Insurance for insurers: how risk is transferred, structured and valued between the ceding company and the reinsurer, from quota share to stop-loss and optimal retention.
Reinsurance is the mechanism by which an insurer (the ceding company) transfers part of its risk to another entity (the reinsurer). It is not a luxury: it is the tool that allows a company to write risks larger than its own capital without compromising its solvency. Its essential functions are:
The fundamental split is between proportional reinsurance (a share of premiums and claims is ceded) and non-proportional reinsurance (the reinsurer pays only above a threshold).
The ceding company and the reinsurer share premiums and claims in an agreed proportion. The reinsurer usually pays back a ceding commission.
The ceding company retains a fixed proportion $a$ of each risk (premiums and claims). Simple and effective for reducing volume, but it also cedes the small risks it could keep.
The ceding company sets a maximum retention $R$ (the line) per policy; whatever exceeds it out of the sum insured $\text{SI}_i$ is ceded. The ceded proportion varies by risk, so the small risks are kept.
The reinsurer pays the ceding company a commission on the ceded premium that recognizes its acquisition and administration costs. It may follow a sliding scale based on the loss ratio.
The reinsurer steps in only when the loss exceeds a retention (deductible), up to a limit. It is the natural cover against large losses and catastrophes.
The reinsurer pays the part of loss $X$ that exceeds the retention $d$, up to a maximum $\ell$ (the "$\ell$ xs $d$" layer). It can be per risk (WXL) or per event (CatXL).
Covers the excess of the annual aggregate loss $S$ over a retention $d$ (often expressed as a percentage of premium). Its pure premium is the integral of the tail of $S$.
After a loss that exhausts the XL layer, a reinstatement restores it for an additional premium proportional to the amount used. It limits the number of events covered.
Layers are priced on experience (burning cost) and on exposure (exposure curves / catastrophe models), adjusting for inflation, trend and tail volatility.
How much risk should be retained? Actuarial theory offers formal answers depending on the objective and the cost constraint.
Karl Borch showed that, for a given reinsurance cost, the contract that minimizes the variance of the retained loss is stop-loss. It is the "most efficient" reinsurance in the mean-variance sense.
At the Pareto-optimal equilibrium, the ratio of marginal utilities is constant across states. If both parties have exponential risk aversion, the optimal sharing is linear (quota share).
Retention is also chosen to keep the probability of ruin below a threshold: ceding risk raises the adjustment coefficient $R$ and reduces $\psi(u)$, at the cost of the ceded premium.
Under the Ley de Instituciones de Seguros y de Fianzas (LISF, Insurance and Surety Institutions Law), ceded reinsurance reduces the solvency capital requirement (RCS) provided the reinsurer is registered (RGRE, the registry of foreign reinsurers) and the risk transfer is effective. The actuary values the effect of reinsurance on both technical reserves and regulatory capital.