In September 2026, Reuters reported that UK insurers had asked the Bank of England’s Prudential Regulation Authority (PRA) to reconsider the design of its first Dynamic General Insurance Stress Test, or DyGIST, following the live exercise conducted in May 2026. The debate matters beyond the UK market because DyGIST represents a different view of stress testing: one that examines how an organisation responds while a crisis is unfolding, rather than only calculating the financial effect of a pre-defined shock.

The PRA selected firms representing about 80% of the UK general insurance market. Over three weeks, participants received successive scenario updates covering a US West Coast earthquake, a Gulf of Mexico hurricane, a UK windstorm, European floods and a cyber event. Companies had to interpret new information, update their assessment and report their decisions as the scenario developed.

From a capital test to a management test

Traditional insurance stress tests usually begin with a defined set of assumptions. Firms model losses, assess their capital position and explain the management actions they could take. That approach is essential for comparing financial resilience across companies, but it can understate the operational difficulty of a real crisis. Information arrives gradually, models disagree, senior experts may be unavailable and decisions must be taken before the full picture is known.

DyGIST was designed to expose that gap. Its live format tested the flow of information between underwriting, actuarial, claims, finance, risk, operations and executive management. It also tested whether governance arrangements that look credible on paper can support decisions under severe time pressure. This explains why some chief risk officers reportedly considered the exercise useful even when other participants criticised its demands.

The distinction is important for boards. A firm may remain solvent in a model while responding poorly in practice. Delayed escalation, inconsistent loss estimates, unclear authority or weak communication with counterparties can turn a manageable event into a larger operational and reputational problem.

The case for proportionality and realism

Industry criticism focused on the volume of reporting, the short notice given for specialist input and the accumulation of several major events over a compressed period. Some firms reportedly cancelled staff leave and mobilised experts at short notice. The International Underwriting Association questioned whether such an extreme sequence would generate enough additional supervisory insight to justify the resources required if the exercise were repeated regularly.

This does not necessarily invalidate the dynamic approach. Stress tests are deliberately severe, and improbable combinations can reveal dependencies that ordinary planning misses. However, an exercise should have a clear learning objective. If participants devote most of their effort to producing regulatory submissions, the test may measure reporting capacity more than decision quality. A more focused structure could preserve the value of surprise while reducing unnecessary operational disruption.

The PRA has said that it will consider feedback on resources and proportionality and expects to publish its findings by the end of 2026. Until then, conclusions about the final supervisory approach remain provisional.

Lessons for multinational benefits programmes

The DyGIST concept also has value outside property and casualty insurance. Multinational employee benefits programmes depend on insurers, network partners, administrators, medical providers, data platforms, banks and employers operating across several jurisdictions. A crisis may combine a health event, provider disruption, cyberattack, payment restriction and sudden demand for employee assistance.

A dynamic exercise for a global benefits network would therefore test more than insured losses. It could examine whether local partners escalate consistently, whether headquarters receives comparable data, whether emergency payments can be authorised and whether client communications remain accurate when information is incomplete. Captive arrangements add further questions concerning aggregation, collateral, liquidity and the timing of reinsurance recoveries.

For boards and programme sponsors, the practical lesson is to combine quantitative stress modelling with live decision exercises. The objective is not to create the most dramatic scenario. It is to identify where information, authority or execution could fail before a real event exposes the weakness.