William T. Barndt
September 2005
Princeton University
## email not listed ##
Please do not cite or circulate without author permission.
4
Yet this story does not end there. When particular business elites choose to oppose an
assault, the form that opposition takes can alter both the likelihood of future assaults occurring and
the likelihood that future assaults will be overcome. When business elites choose to destabilize
governments, these likelihoods remain unchanged (or even increase). These likelihoods can
decrease, however, when business elites instead choose to invest in ad hoc private protections of
political liberty.
2
That overcoming executive assaults depends on the choices and actions of business
elites – rather than the working class, state agents of horizontal accountability, humans rights
groups, social movements, or the public more generally – is a disturbing conclusion that harkens
back to an older tradition of scholarship on democracy and development. Nonetheless, the
inference is clear: If assaults are to be overcome, business elites must be convinced to invest in
defenses against them.
The first chapter of this dissertation situates the problem in the existing scholarships on
regime politics and democratic politics. It then demonstrates that assaults are a persistent problem
in the region, using an original dataset of executive assaults in ten civilian-ruled South American
polities
3
since 1979. Chapter Two develops separate models of the three types of assaults, then
tests them using regression to analyze the above dataset. Chapter Three uses a combination of
statistical tests and process-tracing to analyze the conflicts that surround executive assaults. It draws
on two additional original datasets on assault conflict in Ecuador and Venezuela to identify what
types of people need to ally with the president in order for an assault to succeed – and what types of
people need to ally against him in order for an assault to be defeated. These datasets were compiled
using interviews and archival research during a year of field work in Ecuador and Venezuela.
2
In the absence of viable state protection of political liberty, these efforts provide the foundation for
developing what Helmke and Levitsky (2004) call “substitutive informal institutions.”
3
Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Paraguay, Peru, Uruguay, and Venezuela.
Convention