Skip to Main Content

Knowledge Base

Company Financials / Fundamentals Sources for Lots of Data

Below are commonly used sources in academic research for company fundamentals or directories.

Public Companies

These databases are from different vendors and so have slightly different coverage due to different data collection methods (see this paper that compares them). All are available in WRDS – see WRDS access help.

  • Compustat – From Standard & Poors (S&P). It is the easiest to link to stock data in CRSP. See how to download data from Compustat. Typically used for U.S. and Canadian firms, but some data has worldwide coverage.
  • Osiris – From Bureau van Dijk (BvD). Has worldwide coverage.
  • Worldscope – From LSEG (formerly Thomson Reuters), and previously owned by FactSet. Has worldwide coverage.
  • Amadeus – From Bureau van Dijk (BvD). European coverage only, and because of that it tends to have more variables than the others.

Private Companies

Private company fundamentals and information is harder to find. Some options are:

Startups

Similarly to private companies more generally, information on startups is challenging to collect and access. Some options are:

Establishment-Level Data

These datasets provide establishment-level data (e.g. information about each location of a big box retailer) as a time-series. They cover public and private companies.

  • D&B Historical US – see Redivis access help – From Dun & Bradstreet /Mergent. U.S. data from 1969-2022 (excluding 1981 and 1984). This dataset underlies the NETS (National Establishment Time Series) database, which Stanford does not subscribe to.
  • Data Axle Reference Solutions Historical Data – see Redivis access help – From Data Axle (formerly Infogroup) and formerly called Reference USA. U.S. data only, from 2010-2022.
  • Orbis Historical – see Orbis Historical access help – From Bureau van Dijk (BvD). Has worldwide coverage, with some countries having data back to the 1980s.
    • The 'Shareholders' table allows you to link parents with subsidiaries as a time series back to ~2007.
    • Subsidiary-level information is 'header' data, so to build a time-series you can use the unique cuts of the data we have received over the past ~5 years.




Answered By: Alice Kalinowski
Sep 14, 2023

Related Library Tips

    Resource Use

    Most resources are only available to current Stanford students, faculty, and staff.

    Researchers are responsible for using these resources appropriately. See the eResources Usage policy.

    Accessibility Support

    Ask Us for accessibility support with library resources.