A letter of claim arrives. A man in his seventies has been diagnosed with mesothelioma, and his solicitor has established that he spent four years in the mid-1970s working for a company that no longer exists. The question put to you is narrow: did you write the employers’ liability cover for that employer in those years, and on what terms?
Somewhere in the answer is a paper file, or a microfiche reel, or a box in a third-party storage facility whose index was compiled by someone who retired in 1998. What happens over the following weeks – how quickly you find it, whether you find it at all, and what you can prove about what it says – has consequences that reach a good deal further than the individual claim.
Long-tail liability doesn’t respect the retention schedule
Most business records can be managed on a horizon of six or seven years. Long-tail liability lines cannot. Mesothelioma has a latency period measured in decades, and 2,218 people died of it in Great Britain in 2023 – deaths reflecting asbestos exposure that in most cases occurred before the 1980s[1]. Abuse claims follow a similar shape, arriving decades after the events they concern. In both cases the policy that responds was underwritten by people who have long since left the business, in a system that has been replaced twice, and documented in a medium chosen when the alternative was a filing cabinet.
The exposure sits on the balance sheet regardless. What varies between insurers is whether the evidence supporting it can be produced on demand.
The regulator has already been explicit about this
This isn’t a matter of good practice left to firms’ discretion. Under ICOBS 8.4, a firm with actual or potential liability for UK commercial lines employers’ liability claims must take reasonable steps to conduct effective searches of its records when it receives a tracing request, and must maintain a written policy setting out how it does so[2].
The detail the Handbook goes into is the instructive part. That written policy has to cover where the firm’s historical policies are held or are likely to be held, including records that are archived or stored off site, and the different types of record to be searched – naming electronic files, paper files and microfiche[2]. When a regulator writes microfiche into a Handbook rule, it is not making an antiquarian point. It is acknowledging that a meaningful share of the UK’s long-tail liability evidence sits in formats that cannot be searched, only retrieved and read.
The obligation is on the firm to make that retrieval effective. An archive nobody can index is not a defence, and the industry’s own history here is a caution: ELTO was established precisely because insurers’ individual record-keeping was producing claimants who could not trace cover, after a six-fold rise in tracing enquiries since 2000[3].
Where the cost actually shows up
The claims-handling cost of a slow search is the visible part, and the smallest. Three larger consequences follow from the same root.
Reserving accuracy depends on knowing what you wrote. Allocating a long-tail loss across policy years, applying the right aggregate limits, and identifying which excess layers attach all require the historical wordings. Where the record is incomplete, actuarial judgement fills the gap – and judgement under uncertainty is systematically more conservative than evidence, which means capital held against exposure that may not exist.
Reinsurance recovery depends on evidencing it. A reinsurer asked to respond to a 1970s loss will want the underlying policy documentation. Recoveries that cannot be substantiated are recoveries that get disputed, delayed, or written off, and the amount at stake in a mature legacy book is not marginal.
Legacy transactions depend on it too. Run-off portfolio transfers and Part VII schemes are priced on the acquirer’s ability to assess what it is taking on. A book whose historical records are disorganised, undigitised or partially lost gets priced accordingly, because the buyer is pricing uncertainty as well as liability.
The instinct to treat the archive as a cost is the expensive one
Physical archives are usually managed as an overhead – square footage, storage fees, a line item to be reduced. That framing produces two bad outcomes at once.
The first is that firms retain everything indefinitely because disposing of anything requires knowing what it is, and nobody has the resource to look. Holding decades of unindexed personal data is its own regulatory problem, and it sits awkwardly against data minimisation duties. The second is that the material genuinely worth keeping – the historical wordings, the schedules, the certificates that prove cover – stays as difficult to reach as everything around it.
Both problems have the same solution, and it isn’t more storage or less. It is classification: knowing what the archive contains, so that what carries evidential value can be digitised, indexed and made searchable, and what carries none can be disposed of defensibly. Dajon’s work in this area is exactly that – digitising and structuring legacy archives so that a policy year, an employer name or a reference number returns an answer rather than a search request, and so that retention decisions can be made on the basis of what a record actually is.
The question for the board
The tracing request that lands next year will concern cover written before most of the current executive team joined the industry. The reserve held against it, the reinsurance recovered against it, and the regulatory position defended around it all rest on records that no one has looked at in thirty years.
The useful question isn’t how much the archive costs to store. It’s how long it would take to prove what you covered in 1974 – and what it costs, in reserves and recoveries and transaction value, every year that the answer stays “we’d have to go and look.”
Dajon’s Data Intelligence Solution transforms the unstructured data organisations already hold into structured, analysable intelligence that informs strategic decisions, surfaces emerging risks, and identifies opportunities before they become visible in outcomes. Get in touch to understand what your current data environment could be telling you. www.dajon.co.uk/storage-cost-calculator
References
- Latest annual work-related fatalities published HSE[↩]
- ICOBS 8.4 Employers’ Liability Insurance FCA Handbook[↩][↩]
- About ELTO ELTO[↩]