What gets measured, and why it matters
Many institutional clients turn to scoring models expecting them to provide
clarity and comparability. However, model creators choose which data to include,
how to weight variables, and which timeframes matter. These decisions can
amplify or mute critical signals, especially for sectors with less disclosure.
The result: scores can diverge widely even for similar companies or funds,
raising questions about what is really being measured.
Strengths and shortcomings in practice
Quantitative models excel at processing large data sets and reducing complex
inputs to digestible metrics. But that simplicity can hide the noise. For
example, automated scoring may overlook context-specific issues such as
regulatory updates or nuanced supply chain risks. Blind reliance on the score
risks missing material events outside the standard template.
How to interrogate an ESG score
Not all scores are equally useful. When evaluating a scoring output, I check
whether the methodology is disclosed, if third-party reviews are available, and
how often the inputs are updated. If a model cannot be stress-tested or fails to
adapt to new information, its output should be treated with
skepticism—especially when used for regulatory or stakeholder reporting.