Cardona
MPSA 2004
7
significant regime instability, the two parties have been able to come together to create
power-sharing agreements that preserve institutional order.
6
It is the entanglement of the public forces in this partisan struggle that connects
them to the problem of constitutional governance. We can see this dynamic at play during
the La Violencia period of undeclared civil war that took place between 1946 and 1966 in
Colombia. What emerges from Colombia’s multi-level configuration of fuerza pública,
with numerous points of political control, appears to be a structure that allows for the
prolongation of local-level conflict, without any actor in the system that sufficiently
concentrates power to overthrow the other(s) definitively.
The decision to create decentralized public forces that were beholden to local-
level appointed officials ensured that at the local level, incumbent actors had private
armies that they could use as personal instruments. Paradoxically, this “honeycombed”
structure of the Colombian public forces during La Violencia may have permitted a
greater degree of regime stability than would have been otherwise possible precisely due
to the multiplicity and relative balance of armed actors. In other words, the decentralized
nature of the public forces, as they emerged from the critical juncture, put Colombia on a
path in which the regime was susceptible to stalemate.
7
Had the armed forces been more
centralized the alternative along this path may have been revolution, a state of affairs in
6
The primary example of this is the National Front power-sharing agreement of 1958, in which the
Liberals and Conservatives agreed to alternate in the presidency over the next four terms, and to allocate
administrative posts 50-50 between them.
7
In the interests of space, I have not included here an explicit discussion of the theoretical framework of
critical junctures and path dependence that underlies this analysis. Briefly, the task of a critical juncture
argument in the style of Collier and Collier (1991) or Mahoney (2001) is to demonstrate that the causal
dynamics at play after the critical juncture cannot be deduced from those at play beforehand. That is, the
critical juncture must alter the causal relation of the variables in question in such a way as to constitute a
genuine break between past and future. As for path-dependent arguments, as Thelen (2003) and Boas
(2002) demonstrate, several variants of these arguments exist. The kind that works best for the institutional
design of the public forces is a “layering” model, in which gradual changes in form and function affect the
evolution of an institution over time, but within a set of parameters recognizable from the origin.
Convention