This dissertation
contains three essays in three chapters. The first one introduces a game theory result in
repeated games with imperfect monitoring. The second chapter develops that result in a
political economy model. It proposes a unified model that explains three distinct
phenomena in politics: non-adoption of some ex-ante desirable policies, the election of
ideolgues by non-ideologues, and inertia or the persistence of bad outcomes. In the third
chapter I present an electoral model that shows why strategic considerations may prevent
parties from converging to the middle of the political spectrum. It also sheds some light
as to why ideology may simultaneously be viewed as a coherent set of beliefs about how the
world works and a dogmatic and rigid approach to policy that does not yield to evidence.
In the first chapter I present a model of imperfect monitoring in an infinitely repeated
game where a principal would like an agent to exert costly effort and has the ability to
punish him. In a standard setting, with imperfect monitoring, a scheme where the principal
will "punish the innocent" may arise. That is, in equilibrium the principal will
know the agent is exerting effort; nevertheless whenever he observes a deviation, although
he attributes it to noise in the monitoring mechanism, he will punish the agent in order
to prevent him from having an incentive to deviate. I make two substantial assumptions:
first, that punishment is not costless to the principal; the second assumption introduces
some incomplete information: the principal is not completely certain that the state of the
world is one in which the punishment scheme is worth its cost. Under these two conditions
the principal's ability to implement the scheme unravels.
One way of thinking about it is that in a finite setting the principal will not punish in
the last period and therefore no effort will be exerted. In the second to last period,
since the principal knows there are no gains to be had in the future, he cannot credibly
threaten to punish and therefore the agent doesn't provide effort. By backward induction
we know there will be no effort or punishments. The possibility of a bad state is similar
to having a finite horizon.
The argument is that there exists a bad-luck-path of play in which our monitoring
mechanism will always indicate a deviation and, therefore, a need to punish. This
punishment is costly and, more importantly, it deteriorates the confidence of the
principal. Down the bad-luck-path of play, an end period of enforcement is always reached,
the principal will eventually be uncertain enough about the true state of the world so
that she is not willing to pay the cost of punishment. From this end period of enforcement
I make a backward induction argument and show that there are never enough incentives to
carry out punishments.
The setup presented above, a principal that cannot punish without cost to herself, or
cares about the effects of punishment, and is at least slightly uncertain about the way a
punishment strategy will play out in the world, is readily applicable to political
economy. Indeed, the two assumptions that generate the main result found in the first
essay are natural in this context.
Chapter 2 presents a political economy model in punishments are costly for the government
and in which everybody understands that a credible threat of punishment by the government
will make citizens behave in a desirable way. However, the government, for the reasons
presented above, may be unable to commit to such a policy. Only a government that is
ideological, subjectively certain that it understands how the world works, can commit to
the ex-ante optimal policy. For the ideologue behaves as if there is no incomplete
information. This may lead the electorate to choose ideologues, even when they don't share
their beliefs. This ex-ante optimal delegation has its drawbacks: if the bad scenario is
realized, one where the punishment scheme is not optimal, then the ideological government
will persist in implementing an erroneous policy.
This model is an example of a more general point, that commitment to a policy is important
for its implementation. This model in particular explains why some high powered incentive
schemes that are suggested by mechanism design are not readily applicable in a political
context. The basic structure of the political (democratic) game may be incompatible with
such schemes. The model explains why we may sometimes choose people with extreme beliefs:
because they are uniquely endowed with a capacity to commit. They are well suited to
overcome the difficulties inherent to the political system. Not surprisingly, electing
this type of government might lead to inertia, i.e. the persistence of bad outcomes. The
model gives a good reason why we may have incentives to elect precisely these types of
governments, and why a politician may go to great lenghths to show how firmly grounded in
a certain ideology she is.
In the final chapter, I present a model that sheds light on two related puzzles that have
persistently plagued political economy in general and electoral models in particular.
One is platform divergence. Many of the most commonly used models in political economy
rely on a median voter argument and conclude that party platforms, particularly in a two
party system, will tend to converge on the prefered outcome of the median voter. I think
this puzzle is related to the way we think about parties, we are still a long way from to
understanding political parties and their preferences or ideologies and this model will
also point out some promising results in that direction.
The second puzzle is the definition and role of ideology. Some people who study the social
sciences may agree that it is important, but will still have a hard time agreeing on its
definition. In this essay I point out two particularly contradictory views on ideology.
The first conception is ideology as a consistent set of beliefs about how the world works
(or ought to work). The second one is that of ideology as a rigid set of ideas about
politics or social issues that is not malleable to evidence or discussion.
This essay presents a model in which two informed parties must choose their platforms
before an election. The electorate is less informed about the true state of the world.
Parties have information that is valuable to the electorate and it seems plausible that
competition amongst them might lead them to reveal it in exchange for votes. In this
admitedly simple model, I find that parties, out of strategic considerations, will find it
in their best interest not to reveal information. Each party will present a unique
(differentiated) policy regardless of the state of the world. The basic argument is that
if the electorate relies on just one party to make inference about the true state of the
world it is unlikely that a party will reveal an adverse (to its electoral fortunes) state
of the world. If the citizen must rely on both parties to pin down the true state, at
least one will have an incentive to deviate.
The model pins down a reason why parties have a strategic incentive to adhere to a
differentiated platform that is associated with certain prior beliefs about how the world
works. It also sheds light on the ambiguity regarding the interpretation of ideology. To
the uninformed electorate, parties will appear, in their choice of platforms, as
consistent with a coherent set of beliefs about the world. To the informed observer
parties will sustain their particular platform regardless of the information they possess
concerning its benefits; as a result, they will rightfully be seen as demagogues. A reason
for adherence by a party to a certain (more or less rigid) set of beliefs is thereby
suggested. The spirit of this result is consistent with the assumption used in chapter 2,
where parties are identified with, and commited to a particular set of beliefs. |