The Military in Politics: Complications of Counting

The following guest post is by Jonathan Powell, an Assistant Professor at Nazarbayev University, soon moving to warmer climates at the University of Central Florida.

The Military in Politics: Time for a Post-Mortem?

Andrea Kendall-Taylor and Erica Frantz (KT&F) recently took a fascinating look at authoritarian politics, specifically regarding the means by which autocrats leave office. They focus on the waning of coups d’état and the rise of popular revolts, as well as the potential implications that each form of regime change has for democratization. KT&F suggest that the increase of popular revolts relative to coups has potentially positive effects for democratic transition.

An increasing number of accounts, see here, here, and here, for example, and this most recent effort by KT&F in the Washington Quarterly, supports this interpretation. Coups are a less

frequent means of exit in the contemporary world. However, I’d like to add two very important distinctions before beginning a post-mortem on the coup phenomenon.

Exit versus Entry

First, the means of a leader’s exit isn’t necessarily the same as the successor’s means of entrance. Popular revolts and military coups aren’t mutually exclusive means of political transition. Hosni Mubarak, for example, is considered to have exited via a revolt. While some might consider his resignation to be coerced by the Supreme Council of the Armed Forces (SCAF) and thus qualifying as a coup, Clayton Thyne and I do not need to make that distinction in classifying the case in our dataset. This is because we are completely confident that SCAF’s seizing of power following Mubarak’s resignation was clearly illegal and can be classified as a coup.

Focusing on leader exit undercounts coups by ignoring how power vacuums end. For example, it would be difficult to not consider Captain Moussa Dadis Camara’s seizure of power in Guinea in late 2008 a military coup. The same could be said about the Togolese military’s installation of Faure Gnassinbé in February 2005. None of these, or similar events, are considered coups when relying on data solely reporting leader exits.

So what happens if we look beyond exit? Below I offer two different illustrations of ousters, with each figure reporting a 5-year moving average that reports the ratio of executive transitions that included either a coup or a revolt. The figure on the left illustrates the data utilized by KT&F, and is effectively identical to their original figure that aggregated ratios by decade. The figure on the right uses the same data for revolt but incorporates (successful) coup data from my project with Clay (“P&T Coup”).

In contrast to continuing the decline of the last three decades, considering a wider group of coups shows their ratio of authoritarian transitions has remained steady since the end of the Cold War, though the raw number has decreased. Further, coups have a lower share of ousters than in the Cold War but they still (marginally, at least) outpace popular revolts as a characteristic of political transitions when we move beyond focusing on exit.

Blurred Lines of Dissent

A second important distinction regards the potential to conflate the two types of event. Milan Svolik’s codebook defines revolts as instances when “the leader was forced out of office by a mass protest, uprising, strike, or riot.” This definition itself isn’t an issue and is in line with KT&F’s treatment of revolts. However, Svolik offers one important distinction: “revolts among military units that led to the exit of a leader were coded as a revolt instead of a coup if they involved mass public participation.” In other words, what could otherwise be classified as a coup would be reclassified to a revolt if the coup was accompanied by a substantial mass uprising, riots, strike, etc.

I’ll walk through an example to illustrate this. John Samuel Fitch’s seminal The Military Coup d’état as a Political Process closely examined Ecuador’s coups of the mid-20^th century. Fitch remarks on close similarities between events in 1966 and 1961, both of which he considers a coup (as do Clay and I). Fitch also mentions that “the antigovernment riots and protests of 1966 exceeded the 1961 disorder in frequency and intensity by a considerable margin…” It is likely due to this dynamic that Ecuador’s 1966 putsch is considered to be a popular revolt in spite of the ouster being carried out by the military high command and being classified as a coup by many other observers (see, for example, the ongoing coup list from the Center for Systemic Peace, Samuel Finer’s The Man on Horseback (2002), Morris Janowitz’s Military Institutions and Coercion (1977), Gavin Kennedy’s The Military in the Third World (1975), Gregor Ferguson in Coup d’etat: A Practical Manual (1987), and William R. Thompson’s Grievances of Military Coup-Makers (1973)).

We also should not discount the military’s ability to take on lesser roles during the transition. Svolik’s incredibly comprehensive data also indicate that “the military overtly participated in the leader’s exit from office” for nine different revolts. Though this might seem like a small number, it represents nearly one-third of the revolt exits in his data, while around a half dozen of the cases could directly be coded as coups in the data offered by the Center for Systemic Peace or Powell & Thyne.

This is not to say that the masses do not matter in these hybrid coup/revolt scenarios, or that the data are incorrect, or the conclusions reached by KT&F regarding prospects for democratization are incorrect. Rather, different research questions will lead to different emphases in coding, and drawing out these distinctions could potentially improve the findings of the investigator. A coup during a mass uprising could reflect a response to the popular will, as seen in 1966 Ecuador. Or it could be little more than a pretense for preserving the military’s privileges, as seems to be the case in Egypt. The different pathways states can take makes this a particularly interesting area of inquiry. Hopefully the effort of Kendall-Taylor and Frantz represents only an early step in a research agenda that more carefully considers how specific qualities of political transitions influence the long-term political development of states.