There’s at least a correlation, and strong possibility of a connection
By Bahar Gidwani
One company takes care of its employees, buys only from responsible suppliers, and encourages its managers to behave ethically. Another company has a history of releasing toxic pollutants, periodically closes facilities, and irresponsibly lay off employees, and has been linked to various instances of fraud and price-fixing. Shouldn’t the second company be more exposed to reputational risks than the first?
To find out, we combined CSRHub data on perceived CSR performance and RepRisk data on the level of ESG- related reputational risk exposure for more than 4,000 companies from around the world. And we discovered that we were able to explain 23% of the variation in risk exposure for the 2,000 companies who have revealed the most sustainability data about themselves. (We found almost no correlation between risk exposure and sustainability disclosure for the remaining companies, who have revealed little about themselves.)
For the well-studied companies, it appears that those with the most sources of sustainability ratings (i.e., receive the most attention from social responsible investor (SRI) analysts, participate in the most rating programs, are tracked by the most NGOs or news outlets, or who are studied by multiple certification or crowd opinion sites) have the most risk exposure.
This relationship does not appear to relate to company revenue or market capitalization. Instead, data indicates that sustainability rating sources play a role in discovering and communicating corporate risk events. In addition, companies that have strong records (as measured by CSRHub) in the human rights and supply chain, leadership ethics, and resource management areas seem to have systematically lower risk exposure. Those who have extensive community development, philanthropy, environment policy and reporting efforts or extensive compensation and benefits programs seem to have more risk exposure.
What’s at stake?
Reputational risk is important to many corporate stakeholders. In fact, a 2013 study by Deloitte on behalf of Forbes Insight of 300 corporate executives identified reputation as the highest impact risk area for business strategy. A factory fire, explosion, toxic spill, strike or other negative event can hurt more than a company’s profits. Negative events can discredit a company’s brand, reduce employee morale, encourage community opposition to a new facility, and distract management’s attention. If good CSR performance can “immunize” a company against reputation damaging events, CSR programs can become a type of “reputation insurance” for corporate risk managers.
Seven sustainability elements we identified as related to risk. We also understand why some elements may not relate to risk:
• Risk-reducing factors. Good Human Rights & Supply Chain programs should help a company identify supply chain hot spots and either replace a poor performing supplier or help it improve its internal processes. A tradition that values Leadership Ethics should keep a company from engaging in risk-creating practices. And Resource Management programs that recycle and reuse resources will tend to minimize a company’s ecological footprint and reduce opportunities for environmental problems.
• Risk-increasing factors. Strong Community Development and Philanthropy programs may be connected with increased risk, because companies that have them may be trying to avoid or offset risky activities. Some giving programs may also fail to integrate with a company’s overall CSR strategy or the needs of key stakeholders. Both giving programs and extensive Environment Policy and Reporting programs may be seen as “greenwashing”–only needed when a company has community- or environment-related risk. The positive correlation with Compensation and Benefits may support the theory that it is risky for a company to have aggressive employee incentive programs. These programs may lead to employees pushing to meet unrealistic financial or operating objectives.
• Un-related factors. Six factors did not seem correlated with risk. It seems that company Boards are too removed from operating decisions to have much connection with risk exposure. Diversity and Labor Rights and Energy and Climate Change issues should introduce risk, but perhaps in too abstract a manner to show up in our study. The Product category measures whether or not a company’s products have impact on society—a different issue perhaps from whether they generate corporate risk. We expected Training, Health, and Safety issues to be important for risk. However, many companies may have shifted their risk focus in this area into their supply chain. Finally, the lack of connection with Transparency and Reporting may be due to the timing lag between corporate sustainability reporting and risk-related events.
Is there a causal relationship between CSRHub’s ratings and data source count and risk exposure?
A 24% correlation between a set of sustainability factors and an estimate if risk exposure seems high and strongly suggests there is a connection between these factors. However, as Edward Tufte has put it, “Correlation is not causation but it sure is a hint.” We do not know which of the following is true:
• Perceived CSR performance could be unrelated to reputational risk. Despite the statistics cited above, it is possible that our result is due to random variation and there is no tie between CSR performance and reputational risk. We hope others will test our conclusions using these data sets or ideally, using other similar measures of social performance and corporate risk exposure.
• Reputational risk and CSR performance could both be correlated with some other factor. For instance, although we accounted for market capitalization, we did not adjust for revenue, number of employees, industry group, or geographic region. As a result, our data sets could appear to be correlated with each other, when they in fact just share a common driver. We have tested our results across industry groups and geography, and have not found any major difference in our result. (This is why we have not included these factors in our analysis.) Market capitalization tends to be closely correlated with revenue and number of employees—and we have already shown there is no correlation between market capitalization and either of our two data sets.
• Reputational risk and sustainability are related.
One easy additional test is to estimate the probability that an observed correlation is actually zero (no correlation). This probability is expressed via an “F value.” An F value equal to one would indicate that the chance that the observed correlation is zero is the same as the chance that it is non-zero. With 2,000 data points, an F value above 4 would indicate less than a 5% chance that the observed correlation is zero. The F value for our correlation of all CSR factors and the number of data sources against reputational risk is 18. This suggests there is a vanishingly small chance that there is no correlation between these data sets.
(Bahar Gidwani is Co-founder and CEO of CSRHub.com
Posted March 9, 2015 in 25115in Corporate Social Responsibility