The insurance the bank cannot see: Part 2

Part 2: The documentation duty no one has tested

In July, 113 homes burned down in Krokstadelva, the largest residential fire in peacetime Norway since 1923. In part 1, we looked at what a total loss like this does to the collateral, and why fire differs from flood and landslide: fire risk is not shared through the Norwegian Natural Perils Pool, and it can be priced freely.

Here, in part 2, we look at what can happen to the capital requirement when a bank cannot answer a simple question about a building it has lent against. Most buildings likely have building insurance. But the requirement on banks does not say the bank should assume the collateral is insured. It says the bank should monitor whether it is sufficiently insured.

In part 1, we gave concrete examples of what "not sufficient" can mean in practice. Full-value insurance only covers what is legally required to rebuild, not desired upgrades, and a forgotten notification about an extension leads to a reduced payout. In the event of fire, unlike natural disaster damage, there is no obligation to compensate lost land value, and a gap of a million kroner or more is not unusual. Collateral value across an entire neighborhood can fall, regardless of who actually had a claim to report. The loan runs for 25 years or more, while the policy is renewed, repriced, or in principle can be declined every year. Fire risk is priced freely by each individual insurer, without the solidarity-based pooling the natural disaster scheme provides. If the risk rises, the price will too. Payouts after home fires rose 24 percent in the first half of the year compared with the same period last year, according to Finance Norway. That was before the Krokstadelva fire, and it is hard to imagine this not leading to stricter documentation requirements for customers, and higher premiums for the buildings considered most at risk. Around one in ten households lacks contents insurance, and a minority lack building insurance altogether. And even a fully insured customer can have their settlement reduced if what they disclosed turns out not to match what the burned site actually shows.

None of these gaps is dramatic on its own. Together, they can mean that a seemingly fully insured customer still ends up with a significant bill the bank had not accounted for, because the insurance covers housing costs, not the loan. We follow this further in part 3.

The price of fire risk is becoming its own number

At Telescope, we are seeing growing interest in fire risk of the kind Krokstadelva represents among the insurers we talk to, naturally enough. Since the risk is not pooled with others, it can, like surface water risk, become more differentiated in pricing, and turn into a competitive arena between insurers.

Insurers have paid out 42.3 billion kroner in weather and natural damage claims on buildings and contents over the past ten years, and the last three years have been by far the most expensive on record. Home insurance premiums have risen more than 30 percent over two years, even though both the number of reported claims and the amount paid out fell from 2024 to 2025. According to Norges Bank's staff memo on weather and climate damage (January 2026), around 13 percent of Norwegian homes and 11.5 percent of Norwegian home values sit in areas particularly exposed to natural disaster risk. The memo shows that in high-risk areas in the United States, home insurance premiums have risen around 40 percent faster than the consumer price index. The Norwegian premium increases so far are driven by rising construction costs and weather damage in general, not by differentiated fire pricing, but this could change as a result of the Krokstadelva fire.

For the bank, rising premiums are an early, visible signal that a mortgaged property is becoming more expensive to insure, and eventually harder to insure at all. The insurance gap is growing faster than the losses, and it hits asset quality before it hits the claims statistics.

The regulation already says "monitor"

The capital requirements regulation (CRR) determines how much capital a bank must hold "in the vault" behind every loan. It states, in article 208(5), that the mortgaged property must be adequately insured, and that the bank must have routines in place to monitor that it actually is. As far as we know, supervisors have not enforced this against banks so far, but after the Krokstadelva fire, it suddenly became part of the public debate. Can the Financial Supervisory Authority of Norway then afford not to follow up?

This, then, is two duties, not one: the collateral must be insured, and the bank must monitor it. The word is "monitor", present tense. The EBA's guidelines on the management of ESG risks (EBA/GL/2025/04) point in the same direction, asking banks to assess whether insurance cover could become unavailable or prohibitively expensive for exposed portfolios.

Supervisors do not need a new rule to ask the question. They just need to ask it.

But can the banks answer?

It is unclear, however, how banks would answer if a supervisor asked. There is no central register of current building insurance policies in Norway. A bank cannot look up an address and find out whether it is insured today, or compare the insured sum against the actual rebuilding cost to today's building code, in today's market, in that particular municipality. Insurers know it. Banks almost certainly do not, and there is likely no interface for sharing that information between them, not even between those that belong to the same group, or where the bank is a part owner. Nor does the bank know whether the customer has contents insurance or other cover against lost rental income, which would help liquidity in the period until the building is rebuilt.

If the FSA were to demand documentation of regular monitoring of whether a building is insured, banks likely have three options.

• An annual self-declaration from the customer. Cheap, but not every customer will respond, and the declaration has limited value as documentation.

• A data-sharing partnership with insurers. Perhaps sensible, but demanding. Who wants to both fight over privacy concerns and build the solution?

• Buy cover itself. The EBA addressed exactly this in a Q&A from 2025. A general insurance against losses on collateral can guarantee the bank compensation for losses tied to the collateral being unrealizable due to damage or destruction, but the cost of that cover has to be priced into the margin.

What it costs not to be able to answer

What happens if the FSA asks the bank whether sufficient insurance exists, and the bank cannot give a good answer? Then the collateral cannot be counted as recognized security in the capital requirement calculation.

For non-bankers: a bank must hold its own money behind every loan it grants, and how much depends on how well secured the loan is considered to be. The calculation happens in two steps. First, the loan is multiplied by a risk weight, and for a mortgage with recognized security that weight is low, perhaps 35 percent. A loan of three million kroner then counts as 1.05 million in the bank's calculation base. The bank must then hold common equity tier one capital equal to around 15 percent of that, around 150 000 kroner. If the collateral is not recognized as security, the loan is treated as unsecured. The weight jumps to 75 percent, the calculation base becomes 2.25 million, and the capital need more than doubles. Equity is the bank's most expensive form of funding, and capital tied up behind one loan cannot fund a new one, so an increased capital need has to be recovered through the interest margin. At portfolio level, we are talking billions. And what triggers the jump in our case is neither weaker repayment ability nor falling home prices, but the bank's inability to document that the collateral is adequately insured.

Supervisors are unlikely to disqualify an entire mortgage portfolio overnight. More realistic is a note in the supervisor's assessment of the bank's capital need (the Supervisory Review and Evaluation Process, SREP), a deadline, and a Pillar 2 add-on if the deadline is missed. But if the supervisor asks whether a large number of loans have satisfactory insurance cover, it is demanding for the bank to answer well and quickly, as we have seen above. Any add-on hits the whole portfolio.

The missing interface

This is not a challenge banks face alone.

The Norwegian Insurance Contracts Act gives the mortgage holder protection through co-insurance and notification requirements. But in practice, this has limited value if two parties who both have an interest in the same building, and who are both supervised, have no ongoing channel between them about how the building and its surroundings change over time.

Not every building in Norway burns at the same time. What matters to the insurer, and to the FSA, is how large a share of the exposure could be hit by the same event (in the same way as the bank's concentration risk). To calculate that, the insurer has to divide the portfolio into geographic or logical "boxes" where everything could go wrong at once, so-called accumulation zones. The accumulation figure governs three things: how much reinsurance the insurer buys and at what price, how much capital it must hold under Solvency II, and whether the underwriter is even allowed to write further policies within that "box".

The industry standard for measuring fire risk concentration is a 200-meter radius: the insurer looks at all buildings with insured sums within that circle, and treats them as a single combined risk. Some large insurers use their own, more advanced models, but that is unlikely for most Norwegian ones. That fits a city block better than wind- and vegetation-driven spread up a hillside; the Krokstadelva fire had a larger radius than 200 meters. The spread zone should have been the accumulation unit. And since wildfire is also not part of the natural disaster scheme, Krokstadelva falls between two stools. Banks and insurers need the same underlying data, for partly different reasons, and neither of them has it today.

Five questions for the bank's own portfolio

Risk management and credit teams at banks should ask themselves the following.

1. How many mortgaged properties sit within the same spread zone, and what is the largest combined exposure per zone?

2. When was insurance status last verified for exposures subject to physical climate risk, and how?

3. What does the valuation policy actually say about insurance, beyond the point of loan approval?

4. Is fire included as a scenario in the bank's own assessment of its capital need (ICAAP), or do the physical scenarios stop at flood, surface water, and landslide?

5. If the premium on an exposed property doubles, how does the bank find out?

They may not like the answers they get.

Conclusion

Insurance is a safety net for the borrower (at least to some extent, as we saw in part 1), but not necessarily for the bank. Especially not for a risk that is not shared through the Norwegian Natural Perils Pool, that hits whole neighborhoods at once, and that becomes more likely with every decade of drier summers.

It is worth noting how undramatic the trigger is. Not a housing market crash, not a wave of defaults, but a missing piece of paper about a house that is still standing. The regulation does not ask whether the loan feels safe. It asks whether the conditions for counting the collateral as security are met, and insurance is one of them.

The banks that come out of the next decade in good shape are the ones that know which addresses are exposed before anything burns, factor that into loan terms and pricing, and know how good their insurance cover actually is.

This is part 2 of three. Part 1 covers what a fire settlement does to the collateral, and why fire differs from flood. Part 3 is a worked example that puts numbers on what a settlement that does not go as planned does to the loss rate.