A civil war reshapes an army into something more likely to turn on the state

The argument

Proponents of this position begin with what a civil war does to an army. Fighting one enlarges the military, raises its share of public spending, and expands the range of decisions it makes without reference to civilians. It also pushes officers into choices that are political in substance: which populations to treat as hostile, which irregular forces to arm, how to administer territory held by force. Proponents argue that an army which has spent years governing in this way does not readily return to being an instrument of policy, because it has been making policy. The argument is that war also supplies the grievances and the resources a coup requires. Defeat, or a negotiated settlement the officer corps regards as a surrender of what its soldiers died for, places blame somewhere, and politicians are the available target. Commanders acquire reputations and followings that function as political assets. Units become loyal to the officer who led them rather than to the chain of command. Proponents add that a long war erodes the distinction the civilian state depends upon, between the institution that holds force and the authority entitled to direct it, until officers come to see a government as a faction with opinions about the war rather than the body they are obliged to obey. Holders of the view conclude that conditions treated elsewhere as separate causes of coups are largely products of war. An army with an unusual share of the budget, autonomy from oversight, internal factions, battlefield grievances and officers of independent standing is an army liable to seize power, and civil war is the most reliable way of producing one. Proponents argue that this is why attempts concentrate in states that have recently fought at home, and why an explanation need not treat the two kinds of event as unrelated.

Premises

Counter-arguments

Critics argue that the association runs the other way or through a third factor. States that fight civil wars are states with weak institutions, contested legitimacy and unsettled questions about who may rule, and those features would predict both the war and the coup without either causing the other. They also note that many armies have fought long internal wars without turning on their governments, including some enlarged and heavily politicised by them, while coups occur in states that have had no civil war at all. Others object that the account leaves little room for the decisions of particular officers, treating a coup as something a war does rather than something people choose.

Rejecting the premises

[Rejecting P3] Many armies emerge from long internal wars with the chain of command intact, so politicised commanders and personally loyal units are not a general product of civil war. [Rejecting C] Weak institutions and contested legitimacy would predict both civil wars and coups, so the correlation does not establish that one produces the other.