Insurers buy insurance too
By the time you reach this rung you already understand the central trick of insurance: pool many independent risks and, by the law of large numbers, the *average* outcome becomes far more predictable than any single one. A company sells thousands of policies precisely so that the smooth mass cancels out the bumps. But pooling has a quiet limit. It tames risks that are *independent and similar in size*; it does almost nothing against a risk that is enormous, or one shock that strikes thousands of your policies at once. For those, the insurer needs the same medicine it sells — it needs to insure itself. That second layer of insurance is reinsurance.
The arrangement has its own cast and its own grammar. The company that buys the protection — the original insurer passing risk upward — is the cedant (or ceding company); the company that accepts that risk is the reinsurer. Reinsurers are usually large, globally diversified firms; many sit on the other side of the world from the cedant, which is the whole point — a hurricane in Florida means nothing to a balance sheet whose other risks are earthquakes in Japan and motor crashes in Germany. Risk that cannot be diluted inside one country's book is diluted across the planet.
Retention and cession: the line down the middle
Every reinsurance deal does one thing at its core: it splits each risk into a piece the cedant keeps and a piece it passes on. The part you keep is your retention; the part you hand over is the cession. The whole craft of reinsurance is choosing *where to draw that line*. Keep too much and you are still exposed to the very shocks you fear; cede too much and you give away premium, profit, and the diversification benefit of your own book for protection you may not need.
The line can be drawn two fundamentally different ways, and it is worth fixing the distinction now because every treaty you will meet later is a variation on one of them. You can split *proportionally* — keep a fixed share of every policy, say 70%, and cede the other 30% of premiums and claims alike, as in quota-share. Or you can split by *size* — keep every claim up to a threshold and cede only the amount above it, as in excess-of-loss. Proportional cuts each risk the same way regardless of how big it turns out; excess-of-loss leaves small claims entirely with you and only engages on the large ones. Same idea, retention versus cession, but the line runs in a different direction.
Reason one: capacity — writing risks you could never hold alone
The first classic reason to cede risk is capacity. Imagine a mid-sized insurer asked to cover a 2-billion oil refinery. A single such loss would dwarf the company's entire surplus; prudence — and the regulator — forbid putting that much on one policy. Without reinsurance the insurer simply turns the business away. With it, the insurer can write the whole policy, keep a comfortable slice it can absorb, and cede the rest. Reinsurance lets a small balance sheet underwrite a large risk, the way a contractor takes on a skyscraper by subcontracting most of the work.
Capacity is not only about one giant policy; it is also about *volume*. A young, fast-growing insurer may want to write more business than its capital can safely support all at once. By ceding a steady fraction of everything — a quota-share treaty — it effectively rents the reinsurer's balance sheet, growing the top line today and buying time to build its own surplus. The reinsurer is happy to lend that capacity because, spread across its global book, this slice is just one more small, diversifying piece.
Reason two: stability — smoothing the year-to-year ride
The second reason is stability. Even a well-run insurer has lumpy results: a quiet year, then a year with a few unusually large claims that drags the loss ratio up and the profit down. Owners, regulators, and rating agencies all dislike that volatility — a wildly swinging earnings line looks risky and raises the company's cost of capital. Reinsurance acts as a shock absorber. By ceding the upper part of large claims, the cedant trades a *higher average cost* (it pays away a margin) for a *much narrower spread* of outcomes. You give up a little expected profit in calm years to avoid a frightening loss in bad ones.
Here is a tiny worked picture. Suppose the same hundred-policy book is run for many years. In an ordinary year claims total about 100; once in a while a large claim pushes the total to 180. The expected cost is, say, 108, and the *range* of outcomes is 100 to 180. Now cap every claim through excess-of-loss reinsurance so that the worst possible total the cedant retains is 130. The reinsurer charges a fair price of 9 (the 8 of expected ceded claims plus a 1 margin). The cedant's expected cost rises to 109 — one unit worse — but its worst case has fallen from 180 to 130. That is the trade in numbers: a hair more on average, far less at the top.
Reasons three and four: catastrophe protection and capital relief
The third reason is catastrophe protection, and it attacks pooling's deepest weakness. The whole point of a pool is independence, but a hurricane, earthquake, or flood is the opposite of independent: one event damages thousands of nearby homes *at the same time*. The losses arrive together, fully correlated, and the law of large numbers offers no comfort at all. This is catastrophe risk, and against it an insurer buys catastrophe cover — a treaty that pays when the *combined* loss from a single event crosses a threshold. It is the difference between surviving a bad year and surviving a single bad afternoon.
The fourth reason is the most modern and the least intuitive: capital relief. Regulators require an insurer to hold capital sized to the risk on its books — the economic capital needed to survive a bad year to a high confidence. Cede away part of that risk and the required capital falls with it; the cedant frees up money it can return to shareholders or deploy into new business. In effect, reinsurance can be cheaper than equity: rather than raise expensive capital to back a risk, the company *rents* a reinsurer's capital to stand behind it. When reinsurance is used mainly to manage the balance sheet this way rather than to genuinely transfer loss, regulators watch closely — the deal must move real risk, not just dress up the numbers.
THE FOUR CLASSIC REASONS TO CEDE RISK
1. CAPACITY write risks bigger than you could hold alone
(one giant policy, or more volume than capital allows)
2. STABILITY smooth lumpy year-to-year results
-> shrinks the VARIANCE of retained outcomes
3. CATASTROPHE survive one event that hits thousands at once
-> shrinks the correlated TAIL the pool cannot tame
4. CAPITAL cede risk -> required capital falls -> cash freed
-> rent the reinsurer's balance sheet vs. raising equity
Common thread: trade a little expected cost (the reinsurer's margin)
for a smaller, safer spread of outcomes.Honest limits, and where this leads
Three honest caveats keep this from becoming a free lunch. First, reinsurance has a *price*, and that price is more than the expected ceded loss — the reinsurer charges its own margin for the same reason any insurer does, so on average the cedant pays away a little extra. Second, the protection is only as good as the reinsurer's willingness and ability to pay, which is the counterparty risk we met above. Third, contract wording matters intensely: an exclusion you overlooked, or a dispute over whether several losses count as 'one event', can leave a gap exactly where you thought you were covered.
Finally, the partner need not be another insurer at all. For the very largest catastrophe risks, cedants increasingly turn to the capital markets directly, selling insurance-linked securities such as catastrophe bonds to investors who accept the risk in exchange for a generous coupon. It is the same idea — pass risk to a balance sheet broad enough to absorb it — but the balance sheet now belongs to global investors rather than a traditional reinsurer. We will meet those modern capital-market deals, and the precise mechanics of layers and treaties, in the guides ahead.