Skip to content

Latest commit

 

History

History
91 lines (76 loc) · 3.12 KB

README.md

File metadata and controls

91 lines (76 loc) · 3.12 KB
Raw

wrds: Python Interface for querying WRDS datasets (CRSP, COMPUSTAT)

Author: Eddy (Edwin) Hu

Institution: Rice University

Contact: eddyhu at the gmails

What is it

wrds is a Python package for interfacing with WRDS data (currently local dbs only). It simplifies a lot of the standard data munging activities (querying, merging, cleaning) for CRSP and COMPUSTAT data, and allows the user to quickly access key variables of interest to conduct empirical analyses.

Querying simplified

Suppose you want to pull stock returns and market caps in order to do annual portfolio sorts on size. This is conceptually easy, but involves quite a bit of upfront data munging. wrds does all the data munging and makes accessing the relevant data as easy as:

import sqlalchemy as sa
import wrds

	msf_query = wrds.MSFQuery().query.alias('msf_query')
	q = sa.select([msf_query.c.permno,
               msf_query.c.date,
               msf_query.c.ret_adj,
               msf_query.c.me,
               msf_query.c.vweight])

Which creates a SQL statement:

	SELECT
		msf_query.permno,
		msf_query.date,
		msf_query.ret_adj,
		msf_query.me,
		msf_query.vweight
	FROM ...

the ... is where the magic happens. wrds makes sure that CRSP permnos are correctly lined up with the company names, computes the market cap, adjust returns for delistings, and computes annual buy-and-hold portfolio weights.

Pulling book equity from COMPUSTAT is just as simple:

	funda_query = wrds.FUNDAQuery().query.alias('funda_query')
	q = sa.select([funda_query.c.lpermno,
               funda_query.c.gvkey,
               funda_query.c.datadate,
               funda_query.c.be
               ])

Which creates a SQL statement:

	SELECT
		funda_query.lpermno,
		funda_query.gvkey,
		funda_query.datadate,
		funda_query.be
	FROM ...

wrds merges in CRSP permnos so that the permno-gvkey link is unique, and computes book equity following Fama and French (1993), and Davis, Fama, and French (2000).

Features

  • CRSP Monthly, COMPUSTAT Annual and Quarterly data
    • Aligns accounting fundamentals with market prices
    • Computes delisting returns Shumway (1997)
    • Computes annual buy-and-hold market equity for use as portfolio value weights
    • Convenience functions for accessing Fama-French Factors
  • Computes several anomaly characteristics
    • net stock issuance
    • composie equity issuance
    • total accruals
    • gross profitability
    • net operating assets
    • asset growth
    • investment to assets
    • return on assets
    • Ohlson's O-Score
  • Uses PostgreSQL (via SQLAlchemy) to access/store data
    • Easily configurable for MySQL
  • Returns query results as pandas.DataFrames or record tuples

TO-DO:

  • Implement DSF support
    • basic support added 2014/03/04
    • MOM
    • IVOL
    • BAB
    • CHSDP
  • MSF
    • aggregate market equity by permco