Do not mistake insurance profits for correct pricing
Lee Harris and Toby Nangle recently posed a deceptively simple question in the Financial Times: if the world is becoming more dangerous, why is risk getting cheaper? Their answer focuses on insurance and reinsurance, where premiums have fallen even as climate, technological and geopolitical hazards intensify. Abundant new capital, they argue, is reshaping the insurance underwriting cycle by increasing capacity and compressing prices.
That diagnosis is persuasive. But it points to a more fundamental question: even if capital cycles explain why insurance prices are falling, do today's premiums truly reflect tomorrow's risks?
The traditional underwriting cycle is well known. Strong profits attract capital. New capacity intensifies competition and drives premiums lower. Eventually, large losses erode capital, capacity contracts and prices recover. Some 35 years ago, catastrophic losses wiped out private investors and bankrupted carriers, and institutional capital rebuilt much of the sector's base. Today's softness is partly explained by additional capital from alternative asset managers and sovereign wealth vehicles, whose pull-back after Hurricane Ian amplified the last hard market, and whose staying power in a systemic event remains untested. If this new capital retreats after major losses, the cycle can become more volatile. If it instead steps in when traditional insurers withdraw, pricing may become more stable. For now, abundance is winning: even the record 2025 wildfire losses did not arrest the slide in property-catastrophe reinsurance prices.
Understanding who supplies underwriting capital is crucial. But capital explains why market prices move; it does not determine whether those prices are economically correct. That depends on whether today's premiums adequately compensate those ultimately bearing tomorrow's risks.
Harris and Nangle report an observation that deserves emphasis: insurers can appear highly profitable "in accounting terms" while still destroying value over the longer run. Reported profits drive the cycle but do not settle the price of risk. Insurance can be economically underpriced long before reserves prove inadequate, because profits may look defensible under prevailing assumptions even when new business is written at premiums that fail to reflect tomorrow's distribution of losses.
This is not a failure of actuarial or catastrophe-modelling expertise. The deeper challenge is that many of today's most significant risks are increasingly difficult to price because the assumptions on which pricing depends are becoming less stable.
Climate is the clearest example. We may be gambling with our future by allowing continued greenhouse gas emissions, but climate risk is not roulette. Scenario probabilities can be estimated with care, yet they are not delivered by a stable wheel with known odds, and their limitations in the tail must be explicit. Climate risk is non-stationary: historical experience becomes a weaker guide as the underlying climate itself changes.
Crucially, climate dynamics are also non-linear and prone to cascading, path-dependent effects. Research on climate tipping points highlights "critical thresholds" beyond which the climate system shifts into a different regime. Crossing such thresholds can abruptly alter hazard patterns and loss distributions. Event clustering adds another dimension. Even short of tipping points, inter-annual variability can concentrate acute events in a single year. A very strong El Niño, combining drought, heat and intense rainfall across different regions, is one example. Such combinations can abruptly expose underpricing and stress diversification assumptions across regions and perils, especially if models were calibrated on a past in which such regimes were rare.
Similar issues now arise beyond climate. Cyber threats evolve continuously; geopolitical shocks propagate rapidly through supply chains and markets; artificial intelligence can create highly correlated losses that differ from the largely independent events on which many insurance models were calibrated. The challenge is not only that these risks are larger, but that they evolve and interact faster than those insurance markets have traditionally priced.
When uncertainty ranges are wide, abundant capital and competitive pressure can push pricing towards the optimistic end of what can still be defended. While rogue insurers may prove reckless, this is not required for the market as a whole to drift towards economic underpricing: horizon and incentive mismatches, together with accounting-recognition lags, allow it to persist until experience or revised assumptions force the issue.
This is why accounting profitability should not be confused with economic adequacy. Financial statements look backwards; underwriting prices the future. A market can appear healthy precisely because its assumptions have not yet been tested by experience.
Insurance premiums are more than commercial prices: they are signals about society's assessment of risk. If those signals systematically understate evolving risks, capital may be allocated efficiently according to today's prices while becoming misaligned with tomorrow's realities.
The real question is therefore not whether insurers are profitable today. It is whether today's profits are being earned at premiums consistent with a more dangerous, evolving risk landscape - one shaped by clustered events, shifting hazard patterns, and shocks that can cascade across sectors, markets and countries.
Frédéric Ducoulombier,
Programme Director, Climate Regulation and Policies, EDHEC Climate Institute, EDHEC Business School, Nice, France