Skip to content
All library documents

Using SIC Codes to Build Scalable Equity Peer Groups

Article Quant Q&A · Author: ppt
Read the original

Summary

The document considers how to identify comparable companies for large-scale equity screening when standard sector and industry labels do not match intuitive peer groups. The suggested approach is to use SIC classifications from SEC filings and construct custom industry groupings from those codes. This offers more flexibility than relying only on broad categories supplied by a financial data service.

The answer points to an industry portfolio scheme that groups several large technology companies together, illustrating that researchers can aggregate classification codes into categories suited to their analysis. It also notes GIC as an alternative classification, while describing it as proprietary. The exchange does not provide an implementation, test peer similarity against financial outcomes, or address how to handle diversified firms and changing business lines. SIC-based grouping is presented as a useful starting point for peer construction rather than a definitive measure of economic comparability.

Key ideas

  • SEC filing data can be used to retrieve companies’ SIC classifications.
  • SIC codes can be aggregated into custom industry groups for peer comparisons.
  • Custom groupings may better fit a screening model than default sector labels.
  • A published industry portfolio scheme illustrates how multiple firms can share a broader group.
  • Industry codes provide a starting point but do not prove that companies are economically similar.

Tags

EquitiesFactor investingStatisticsUS markets
Full text
# How could I identify peers of companies (stocks) at scale?


# How could I identify peers of companies (stocks) at scale?












Context:

I'm making a small script to screen through thousands of companies at scale. A feature that I want to include would be relative metrics, e.g. is this company growing faster than its peers, has better margins, and so on.

I can find the sector and industry in various databases, but it's not always very accurate. For example on Yahoo Finance:











While these classifications are true, I'd probably group these companies together.

Google autocomplete would often match my intuition better. For example, it gives me:

- MSFT vs Apple

- GOOG vs Microsoft

Question:

How can I identify peers of companies at scale? For example:

- Does some database for this exist?

- Can I get google autocompletes at scale and map these somehow easily to companies?

- Or is there some other clever way to find these?

## Answer by Mild_Thornberry (score 1)

https://quant.stackexchange.com/a/63821

I don’t know what “at scale” means, but it sound like GIC or SIC code is what you’re looking for. I think GIC is proprietary since it’s from MSCI, but SIC can be found on the EDGAR database.

https://www.sec.gov/info/edgar/siccodes.htm

Edit: To be clearer, using SIC codes gives you the flexibility to construct your own industries. Check out the Kenneth R. French data library as an example. The 12 industry portfolio groups GOOG, AAPL, and MSFT under “BusEq”. That is why I think SIC codes are what you’re looking for. It starts with using the SEC/EDGAR database as a resource.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.