Solvency II and the Standard Formula

An insurer's whole business model is collecting premiums today against a promise to pay claims later, so the regulator's core question is simple: if a genuinely bad year hit tomorrow, does this company still have enough assets to cover what it owes? Solvency II is the European Union's answer, in force since 2016. It requires every insurer to hold capital equal to the loss it could plausibly suffer over one year at a 99.5% confidence level — roughly a one-in-200-year event — and it gives most insurers a standardized recipe for computing that number rather than requiring every firm to build its own risk model from scratch.

The two capital requirements

Solvency II defines two thresholds. The Solvency Capital Requirement (SCR) is the "healthy" target: capital sized to survive a 1-in-200-year shock. The Minimum Capital Requirement (MCR), set lower, is the point at which a regulator steps in hard — restricting the firm's business or forcing a wind-down. An insurer holding capital between the two is under intensifying supervisory attention; falling below the MCR risks losing its license to write new business.

What the Standard Formula does

Building a bespoke internal model of every risk an insurer faces is expensive and only a handful of large insurers do it (subject to regulatory approval). Most insurers instead use the Standard Formula, which breaks total risk into modules — market risk (equities, interest rates, property, spreads), underwriting risk (mortality, longevity, catastrophe, lapse), credit risk, and operational risk — stress-tests each module with a prescribed shock, and then combines the module-level capital charges using a correlation matrix rather than simply adding them up. The correlation matrix matters because it recognizes that not all bad things happen at once: an equity crash and a spike in mortality claims aren't perfectly correlated, so summing worst-case losses across every risk module would overstate the capital an insurer actually needs. The formula effectively asks "what's the worst combined outcome," not "what's the sum of each risk's own worst case."

For example, a life insurer's equity book might face a prescribed shock of roughly a 39% fall in developed-market equities under the market risk module, while its longevity book faces a shock of, say, a 20% permanent improvement in life expectancy (bad for an annuity writer, since it pays out longer). Each shock produces a stand-alone capital number; the correlation matrix then determines how much of those two numbers can offset each other when producing the firm-wide total, because equity crashes and longevity improvements aren't things that typically happen together.

The regime also changes how liabilities are valued in the first place: insurance liabilities are discounted using a prescribed risk-free curve plus adjustments (the "volatility adjustment" and "matching adjustment") designed to reduce artificial swings in reported solvency purely from short-term bond market noise, which matters enormously for long-dated annuity books.