How an Insurer Earns, and What the Premium Does Not Say

An insurer earns in three separate places. The underwriting result, being the premium left standing once claims and the price of running the business have gone. The return on money it sits on across the stretch between a premium coming in and a claim going out. And on some contracts, charges taken for administering them. Premium is money received against an obligation rather than money earned, so the premium line shows none of the three.

Why is an insurer harder to read than a lender?

Put two reported results side by side. Suvarna Commercial Bank Limited, an invented commercial bank, earned interest of Rs 18,600 crore in a stated year and paid out interest of Rs 11,160 crore. Net interest income was Rs 7,440 crore. Both of those numbers belong to the same twelve months. The money came in that year, the money went out that year, and the subtraction between them is a subtraction between two things that actually met.

Now take Chandrika Life Insurance Limited, an invented life insurer. Being a life insurer is the distinction that decides everything that follows. Chandrika Life Insurance received total premium of Rs 15,600 crore in its stated year. Some of the obligations that premium bought will fall due next year. Some will fall due in twenty years. Some may fall due in forty. The largest figure an insurer publishes is a gross inflow, and the thing it bought with that inflow is not on the same line, not in the same period, and on a life book quite possibly not in the same decade.

The distance between an inflow and the obligation it bought is the whole difficulty, and every question that follows is a version of it. Where does the money come from, what does the reported result actually contain, and how far apart in time are the two halves of the trade.

Where does an insurer's money actually come from?

How Insurers Make Money

The shape is the same in a wedding hall, and nobody needs a textbook for a wedding hall. Think about one for a moment. A hall takes a booking eighteen months out and collects the whole amount on the day the booking is made. Three quite different things then happen to that money. Part of it will be consumed by the actual cost of giving somebody a hall for a day, and what survives that is the hall's real result on the booking. Part of it sits in a bank for eighteen months and earns something while it waits. And the hall charges a separate fee for the paperwork and the coordination. The fee stands whether the function is lavish or plain.

An insurer has exactly those three, and the reason to keep them apart is sharper here than anywhere else in the way financial institutions are read.

First, the underwriting result. This is what is left of the premium once the claims have been met, once the cost of getting the business and running it has been paid, and once whatever had to be added to the holding against future claims has been added. The underwriting result is the result of the actual insurance trade: a price was quoted against a risk, and this is what the trade produced.

Second, the return on the money held between premium and claim. Premium arrives first and claims fall due later, so there is a pool sitting in the middle. The pool in the middle is invested. On a short contract the gap is months and the return is a footnote. On a life book the gap is measured in decades, and the return on the pool can be larger than the underwriting result it sits beside.

Third, on some contracts, charges taken for administering the arrangement. These are amounts deducted for running a contract rather than for carrying its risk. The charges behave like a fee on a balance, which is a different animal again. Whether the claims came in above or below the price makes no difference to them.

Why does the separation matter more here than for a lender? Because the three move in different directions when conditions change. A stretch of unusually heavy claims hits the first and leaves the third untouched. A move in what the pool earns hits the second and leaves the first untouched. Adding the three together before looking at them destroys the only information that shows what kind of year the insurer actually had. It is also why a manager of somebody else's money, such as Vaidehi Asset Managers Limited with its Rs 1,80,000 crore under management and a blended charge of 0.55 per cent a year producing revenue of Rs 990 crore, looks superficially like the third source and is a genuinely different business. The comparison between an insurer's administration charges and a manager's fee on somebody else's money is worked out separately.

What is the difference between gross and net premium, and where does it bite?

Gross vs Net Premium

A contractor wins a job worth Rs 40,00,000/- and sublets the tiling for Rs 4,00,000/-. His order book says forty lakh. The work he is actually on the hook for is thirty six lakh. Both figures are true, and both describe something real. The moment he divides his own costs by the forty lakh figure, he has made himself look better than he is by exactly the work he handed to somebody else.

An insurer does the same thing with risk. Gross premium is everything it took in. Net premium is whatever survives once the premium cededPassed on to a reinsurer along with the corresponding slice of the risk, in exchange for a share of the premium. How that arrangement is actually structured is worked out separately. to a reinsurer has gone out of the door. Chandrika Life Insurance received total premium of Rs 15,600 crore. Suppose, purely as a supposition, that one rupee in every ten went out as ceded premium. That is Rs 1,560 crore leaving, and net premium of Rs 14,040 crore. No figure for premium ceded by this insurer is on record, so the cession is called a supposition wherever it turns up below.

Now the part that catches people who already know the definition. The difference bites in three separate places, and only the third one does real damage.

The difference bites on the premium line, where the gross figure describes the business the insurer wrote and the net figure describes what it kept for its own account. Neither is more correct. The two figures answer different questions.

The difference bites on the claims line as well. The same split exists there. If a tenth of the premium went out, a corresponding share of the claims comes back from the reinsurer. On the supposition above, Rs 672 crore of the Rs 6,720 crore of claims is recovered, leaving Rs 6,048 crore net.

And the difference bites on every ratio built from those two lines. The damage is done there. Divide net into net, Rs 6,048 crore by Rs 14,040 crore, and the reading is 43.08 per cent of net premium. Divide gross into gross, Rs 6,720 crore by Rs 15,600 crore, and the reading is 43.08 per cent of total premium received. The same proportion walked off both lines, so the two readings are identical. Now mix them. Take Rs 6,048 crore of net claims and divide it by Rs 15,600 crore of total premium received. Out comes 38.77 per cent of total premium received, and the insurer looks 4.31 percentage points better than it is. Nothing improved. The improvement is exactly the risk it passed to somebody else, counted on the top line and not on the bottom.

What does the premium account show once claims and running costs come out?

Here is the subtraction, worked rather than described. Chandrika Life Insurance Limited, a life insurer, one stated year. Total premium received was Rs 15,600 crore. Of that, Rs 5,200 crore came in as new business premium and Rs 10,400 crore as renewal premium, so renewal ran at exactly twice new business. Claims met came to Rs 6,720 crore. Getting and running the arrangement cost Rs 2,496 crore.

Claims come to 43.08 per cent of total premium received. Expenses come to 16.00 per cent of that same base. Added, the pair take 59.08 per cent of it, leaving Rs 6,384 crore standing behind them at 40.92 per cent of total premium received. Every one of those four percentages carries its base inside the same sentence, and that is not fussiness. Total premium received, net premium, earned premiumPremium matched to cover the insurer has already provided, as distinct from premium taken in for cover still ahead. Working with it is covered under the underwriting ratios. and policyholder funds are four different denominators, and the same rupee amount of claims produces four different percentages over them.

Why is what is left over not the insurer's profit?

A school collects a full year of fees in April. On the last day of April the bank balance is magnificent and the school has taught for four weeks. The obligation that balance was collected against is eleven months from being discharged, so nobody would call the balance the year's profit. Everybody understands this instantly about a school. Almost nobody applies it to an insurer, and the reason is that an insurer's version of the obligation is not visible on the same line.

The subtraction in the table above is missing one deduction, and on a life book it is the biggest item on the account. The missing deduction is the amount that had to be added during the year to the reserveMoney set aside now for claims nobody has made yet, which the contracts already written will eventually throw up. How it is measured and reported is covered separately. held against claims not yet made. A policy sold this year takes premium now and creates an obligation that may not fall due for four decades, and the money to meet it has to be carved out of that same premium rather than found somewhere else later on. An actuaryThe person whose work it is to put a number on obligations that have not fallen due yet, using assumptions about how long they will run and how often they will arise. works out how much, and the answer feeds straight into the account as a deduction.

So the Rs 6,384 crore is premium that has not yet met its obligation. The residual is not profit that has been earned. The subtraction that produced it is correct in every step. The name on the answer is what is wrong.

And the profit for the year cannot be worked out from what is on record. No reserve movement for Chandrika Life Insurance is on record. The row in the table above says exactly that, and the drawing below leaves the space physically empty. The Insurance Regulatory and Development Authority of India (IRDAI) settles how that reserve is valued and reported, at irdai.gov.in, and the setting shifts.

How does underwriting quality show up in the numbers, and when?

How Underwriting Quality Affects Insurer Economics

Picture a builder who quotes cheaply for a foundation. The house goes up, the customer is delighted, and the crack appears in the fourth monsoon. By then he has quoted the same way on six more houses. Nothing in the first three years told him to stop. The crack in the fourth monsoon is the whole of underwriting quality in one image, and three distinct effects have to be kept apart.

The first is the lag. A book written this year is priced on an estimate of what the claims will be. The claims that test the estimate arrive over the years that follow. So the quality of a pricing decision becomes visible long after the decision was taken, and long after the same estimate has been used to price the next year's book and the year after that.

The second is the double effect, and it is the one that fools people. A book priced too low is attractive, so more of it gets written than should have been. Writing more of an underpriced book does two things at the same moment. The extra business enlarges the obligation the insurer has taken on, and it enlarges the money held against that obligation. Premium goes up. Policyholder funds go up. The insurer gets visibly bigger at exactly the moment it is getting worse, and every size measure on its published result moves in the flattering direction.

The third is compounding on a contract whose price cannot be reset. On an arrangement that renews every year, a pricing error is corrected at the next renewal and the damage is one year deep. On a long contract the price is fixed at the start, so an error made once is carried for the whole term. The contrast between a life book and a short book is not a difference of degree. The difference is in how long a mistake lives.

So what can an outside reader infer about underwriting quality from a single year of an insurer's figures? Very little, and that is not a hedge, it is the arithmetic. A larger premium line, a larger claims line and a larger pool of policyholder funds are all consistent with a book that grew because the product was good and with a book that grew because the price was wrong. The two look identical for as long as the lag lasts.

Two things are needed instead, and neither is in the figures above. The first is claims sorted by the year in which the policies producing them were written. Sorted that way, each year's pricing decision can be judged against the claims it actually threw up rather than against everybody's. The second is a reading of how much of the earlier book is still in forceThe condition of a contract that is still running and still providing cover, as distinct from one that has ended, matured or lapsed.. A book quietly draining away tells a very different story from one of the same size that is staying put. Neither sorting exists in this record, and disclosure about policies staying in force, across whichever stretches it is measured on, belongs to IRDAI at irdai.gov.in.

What does the money held contribute, and who does it belong to?

Chandrika Life Insurance is holding Rs 72,000 crore of policyholder funds. Its own net worthWhat the insurer itself is left with once everything it owes has been taken off what it holds. Its own capital, as against the money it is holding for other people. is Rs 7,200 crore, exactly a tenth of that. Add the two and Rs 79,200 crore is at work as invested assetsEverything the insurer has put to work, taken together, whether the money underneath it is owed forward to policyholders or belongs to the insurer itself..

No rate is put on that money here, anywhere, at any point. No investment return for Chandrika Life Insurance is on record. The arithmetic is therefore done per percentage point instead, a division rather than an expectation.

One percentage point a year on Rs 79,200 crore is Rs 792 crore. Split it by pool, and the split is the point: Rs 720 crore of that arises on the Rs 72,000 crore of policyholder money and Rs 72 crore on the insurer's own Rs 7,200 crore. The pools themselves are ten to one, so the two amounts stay ten to one at every setting.

Setting one percentage point against the underwriting residual is the fact that reorganises a reader's picture of the business. Rs 792 crore a percentage point against Rs 6,384 crore of premium standing means that a little over eight points on the money held would produce, on its own, an amount the size of everything the underwriting side has left standing after claims and costs. On a book of long contracts the money held is not a sideline to the insurance. The money held is half the business.

Which raises the question the arithmetic cannot answer. The Rs 720 crore arose on money that is owed to policyholders. Nine rupees in every ten of what the invested pool produces arises on money that is not the insurer's. How any surplusWhat arises when the amount an insurer set aside for an obligation turns out to be more than the obligation actually needed. gets shared between policyholders and shareholders belongs to IRDAI, published at irdai.gov.in, and the answer shifts. The drawing above carries an empty row exactly where that split would sit, and the simulation below carries the same empty row, at every setting, for the same reason.

At the default setting the drawing reproduces the arithmetic in the paragraphs above exactly: one percentage point a year on invested assets of Rs 79,200 crore gives Rs 792 crore arising in all, being Rs 720 crore on the Rs 72,000 crore of policyholder money and Rs 72 crore on the Rs 7,200 crore of the insurer's own, against a reference line at Rs 6,384 crore. Push the control to five points and the three readings become Rs 3,960 crore, Rs 3,600 crore and Rs 360 crore. Push it to eight and the total is Rs 6,336 crore, just short of the line. The bars pass the line at 8.1 points on this control, at Rs 6,415.20 crore. The pools underneath are ten to one, so the two parts stay in that same proportion the whole way up, and no setting of the rate changes it.

What can somebody outside the insurer see, and what can they not?

A published set of an insurer's results sorts into what it tells a reader and what it does not, and that sorting is the most portable finding on the subject.

So an insurer read on its premium line alone has been read on the one number that is a gross inflow rather than an earning. That is the sentence to carry away. Everything else here is the working behind it.