Sources: Jessica Weeks on Autocracies and Conflict



At a conference in Tel Aviv last week, Jessica Weeks (Cornell University) presented an interesting paper, "Strongmen and Straw Men: Authoritarian Regimes and the Initiation of International Conflict."

The standard finding in the democratic peace literature is that autocracies are more prone to international conflict than democracies. Weeks argues that not all autocracies are created equal. In particular, they vary along at least two dimensions: whether leaders enjoy a lot of discretion (personalist rule) or are checked in various ways, for example by political parties; and whether the leadership is civilian or military. The result is a fourfold typology of autocracies.

Civilian leadership Military leadership
Personalist, unconstrained leaders Bosses 

(Kim Jong Il, Mao, Castro, Qaddafi)


(Nasser, Idi Amin)

Non-personalist, elite-constrained leaders Machines 

(Chinese and Soviet Communist parties)



Weeks argues that personalist leaders and those with military backgrounds are more likely to take risks and initiate international conflicts. The former are more conflict-prone because they are unchecked; the latter because they bring military views of the world to the table. Using a massive dataset of interstate conflicts from 1946-1999 (nearly a million dyads in one baseline model) she finds that bosses, strongmen, and juntas are indeed more prone to conflict than authoritarian machines subject to some checks or "audience costs."

Why is Weeks’ paper relevant to our parochial interest in North Korea? Because some authoritarian regimes really are worse than others. We have long argued that North Korea could benefit from pursuing a reformist, Chinese-style economic path; such a policy shift would probably require greater empowerment of the cabinet and technocrats. Similarly, the Kim regime might be less bellicose if the leadership was checked by a more robust and functioning party.

Not all the news is hopeful, however. Weeks finds that strongmen and juntas are also more conflict prone than machines or democracies. Depending on the transition path, we could be no better off than we are, particularly given the ongoing emphasis on “military first politics.”

More From