Tools

Settlement When a Loss Outlasts the Indemnity Period

A business interruption policy stops paying at the maximum indemnity period, and reduces what it does pay in proportion if the sum insured was short of what the loss basis required. If the recovery takes two years and the cover runs for one, how much of the loss is actually inside the policy — and which of the four reductions is the one that cost you?

Every business interruption calculator answers the question you ask before the loss: how much cover should I buy. This one answers the question you ask afterwards, which is a different question and has no calculator. Two things happen to a long interruption and they compose. The loss stops being covered at the maximum indemnity period, so a recovery that outlasts the cover has a tail that is outside the policy whatever the sum insured says. And what remains is then reduced in proportion if the declaration was short of what the loss basis required — which it can be even when it was worked out carefully, if it was worked out on a narrower definition of gross profit than the loss is measured on. The reassuring half is that the tail is smaller than the calendar suggests: an interruption is worst at the start, and the cover is the start.

What this tool does not decide

How much cover you should buy. That is the forward question and it has many calculators. This one is about a loss that has already happened and the policy you already have, and the two arithmetics are not the same.
What your policy says. Whether it contains a condition of average at all, how its uplift is worded, what its definition of gross profit includes and whether any part of the loss is excluded are questions about your policy document. This page reads no wording and names no insurer.
What you will actually be paid. An insurer settles on that document and on an adjuster's reading of your accounts. If the figure here and the figure you are offered disagree, the reason is in those two places and this page cannot tell you which.
How long your recovery takes. That is your own experience or your own estimate. This page forecasts nothing, and the two run-off shapes it offers are elementary shapes rather than profiles it publishes.
Whether your accounts support the gross profit figure. Gross profit in the insurance sense is turnover less the costs that stop when trading stops, and what an adjuster accepts is a matter for your accounts.
Anything about tax, in any jurisdiction, at any time.

Your figures

The interactive calculation runs in your browser and needs local JavaScript, which is currently switched off. Activating the Calculate button does nothing at all: no request is sent and nothing is submitted.

Everything else on this page works without it. The model, the formulas, the timing convention, the worked example, the assumptions and the limitations are all in the methodology below, and the companion explainer covers the mechanism in full.

Your policy and your loss, in eight numbers

All eight figures are yours and this page publishes none of them. The sum insured, the indemnity period, the uplift and the basis your declaration was made on are read off your own policy schedule; the rate of gross profit, the length of the recovery and how much trading you lost in the first month are your own accounts and your own experience. This page names no insurer, holds no policy wording, assumes no uplift and forecasts no recovery — which is what makes it correct wherever this mechanism exists rather than correct under one wording, and why nothing on it can go out of date.

From the policy schedule

The sum insured, as declared

The figure on the schedule. Enter 0 if nothing was declared — that is a real case and the page will show you exactly what average does to it. There is no currency here and no field for one: every amount on this page comes back in the unit you enter the gross profit in.

The maximum indemnity period

A whole number of months, from the schedule. Twelve is the common case and is deliberately not prefilled here: this page is about what happens when a recovery does not fit inside the period you actually have, and putting an insurer's ordinary term in the box would be publishing it as though UBWHY had established it.

The declaration-linked uplift the policy grants

Enter 0 if the policy grants none. Above zero it does two separate things, and this page publishes both rather than folding them together: it removes average entirely, so a short declaration costs you nothing in proportion, and it raises the ceiling from the sum insured to the sum insured plus the uplift. Zero is one of the two cases and not a lesser one — it is the case in which average applies.

From your own accounts and your own experience

The rate of gross profit a year at the time of the loss

Gross profit in the insurance sense — turnover less the costs that stop when trading stops — at the rate it was running when the interruption began. It is the basis the loss is measured against and the basis the required sum insured is measured against, which is why it is the one figure here that cannot be zero. Every amount on this page comes back in the unit you enter this in.

How long the business takes to get back to normal

A whole number of months. Your own experience if the recovery is over, your own estimate if it is not — this page forecasts nothing and has no recovery profile of its own. A recovery shorter than the indemnity period is a valid answer and is the case in which nothing is truncated at all.

How much of normal trading was lost in the first month

A total shutdown is 100%. A business trading at reduced volume is less. This is the depth of the interruption at its worst, and the shape below says what happens to it after that.

How the shortfall runs off over the recovery

Two elementary shapes and no third. Tapering is the ordinary shape of a recovery — worst at the start, easing steadily — and it is the shape on which the covered share of the loss exceeds the covered share of the time. Level is the boundary case, on which the two shares are the same figure by construction. Neither is a forecast and neither is what any particular business does.

Result

A loss that outlasts the maximum indemnity period is cut off at that boundary, and the part beyond it is uncovered whatever the sum insured says. The arithmetic almost everybody does next is to divide the months — twelve of cover against twenty-four of recovery, so half the loss. That is wrong, and it is wrong in the direction that surprises people: an interruption is front-loaded, because the business is trading at its worst at the start and better every month afterwards. On a shortfall that tapers, the first twelve months of a twenty-four month recovery carry about three-quarters of the loss and not one-half. This page publishes both shares and the gap between them, because the gap is the part a reader cannot get to unaided and the part that changes what they think their policy did.

Four separate things reduce what a policy pays, and they compose rather than compete. The loss beyond the indemnity period is simply outside the cover. What remains is then reduced in proportion if the declaration was short of what the loss basis required — that is average, and it bites in the same proportion on a small claim as on a total one. What survives that is capped at the sum insured. And a declaration-linked uplift, if the policy grants one, removes the proportional reduction entirely and raises the cap at the same time. This page publishes all four separately and says which one actually decided the answer, because the four have completely different lessons: one of them means declaring more would have helped twice, one means it would not have helped at all, and a single reduced figure tells a reader neither.

Enter your eight figures above and select Calculate. Nothing is sent anywhere: the calculation runs in this browser, and no value is stored, shared or placed in the address bar.

The share of the loss that is covered, the share of the time that is covered, and the gap between them

Three rows and never one. The share of the loss inside the indemnity period is published beside the share of the recovery it corresponds to, because they are different figures and reading the second for the first understates what your policy covers.
Which share	What it comes to
The share of the recovery the indemnity period covers	Not yet calculated
The share of the loss that falls inside it	Not yet calculated
The difference, which is what front-loading is worth to you	Not yet calculated

The four things that reduce a settlement, and which of them bit here

Four rows and never one. Everything else that computes this reports the four as a single reduced figure, and the four have four different lessons — one of them means declaring more would have helped twice, and one means it would not have helped at all.
The reduction	What it comes to	What it did
The tail beyond that boundary, outside the cover whatever the sum insured	Not yet calculated	Not yet calculated
The proportion average leaves you with	Not yet calculated	Not yet calculated
The most the policy can pay	Not yet calculated	Not yet calculated
Is the policy declaration-linked?	Not yet calculated	Not yet calculated

The figures

The normal rate of gross profit in a month: Not yet calculated
Everything the interruption cost, over the whole recovery: Not yet calculated
The part of it falling at or before the maximum indemnity period: Not yet calculated
The tail beyond that boundary, outside the cover whatever the sum insured: Not yet calculated
What should have been declared, measured on the loss basis: Not yet calculated
The proportion average takes away: Not yet calculated
The settlement before the limit is applied: Not yet calculated
What the policy pays: Not yet calculated
What the business carries itself: Not yet calculated
Which of the four actually decided it: Not yet calculated

Every amount here is in the same unit as the figures you entered, and this page never names one. Months are whole months from the start of the interruption, and there is no calendar, no date and no jurisdiction anywhere in the arithmetic. The monthly rate of gross profit is a twelfth of the annual figure you entered, evenly, which is the convention the whole calculation rests on and is a simplification about seasonality rather than about insurance.

This is arithmetic on eight figures you described, not a reading of any policy and not advice. It applies one proportional condition of average to one loss over one recovery. It does not read your wording: whether your policy has a condition of average at all, whether its uplift works the way this page models it, what its definition of gross profit includes, and whether any of the loss is excluded for a reason nothing here knows about are all questions about your policy document. Your insurer’s settlement will be based on that document and on an adjuster’s assessment of your accounts, not on this page. The two elementary run-off shapes offered here are shapes, not forecasts, and neither is what any particular business does.

This is not what you will be paid. It is what the mechanism produces on the figures you entered, which is a different thing from what an insurer will settle after reading your wording and your accounts. Nothing here is advice about cover, about a claim, about an insurer or about a broker, and there is no field on this page that could carry a recommendation about any of them. If the figure here and the figure you are offered disagree, the reason is in your policy document and in your accounts, and this page cannot tell you which.

How this is worked out

Why the share of the loss is not the share of the time

A maximum indemnity period is a boundary in time: the policy pays for losses arising in the months at or before it, and nothing after. The obvious next step is to divide the months — twelve of cover against twenty-four of recovery, so half the loss — and it is wrong.An interruption is front-loaded. The business trades at its worst in the first month and better every month afterwards, so the early part of a recovery carries far more than its share of the loss — and the cover is the early part. On a shortfall that tapers in a straight line, the first twelve months of a twenty-four month recovery carry about three-quarters of the loss and not one-half. That is a property of the shape rather than of any policy, and it is the one direction of error people do not expect: reasoning from the calendar makes readers write off money they were entitled to.

Three readings and three different errors. Two of them understate what the policy covers and one overstates it, which is why a page that corrected only one of them would leave a reader wrong in the other direction.
The reading	What it is	Why
Twelve months of cover against twenty-four of recovery, so half the loss	Understates it, usually badly	The loss is front-loaded. On a tapering shortfall the first half of a recovery carries about three-quarters of the loss, because the business is trading at its worst at the start and the cover is at the start
The sum insured is what the policy pays, so the settlement is that	Overstates it whenever the declaration was short	The sum insured is a ceiling, not a payment. What is paid is the loss inside the boundary, reduced in proportion if the declaration was short of what the loss basis required, and only then capped
The declaration was made carefully, so average will not apply	True only if the two bases match	A declaration made on a narrower definition of gross profit than the loss is measured against is short by the ratio between them, however carefully the narrower figure was worked out. That is the wages-in-or-out problem and it is invisible until a claim

The arithmetic

Write G for the annual rate of gross profit, I for the sum insured,M for the maximum indemnity period, R for the recovery, s for the share of normal trading lost in the first month, u for the declaration-linked uplift and d for the share of the loss basis the declaration was made on:

shortfall(k)  = (G ÷ 12) × s × (R − k + 1) ÷ R     tapering
              = (G ÷ 12) × s                       level

total loss    = sum of shortfall(k) for k = 1 … R
covered loss  = sum of shortfall(k) for k = 1 … min(M, R)
uncovered     = total loss − covered loss

required      = G × M ÷ 12 ÷ d
average       = 1                     if u > 0
              = min(1, I ÷ required)  otherwise
limit         = I × (1 + u)

settlement    = min( covered loss × average, limit )

Four reductions in five lines, and they compose rather than compete. The last line is what a policy pays; the four lines above it are why, and everything found that computes this reports only the last one.

The four things that reduce a settlement

Four reductions and four different lessons. The last column is why they are published separately: knowing the settlement was reduced tells you nothing about whether declaring more would have helped, and the four answers to that question are opposite.
The reduction	What it does	What it means for next time
The loss beyond the maximum indemnity period	Cut off at the boundary. Outside the policy whatever the sum insured says	Declaring more would not have recovered a unit of it. The boundary is a length of time, and a larger number does not extend it
Average, for a declaration short of the loss basis	Reduces what fell inside the boundary, in proportion	The same proportion comes off a small claim as off a total one. This is the one still worth acting on, and the required sum insured on the loss basis is published so the gap can be seen
The limit the sum insured imposes	Caps whatever survives average	Where this bound, declaring more would have helped twice — once against average and once against the cap. It is easily mistaken for the indemnity period cutting you off, and it is a different failure with a different remedy
A declaration-linked uplift, where the policy grants one	Removes average entirely and raises the ceiling, both	Two effects and not one. A page that folded them together would leave you unable to see which of them saved you, and on a badly under-declared policy the first is worth far more than the second

Why the required sum insured is measured on the loss basis

Average compares what was declared against what should have been declared, and "should have been" is measured on the definition of gross profit the loss is measured on. If a declaration was made on a narrower definition — leaving out a cost the loss basis includes, which is usually wages — then the required figure is larger than the one the declaration was worked out against, by exactly the ratio between the two bases.A reader can therefore be under-insured on a declaration they calculated carefully and correctly, on a basis nobody told them was the wrong one. This page asks for that ratio rather than for a wages amount, because it has no view on what any policy's definition includes and is not entitled to one.

Why the tapering shape never reaches zero

The obvious way to write a straight-line run-off is one that arrives at nothing in the final month of the recovery. It is wrong about the definition rather than about the arithmetic:a shortfall of zero in the last month says the business had fully recovered a month before it had, which is the opposite of what a recovery period means. The form used here runs from the full initial rate in the first month to one Rth of it in the last, is defined for a recovery of a single month, and never reaches zero inside the recovery.

What this page is not

It is not the forward question. Every business interruption calculator this Asset was qualified against answers how much cover should I buy, and this one deliberately does not — it answers how much of a loss that has already happened is inside the policy I already have. It is also not a settlement offer: an insurer will settle on your wording and an adjuster's reading of your accounts, and if that figure and this one disagree the reason is in those two documents and this page cannot tell you which.

The limits this calculation imposes on itself

Every figure is yours. This page publishes no sum insured, no indemnity period, no uplift and no recovery profile, and there is no field that could carry one. Nothing on it can go out of date because nothing on it is anybody’s published parameter.
It does not read your wording. Whether your policy contains a condition of average at all, how its uplift is worded, what its definition of gross profit includes, and whether any part of the loss is excluded for a reason nothing here knows about are all questions about your policy document.
The two run-off shapes are shapes and not forecasts. Neither is what any particular business does, and the page has no third — offering one would be publishing a recovery profile this site has no authority to publish.
The monthly rate of gross profit is a twelfth of the annual figure, evenly. That is a simplification about seasonality rather than about insurance, and a business whose interruption fell across its busiest quarter has a loss this shape does not describe.
This is not the forward question. Every calculator this Asset was qualified against answers how much cover to buy, and this one deliberately does not — it answers how much of a loss that has already happened is inside the policy you already have.

What is not modelled

any insurer, policy wording, broker, jurisdiction, statute or date
whether your policy contains a condition of average, or how its uplift is worded
what your policy’s definition of gross profit includes — you supply the share, not the definition
any exclusion, deductible, time excess, waiting period or additional increased cost of working
any premium, any probability and any view on how much cover to buy
what an adjuster would accept as the rate of gross profit or as the length of the recovery
seasonality — the monthly rate is a twelfth of the annual figure, evenly
more than one loss, and more than one policy

The distinction the second half of this page depends on — that a figure on a document is not money arriving when the shortfall does — is taught in the explainer:liquidity is not the same as net worth. That explainer is about a household's own balance sheet and this page is about a policy; what they share is that a sum insured on a schedule is a promise about a ceiling, and what a business carries month by month while trading badly is a different quantity entirely.

Go deeper

Whether average bites only on a catastrophe
Settlement share across underinsurance
Six levels of cover against five loss sizes in one reference table: the share of a claim that is paid, the share the holder bears, and why both are the same at every loss size.
This page applies average to one loss. That reference answers the question it raises and cannot settle: the proportion average takes is the same for a small claim as for a total one, so a short declaration is a permanent cost rather than a catastrophe-only one — which is what decides whether raising it is worth doing.
Read the reference: Settlement share across underinsurance
What a figure on a schedule is and is not
Why Liquidity Is Not the Same as Net Worth
Why an asset can add to your net worth and still be no help in an emergency, what access time and sale friction actually cost, why a usable share is your assumption rather than a valuation, and why holding more cash has a price of its own.
A sum insured is a promise about a ceiling. What a business carries month by month while trading badly is a different quantity, and the difference between having value and having money when it is needed is what that explainer is about.
Read the explainer: Why Liquidity Is Not the Same as Net Worth

Calculation model and corrections

Calculation model: Indemnity Period Truncation v1.0
Last reviewed: 29 August 2026
Why the share of the loss is not the share of the time: Because an interruption is front-loaded. The business trades at its worst in the first month and better every month afterwards, so the early part of a recovery carries more than its share of the loss — and the cover is the early part. Dividing the months understates what is covered, which is the one direction of error people do not expect
The four things that reduce a settlement: The loss beyond the indemnity period, which is outside the cover whatever the sum insured; average, which reduces what remains in proportion if the declaration was short of what the loss basis required; the limit, which caps what survives that; and a declaration-linked uplift, which removes average entirely and raises the limit at the same time. They compose. This page publishes all four and says which one decided the answer
Whose figures these are: Yours, all eight of them. This page publishes no sum insured, no indemnity period, no uplift and no recovery profile, and there is no field that could carry one. No insurer is named, no wording is read, and nothing here can go out of date because nothing here is anybody’s published parameter
What the required sum insured is measured on: The loss basis, not the basis your declaration was made on. That is the whole of the wages-in-or-out problem in one line: a declaration made on a narrower definition of gross profit than the loss is measured against is short by the ratio between them, and average bites on a declaration you believe was correct
What a declaration-linked uplift does: Two separate things, and they are published separately. It removes average entirely — not partially, entirely — and it raises the ceiling from the sum insured to the sum insured plus the uplift. Folding the two together would leave a reader unable to see which of them saved them
The two run-off shapes, and what they are: Elementary shapes rather than forecasts. The tapering one falls in a straight line from the first month to the last and never reaches zero inside the recovery — a profile that reached zero at the final month would say the business had fully recovered a month before it had. The level one runs at the same rate throughout and then stops. Neither is what any particular business does, and the page has no third
The monthly rate: A twelfth of the annual rate of gross profit you entered, evenly. That is a simplification about seasonality and not about insurance, and it is the convention every figure on this page rests on. A business whose interruption fell across its busiest quarter has a loss this shape does not describe
What this is not: Not the forward question. Every calculator this Asset was qualified against answers how much cover to buy; this one answers how much of a loss that has already happened is inside the policy you already have. The two are different questions and the second has no published surface
The unit: Whatever unit you entered the gross profit in, throughout. This page names no currency and there is no field for one. Months are whole months from the start of the interruption, and there is no calendar, no date and no jurisdiction anywhere in the arithmetic
Excluded: Every insurer, wording, broker, jurisdiction and date; every exclusion, deductible, time excess and waiting period; additional increased cost of working; any premium and any probability; seasonality; and any view on how much cover to buy
Rounding: Display only; intermediate values remain unrounded

Correction history

The tapering profile was first written to reach zero at the final month of the recovery, which is the obvious way to write a straight line down to nothing. It is wrong, and it is wrong about the definition rather than about the arithmetic: a shortfall of zero in the last month of a recovery says the business had fully recovered one month before it had, which is the opposite of what a recovery period means. The discrete form used instead runs from the initial rate in the first month to one Rth of it in the last, is defined at a recovery of a single month, and never reaches zero inside the recovery.
The corpus classified this candidate as requiring a new model rather than an extension, and the classification was tested rather than accepted. It held: truncation at an indemnity boundary followed by proportional settlement is not the shape of any locked model on this site, and the published under-insurance settlement identity applies average with no time boundary at all. A sibling candidate in the same batch was classified an extension of an existing model, tested the same way, and did not hold — it would have published a quantity of zero on every input with nothing failing. That is the reason both classifications were tested rather than trusted.