October 12, 2026 · BriMindInvest Research Team · 19 min read
Insurance is the only major sector where the customer pays in full, years before anyone knows what the product cost to deliver. That inversion creates float — other people's money, held and invested, often at negative cost. It also makes reported earnings an estimate rather than a fact, which is why the standard valuation toolkit fails here. This guide covers the mechanics first, then eleven companies across four genuinely different business models.
A manufacturer buys materials, pays workers, builds a product, ships it, and then collects revenue. Costs are known before the sale. Margin is arithmetic.
An insurer sells a promise. The premium arrives on day one. The cost of honouring that promise might be zero, or it might arrive seven years later in the form of a liability claim nobody anticipated. Insurance is the only large sector that books revenue before it knows its cost of goods sold.
Two consequences follow, and almost everything else in this guide is downstream of them.
Between collecting premiums and paying claims, the insurer holds the cash. It invests it. The investment income belongs to the insurer, not the policyholder. This pool is called float, and it is the engine of the entire sector.
To report earnings, an insurer must estimate what it will eventually pay on claims that have occurred but have not yet been settled — and on claims that have occurred but have not yet even been reported. That estimate is called the reserve, and management sets it. Move the estimate slightly and this quarter's profit moves a lot.
You are not reading a historical record when you read an insurer's income statement. You are reading management's current opinion about events that have not finished happening. The quality of that opinion — whether this management team has historically been too optimistic or too conservative — is the single most important qualitative judgment in insurance investing, and it is knowable from the disclosures.
Float is the money an insurer holds between receiving premiums and paying claims. For a large carrier it runs to tens of billions of dollars, and critically, it is persistent: as old policies pay out, new ones are written, so the pool refills continuously. A stable or growing book of business means the float never has to be repaid in aggregate.
The decisive question is what the float costs.
A business that is paid to hold other people's capital, and that can invest that capital for its own account, compounds in a way an ordinary balance sheet cannot. That is the entire thesis for high-quality P&C underwriters. It only works if underwriting discipline holds — which is why the combined ratio matters so much.
Not all float is equally valuable. Personal auto claims settle in months, so the money is held briefly and must be invested conservatively. Long-tail commercial casualty claims — professional liability, workers' compensation, product liability — can take a decade to settle, so the float is held far longer and can be invested further out the risk curve.
Long-tail float is more valuable per dollar and more dangerous per dollar, because a decade is long enough for the original estimate to be badly wrong.
The combined ratio is the percentage of each premium dollar consumed by claims and expenses. Below 100% the insurer made money underwriting before any investment income. Above 100% it lost money underwriting and is relying on its portfolio.
Seven dollars of underwriting profit on a hundred dollars of premium sounds thin. It is not, because it ignores the second income stream. That same insurer is also holding and investing float generated by this and every prior year's policies. The underwriting profit is the fee it collects for the privilege of holding the investable money.
A 97% combined ratio can be built very differently, and the construction tells you about durability:
Direct-to-consumer carriers generally run higher expense ratios because they pay for advertising rather than agent commission, and the best of them offset it with a materially better loss ratio from superior segmentation.
Always compare within a line of business. A specialty commercial underwriter and a personal auto carrier running the same combined ratio are not performing equally.
This is the section most sector guides skip, and it is the one that separates people who can read an insurer from people who cannot.
When an insurer sets a reserve, it is guessing. Later, reality arrives. If claims settle for less than reserved, the surplus flows back into earnings as favourable prior-year development. If they settle for more, the shortfall is charged against current earnings as adverse development.
An adverse development charge is not simply a bad quarter. It is a formal statement that profits reported in earlier years were overstated — that the margin you were shown did not exist. And reserve errors are rarely isolated: a team that underestimated one accident year usually used similar assumptions on adjacent years, so the first charge is frequently followed by more.
The reverse carries a warning too. Consistently large favourable development is pleasant but not free: it means the company has been over-reserving and therefore understating past earnings. That is the conservative failure mode, and markets reward it — but a company that releases reserves every quarter to hit an earnings target is using the reserve as a lever, not an estimate.
Social inflation — rising jury awards, litigation funding, broader liability theories — has made long-tail casualty reserving harder across the industry. Carriers with large casualty books carry more of this uncertainty than their combined ratios suggest.
This is the most common analytical error in the sector. These four models have different risk profiles, different valuation anchors, and different sensitivities. Buying one of each is not diversification within a theme.
They take the actual risk. Auto, home, commercial property, liability. Earnings are volatile because catastrophes are volatile, and the float is the prize. Valued on price-to-book against sustained return on equity. Most exposed to weather, to reserve error, and to the pricing cycle.
They place insurance and collect a commission. They take no underwriting risk whatsoever. No claims, no reserves, no catastrophe exposure, minimal capital. When industry premiums rise, their revenue rises with no corresponding increase in liability. This is a fundamentally different — and structurally higher-quality — business, which is why brokers trade at multiples closer to professional-services firms than to insurers. They are valued on earnings and free cash flow, where P/E is actually appropriate.
They sell contracts lasting decades and are primarily a bet on interest rates, longevity assumptions and credit quality. Rising rates generally help them by raising reinvestment yields on an enormous bond portfolio; falling rates compress spreads on guaranteed products. The liabilities are long and the assumptions are difficult to verify from outside. Valued on book value, with close attention to the quality of the investment portfolio.
Insurance for insurers. They absorb the tail risk the primary carriers do not want, which means extremely lumpy results: long stretches of profitability punctuated by severe loss years. Reinsurance pricing hardens sharply after major catastrophe losses, which makes the sector the most cyclical corner of an already cyclical industry.
A rising-premium environment is unambiguously good for brokers and good for disciplined underwriters, but it usually arrives because underwriters just absorbed large losses. The broker captures the upside of hard pricing without having paid the claims that caused it. If you want exposure to the insurance cycle without the catastrophe tail, that asymmetry is the reason brokers exist in portfolios.
These are the large, liquid, widely held names that define the sector. We describe what each business actually is and what the live debate about it is — and we deliberately do not hardcode valuations or combined ratios, because those change quarterly and a stale number is worse than no number. Each name links to its analysis page for current figures.
The case: The best data operation in personal lines. Decades of telematics and segmentation let it price risk at a granularity competitors cannot match, which shows up as a structurally lower loss ratio rather than lower expenses.
What to watch: Growth is volume-led, so it gives back margin to take share. Combined ratio can look worse precisely when the business is doing well.
The case: Large captive agent network and a recognised brand, with a long repositioning toward direct distribution and tighter underwriting in catastrophe-exposed states.
What to watch: Homeowners exposure concentrates weather risk. Rate filings are a regulatory negotiation, so margin repair lags inflation by quarters.
The case: Disciplined commercial underwriter with a long record of reserving conservatively. Commercial lines reprice faster than personal lines, which makes the earnings path less politically constrained.
What to watch: Commercial is cyclical. When pricing softens across the industry, discipline means losing business rather than losing money.
The case: Specialty and large-account commercial lines where underwriting expertise is the product. The high-net-worth personal book is unusually defensible — wealthy clients rarely shop on price.
What to watch: Global footprint adds currency and geopolitical exposure that domestic-only peers do not carry.
The case: Strong small-commercial franchise plus a group benefits arm whose economics behave differently from P&C, smoothing the cycle somewhat.
What to watch: Smaller scale than the mega-caps in an industry where scale genuinely helps spread catastrophe risk.
The case: Earns commission on premium placed and takes no underwriting risk at all. When premiums rise industry-wide, revenue rises without any corresponding increase in claims exposure.
What to watch: Priced as a quality compounder. Multiple compression is the main risk, not operations.
The case: Same capital-light model, with a large reinsurance broking arm that sits at the centre of how catastrophe risk gets distributed globally.
What to watch: Organic growth depends on the pricing cycle it brokers; a prolonged soft market compresses commission growth.
The case: Serial acquirer of small regional brokers, consolidating a genuinely fragmented middle market. Roll-up economics on top of a capital-light base.
What to watch: The thesis depends on continued disciplined M&A. Acquisition multiples rising is the thing to monitor.
The case: Supplemental policies are small-ticket, high-persistency and sold at the worksite. A large share of earnings comes from Japan.
What to watch: Yen translation materially affects reported results. Currency is a real variable, not a footnote.
The case: Group benefits sold through employers are sticky and far less rate-sensitive than individual annuities.
What to watch: Life insurers are leveraged to interest rates and to long-dated actuarial assumptions that are hard to audit from outside.
The case: GEICO, Berkshire Hathaway Reinsurance and National Indemnity generate float that funds an operating-company and equity portfolio. The purest expression of the float model in public markets.
What to watch: You are buying a diversified holding company, not an insurance pure-play. The insurance thesis is only part of what you own.
These eleven are included because they are the most liquid and widely analysed names across the four models — which makes them a sensible place to learn the sector. Inclusion here is not a view that any of them is attractively priced today. Run each through the framework above and through current numbers before acting.
Insurance pricing moves in long, self-correcting waves, and the mechanism is genuinely elegant.
The investment consequence is counterintuitive: the best time to own quality underwriters is usually just after the industry has taken large losses, provided the specific company you own did not take an outsized share of them. The worst time is late in a soft market, when everything looks calm, margins look fine, and the underpriced policies being written today have not yet produced their claims.
For an underwriter, the P/E ratio is built on a number that is partly an estimate, and it swings violently with catastrophe timing. A carrier can report a terrible year because of one hurricane season and an excellent one because of a quiet year, with no change whatsoever in the quality of the business. Capitalising either figure gives a nonsense answer.
Book value is more stable, because it reflects accumulated capital rather than a single period's outcome. The correct framing pairs it with returns:
Two refinements worth applying. First, use growth in book value per share plus dividends over five to ten years as the real scorecard — it is the cleanest measure of whether an insurer is compounding capital. Second, check whether book value has been distorted by unrealised bond gains or losses, which move with interest rates and can swing reported equity sharply without anything changing operationally.
For brokers, none of this applies. They hold no reserves and carry little capital, so earnings and free cash flow are the right lens and P/E is a perfectly sensible metric. Applying insurance-underwriter logic to a broker is its own category of mistake.
The single most common way an insurer destroys value. Reserves are an estimate of claims not yet paid. Set them too low and today's earnings look excellent — right up to the year the estimate has to be corrected. Adverse development is the formal admission that prior profits were overstated.
When industry pricing falls, the disciplined response is to write less business. Managers compensated on premium growth often do the opposite, winning share precisely when each policy is worth least. The losses arrive two to four years later.
An insurer can be perfectly profitable for a decade and lose that decade in one season if its book is geographically concentrated. Reinsurance exists to cap this, but reinsurance is itself repriced after large loss years.
Float has to be invested somewhere. The temptation in low-rate periods is to move into credit and illiquid assets. That converts an underwriting business into a leveraged credit fund, and the two risks correlate badly in a crisis.
Asbestos, opioids, talc, and environmental claims surfaced decades after the policies were written. Any insurer with legacy casualty exposure carries a tail that is genuinely unknowable from the outside.
Notice that four of these five are failures of discipline rather than failures of luck. Insurance rewards companies that are willing to shrink. A management team that treats premium growth as the primary objective is, in this sector specifically, describing a risk rather than an achievement.
Every insurer runs a large bond portfolio funded by float. Higher rates raise the income earned on that portfolio, and the benefit builds gradually as maturing bonds are reinvested at higher yields — which means a sustained higher-rate environment improves insurance earnings for years after rates stop rising.
The transition itself is less pleasant. When rates rise quickly, existing bonds fall in value, which reduces reported book value through unrealised losses. Since insurers typically hold to maturity, those losses largely reverse over time — but the optical damage to book value is real while it lasts and can make a healthy insurer look cheap or impaired depending on which direction you are reading it from.
Life insurers and annuity writers are the most rate-sensitive of all, because their liabilities extend furthest. For them, the level of rates is close to the whole investment thesis.
Rising catastrophe losses have changed the sector structurally. The important shift is not that storms happen — they always did — but that the pricing response is increasingly constrained by regulation in exactly the places where risk is rising fastest.
A reasonable way to approach the sector, assuming you want exposure rather than a single bet:
On dividends: large P&C insurers and brokers are often durable dividend payers, because stable float-driven earnings support steady distributions. But the dividend is only as safe as the reserves behind it — a carrier taking adverse development has a strained balance sheet whatever its payout ratio implies. If income is the objective, our dividend growth guide covers how to assess payout durability more generally.
Insurance earnings are an accounting estimate, swinging on reserve assumptions and catastrophe timing. A low P/E often means the market disbelieves the reserves. Price-to-book, read alongside sustained ROE, is the right primary lens.
A 95% combined ratio achieved by underreserving is worse than a 101% achieved honestly. Always check whether prior-year reserve development is favourable or adverse before you credit the margin.
Brokers take no underwriting risk and benefit from rising premiums. P&C underwriters carry the risk. Life insurers are primarily an interest-rate and longevity bet. Owning one of each is not diversification within a theme — it is three different businesses.
The reflex trade is to buy insurers after a hurricane because pricing will harden. That is sometimes right, but only for companies whose actual losses were contained. Harden-the-market logic does not help a carrier that just absorbed the loss.
Rate increases in personal lines require approval from state insurance departments. In politically sensitive states, a carrier can be structurally unable to price for the risk it is carrying — and its only remaining option is to stop writing business there.
Every company in this guide has a full analysis page with current valuation, profitability, balance sheet and historical financials — so you can apply the book value and ROE framework above to live numbers instead of the ones in this article.
Insurance is the only sector where revenue arrives before cost is known. That inversion creates float — a persistent pool of other people's money that the best underwriters are effectively paid to hold — and it makes reported earnings an estimate rather than a fact.
Which means the usual toolkit fails. Use price-to-book against sustained ROE, not P/E. Read the combined ratio across a full cycle, not a single year. And check reserve development before you believe any of it, because that is where an optimistic management team hides the difference between reported profit and real profit.
Finally, decide which of the four businesses you are actually buying. A broker, a P&C underwriter, a life insurer and a reinsurer share a regulator and almost nothing else. The most common way to get this sector wrong is to treat them as one trade.
This article is educational and is not investment advice. Company descriptions reflect business models and structural characteristics; current financial metrics change quarterly and should be checked on each company's analysis page before making any decision.
Get AI prediction signals, unlimited stock comparisons, portfolio analytics, and personalized watchlists — free for 14 days, no credit card required.
14-day free trial · No credit card required · Cancel anytime
Not ready to sign up? Get one free email a week instead.