First-Loss Sum Insured in Burglary Insurance: When Partial Cover Is the Right Call

Consider a distribution warehouse holding Rs 100 crore of packaged consumer goods at peak. If it insures that stock against fire, the full value is the right sum insured, because a fire can take the entire building and everything in it. If it insures the same stock against burglary at the same Rs 100 crore, it is buying cover it can never fully use.

A burglar cannot remove Rs 100 crore of packaged goods. A theft is bounded by physical reality: how the intruders get in, how much they can carry or load, how long they have before detection, and what a receiver will actually take. In a large stockholding, the realistic maximum a single burglary can remove is a fraction of the total, perhaps a few percent, not the whole. Insuring the full value against a peril that can only ever remove a slice means paying premium on a large tranche of sum insured that no burglary claim will ever reach.

This is the gap that first-loss insurance was designed for. Rather than insure the full value at risk against a peril that removes only part of it, the buyer insures the realistic maximum loss from a single event, and pays for that. Done properly it is not under-insurance and it is not a cut corner. It is matching the sum insured to the shape of the peril, which is what an efficient burglary programme does and a lazy one does not.

A first-loss policy separates two numbers that a conventional policy holds together: the total value of the property, and the amount the insurer will pay for any one loss.

The full value at risk is the actual total value of the insured stock or property, Rs 100 crore in the example. The first-loss sum insured, sometimes called the first-loss limit, is the maximum the insurer will pay for a single loss, set deliberately below the full value as a chosen percentage of it. If the buyer judges that no single burglary can remove more than Rs 10 crore, it sets a first-loss limit of Rs 10 crore, which is 10 percent of the full value at risk.

The name captures the mechanism: the insurer covers the first slice of loss, up to the chosen limit, and the buyer accepts that a loss beyond the limit (which it believes cannot happen from one event) is its own. The insurer still needs to know the full value at risk, because that figure governs how the cover is rated and how average is applied, but the payout is capped at the first-loss figure. So the buyer holds full protection against the loss it can actually suffer, and stops paying for a layer of sum insured that the peril cannot reach. The distinction between the full value at risk and the first-loss limit is the whole of the structure, and misunderstanding it is where buyers get into trouble at claim.

The instinctive assumption is that a first-loss limit of 10 percent of full value should cost 10 percent of the full-value premium. It does not, and the reason is the same reason first-loss cover makes sense at all.

The first-loss limit sits over the most exposed part of the risk. It is the layer that every realistic burglary hits, because every realistic burglary produces a loss inside that band. The upper layers of sum insured that a full-value policy also carries are almost never touched, so they cost very little per rupee of cover. When the buyer keeps only the first-loss layer, it keeps the frequently exposed part and gives up the rarely exposed part, so the premium falls by much less than the sum insured does.

Insurers price this with a first-loss scale: a table that expresses the first-loss premium as a percentage of the full-value premium for each first-loss percentage of value. The scale is always steeper than a straight line. A first-loss limit of 10 percent of value might be rated at, for illustration, something like 40 to 50 percent of the full-value premium, and a limit of 25 percent of value at a still higher share. The exact scale is the insurer's, and it is negotiable, but the shape is fixed: the smaller the first-loss percentage, the greater the premium per rupee of cover, because the retained layer is the busy one. Understanding this stops a buyer from expecting a 90 percent premium saving from a 90 percent cut in sum insured, and it stops an insurer from quoting first-loss cover at a pro-rata rate that fails to reflect where the exposure sits.

Take the warehouse holding Rs 100 crore of stock, and assume for illustration a full-value burglary rate that produces a full-value premium of Rs 10 lakh (a rate of 0.10 percent on the Rs 100 crore). These figures are illustrative; real rates depend on the risk, the protections and the market.

The risk assessment says that given the access points, the load-out time before the alarm and patrol respond, and what the stock is, the largest credible single theft is around Rs 8 to 10 crore. The buyer sets a first-loss limit of Rs 10 crore, which is 10 percent of the full value.

1. Full-value cover: sum insured Rs 100 crore, premium Rs 10 lakh, of which the vast majority pays for a sum insured no burglary will reach.
2. First-loss cover at Rs 10 crore: the premium is not 10 percent of Rs 10 lakh (Rs 1 lakh). Applying an illustrative first-loss scale of, say, 45 percent, it is 45 percent of Rs 10 lakh, or Rs 4.5 lakh.
3. The trade: the buyer pays Rs 4.5 lakh instead of Rs 10 lakh, saving Rs 5.5 lakh, and retains full cover for any burglary up to Rs 10 crore, which is the largest loss it believes a single event can produce.

The saving is real and the cover is intact for the loss that can actually happen. What the buyer has given up is indemnity for a loss above Rs 10 crore from a single burglary, which its own risk assessment says cannot occur. The quality of that risk assessment is the whole basis of the structure. Set the first-loss limit sensibly above the credible maximum single theft, and first-loss cover is efficient. Set it below the loss a determined, well-organised theft could actually achieve, and the saving becomes a retained exposure the buyer did not mean to keep.

The most misunderstood feature of first-loss cover is that it does not free the buyer from declaring the full value at risk, and it does not switch off average. It changes where average applies, not whether it applies.

The insurer sets the first-loss scale, and therefore the premium, on the ratio of the first-loss limit to the full value at risk. That calculation only works if the declared full value is honest. So a first-loss policy typically requires the insured to declare the true full value at risk, and applies average against that declared full value, not against the first-loss limit. If the buyer understates the full value to lower the base on which the scale is applied, average reduces every claim in proportion to the understatement.

The discipline is therefore two-sided. The full value at risk must be declared accurately and updated as stock levels move, exactly as it would be on a full-value policy, and separately the first-loss limit is chosen as the credible maximum single loss. Buyers who conflate the two, treating the first-loss limit as if it were the value they need to declare, walk straight into average at the claim. The first-loss limit governs the payout ceiling; the declared full value governs whether average applies. Both have to be right.

First-loss structuring is not a universal tool. It works precisely where the peril can only ever remove a fraction of the total value, and it fails where the peril can take the whole.

The classic first-loss lines share that partial-loss character:

• Burglary and theft of stock in a large holding, where a single break-in removes a bounded quantity, is the textbook case.
• Money insurance, where the cash in transit or on premises at any one time is a fraction of the annual throughput, so the first-loss limit is set to the maximum exposure at a single moment rather than the annual sum handled.
• Plate glass and similar covers, where a single incident damages a defined, limited portion of the total glazing.

Fire is the counter-example, and the contrast is instructive. A fire in a warehouse can destroy the entire stock, so the largest credible single loss is the full value, and a first-loss limit set below it is simply under-insurance dressed up as structure. This is why fire cover is written on the full reinstatement value with the average clause enforcing it, and why applying first-loss thinking to fire is a mistake.

The test is always the same question: what is the largest loss a single event of this peril can credibly produce? Where that answer is a fraction of the total value, first-loss cover matches the sum insured to the peril and saves premium honestly. Where the answer is the whole value, only full-value cover will do.

The entire value of a first-loss structure rests on one judgement: the credible maximum loss from a single event. Get that right and the cover is efficient and sound. Get it wrong and the saving is a hidden retention. A disciplined method sets the limit rather than guessing it.

The assessment should reflect the physical and operational reality of the site, not a round percentage. How the stock is stored and its portability, the access and egress a thief would use, the load-out time available before the alarm, monitoring and patrol respond, the resale market for the goods, and any history of theft in the trade and the location all bound the credible single loss. High-value, portable, easily fenced goods (electronics, branded consumables) support a higher first-loss limit than bulky, low-value, hard-to-move stock, because a theft can remove more value faster.

Once the credible maximum is estimated, the first-loss limit should sit comfortably above it, not on it, to allow for a worse-than-expected event and for growth in stock values between reviews. The limit should then be revisited when stock levels, product mix or site protections change materially, because a first-loss figure that was prudent for last year's stock can be short for this year's. And the full value at risk must be declared and kept current alongside the limit, so average never becomes the insurer's answer to a claim.

First-loss cover, understood this way, is one of the cleaner efficiencies available to a commercial buyer: it stops the waste of insuring a value the peril cannot reach, without giving up any of the protection the buyer actually needs. The only conditions are an honest full-value declaration and a well-judged limit, and both are within the buyer's control.