Banks each measured market risk with their own in-house models, so a 'risk' figure meant different things at different desks. There was no comparable, auditable number for the whole market.

In 1994 J.P. Morgan released RiskMetrics: a public methodology and a public dataset of volatility and correlations that let anyone compute Value-at-Risk the same way. The open standard made risk a number regulators and counterparties could compare.

The move was a deliberate bet that becoming the common language would matter more than keeping the model private; it worked so well that the methodology and its data business were later spun out as RiskMetrics Group, later acquired by MSCI.

Give the method away to become the benchmark

An open standard makes a risk figure comparable across firms

Free publication removes the adoption cost that blocks a private model

The method's owner sits where every user has to look

Regulators and counterparties need a number they can trust

When the real business is serving a market that needs a standard, giving the spec away can beat selling a private one; trade margin in the model for the position everyone reads.

VaR became the dominant market-risk measure for thousands of institutions, used by traders and regulators. It also attracted criticism as a way to hide tail risk and was later blamed for missing extreme events, but the RiskMetrics move remains a classic example of standard-setting via publication.

FOLLOW THE EVIDENCE

The sources

  1. RiskGrades Technical Document MSCI · msci.com ·
  2. RiskMetrics Wikipedia · en.wikipedia.org