) have highlighted the instability and vicissitudes of imperial rule in Byzantium. Therefore (inspired also by Lilie), we made a new statistical analysis of the dynamics of the imperial regime. What were the chances for an emperor to stay in power for a longer period?
Assume that you are a Venetian banker asked by the Emperor in Constantinople for a loan which he promises to pay back over the next 20 years of his reign. Understandably, you would like to know the probability that your potential debtor will be actually able to do so – and to estimate the risk of an earlier change of ruler who may declare the obligations of his precursor null and void. To anticipate the results of our calculations: the chances that your client will make it the entire 20 years are bad.
We based our calculations on all rules in the Byzantine Empire for the period between the death of Emperor Constantine the Great (337 CE) and the Fall of Constantinople to the Ottomans (1453 CE); for the time of Latin rule in Constantinople between 1204 and 1261 CE, we used the data for the Emperors of Nicaea. Furthermore, we differentiated between rules with a non-violent and a violent start (the use of force to re-place the former emperor). You can find an overview of statistical results below.
Fig.: Statistical properties of the distribution of the durations of rules in the Byzantine Empire, 337-1453 CE (J. Preiser-Kapeller, 2015)
Although the mean duration of rules is more than 12 years, the lengths of reigns are distributed very unequally; 25 % of all emperors made it not beyond their third year, for instance. On the other hand, only 25 % ruled beyond their 19th year. A visualisation of the frequency distribution of rule durations makes these inequalities even more visible (see below); in contrast to an equal, so-called “normal distribution” (the red line), we observe a multitude of reigns below the mean of 12 years and a “long tail” of some few longer lasting rules.
Fig: The distribution of frequencies of durations of rules in the Byzantine Empire, 337-1453 CE (J. Preiser-Kapeller, 2015)
But as the data above indicated, chances for a long reign were even smaller if you started your rule with violence; in this case, the mean duration of rules is only ca. 10 years, and 25 % of emperors did not make it beyond their second year. These differences between rules with violent and non-violent starts become also visible in a so-called “survivorship-plot” (see below): the red line indicates the durations of all rules, the blue line of those with non-violent starts and the green line of those with violent starts. Clearly, we observe a divergence of survival chances between non-violent and violent rules after the seventh year (a “Seven Year Itch” for usurpers?) and a more rapidly declining rate of survival for violent rules afterwards.
Fig.: “Survivorship”-plot of all reigns (“total”), reigns initiated by violence (“violent”) and not initiated by violence (“non-violent”) for Byzantium, 337-1453 CE; the lines indicate the (declining) number of reigns which lasted up to that number of years, starting from 0 years.
You as Venetian investor therefore should keep in mind that your potential client has a 30 % chance to make it to his 20th year of reign if he has started his rule without violence – but only a 19 % chance if he did so.
The occurrence of ruler change and especially violent ruler change was of course not equally distributed across the Byzantine centuries; as every student of Byzantine history knows, there were periods of crisis and internal turmoil when emperors were replacing each other very quickly year after year. The graphs beyond illustrate this chronological distribution for all changes of rulers and for violent ruler changes.
Fig. Years of ruler change per decade in the Byzantine Empire, 337-1453 CE (J. Preiser-Kapeller, 2015)
Fig. Years of violent ruler change per decade in the Byzantine Empire, 337-1453 CE (J. Preiser-Kapeller, 2015)
Furthermore, in an earlier paper (https://www.academia.edu/3255595/Games_of_Thrones._The_temporal_dynamics_of_ruler_change_in_the_Roman_and_Post-Roman_World_0-800_CE_; in this paper you also find more technical background on our calcuations) we have compared the dynamics of ruler change in Byzantium with other polities across Eurasia for the early medieval period – and in some neighbouring polities, chances for a long reign were even smaller (only 25 % of all Caliphs made it beyond their 12th year, for instance). You as an investor therefore should be aware that a bet on a long rule of your royal debtor would have been a risky speculation in almost any case.
Fig: A violent start for a (relatively long-lasting) rule: the murder of Emperor Michael III by Basil the Macedonian, 867 CE (Skylitzes Matritensis)
Johannes Preiser-Kapeller, ÖAW – RGZM
Allison, P. D.: Event History Analysis: Regression for Longitudinal Event Data (Quantitative Applications in the Social Sciences). Sage, London 1984.
Berg-Schlosser, D./Cronqvist, L.: Aktuelle Methoden der Vergleichenden Politikwissenschaft. Einführung in konfigurationelle (QCA) und makro-quantitative Verfahren. Opladen & Farmington Hills 2012.
Blaydes, L./Chaney, E.: The Feudal Revolution and Europe’s Rise: Political Divergence of the Christian West and the Muslim World before 1500 CE. American Political Science Review, February 2013, 1-19 (doi:10.1017/S0003055412000561).
Box-Steffensmeier, J. M./Jones, B. S.: Event History Modeling. A Guide for Social Scientists. Cambridge 2004.
Eder, W./Renger, J. (eds.): Herrscherchronologien der antiken Welt. Namen, Daten, Dynastien (Der Neue Pauly Supplemente 1). Stuttgart – Weimar 2004.
Kaldellis, A.: The Byzantine Republic People and Power in New Rome. Cambridge, Mass. 2015.
Kantz, H.; Schreiber, Th.: Nonlinear Time Series Analysis. Cambridge ²2004.
Kokkonen, A./Sundell, A.: Delivering Stability - Primogeniture and Autocratic Survival in European Monarchies 1000-1800. QoG Working Paper Series 2012:3, Department of Political Science, University of Gothenburg April 2012.
Lilie, R.-J.: Der Kaiser in der Statistik. Subversive Gedanken zur angeblichen Allmacht der byzantinischen Kaiser, in: Ch. Stavrakos – A. Wassiliou-Seibt et al. (eds.), Hypermachos: Studien zur Byzantinistik, Armenologie und Georgistik ; Festschrift für Werner Seibt zum 65. Geburtstag. Wiesbaden 2008, 211-234.
Preiser-Kapeller, J.: Complex historical dynamics of crisis: the case of Byzantium, in: S. Jalkotzy-Deger – A. Suppan (eds.), Krise und Transformation. Vienna 2012, 69–127.
Preiser-Kapeller, J.: Games of Thrones. The temporal dynamics of ruler change in the Roman and Post-Roman World (0-800 CE), Working Paper: https://www.academia.edu/3255595/Games_of_Thrones._The_temporal_dynamics_of_ruler_change_in_the_Roman_and_Post-Roman_World_0-800_CE_
Thome, H.: Zeitreihenanalyse. Eine Einführung für Sozialwissenschaftler und Historiker, Munich – Vienna 2005.