How Life Insurers Make Money: Float & Embedded Value

A reader wrote to me last results season, genuinely confused. "HDFC Life trades at a P/E of 67," he said. "That's richer than most consumer-staples darlings. Is a life insurer really worth more than the value of a year's profit sixty-seven times over?" Fair question. He'd done the arithmetic right — at a market value near ₹1,28,000 crore (as of 29 Apr 2026) against about ₹1,911 crore of FY26 net profit, the P/E (price ÷ yearly profit) really is roughly 67 (Inve data, 2026). Market value and the P/E and price-to-EV derived from it move with the daily share price, so treat them as a snapshot, not a fixed number.

The arithmetic was fine. The tool was wrong. He was reading an insurer the way you'd read a packaged-goods company — and an insurer does not earn money the way a company that sells you a product today does. Until you see how the money actually arrives, the profit line will mislead you every single time.

The shopkeeper who collects rent decades before he owes anything

Here is the homely picture to hold for the rest of this piece. Imagine a shopkeeper who, instead of selling you something today, collects a small payment from you every year for twenty years — and only has to hand money back, to you or your family, far in the future, if and when something specific happens. In between, he is sitting on an enormous pile of other people's money. He can invest that pile. The returns it earns are largely his to keep.

That is a life insurer, almost exactly. You pay premiums now; the claim — death cover, or a maturity payout — comes years or decades later. The gap is the whole business.

Warren Buffett, who built Berkshire Hathaway on exactly this, named the pile float. His definition is worth memorising: "To begin with, float is money we hold but don't own. In an insurance operation, float arises because premiums are received before losses are paid, an interval that sometimes extends over many years." (Berkshire Hathaway 2002 Chairman's Letter). Money we hold but don't own. Read it twice. That is the single most important sentence in insurance investing.

How big does this pile get? Across all Indian insurers, the money they manage on behalf of policyholders is roughly ₹74 lakh crore, against total premiums of about ₹11.93 lakh crore collected in FY25 (Deccan Herald opinion column, citing industry figures). Most of an insurer's value sits in managing that mountain well — not in this year's reported profit.

Why the profit line tells you almost nothing

Look at HDFC Life's own numbers and the mismatch jumps out. For FY26 it ran about ₹99,432 crore of total income — premium income plus investment and other income, summing the four quarterly total-income lines (Inve data, as of 29 Apr 2026) — but reported about ₹1,911 crore of net profit (the company's own FY26 figure is ₹1,912.3 crore; EquityBulls puts FY26 total premium income alone at ₹79,493 crore). The profit is roughly two paise on every rupee of income. If you valued it on that thin sliver — the way the P/E forces you to — you would conclude it's wildly overpriced and walk away.

You'd be making the shopkeeper's mistake: judging him by the cash in his till this evening, while ignoring the twenty-year pile growing in his back room. Most of the money a policy will ever make the insurer is locked in the future — earned slowly as that customer keeps paying, and as the float compounds. Accounting profit barely captures it. So insurers report a different number built precisely to capture it: embedded value.

Embedded value (EV) is the insurer's net worth plus the present value of all the future profit already baked into policies it has sold. Think of it as the back-room pile valued today. HDFC Life's EV stood at ₹55,423 crore for the year ended March 2025, up 17% year-on-year (HDFC Life FY25 press release). Against that, the ~₹1.28 lakh crore market value implies a price-to-EV of roughly 2.3x. But note what EV actually is: a management estimate, built on movable assumptions — the assumed investment return, how long customers stay, future expenses. Change those inputs and the number moves. So 2.3x is a different lens on the same business, not a "cheap" or "expensive" verdict — don't swap one naive multiple for another.

The number that decides whether the float survives: persistency

Float only compounds if customers keep paying. A life policy is a twenty-year relationship; if the customer quits in year three, the insurer loses most of the future profit it had pencilled in. So the metric an owner watches like a hawk is persistency — the share of customers who renew, measured at fixed ages of the policy.

HDFC Life reported 13th-month persistency of 87% for FY25 (India Infoline) — meaning 87 of every 100 customers were still paying a year after they bought. The longer-dated 61st-month number (still paying after five years) sat in the low 60s. The gap between those two tells you how sticky the book really is: it's easy to keep a customer for a year, hard to keep one for five. A falling 13th-month number is an early warning that the float is leaking before it ever gets a chance to compound.

This is also where the new-money lens comes in. VNB — value of new business — is the embedded value created by this year's fresh policies. It answers: is the company still feeding the pile profitably, or just running off old business? An insurer with a shrinking VNB is a back room that's slowly emptying, however fat the till looks today.

Reading management: where the float story meets the guidance story

Because so much of an insurer's value lives in future assumptions — how long customers stay, what the float earns, how fast new business grows — management's commentary matters more here than in almost any other sector. The numbers you're trusting are partly their projections. So the question shifts from "what did they earn?" to "do they do what they said they'd do?"

This is exactly where tracking guidance over time earns its keep. HDFC Life, for instance, guided in its Q3 FY25 commentary for VNB growth "upwards of 15%" for the year — a target our Promise Tracker records as missed (Inve data, 2026). Separately, management's Q4 FY25 framing of doubling VNB "every 4-4.5 years" (an implied ~16-17% CAGR) was, by the Q4 FY26 call, revised down — management guided a normalised three-year VNB CAGR of around 9-10%, a clear walk-back from the doubling target (Inve data, 2026). Neither is a scandal — VNB is sensitive to interest rates and product mix, and managements are allowed to recalibrate. But for an owner whose valuation rests on the growth of the pile, a guided growth figure that gets quietly walked back is the single most important thing to catch. (None of this is a view on the share; it's a worked example of what to read.)

The same four lenses apply to every listed life insurer worth studying — LIC, SBI Life and ICICI Prudential Life all live or die by the same float, persistency and embedded-value mechanics.

That's the practical edge. You don't value an insurer off its P/E. You value it off the float, the persistency that protects it, the embedded value that prices it, and the credibility of the management telling you how fast it will grow — and you check that last one against what they actually delivered, quarter after quarter, not against the press-release headline. If you hold several financials, doing that by hand across a portfolio is the real grind; reading the concall summaries is how owners keep up without reading every transcript.

Where this lens can mislead you

One honest caveat, because the float story can be seductive. Float is only an asset if it's cheap and invested well. An insurer that wins customers by underpricing risk, or that earns poor returns on its pile, has manufactured a liability dressed as an asset — Buffett is blunt that bad underwriting makes float costly, not free. And embedded value is built on assumptions; change the assumed investment return or persistency and the number moves. So don't treat EV as gospel — treat it as management's best estimate, and ask whether their past estimates held up. We haven't modelled HDFC Life's investment book or its assumption changes here; that's the next layer of work, not something one article settles.

The point of this piece is narrower and more durable: stop reading a life insurer through a product-company's lens. The till is not the business. The back room is.

Frequently asked questions

Related reading

• How a Bank Makes Money: CASA, NIM and Bad Loans
• Order Book: How a Capital-Goods Business Earns
• How a Cement Company Makes Money in India