ESG with a risk lens
and transparency

Over 17 years of deep domain expertise in ESG risks
Exclusively with an outside-in perspective to assess whether companies walk their talk

Born out of credit risk management, the purpose of RepRisk’s dataset is not to provide ESG ratings, but to systematically identify and assess material ESG risks. We have always taken an outside-in approach to ESG risks, by analyzing information from public sources and stakeholders and intentionally excluding company self-disclosures. It is now well-accepted that self-reported information is not reliable data – especially when it comes to risks.

Over a decade of experience has shown that RepRisk’s unique perspective serves as a reality check for how companies conduct their business around the world – do they walk their talk when it comes to human rights, labor standards, corruption, and environmental issues? This perspective, together with a transparent, rules-based methodology and daily updates, ensures that our clients have consistent, timely, and actionable data at their fingertips.

Combining the best of both worlds
AI and machine learning empower the size and scale of our dataset, while human intelligence adds depth and relevance

Advanced machine
learning

Since 2007, RepRisk has produced the largest,
high-quality annotated (human-labeled)
dataset that allows us to train our machine
learning algorithms to be more accurate
and effective in identifying ESG risks

Actionable
insights

Our clients have access
to a consistent time series
of high-quality data that
can be used for rigorous back-
testing and quantitative analysis

Human
intelligence

Our highly-trained team of 150+ analysts
curates and analyzes each risk incident
according to our transparent
rules-based methodology that
ensures data depth and quality

Advanced machine
learning

Since 2007, RepRisk has produced the largest, high-quality annotated (human-labeled) dataset that allows us to train our machine learning algorithms to be more accurate and effective in identifying ESG risks

Human
intelligence

Our highly-trained team of 150+ analysts curates and analyzes each risk incident according to our transparent rules-based methodology that ensures data depth and quality

Actionable
insights

Our clients have access to a consistent time series of high-quality data that can be used for rigorous back-testing and quantitative analysis

A rules-based and transparent research process
Helps ensure consistent data over time by
translating big data into curated research and metrics

Screening and identification

RepRisk screens, on a daily basis, over 150,000 public sources and stakeholders in 23 languages to systematically identify any company or project associated with an ESG risk incident, per RepRisk’s research scope.

2,000,000+ documents are aggregated through advanced text and metadata extraction from unstructured content and undergo multilingual de-duplication and clustering processes, reducing incoming documents to approximately 150,000 daily.

Analysis and curation

150+ analysts review and approve the results of the screening process of automated tagging, relevancy scoring, and news analytics. Each risk incident is analyzed according to the following three parameters:

  1. Severity (harshness) of the risk incident or criticism.
  2. Reach of the information source (influence based on readership/circulation as well as by its importance in a specific country), based on RepRisk’s rating.
  3. Novelty (newness) of the issues addressed for the company and/or project, i.e. whether it is the first time a company/project is exposed to a specific ESG Issue in a certain location.

Quality assurance

Before a risk incident is published in RepRisk's dataset, it undergoes a quality assurance check and approval performed by a Senior RepRisk Analyst to ensure that the overall analysis process is in line with RepRisk’s strict, rules-based methodology.

The goal of the quality assurance process is to ensure that the complete analysis process has been completed in line with RepRisk’s strict, rules-based methodology.

Quantification

The final step in the process, the quantification of the risk, is done through data science. There are proprietary standard and customized risk metrics.

The RepRisk Index (RRI) dynamically captures and quantifies reputational risk exposure related to ESG Issues. The RepRisk Rating (RRR), a letter rating (AAA to D), facilitates benchmarking and integration of ESG and business conduct risks. The UN Global Compact Violator Flag identifies companies that have a high risk or potential risk of violating one or more of the ten UNGC Principles.

Uniquely positioned to take advantage of the latest
Advancements in natural language processing

1. Training data

17+ years of highly accurate, domain-specific, human-labeled data

2. Advanced machine learning

Trained and refined machine learning (ML) models filter out irrelevant data i.e., “the noise”

3. Human intelligence

Human analysts review, analyze, and approve the results of the screening process and work together with generative text models to develop the highest quality risk incident summaries as fast as possible.

4. Feedback loop

The ML models learn from the analysts and accuracy improves through a feedback loop

The latest machine learning models
Identify and classify ESG risks consistent with
how key international standards and norms define ESG

Big data 2,000,000+ documents screened daily from 150,000+ sources in 23 languages Documents scraped from online sources and fed to machine learning (ML) applications Text classification ML reducer Named entity recognition Deduplication ML applications predict relevant and unique ESG risk incidents Irrelevant results discarded and predictions fed to the multilingual queue Results sent to the ML reducer Multilingual queue Human analysts Approved Documents sorted in priority order A team of 150+ human analysts: Two analysts review Confirm and correct ML predictions Assess severity, reach, and novelty Write risk incident summaries along with generative text models ESG Risk Platform, ESG Risk Monitor, or via RepRisk Data Feed Results are published to client-facing solutions
Big data 2,000,000+ documents screened dailyfrom 150,000+ sources in 23 languages ESG Risk Platform, ESGRisk Monitor, or viaRepRisk Data Feed
“We have set up our complete universe of portfolio companies to be monitored by RepRisk. To us, this is a great step forward in implementing ESG considerations in private markets, as we are able to monitor both the ESG issues of our portfolio companies as well as how the respective GP deals with these issues.”
LGT Capital Partners

Environment

Environmental Footprint

  • Climate change, GHG emissions, and global pollution
  • Local pollution
  • Impacts on ecosystems and biodiversity
  • Overuse and wasting of resources
  • Waste issues
  • Animal mistreatment

Social

Community Relations

  • Human rights abuses and corporate complicity
  • Impacts on communities
  • Local participation issues
  • Social discrimination

Employee Relations

  • Forced labor
  • Child labor
  • Freedom of association and collective bargaining
  • Discrimination in employment
  • Occupational health and safety issues
  • Poor employment conditions

Cross-cutting Issues

  • Controversial products and services
  • Products (health and environmental issues)
  • Supply chain issues
  • Violation of national legislation
  • Violation of international standards

Research coverage in 23 major business languages
Means RepRisk identifies risks at the local level – so you can know early and know more

RepRisk language coverage coming soon

+95% of the world’s GDP (2021) is covered by the 23 languages in which RepRisk conducts its research, based on the official language indicated per country.