Big unsolved problems
- intertemporal pricing: equity premium, peso
problem (forward exchange rate)
- insurance: consumption correlation; insurance
pooling/adverse selection mechanism design
problem
- wealth and poverty over time and in cross-section
- the demographic transition
- intermediaries and specialization
- losses from rent-seeking
- "bubbles"
Smaller but significant issues
- x-inefficiency in monopoly
- tradeoff between numbers, observability and
patience in repeated games
General advice on writing a good thesis
There are many answers to the question of
what might constitute a good dissertation.
This is an effort to provide one answer to
that question. Part of writing a dissertation
is learning the skills required to be a practising
economist. Nowadays economists write scientific
papers and not books, and they often collaborate
with other researchers in conducting scientific
research. So a good thesis can constitute
2-3 essays, each written in the form of a
paper that might be submitted to a journal.
Collaborative essays with other students
are especially good.
A good essay would have three parts: a stylized
fact or facts to be explained; a theoretical
model to explain the facts and a determination
of how well the model succeeds in explaining
the facts.
- The literature on the stylized fact or empirical
puzzle should be reviewed and ideally, data
should be analyzed to verify the fact. As
a rule, stylized facts vary in how puzzling
they are. The more puzzling the fact, and
the more satisfying the answer, the better
the paper.
- The adequacy of existing explanations of
the fact should be discussed. What new
insight
do you have to contribute to the analysis?
You may have either a new idea, or wish
to
provide a carefully worked out model of
an
idea that has been discussed informally.
Focus on those features of the economy
that
are of greatest relevance to the facts
you
wish to explain. Do not add elements to
the
model merely because they exist in real
life:
the goal is to explain the facts as simply
as possible. If you wish to depart from
the
hypothesis of rationality, proceed with
care.
Although weakening of rationality may be
useful, try to avoid arguing that economic
actors behave in capricious or foolish
ways,
or designing modes of behavior that conflict
with the basic fact that most people behave
relatively sensibly in making day to day
decisions. Assuming that a small part of
the population is not rational, or that
individuals
are only approximately rational is plausible,
provided that this assumption plays some
real role in the analysis.
- There are several methods that you can use
to derive conclusions from your model. A
theoretical analysis will typically focus
on important special cases and prove theorems
concerning the working of the model. A simulation
analysis can generate summary statistics
for comparison with data. Econometric analysis
can provide useful summaries of the data.
Generally some combination of methods is
used. When econometric analysis is used attention
should be focused on the economic (as opposed
to statistical) significance of the parameters,
and whether the model does a good job of
fitting the data along economically relevant
dimensions (warning: statistical goodness
of fit statistics rarely provide a useful
answer to this question).