Executive summary
- Expected loss is the starting point, not the answer; it is the least uncertain and least decision-relevant element of the chain.
- Capital is the price of uncertainty, and it should be charged to the business that generates it.
- Technical premium equals expected loss plus expenses plus the cost of capital plus a margin for model uncertainty.
- Limits translate appetite into constraints that bind at the point of underwriting rather than at year end.
- Risk appetite expressed only in narrative terms cannot be monitored and therefore does not constrain anything.
The chain, link by link
- Expected loss: the probability-weighted average cost of claims for the exposure, before expenses and before any allowance for volatility.
- Uncertainty: the dispersion around that expectation, including process, parameter and model components.
- Capital: the resources required so that the probability of failure over the chosen horizon stays within the tolerance the board has set.
- Cost of capital: the return required on those resources, which becomes a cost the price must recover.
- Technical price: expected loss plus expenses plus cost of capital plus a margin for what the model may be missing.
- Limits and appetite: constraints on accumulation, single-risk exposure and portfolio composition that keep aggregate outcomes within tolerance.
Each link depends on the one before. If the expected loss is understated, every subsequent figure inherits the error. If the uncertainty is understated, capital is too low and the price is too cheap in exactly the segments that are most volatile.
Capital as the price of uncertainty
Two portfolios with identical expected losses can require very different capital. The one with heavier tails, greater accumulation potential or weaker data consumes more. Charging both the same loading means the stable portfolio subsidises the volatile one, and the subsidy is invisible because it never appears as a line item.
P = E[L] + X + r × C + M
- P
- — technical premium
- E[L]
- — expected loss for the exposure
- X
- — allocated expenses, including acquisition and claims handling
- r
- — required return on capital, net of investment return on the capital held
- C
- — capital allocated to the exposure, reflecting its marginal contribution to portfolio tail risk
- M
- — margin for model and parameter uncertainty
The marginal formulation matters. Capital should be allocated on the basis of what a risk adds to the portfolio tail, not on a standalone basis. A risk that diversifies against the existing book consumes less capital than the same risk added to a book already concentrated in that peril, and the price should reflect that.
Making risk appetite operational
A risk-appetite statement that says the insurer has a low tolerance for catastrophe exposure cannot be monitored, breached or enforced. An operational statement specifies a metric, a threshold, a monitoring frequency and an owner.
| Narrative appetite | Operational expression |
|---|---|
| Low tolerance for solvency volatility | Solvency ratio remains above a stated floor in the one-in-twenty adverse scenario, reviewed quarterly |
| Limited catastrophe exposure | Modelled 1-in-200 net occurrence loss does not exceed a stated percentage of own funds |
| Diversified portfolio | No single peril-region contributes more than a stated share of modelled capital |
| Prudent reserving | Booked reserves sit within a defined percentile band of the stochastic distribution, with deviations reported |
The translation exercise itself is valuable. It forces the board to decide what the narrative actually means, and it frequently reveals that different members had different numbers in mind for the same words.
Where the chain breaks
- Pricing and capital teams using different assumption sets, so the price does not recover the capital the capital model says is required.
- Limits set at levels never approached, which provide comfort without constraint.
- Appetite monitored annually while underwriting decisions are taken daily.
- Underwriters given authority in premium terms rather than in exposure or modelled-loss terms.
- Model uncertainty margin omitted entirely, so the price implicitly assumes the model is correct.
Practical example
Consider a specialty risk with an expected loss of 100, allocated expenses of 25, and marginal capital of 300 at a required net return of 8 per cent, giving a capital charge of 24. Adding a model-uncertainty margin of 10 per cent of expected loss produces a technical premium of approximately 159.
If the market clears at 130, the risk is being written below technical price. That is a legitimate commercial decision — for strategic reasons, portfolio balance or relationship value — but it should be a recorded decision with the shortfall quantified, not an unrecognised outcome of pricing on expected loss plus a flat loading. Aggregating the shortfall across the book gives management a number for what the market cycle is costing.
Limitations and caveats
- Marginal capital allocation depends on the portfolio composition at the time of calculation and changes as the book changes.
- Required return on capital is a management choice, not an observable market price, and reasonable people set it differently.
- Model uncertainty margins are judgemental; there is no accepted method for calibrating them.
- Expected loss estimates for emerging exposures may be based on very limited data.
Conclusion
The insurance decision chain is conceptually simple and organisationally difficult. Each link is owned by a different function, and the connections between them are where information is lost.
An insurer that can trace a single risk from expected loss through capital to price to limit, using one consistent set of assumptions, has a functioning framework regardless of how sophisticated its individual models are.
References
- Insurance Core Principles — International Association of Insurance Supervisors
- Solvency II Directive 2009/138/EC — European Commission, 2009
- EIOPA Guidelines on own risk and solvency assessment — European Insurance and Occupational Pensions Authority
Author bio
Jonas Mohamed Osman Abdelghafour, known as Yonas Osman is an actuary, FRM and financial risk professional specialising in banking, insurance, model risk, capital modelling and quantitative risk management.