We do not have a Ph.D. program at Vanderbilt currently, so I am open to working with Ph.D. students at other universities. For instance, I have a paper (and hopefully soon another) with Chi Zhang who did his Ph.D. at Temple University, and is now faculty at UMass-Lowell.
I also periodically get emails from students who want to work with me. If you are in this category, please do the following:
- Replicate two papers that I cite at least once in my papers. They should come from a finance journal, the higher ranked the better. This will take you a while (several hours or perhaps a day or two), but this is helpful to you even if we do not work together because A. you see the details of how to generate real results and B. you begin to build an inventory of computer code you can re-use.
- If you want to email me first asking if you have chosen “good” papers, feel free to do so.
- You do not need to replicate the whole paper. Replicate at least summary statistics to show that you have the same sample as they do and one or two of their main tables of results – i.e. the results most important to the paper’s contribution.
- Be sure you have the data you need. There should be a “Data” section of some sort describing what data they use and how they get it. You will likely need access to Wharton Research Data Services (WRDS). This may not be necessary in all cases, but it will make things easier. You will not be able to replicate a paper with proprietary data, so check that ahead of time.
- You should use SAS, MATLAB, R, or Python. I’m pretty good at SAS, MATLAB and R, and I’m planning on learning Python.
- Send me both your code and your results. This could be a zip file with the code and an excel file of tables, or it could be an Rmarkdown document with your code and tables generated, etc. Please label your output so I can easily compare it to the paper (i.e. “This table replicates Table V in Blocher et al (2013)”)
- Do at least one robustness check. This means, change the analysis somehow and see if the result is still there. It could be changing frequency (Daily instead of Monthly) or time period (split the sample into an older and newer period) or perhaps changing the way something is measured (i.e. instead of dividing by Total Assets, divide by Total Book Equity).
- Note: the result might disappear. That is fine – tell me why you think that happened (after you check to be sure you didn’t make a mistake).
- In your email, please tell me the two papers you replicated (give full citations) and explain to me why you chose the paper you did (i.e. why it is interesting to you). Some of the reason may be data availability; this is understandable.
- These replications are the main portion of what I will consider when deciding if I will work with you. The tell me whether you are capable of executing a research analysis with skill and have basic programming ability.
- Include your CV as well as a list of your interests.
- Note, you are more likely to get a positive response if your interests correlate with mine. Don’t tell me you are interested in Labor and Health Economics and you want to work with me.
If I like your work, I’ll contact you and let you know. I may or may not have something immediately for you to work on. It depends on what ideas I have at the moment that interest me (and whether I’m teaching or not). My busiest teaching time is from October-December, so I will be slowest to respond then.