Against the Gods Ch. 4: Insurance — Turning Risk Into a Commodity

Investment Background

The first three chapters covered ideas. This chapter covers the first time those ideas changed the world at scale.

Insurance is the first industry built entirely on probability.

And its importance to retirement planning exceeds anything else in this book — because the two largest risks in retirement can only be handled by insurance, not by investing.

The Wall Street Translation

Origins in a Coffee House

In the 1680s, a coffee house stood by the Thames in London, run by a man named Edward Lloyd.

Ship owners and merchants gathered there to exchange news about shipping — which routes had pirates, which season brought storms, which vessel was overdue.

Gradually a transaction formed there: wealthy men willing to bear risk signed their names beneath a document, along with the share they would carry. That is the literal origin of the word "underwriter" — one who writes underneath.

That coffee house became Lloyd's of London.

The origin story contains the whole mechanism of insurance, and is worth taking apart.

What Insurance Actually Does

The most common misconception is that insurance reduces risk.

It does not. A ship's probability of sinking does not fall because it is insured.

Insurance does two entirely different things:

One, transfer. Risk moves from someone who cannot bear it to those who can.

To one ship owner, losing a ship is ruin. To a group of underwriters each carrying 1% of a hundred ships, losing one ship is a predictable cost.

The identical loss means completely different things to different balance sheets.

Two, pooling. The mathematically more interesting part, and it uses Chapter 3's law of large numbers directly.

Whether one particular ship sinks is entirely unpredictable.

How many of a thousand ships will sink is quite predictable.

That is insurance's central magic: unpredictability at the individual level becomes predictability in aggregate.

An insurer need not know which ship will sink. It need only know roughly how many will.

A Key Corollary

The mechanism yields a direct and highly practical corollary:

Insurance works only where risk can be pooled.

If one event affects every policyholder simultaneously, pooling fails.

Which is why certain risks are hard to insure:

Risk	Poolable	Insurance market
One person dying early	Yes — independent events	Mature, cheap
One house burning	Yes	Mature
One person living too long	Yes — this is an annuity	Exists
A nationwide financial collapse	No — affects everyone at once	Essentially none
Hyperinflation	No	Very limited

The last two rows matter enormously to retirement investors, and we develop them below.

The Two Risks in Retirement Planning

This is the chapter's most practical section.

Among retirement's risks, two cannot be solved by investing — and both happen to be poolable, which is why they belong to insurance.

Risk One: Longevity Risk

A peculiar risk, because it arrives through good news.

You do not know how long you will live. Plan to 85 and live to 97, and you exhaust your money at your most vulnerable age.

Why investing cannot solve it: you can build an enormous portfolio for an extremely long life, but that means excessive frugality through years you may never see — and if you die early, that money is wasted caution.

Pooling can solve it.

That is the annuity mechanism: a group jointly bears the uncertainty of who lives long. Early decedents' funds subsidize payments to the long-lived.

Both sides of annuities must be stated fairly:

• They genuinely solve a problem investing cannot.
• They are also frequently sold with high commissions, high fees, and complex terms — exactly what the incentive model in Chapter 3 of Poor Charlie's Almanack predicts.

So the conclusion is: annuities are sound as a concept; specific products need the scrutiny of that chapter. Our recommendation to delay Social Security is fundamentally the purchase of a government-provided, inflation-adjusted annuity — usually the best-priced one available.

Risk Two: Long-Term Care

One of the two risks Chapter 6 of Retirement Decumulation Mechanics explicitly names as unsolvable within that book.

The reason is the same as longevity: most people never need long-term care, a minority need it and the cost is enormous. That distribution is exactly the shape insurance was designed for.

An Important Qualification

Insurance is not universal, and understanding its boundary matters as much as understanding its mechanism.

One, insurance has a cost, and the cost is real.

Insurers must cover claims, operations, and profit. So in pure expected-value terms, buying insurance is a negative expected value transaction.

Then why buy it?

Because expected value is not the only criterion — precisely the qualification at the end of Chapter 2. When a loss removes you from the game, the value of avoiding it exceeds its expected cost.

Which gives a clear rule:

Insure losses you cannot absorb. Do not insure losses you can.

That rule explains many specific recommendations:

• Should buy: homeowner's, health, term life (when you have dependents), liability. Their maximum loss is catastrophic.
• Usually should not: extended warranties, phone insurance, trip cancellation. Their maximum loss you can simply absorb.

Every policy is negative expected value. The difference is whether it prevents an unrecoverable outcome.

Two, some risks cannot be insured at all.

Return to the table above: systemic risk cannot be pooled.

Nobody can sell you a policy protecting you from thirty years of poor global equity returns, because the insurer would be bankrupt when it triggered.

For those risks the only response is buffers and diversification — not insurance.

This is the practical application of Chapter 1's distinction: poolable risk gets insurance; unpoolable uncertainty gets buffers.

Executable Trading Rules

1. Decide whether to insure by "can I absorb the maximum loss," not by probability. The chapter's most important line. A low-probability, absorbable loss (a lost phone) needs no insurance. A low-probability, unabsorbable one (long-term care) does.

2. Understand delaying Social Security as buying an annuity, not as an investment decision. That reclassification changes your criterion. You are not maximizing expected value; you are eliminating a tail risk. Our Social Security Optimizer computes the breakeven, and this chapter explains why the breakeven is not the only consideration.

3. For any insurance product, ask about the incentive structure first. See Chapter 3 of Poor Charlie's Almanack. The higher the upfront commission, the more independent verification it needs.

4. Do not try to insure against systemic risk. Complex products promising protection in a market collapse are typically expensive and may fail when actually needed (counterparty risk). Buffers and diversification are more reliable and cheaper.

5. Accept that insurance is negative expected value, and buy it anyway. Not a contradiction. Expected value applies to repeatable bets; insurance handles non-repeatable catastrophes.

Relevance to a Retirement Portfolio

This chapter fills an important framework gap on this site.

Most retirement content handles investment questions: what to hold, how much to withdraw, when to rebalance.

And some of retirement's risks are not investment questions at all.

Risk	Correct tool
Market volatility	Investing (diversification, cash buffer)
Inflation	Investing (equities, inflation-linked bonds)
Sequence-of-returns risk	Investing (buffer, dynamic withdrawal)
Longevity	Insurance (annuities, delayed Social Security)
Long-term care	Insurance (or a dedicated self-insured fund)
Early death (for dependents)	Insurance (term life)

Using investment tools to solve an insurance problem is a common and expensive error.

Its specific form: someone tries to handle longevity risk and care costs by holding an oversized portfolio. That is mathematically possible, but it requires far more savings than handling the same risks through insurance.

Because insurance exploits pooling and individual savings cannot.

Which is exactly what this book's history demonstrates: what those seventeenth-century London underwriters discovered is a mechanism no individual, however wealthy, can replicate alone.

Chapter 5 covers the story's turning point: when twentieth-century economists discovered that the entire edifice they had built contained a hole no mathematics could fill.