This article called “Publication, Publication” argues that a great way to write a publishable paper is to start with the replication of a published article. This replication exercise will count towards 80% of your final grade. You may either work alone, in which case you need to replicate one paper, or work in pairs, in which case you need to replicate two. For choosing papers, you have three options:
One is to choose one paper in the below indicative list of papers to replicate; I have a weak preference for this option.
If you are not interested to replicate any of these papers (for instance, because this is not the type of research that you would like to be writing), you may choose any alternative paper from the non successfully replicated paper in this list - they have either missing data or code: Andrew C. Chang and Phillip Li. “Is Economics Research Replicable? Sixty Published Papers From Thirteen Journals Say “Often Not”", Forthcoming at Critical Finance Review. The list of papers is in the online appendix of that paper - again, you should choose one of the paper for which the data is public, but which failed to replicate for reasons of missing data, or code, or both.
You may replicate any other empirical macro-finance paper. However, in this case, the code and data should not be available from the authors’ website - or at least, the results should not replicate using only what the authors have provided. The idea is for you to engage in some replication of the results. You should send me the paper you would like to replicate for approval. In particular, I will not allow papers which are too “financy”, “marginal” or “inside baseball”, in the sense that it should have implications for macroeconomics as well. The paper should be as general interest or policy relevant as possible. For a finance paper, there should be some implications for investment behavior - it should be Macro-Finance, not pure asset pricing.
Note that some of these papers do not always have regressions in them - some are just “facts” papers. However, this papers are sometimes very interesting. Descriptive statistics are sometimes useful, and it’s also useful to investigate how robust they are.