RepRisk – Looking Under the ESG Hood
Many companies claim to have sustainable business and ethical practices, here’s why RepRisk will not take their word for it.
ESG, or Environmental, Social, and Governance, is a relatively new term for businesses, but already it is a term at the core of many financial agreements. It is incorporated into investments, due diligence and know-your-client procedures, indexes, pensions, and credit risk management to name a few. Knowing what a company’s ESG business practices look like can help others decide where to invest, where to operate, and what strategic decisions must be made.
The challenge is in evaluating a company’s ESG business practices – how does one discern between companies with meaningfully sustainable operations and those that just have good PR?
“Think of us as an ESG technology company,” explains Dr Philipp Aeby, CEO of RepRisk. “We leverage technology like machine learning and geospatial analytics to assess ESG risks. We can provide relevant data thanks to that mix of artificial and human intelligence, providing incredible coverage for companies and infrastructure projects with relevant information used directly by our clients.”
RepRisk uses this technology to evaluate businesses for their impact on biodiversity, child labour, corruption, and a host of other ESG issues and risks. That data can in turn be directly applied by market practitioners in their financial decision-making.
“We don’t rely on companies to report and disclose their behaviour when it comes to ESG,” Aeby says. “We get information from outside the company such as media reports, think tanks, governmental sources, and NGO reports.”
RepRisk’s technology scans these sources in 23 languages to identify any relevant risks associated with the companies it examines.
“What makes our data generation methodology unique is our human analysts. The analysts annotate the documents our machine learning and AI take in and feed them back to the machine – further sharpening the results. It is a kind of symbiosis,” Aeby tells us.
As an example, if you had a phone manufacturer with labour issues in India, RepRisk would use machine learning to gather sources, identify issues – such as the company not paying a living wage and determining if it violates ILO standards – and quantify the risk, building a complete risk profile of that company.
“We sum it all up, linking a company or infrastructure project to an ESG risk incident, such as child labour or deforestation,” Aeby says. “On top of that, we take these data elements and build a number of standard and customised metrics. So, if a bank wants to onboard a client and wants to know if there are any issues we’re aware of, we can say ‘be careful, there have been corruption charges in the past’ or ‘these issues are systemic, use enhanced due diligence’. We allow investors to invest in a way that better manages risks.”
This information can be provided in a data format to be inputted into their clients’ own systems, or through a software-as-service analytic platform, the RepRisk ESG Risk Platform.
Changing Mindsets and Data Sets
“Our challenge now is to showcase how our technology is a fantastic way to address the ESG data gap, a knowledge gap over what the real conduct of companies is,” Aeby says. “A company might have a policy stating human rights are important, but their conduct could be very different. This requires a change in mindset because traditionally this was evaluated through company self-disclosure and reporting. Now technology can be leveraged to check their claims.”
RepRisk scans 500,000 documents from 100,000 sources every day, in 23 languages, but data alone will not get results. RepRisk leverages human curation to ensure relevant, actionable intelligence.
“What I love about RepRisk is that we have to excel in a number of quite different tasks,” Aeby says. “If you want to do machine learning you need a lot of labelled documents and this is what our analysts have been doing that for 15 years. That is a different skill than selecting and adapting the best machine learning algorithms for specific uses cases, training, validating and constantly re-training the corresponding classifiers, or maintaining the whole IT big data infrastructure. The machine learning infrastructure alone requires a mix of data scientists, data ops people, developers and machine learning engineers. In any case, we have people with a lot of curiosity and passion in all areas.”
A pioneer in ESG datasets, RepRisk’s next move will be in spatial risk development.
“Our next step is combining these data sets with geospatial data,” Aeby explains. “If you have a gold mine with a tailings dam spilling over into protected areas you might have a report that suggests exactly that. But what if you could check the position of the mine and the protected area to see if this dam is adversely affecting it? We are going to link our documentary data with GPS coordinates and gather corresponding data sets.”
Going to the Source
RepRisk provides an invaluable early warning system for market practitioners, but Aeby believes there is more to be done to achieve a truly sustainable economy.
“90% of our business is with financial institutions, big pension funds, insurance companies, and banks or private equity,” Aeby says. “But ultimately we need to engage with companies directly.”
There are companies out there genuinely working to make positive change. What if, Aeby argues, instead of evaluating those companies for investors, RepRisk was showing those companies how to do better? How can you monitor and account for unintended consequences of even well-intentioned practices?
“There is a lot of talk about when it comes to transitioning into a low carbon economy and what that means in terms of protecting biodiversity,” Aeby says as an example. “If you are a jet engine manufacturer that wants to be carbon neutral you might transition to biofuels, but someone has to produce those biofuels, which involves a lot of monocultures and takes a lot of land to cultivate. We can help companies see what potential, inadvertent, or unwanted side effects their activities might drive. We are moving from evaluating companies to working with the companies directly to show them the consequences of their actions.”
Machine learning and artificial intelligence are buzzwords across the tech sector, but RepRisk demonstrates real concrete applications for that technology. It is an appealing prospect for data scientists and technologists.
“The people we hire share our mission to provide transparency in business conduct. We make companies accountable and the world a better place,” Aeby says. “We’re doing it using the latest technology.”