Data and Code 

Useful websites:

1. WRDS SAS applications: https://wrds-www.wharton.upenn.edu/pages/support/applications/ 

2. WRDS SAS Macros: https://wrds-www.wharton.upenn.edu/pages/support/research-wrds/research-macros-wrds/#analytical-macros 

3. SAS for Finance PhD students: http://pages.stern.nyu.edu/~adesouza/sasfinphd/index/ 

4. Brian Bushee's institutional investor classification data: http://acct.wharton.upenn.edu/faculty/bushee/IIclass.html 

5. Judson Caskey's data and Stata coding: https://sites.google.com/site/judsoncaskey/data 

6. Jie Jay Cao's website on coding and databases: https://sites.google.com/site/jiejaycao/home/tools 

7. Kai Chen's website on Python, SAS, and Stata: http://kaichen.work/?page_id=7 

8. Ed deHaan's data and code: http://faculty-gsb.stanford.edu/dehaan/SAS.html 

9. Hoberg-Phillips data library on text-based industry classification: http://hobergphillips.tuck.dartmouth.edu/ 

10. Mark Ma's SAS code and research note on two-way clustering: https://sites.google.com/site/markshuaima/home/two-way-clustered-standard-errors-and-sas-code 

11. Daniel Taylor's data and coding: http://acct.wharton.upenn.edu/~dtayl/code.htm 

Notes on analyst-related data: 

1. SAS Macro from WRDS linking IBES Ticker with CRSP Permno

2. Researchers start to explore the rich information in analyst reports, including myself - see my JAE 2018 paper on analyst ownership in covered stocks,  my RAST 2020 paper on analysts' treatment of stock-based compensation in their DCF models, and two working papers on identifying analyst initiations and analysts' non-GAAP exclusions. Reports can be downloaded from ThomsonOne (IE only) or Investext platforms. 

Notes on short-selling related data:  

1. Short-interest data. 

Short-interest refers to the number of shares that have been sold short but have not been covered yet. Usually we divide it by total shares outstanding to get short-interest ratio. The short-interest data is available in Compustat (comp.sec_shortint). 

2. Short-volume data. 

If we accountants think short-interest is like a B/S account, short-volume (or short-selling volume or short-sale volume) is like an I/S account - it refers to how many shares are sold short during a certain period of time. Put it simply, it is part of the total trading volume, which includes both short-selling volume and long-selling volume. Daily short-volume data is available on FINRA website from August 2009 here: http://www.finra.org/industry/trf/monthly-short-sale-transaction-files-2009. 

If you are interested in the short-volume data, I would recommend that you read this very interesting paper titled "Short Covering" by Blocher and Ringgenberg very closely. Their Table A1 lists all sources of short volume data for US equities. 

3. Short-trade data. 

Thanks to Reg SHO, short-selling data at the trade level are available at TAQ Reg SHO database from January 3, 2005 to July 6, 2007 for all short sales that were reported to the NYSE. I have not used this database. If you are interested in this database, please refer to prior studies such as this one by Engelberg, Reed, and Ringgenberg (2012 JFE).  

4. Equity-loan market data.

Researchers gradually realize that actual short-selling is a joint outcome of both demand and supply of equity loan market. Studies of lending supply mostly use data from Markit. If you are interested in this database, you can take a close look at this excellent paper by Beneish, Lee, and Nichols (2015 JAE). 

5. Reg SHO pilot program. 

The SEC conducted a controlled experiment from 2005 to 2007 to evaluate whether short-sale price tests can be removed. The experiment has been used by more than 30 studies in finance and accounting as an exogenous shock in short-selling constraints/threats, including my JAE 2017 paper on the role of short-selling threats in audit pricing. If you are interested, here are the List of pilot and control stocks and the SAS Code to identify them. 

Notes on institutional ownership data: 

1. Thomson-Reuters 13F data (S34) have serious issues in recent years. See WRDS Research Note, which advises researchers to use SEC 13F Holding data by WRDS.