CSRHub’s goal is to offer broad, consistent ratings of CSR and sustainability performance on as many companies as possible. We have received data from our loading... sources on more than 100,000 companies and organizations. Eventually, we hope our system will allow us to publish ratings on all of these companies—and many hundreds of thousands of publicly-held companies, privately-held companies, not for profits, and government organizations.
We provide ratings across the four categories of data we track (Community, Employees, Environment, and Governance) for over loading... companies. For some of the companies in our system, we often have only one source of data on a company or information on only a few areas of a company’s sustainability performance. In several thousand cases, we believe this data is sufficient for us to publish partial ratings—data on one or more of our twelve subcategories. These partial ratings generally allow us to also provide at least one category rating for these partially rated companies.
Here are the steps we take to determine which ratings to publish:
SUBCATEGORY RATINGS WILL SHOW OR NOT SHOW, DEPENDING ON THESE TWO RULES:
- Don’t show these subcategory ratings when we don’t have at least 1.5 weight of data. Some of our major sources by themselves have weight of 2.0, other sources have weight as low as 0.33. The exact total of weight depends on how many sources we have, how many data elements a source offers on the subcategory (some sources offer multiple data items that we can map to a single subcategory) and how we have weighted the source. Source weights are driven by our normalization process. We perform a regression-like analysis that indicates how well each source correlates with the values our other sources offer for each company that the source covers. Those sources who track well with the aggregate of the rest of our data get a greater weight.
- Filter out subcategory above 95 or below 5. We do not believe that it is possible for a company to be completely perfect (or completely bad) on any aspect of social performance. This is especially true since our system assumes a beta distribution between 0 and 100. Therefore, values that approach either extreme become statistically improbable (likelihood of less than 0.0001%).
COMPANIES WILL SHOW AS PARTIALLY RATED UNLESS THEY PASS THESE FILTERS AND RECEIVE AN OVERALL SCORE
- Only Partial Ratings for any company that does not have at least one subcategory in each of the four categories. Overall scores are adjusted by each user’s personal ratings profile. We cannot apply these adjustments fairly, if we are missing a category.
- Block full ratings for companies that do not have at least five subcategories. While it would be possible to get a category score with only four subcategories (one per category), we have found that we need at least five total subcategories to get stable overall scores.
- Require fully rated companies to have at least one broad, well-established source and at least three additional more narrow sources. We have a number of sources who use teams of analysts to gather and analyze their data. These sources report across several thousand companies, so we can do a good job of normalizing their input and removing any biases. By combining input from at least three other sources, we ensure that no single voice is the sole determinant of a company’s overall rating.
- Test the weight (as defined above) per subcategory across all of the reported subcategories for a company. If the average across all subcategories is not at least 4.0 and the total weight for the whole company is not at least 40, the company should not be fully rated. This rule strips out situations where a few highly specialized sources may allow five or six subcategories to be rated, but do not provide enough breadth of data to allow our normalization functions to work properly.
Our site reveals (to registrants) all of the data sources we use for each company. We also reveal (to subscribers) most of the data elements and data values we process. We do not share the weights we assign to each source or the full details of our score conversion, normalization, and back-testing system. However, users should be able to easily inspect enough information to understand why some companies are partially rated and why some companies do not have all subcategories rated.
When you search for ratings on our site, you can choose via a check box on the search page to turn off viewing partially rated companies. However, we integrate the values from partially rated companies into:
- Average ratings for countries and industries. In general, partially rated companies have slightly better than average scores for the area they report. Therefore, including them in our averages raises those averages slightly. We believe this makes sense. Companies tend to report their more positive information and suppress data that could make them look bad.
- Counts of the number of companies in a country or industry. Our mission is to encourage transparency for existing CSR data and to create a feedback loop that stimulates companies to release more and better data. By showing that more companies have revealed at least some data, we hope to convince “laggards” that they need to start participating in sustainability reporting activities.
We often do not have as much collateral information on partially rated companies as we do on the better-studied fully rated ones. So, we may not know as clearly which special issues apply to them or information such as their web site address, the CSR web site address, or other descriptive information. We are working to improve this part of our data—and to add more sources of information that will allow us to partially rate more companies and to fill in the holes that remain on those we already rate.
See the CSRHub rating rules.