List of Potential Final Projects#
The project that you are assigned will come from one of the following three categories: papers, data series from “Evidence from Many Asset Classes”, or “other”.
Papers#
Lewellen (2015), Critical Finance Review. “The Cross-section of Expected Stock Returns”
Replicate Table 1, Table 2, and Figure 1
Data used: CRSP and Compustat
-
Replicate Table 1
Data used: Bank Call Reports. See https://wrds-www.wharton.upenn.edu/pages/get-data/bank-regulatory/
-
Replicate Table A1 and Figure A1
Data used: Bank Call Reports. See https://wrds-www.wharton.upenn.edu/pages/get-data/bank-regulatory/
Koijen and Yogo, A Demand System Approach to Asset Pricing (2019). Journal of Political Economy
Replicate Table D1
Data used: Securities and Exchange Commission Form 13F, via Thomson Reuters. See https://wrds-www.wharton.upenn.edu/pages/get-data/thomson-reuters/
Li Wang (2023), Valuation Duration of the Stock Market
Replicate Table 1
Data used: S&P500 Futures (authors use Datastream, you may use the Bloomberg terminal), Fama-Bliss database, CRSP, and S&P Global via Goyal and Welch (2007).
Peng Wang (2023), Factor Demand and Factor Returns
Replicate Table 1 and Table 2
Data used: CRSP, Securities and Exchange Commission Form 13F, via Thomson Reuters. See https://wrds-www.wharton.upenn.edu/pages/get-data/thomson-reuters/
-
Replicate Table 1 and part of Table 2 (the betas and the R-squared, but not the standard errors)
Data used: S&P 500 futures (Source: Bloomberg) and Fama-Bliss database
Nagel, Stefan. “Evaporating liquidity.” The Review of Financial Studies 25, no. 7 (2012): 2005-2039
Replicate Table 1 and Table 2
Data used: CRSP
Some helpful code may be found online: https://voices.uchicago.edu/stefannagel/code-and-data/
Siriwardane, Sunderam, Wallen (2022). Segmented Arbitrage
Replicate Figure 1 and Table 1
Data used: Bloomberg
Bao, Pan, and Wang (2010). The Illiquidity of Corporate Bonds
Replicate Table 1 and Table 2
Data used: FINRA TRACE
Some code is available online. See here.
-
Replicate Table 1
Data used: FINRA TRACE
Coda and data is available online: https://openbondassetpricing.com/code/
Since code and data is available, part of this project is cleaning up the code and making it as presentable and well-documented as possible.
Data Series from “Evidence from Many Asset Classes”#
He, Kelly, and Manela (2017) above is a well known paper that examines the effects that intermediaries’ balance sheets have on asset prices. This paper happens to test this theory on a variety of asset classes, going beyond just debt and equity. The projects in this section will be to replicate the test asset returns across this variety of asset classes. Each project below focuses on a different asset class. The returns of these tests assets are provided here. The data provided in that file goes up until 2012. Your project will be to replicate that data up until 2012 and then provide updated returns going as close to the present as possible. The derivation of the returns is defined by a separate paper. These are listed below.
-
Replicate the corporate bond columns from He_Kelly_Manela_Factors_And_Test_Assets_monthly.csv
Data used: Bond Returns by WRDS, but maybe also FINRA TRACE, Mergent FISD/NAIC, TRACE, Lehman Brothers Fixed Income Database.
You may use data from openbondassetpricing.com/ and here if it’s helpful.
-
Replicate Table 2 and Table B1. Table 3 or 4 may be difficult, but may be attempted as a stretch goal.
Data used: OptionMetrics. CBOE Options data or Option Suite by WRDS may be helpful.
Palhares, Diogo. Cash-flow maturity and risk premia in CDS markets. The University of Chicago, 2013.
Replicate the Credit Default Swap columns from He_Kelly_Manela_Factors_And_Test_Assets_monthly.csv
These columns are described in He, Kelly, and Manela. (2017), but the definition of returns on CDS contracts comes from Palhares (2013).
Data used: Markit
Borri and Verdelhan (2011), “Sovereign Risk Premia”
Replicate Table 1 and the Sovereign Bond Columns from He_Kelly_Manela_Factors_And_Test_Assets_monthly.csv
Data used: Datastream or Bloomberg
-
Replicate Table 1 and the first 6 FX columns in He_Kelly_Manela_Factors_And_Test_Assets_monthly.csv
Data used: Datastream or Bloomberg
-
Replicate Table 1. It may be difficult to replicate the commodities columns in He_Kelly_Manela_Factors_And_Test_Assets_monthly.csv, but you may try.
Data used: Futures Contract data from Datastream
Other#
Monash Time Series Forecasting Archive
Replicate Table 1 and Table 2
Code and data are already available online. (See here and here.) This project is to refactor the code of the archive so that it all runs via PyDoit (i.e., the data is automatically downloaded and the tables are all automatically generated).
Some prior familiarity with the R programming language would be helpful for this project.
Do short sellers respond to ESG ratings? (extra credit available)
Work with sponsoring adviser to explore the following question: “There has been a recent focus on ESG (Environmental, Social, and Governance) risks in finance, and several papers have studied the relationship between a firm’s ESG rating and its subsequent stock returns. This project aims to study whether short sellers appear to incorporate information in ESG ratings in their investment decisions, which has not been studied. The project will involve merging and analyzing data from 2 main sources: S&P Global / Markit Securities Lending data and RepRisk ESG data.”
Merge Markit securities lending data with RepRisk Data Feed (Risk Incidents)
For each ESG incident/severity/novelty/news’ reach, produce summary statistics of the following variables: short interest ratio (short interest shares / shares outstanding), loan supply ratio (shares available to be lent / shares outstanding), loan utilization, and loan fee.
Data used: Markit securities lending and RepRisk
Replicate several portfolios in in Ken French’s Data Library
Reconstruct the “bivariate sorts on size, e/p, cf/p, and d/p”. these are the portfolios formed on size and earnings/price, size and cashflow/price, and size and dividend yield. Use the unaggregated CRSP and Compustat data and reconstruct the portfolios that match.
Reconstruct the “bivariate sorts on Operating Profitability and Investment”. This consists of three files, each with 25 portfolios in it. Reconstruct these from unaggregated CRSP and Compustat data.
Reconstruct each of the “5 industry portfolios” and “49 industry portfolios” from “scratch.” That is, use the unaggregated CRSP data and reconstruct the portfolios that match.