By Sanford Bragg February 11, 2021
After more than a dozen years in research management, Shai Hill left his European Director of Research job at Macquarie Group in 2018 to try something different: an AI-driven data firm focused on ESG metrics. Launched five months ago, Integrum ESG is gaining traction as a differentiated alternative to established ESG ratings providers.
From sell-side to data provider
After leading a team of 40+ sell-side analysts for over a decade, Shai Hill was ready for a new challenge. “There are so many legacy issues in how sell-side research teams are set up, and I could see the advantages in starting with a blank sheet of paper,” Hill said in an interview. “One goal was to eliminate the subjective inconsistencies between analysts to ensure our conclusions were objective, quantitative, and transparent. Another was to embrace A.I. – I’d rather have analysts focus on the intellectual challenges and let the machines do the data gathering.”
Integrum ESG offers three varieties of sustainability data: ESG risk metrics, impact data and sentiment data. Data is delivered either through an API or a customizable dashboard, a SaaS cloud-based platform which displays company- and country-specific ESG data and ratings.
Emphasis on transparency
The ESG sustainability metrics are based on the Sustainability Accounting Standards Board (SASB) standards, while governance metrics utilize a framework from European proxy firm Minerva Analytics. Metrics vary by sector to reflect those that are most material. For example, data security is more important to an insurance company whereas carbon emissions are critical for an aluminum manufacturer. “Portfolio managers don’t want to see 100 ESG metrics per company when only 20 will make any difference,” said Hill. Users can view the sub-metrics that provide additional granularity for a given metric.
Like financial data platforms such as S&P’s Capital IQ, the source data for each metric is provided. The dashboard gives explanations for the metric score and includes extracts from relevant source documents. If users have a different perspective on a sector, they have the ability to customize the weightings of the metrics, or add and subtract metrics relevant to a sector.
Besides the ESG risk metrics, Integrum also offers impact measures which use Cambridge University’s framework for assessing how much a company is advancing UN Sustainability Goals. As with other metrics, the provenance of the impact metrics is documented, and the weightings can be customized.
Machine learning and NLP
The data and scores for each metric are derived from machine learning (ML) and natural language processing (NLP). Hill says it took two years to train the ML models for the risk and impact metrics. Development of the firm’s artificial intelligence is managed by Integrum’s CTO, who was previously with the Bank of England, Barclays and CSFB.
The third overlay offered by Integrum is ESG sentiment, which tracks intraday swings in positive and negative ESG sentiment. Sentiment is derived from over 850,000 news and social media sources in 187 languages and allows alert notifications.
Integrum currently covers over 1,700 publicly traded companies, including all components of the MSCI World Index. The firm recently added sovereign ESG coverage for 180 countries, including every sovereign bond-issuing country.
“Some of the larger ESG ratings firms are well-entrenched,” said Hill. “But we started Integrum ESG because the buy-side is frustrated with the opaque, rigid, black-box nature of the ESG assessments they’re currently purchasing.” Hill argues that given the rising importance of sustainable investing, asset managers should not rely on one or two ESG providers. A veteran of MiFID II unbundling, he proposes asset managers kick out one or two Tier 2 brokers to make room in the budget for Integrum’s service.
Although launched only five months ago, the firm has 12 institutions as clients, both hedge and long-only. A firm-wide subscription costs $5,950 per quarter.
The company has been financed through two angel raises, in December 2019 and December 2020. The firm currently has 8 employees and is recruiting for four positions: a ML developer, an analyst and two salespeople.
We’ve been asking sell-side research directors about their use of AI in the research process and nearly all maintain that they have AI projects underway. However, it is a safe bet that none of the sell-side firms are planning an implementation as radical as the one that Shai Hill has managed. As one research director expressed it, the preferred investment is people not machines.
That mentality is starting to shift in the ESG space, although the traditional ESG ratings firms still rely heavily on subjective analyst input. Integrum ESG is one of a new class of ESG data providers relying primarily on artificial intelligence. What is impressive about Integrum is that it has been able to spin up three distinct varieties of ESG data (risk, impact, sentiment) in a relatively short time. But then again, that is one advantage of leveraging machines. We suspect Shai Hill’s venture is a portent for the future of the broader research space.
- Robust Growth of MSCI ESG Research Validates Ongoing Investment (12)
- VisasQ Partners with DeepBench (12)
- Moody’s Exec Says Alt Data Should Drive Transparency for ESG Risk (12)
- Indie Boutique CreditSights Embraces Machine Learning (12)
- ESG Data is Inconsistent, Confusing, Opaque Yet Increasingly Vital (12)
- Mergers and AI are Curing ESG Data’s Ills (12)
- AI-Driven ESG Provider Clarity AI Expands US Footprint (12)