Skip to content

Latest commit

 

History

History
71 lines (60 loc) · 4.57 KB

thesis_contribution.md

File metadata and controls

71 lines (60 loc) · 4.57 KB

On the contribution of thesis

  • Academic contribution is now considered less important in defense.
  • However, your thesis should contain new elements.

What can be new elements?

  • What can your thesis add to the knowledge of people around you (vs human being)?
  • The new elements do not have to be great, ingenious, creative, new, etc.
  • You should provide answers to research questions
  • Asking a right question usually lead to good research.
  • But research questions can be much broader than you think. (see below)

Type of new elements

1. New empirical finding / theory / model / method / etc

  • New hypothesis, channels, econometric method, etc
  • However, there is nothing new under the sun.
  • Most of new findings are incremental or build upon existing findings.
  • There is always context why such new result came out.

2. Replication

  • Tweak on an existing (established/well-known) research
  • Extension of empirical evidence: China market data, etc
  • Application of theory/model paper to empirical data (mix and match)
  • But you need to justify why your topic is meaningful? Research question again!
    • Why testing against China data is meaningful? Any element in China data to strengthen/weaken the original paper?
    • The same approach has been tried before?

3. Horse racing / survey paper

  • Extensive summary of existing methods
  • Systematic comparison of the performance of existing methods / hypothesis
  • When and why such one method is better than the ohters.
  • Often, this of research leads to new ideas/finding.

Example 1:

  • One important question in market microstructure is where does the price discovery occurs when similar products are traded in several markets. (E.g., stock vs futures)
  • There are two important methods to measure the price discovery shares:
  • There is a paper systematically comparing the two is as famous as the two original ones:

Example 2:

  • SABR volatility model is one of the most popular stochastic volatility model to explain volatility smile.
  • The original paper provided an approximate formula to quickly compute the Black-Scholes volatility. However, this formula is not very accurate.
  • Several improvements have been proposed. Each claim that the method is superior to the original method. However there is no research comparing the new methods.
  • So I posted a question on Quant StachExchange. No answer. So I proposed this as a thesis topic.
  • Zhang, Z., 2019. Who is the winner of the analytic approximations for the SABR model? (mathesis). Peking University HSBC Business School, Shenzhen, China.
  • Drawback: lacks

Suggestion for Choosing Topic

  • Topic does not have to be so big. You may set a small goal, but plan to achieve it thoroughly. Then, ideas will follow soon.

    • Though beginning seems humble, future will be prosperous
    • Start from where you are, not from where you want to be
    • Stay hungry. Stay foolish.
  • Find topic around you:

    • Any topic related to courses you took?
    • Any topic you were curious?
    • Any topic you think is not correct?
    • Any topic you want to thoroughly understand (i.e. survey)?
    • Can you improve any course project you did?
  • Be creative and critical in choosing topic. Keep asking questions!

  • Regardless of the "contents" of the thsis, you have to try your best to see your paper (make it look good).

    • Argue the reseasrch question is important. (Even if it is not so important.)
    • Argue your approach is meaningful. (Even if it is not so meaningful.)
    • Argue you made a contribution. (Even if there's not much contribution.)
    • In order to make better arguments above, you have to think/read/study very hard.
    • It will be a great learning process you take away with graduation and a great skill set for your career.

Good luck!