regimes—partial autocracies and partial democracies—are substantially more vulnerable to crisis
than their more “coherent” counterparts.
The real surprise came from the predictive power associated with the addition of a single
element of this scheme. By distinguishing partial democracies according to the presence or
absence of factionalism in political participation, we found we could substantially improve the fit
of our models. As measured by Polity, factionalism occurs when political competition is
dominated by ethnic or other parochial groups that regularly compete for political influence in
order to promote particularist agendas and favor group members to the detriment of common,
secular, or cross-cutting agendas. Factionalism typically occurs in new democracies, where
party systems are weak and political participation is more likely to flow through networks rooted
in traditional identities or other parochial interests, but it can emerge in established democracies
as well. Factionalism is also the most common form of participation in autocracies that do not
repress political competition, either by design or incapacity. Recent examples include Bolivia
and Ecuador, where the mobilization of indigenous movements has produced stand-offs over the
appropriate uses of mineral wealth; Venezuela, where polarization along class lines has produced
mass protest and a failed coup in recent years; and Bangladesh, where personalist rivalries have
dominated national politics since a return to democracy in 1991.
The link between factionalism and instability may sound unsurprising, so much so that
one might wonder whether factionalism itself is simply part and parcel of the instability that
ensues. Our analysis suggests that it is not. Only about half of countries coded with factionalism
develop instability in our data; thus knowing a country is factional in its political competition
still gives you no better than a 50-50 guess about its stability status two years hence. It is only
when factionalism is combined with a relatively high level of open competition for office