LSEG Rebuilds ESG Scoring From Scratch

London Stock Exchange Group (LSEG) has launched a new suite of ESG scores and sustainability analytics that represents its most significant overhaul of ESG data methodology since the Refinitiv acquisition in 2021. The new Sustainability Ratings and Data product introduces 220 standardised indicators, a threshold-based scoring scale running from 0 (“not aware”) to 5 (“leading”), and a redesigned materiality matrix that applies double materiality at the business-segment level. The methodology is aligned with ISSB, GRI, SASB and ESRS frameworks, and the product is available across LSEG Workspace, DataStream, API, bulk file delivery and cloud platforms including Snowflake and Azure.

Elena Philipova, Director of Sustainability Solutions at LSEG, tells Market & Alt Data Insight that the decision to start fresh was driven by the widening gap between what clients need and what existing models could deliver. “The gap between customer expectations and the available supply of data and analytics has grown progressively wider,” she says. “That is the primary reason we took the opportunity to build a brand new scoring methodology, rather than simply uplifting or tweaking our existing legacy LSEG model.”

Philipova identifies three drivers behind the rebuild: the rapid growth and maturity of ESG disclosure data, which has outpaced the models designed to assess it; shifting regulatory requirements that are fundamentally changing client workflows; and a broader credibility crisis around ESG scores as a category, which LSEG believes demanded a response.

The transparency bet

At the centre of LSEG’s pitch is a claim about transparency and modularity. The new product adopts a modular architecture – with a core ESG scoring layer and a “Plus” tier incorporating controversies, sovereign ESG risk and positive environmental impact signals such as green revenues – designed to let clients select the components relevant to their mandates rather than accepting a single composite score.

“A significant problem with most ESG scores today is that you either take the score as it is or you build your own proprietary model,” Philipova says. “That is precisely why many investors have been forced to create their own. Our modular approach lets clients choose the components that align with their individual needs, mandates and requirements, and genuinely move beyond one-size-fits-all.”

LSEG is also publishing the materiality weightings that underpin the scores – a move Philipova claims goes further than any competitor. “If a theme has been deemed not relevant for a company in a particular sector, we don’t just publish the theme score – we also publish the materiality score for customers to use in their own workflows, or to adjust if they disagree,” she says. “That level of transparency and structured, replicable data is something that no other scoring methodology in the industry provides.”

The shift from percentile-ranked to threshold-based scoring is methodologically significant. Academic research has documented how peer-comparison models can produce score inflation when new, smaller companies with sparse disclosure enter the scoring universe, a criticism that has been levelled at the legacy Refinitiv model among others. Threshold-based scoring is designed to address this by anchoring assessments to absolute performance levels rather than relative position within a cohort. But the thresholds themselves become the new design choice that requires scrutiny: who sets them, how frequently they are recalibrated, and whether the rubric will be published in full are questions LSEG has yet to address in detail.

The transition reality

For data teams, the most pressing questions are not about methodology but about migration. Philipova confirms the operational detail. “We have launched the new scores in parallel to the legacy Refinitiv ESG scores on LSEG Data & Analytics products,” she says. “The dual-run period will be significant – extending until at least the end of 2027 – to ensure a smooth client transition and regulatory readiness.”

That gives clients a runway of nearly two years, but it also means firms will be managing two concurrent ESG scoring methodologies from the same provider during that period, with different scales, different indicator counts and different underlying logic. For data management teams, this is not a simple switchover; it is a period of active coexistence that will require mapping, reconciliation and potentially dual reporting.

The FTSE Russell side of the equation is less settled. “FTSE Russell will consider the use of the new LSEG ESG scores across index families,” says Philipova. ”FTSE Russell will conduct analysis of the impact to indices of a transition to the new LSEG ESG scores, which will be shared with index clients. The possible transition for live FTSE Russell indices is expected no earlier than 2027.” This suggests the beginning of a consultation process, not a confirmed migration, meaning that there is a scenario in which the FTSE Russell ESG model and the new LSEG Sustainability Ratings coexist for longer than the Refinitiv scores do, meaning LSEG could be running three ESG models concurrently well into 2027 and potentially beyond.

On historical data, LSEG has confirmed an explicit break with the legacy scores. “The new ESG scores are available from fiscal year 2022 onwards, so there are three to four years of time-series data available,” Philipova says. “This is intentional, as it captures the latest reporting trends and data availability, with emerging themes such as biodiversity having low data availability for prior years.” The rationale is defensible. Applying a methodology designed around current disclosure standards to an era when those standards did not exist would compromise the model’s integrity. But for some quantitative and systematic teams, three to four years of history may be too short for the most robust back-testing, factor analysis or long-horizon risk modelling. Firms that need deeper history will have to continue using the legacy scores for that purpose, creating an analytical seam: one methodology for historical analysis, a different one going forward.

Scores, not ratings

A notable feature of the launch is LSEG’s explicit positioning of the new product as scores, relying on transparent, rules-based methodology that “do not incorporate analyst judgement, which clearly differentiates them from ESG ratings.”

The distinction is worth watching. ESG ratings regulation is advancing in the EU, UK and across Asia-Pacific, and how providers are classified – whether as offering “ratings” or “scores” – will carry implications for oversight, transparency mandates and conflicts-of-interest rules. LSEG’s rules-based, no-analyst-judgement methodology places it naturally on one side of that line, while providers whose models incorporate qualitative assessment, such as MSCI and Sustainalytics, may fall on the other. As regulators refine their definitions, the question for the market is how clearly that boundary holds, particularly given that many clients will deploy both “scores” and “ratings” in similar workflows and for similar purposes.

Competitive context

The launch sits within a market that is consolidating rapidly. Deutsche Börse recently moved to acquire the remaining 20% of ISS STOXX, bringing its own ESG data stack fully in-house. MSCI is widely regarded as the market leader in ESG ratings with deep buy-side penetration. Bloomberg, S&P Global and Morningstar’s Sustainalytics all remain active competitors in the ESG data space.

Against that backdrop, LSEG operating two overlapping ESG models from different heritage businesses was a known competitive liability. This launch begins to rationalise that, but the follow-up detail confirms the rationalisation will take time. With a dual-run extending to at least end-2027 and the FTSE Russell index transition still under consideration, LSEG will be managing significant product complexity for a considerable period. The Core/Plus tiered architecture mirrors how competitors already segment their ESG product lines, so the pricing structure is not novel in itself; the transparency and modularity claims are where LSEG seems to be trying to carve out differentiation.

The AI question

With the press release referencing “AI-ready ESG workflows,” Philipova describes AI’s near-term role as focused on data quality, observability and anomaly detection rather than analytical intelligence.

“The opportunity we see is to connect the quality-control processes that today operate on both sides of the fence,” she says. “Rather than clients discovering data anomalies independently, we want to be alerting them proactively – making the entire process deliver results far more quickly.”

What data teams should watch

For market data and ESG data teams, the practical items to monitor in the coming months are now reasonably clear: the dual-run period through 2027 and how entitlements and licensing work during the coexistence of old and new scores; the FTSE Russell impact analysis and consultation process, which will determine whether index clients face a further methodology change; the fiscal year 2022 start point for historical data and whether LSEG intends to extend it backwards over time; and the licensing and cost implications of the Core/Plus modular architecture, particularly for firms currently mid-cycle on procurement or renewal.

Underlying all of this is a broader question about what ESG scores should be expected to do. Philipova is candid on this point. “At their core, ESG scores combine very complex, diverse topics into a single number,” she says. “If the decisions behind that are not fully transparent, users struggle to understand what a score actually represents and why scores differ across vendors. We believed it was a pivotal moment to make a clear statement about what ESG scores are designed for, what their strengths are – but also what their weaknesses are and what they should not be used for.”

That willingness to acknowledge the limits of the product category is, in its own way, as significant as the product itself. Whether the new methodology delivers on its transparency promise will become clearer as clients begin to work with it in practice. In the meantime, the migration ahead is real, the timeline is long, and the hardest operational questions – around entitlements, historical continuity and index integration – are still being worked through.