Economic Research: Creating, Writing, Presenting

1
Economic Research Creating, Writing, Presenting

• Maarten C.W. Janssen

2
Aim, Background of the Course

• Assumption 1 You want to become known for your
  research in the international scene
• Assumption 2 You have the necessary skills and
  knowledge to do research in your field
• But this is by far not enough (only 50 of work)
• You have to do your own marketing
• Competition for attention
• Backward induction

3
You can be very smart, but

• You are (very) unlikely to be in the position to
  create a new line of research
• Research is building on others, linking to
  others, get feedback from others
• Network is important, internally in the school
  (among PhD students), externally in your field
• Collaboration can be inspiring
• Being recognized is about how you connect to
  others others will help you by giving their
  feedback

4
What is my message and for who?

• Why is my work important?
• Do I contribute to existing literature? If so,
  what? And why is this literature important?
• Do I have a methodological contribution? What are
  the (potential) applications?
• Do I contribute to an understanding of some real
  world issues? What other contributions are there?
  Are they complementary? Is my contributions
  better, more plausible, why?
• Which people would be most interested in my work?
  Do I have different lessons for different
  audiences?

5
Lessons to be Learnt from JET editorial report 1

• Does my paper really matter?
• There is some literature (quality chosen, price
  signals quality?) Is my modeling more
  appropriate?
• What is wrong with previous papers? Do I have
  evidence of this?
• If it is a theory paper, it should have useful
  machinery
• Presentation is way too awkward
• Equations too dense
• Not adequate notation
• What is a real result, what just some technical
  lemma?

6
Creating

• Start naively
• I have a data set
• I saw some paper and checked whether results hold
  under alternative assumption (robust)
• I was struck with a newspaper article, everyday
  life
• In the US without US adaptor
• Taxi driver does not want to take you
• Sitting with wife in the car, to buy gasoline or
  not?
• Adverse selection is not a problem (bachelor
  student)
• Anything is good, but soon you have to ask what
  is my message and for who?
• You may change the answer all the time, but you
  need to know where you intend to go to direct
  your research efforts
• Remain open to alternative messages,
  interpretations
• Read nontechnical, non economic literature
  institutional detail
• Can I publish about Russian economy in
  international journals?

7
During research

• Think about what are the possible outcomes?
• Theoretical research Any equilibrium has certain
  properties? Or, there is an equilibrium with
  certain properties?
• Empirical research economic and statistical
  (in)significance?
• Very different things
• Is any outcome interesting? But for different
  reasons?
• Expected cost/benefit analysis

8
Start simple

• Theoretical research what is the simplest model
  that you can think of that captures the
  phenomenon you want address?
• What is essence?
• US Electricity adaptor
• Adverse selection
• Empirical research look at the data, do also
  simple descriptive statistics.
• Helps enormously to focus main message, get it
  across.
• Research is then mainly a robustness check
  (helping you to sharpen and deepen main results)
• Adverse selection

9
What if you get the opposite result of what you
expected?

• Ask yourself always why you get a certain result?
  
• Can I explain the result without going into
  technical details?
• Is the result correct? (often, when you cannot
  explain there is a mistakesimulations?)
• If so, what causes this unexpected result?
• Is it an unimportant assumption that can be
  replaced by another one?
• Or does my intuitive result only hold under some
  conditions (and can I understand these conditions
  intuitively)
• Learn from intermediate results you get, relate
  back to main message? Should it be reformulated?
• What are the crucial assumptions, what not?

10
Role of Simulations, Numerical results

• Depend on area
• Computational heavy (dynamic macro, or not)
• But generally,
• It is not considered as a substitute for a proof
• It is nevertheless increasingly useful
• Plot f(x) lt 0, but not for f(x a,ß,?) lt 0
• Get an idea whether your intuitive result is
  correct or whether there are counterexamples
• Aid understanding by giving a flavour of
  magnitude of effect (economic significance)
• Draw pictures
• Check robustness (where model is analytically
  untractable)
• Be open and honest about which results are
  analytic and which you obtained with simulations
• Do not hide unclear numerical analysis in proof
  in appendix, (especially not if you do not have
  clear intuitive explanation for your claim)

11
Read, but not too much

• You have to be able to relate what you have been
  doing to what others have done
• You have to know that
• Especially in an oral presentation you cannot
  hide behind others this is a crazy assumption,
  but others also make it
• Reading too much may prevent you to be original
  (as you tend to copy things you have read)
• People do not want to hear (or read) endless
  literature review (Murayev Boulatov, finance
  presentation)
• Read at different stages differently
• In beginning did someone do what I intend to do
• Later how exactly do I differ? Which crucial
  assumptions, timing of events? What if I
  incorporate some of their assumption in my model?
• You should be able to defend your crucial
  departures from the literature
• Yankelevich presentation behaviour of shoppers

12
Talking, talking, talking

• You only understand your own research when you
  can explain it.
• When you have to explain to someone else, you
  have to distinguish between main and side issues.
• Often prevents loopholes in argumentation.
• Talk during lunch, come to the office
• Organize your own (brown bag) PhD research
  seminars where you discuss among each other!

13
Presenting, Writing

• Rule 1 (almost) never follow the line of
  research that lead to the results
• Rule 2 be enthousiastic about your results
• If you say I did some exercise and this is the
  result you do not motivate people to pay
  attention to you.
• Be honest if there is a critical assumption, you
  cannot get rid of, or if you do not know what
  happens if you change it, say so
• But give it a positive turn (future research, I
  would be a hero in this branch of literature if I
  could)

14
Presenting, Writing

• Learn from others
• Why did I like this paper, presentation?
• Why did I get board?
• How do I read other peoples work?
• When do I loose concentration during a
  presentation?
• Develop your own style

15
Peoples time span of attention

• 1 minute people want to know whether it is
  worthwhile to listen, read further
• What is the topic, what can they gain?
• Title (abstract) is very important
• 5-10 minutes which main results do you have,
  what do you add to what is known
• Introduction should contain this
• 1-2 hours get into more details, how do you get
  these results, which methodology, which data,
  model?

16
Titles

• Do they attract attention?
• Convey topic of the paper, main result?
• Examples
• Going where the Ad Leads you
• Do Auctions Select Efficient Firms?
• Can we Rationally Learn to Coordinate?
• Non-exclusive Conventions and Social Coordination
• Signaling Quality Through Prices under Oligopoly
• Job market papers at HSE
• Can you trust your broker?
• If you pose a question,
• It should be attracting attention
• You should answer it clearly, and it should have
  element of surprise

17
Introductions I

• Most critical part of a paper, usually rejection
  is based on Introduction only
• Requires writing, rewriting, .., 5-6 times
• How to start, where to explain your insights
• Get as quickly as possible to your results
• Start with the most key papers and what they lack
  if you have a pure theory, methodology
  contribution
• Start with describing real world issue and what
  you aid in our understanding of it (do literature
  review then afterwards, and sometimes even before
  the conclusionsif your breakthrough can only be
  really described after you have presented it)
• Adverse selection paper as example

18
Introductions II

• Describe main features of your set-up concisely
• Describe your main results
• Why they are interesting (adds to methodology,
  adds to understanding)?
• Why you get it (you should be able to explain the
  main mechanism(s), not I worked hard to crack the
  math puzzle and this is the answer
• Only a few readers will ever go through all your
  proofs (even referees seldom do) you have to
  write so that people understand it and believe
  result once you have made them think about it.
• Relate to several literatures (if possible), to
  show breadth.

19
What comes next?

• Model?
• Often yes, but try to be creative
• A simple 2x2 example capturing the main features
  of paper (Kandori, Malaith, Rob, 1993 Ectrica
  2010 EJ paper)?
• Extensive Literature Review?
• Description Institutional Context or Data?
• Literature review should not be a list of short
  statements about other papers.
• Make a story with your paper and results as the
  central theme. Start with fundamental papers, say
  how other papers have built on it and how you
  continue (or break with that line)

20
Once you get to a formal model

• Take the lessons of Thompson to heart
• (Here is one example where an oral presentation
  is different from a written one)
• Notation
• Follow existing literature as much as possible (f
  ma)
• Use natural symbols
• Is notation essential (if used only once or
  twice)
• Explain assumptions and their interpretation (are
  they usually met? What is their role .)
• And if you can do without the word assumption, it
  is better.
• We assume there are two periods. In each of these
  periods, individuals are assumed to work and we
  introduce the assumption that the utility
  function is concave. Or
• Consider a two-period economy where individuals
  have a concave utility function and work in
  every period.

21
Present Formal Results

• In main body of the text, people want to see
  things that stick out (propositions, hypotheses,
  results, statements)
• Indicate them as such (with open space)
• Take care these are very precisely formulated
• Make sure that they correspond to the main
  results of the paper as explained in the Intro.
  Déjà vu, but in a more formalized manner
• I have seen main results being presented
  somewhere in a footnote (and vice versa)
• And spend at last one paragraph introducing the
  propositions and (after stating them) explaining
  them.
• Think about proofs in appendix, properly bordered
  in the text, or informally presented as part of
  the flow of the text (what is audience, what is
  purpose of proof?)