Self-insurance vs captive vs conventional insurance: the decision spectrum
The spectrum, end to end
Conventional transfer. The default position: the insurer takes the risk from the ground up (above a modest excess), prices it, and absorbs the volatility. You buy certainty of cost. What you give up is the underwriting margin in good years, and — in a hard market — exposure to pricing driven by the market cycle rather than your own record.
Higher excess / self-insured retention. The middle of the spectrum. The business keeps the first slice of every loss — a larger excess or deductible, or a formal self-insured retention under which it also handles claims within the retained layer — and insures above it. Premium falls because the insurer no longer prices the frequent, predictable losses; the business budgets for those itself. This is where most of the practical gains live for mid-sized firms, because it needs no new entities, no licences and no capital — just a clear-eyed view of what the claims record supports.
The captive. The formal end: retained risk moves inside a licensed insurance company the group owns, with capital and reserves behind it and reinsurance above it. We explain the mechanics in captive insurance explained. The captive adds discipline, data and reinsurance access to what is, underneath, the same decision as the middle of the spectrum: keep the losses you understand, transfer the ones you don’t.
Published minimum PI requirements by UK regulator or professional body
How far along the spectrum a regulated firm can move is constrained by the minimum cover its regulator requires.
| Regulator / body | Minimum limit of indemnity | Basis | Run-off required | Excess cap |
|---|---|---|---|---|
| SRA (solicitors) | £3m for a relevant recognised body or relevant licensed body; £2m in all other cases | Any one claim; no monetary limit on defence costs | Six years after cessation | Not fixed in the Minimum Terms |
| ICAEW (chartered accountants) | £2m; or 2.5 × gross fee income (minimum £250,000) where gross fee income is under £800,000 | Any single claim and in the aggregate | At least two years, then all reasonable steps for a further four | Aggregate excess capped at the higher of £3,000 or 3% of gross fee income |
| ACCA | Total income under £600,000: greater of 2.5 × relevant total income or £100,000. Total income £600,000 or more: at least £1.5m | Per ACCA Global Practising Regulations | Not published as a fixed period in this source | Uninsured excess restricted to £20,000 per principal |
| AAT (licensed members) | Sole traders: greater of 2.5 × gross fee income or £50,000. Partnerships and limited companies: greater of 2.5 × gross fee income or £100,000. Maximum required cover £1m where gross fee income exceeds £400,000 | Full civil liability, fully retroactive | Not published as a fixed period in this source | Set at a level the member can meet at all times |
| RICS (chartered surveyors) | Turnover £100,000 or less: £250,000. £100,001–£200,000: £500,000. £200,001 and above: £1m | Each and every claim (or aggregate plus unlimited round-the-clock reinstatement); defence costs in addition to the limit | Six years; consumer run-off £1m in all for six years | Greater of 2.5% of the sum insured or £10,000, for limits up to £10m |
| ARB (architects) | £250,000 | Each and every claim, except fire safety, cladding, asbestos and pollution which may be aggregate | Six years, or five years in Scotland, at the same level as the last year before cessation | Not published |
| FCA — insurance intermediaries (MIPRU 3.2) | €1,300,380 for a single claim; in aggregate the higher of €1,924,560 or 10% of annual income, capped at £30m | Per year | Not set in MIPRU 3.2 | Higher of £2,500 or 1.5% of annual income (no client money); higher of £5,000 or 3% (client money held) |
| FCA — IDD insurance intermediaries (IPRU-INV 13.1) | Relevant income up to £3m: at least £500,000 single claim and aggregate. Relevant income over £3m: at least £650,000 single claim | Per policy | Not set in IPRU-INV 13.1 | Excess over £5,000 requires additional capital resources |
Sources: SRA Minimum Terms and Conditions (sra.org.uk); ICAEW PII Regulations effective 1 September 2024, regs 3.2, 3.3, 3.7 (icaew.com); ACCA Professional Indemnity Insurance Regulations (accaglobal.com); AAT professional indemnity insurance requirements (aat.org.uk); RICS Professional indemnity insurance requirements, UK and Republic of Ireland, 2 July 2025 (rics.org); ARB PII Guidance (arb.org.uk); FCA MIPRU 3.2 (handbook.fca.org.uk) and IPRU-INV 13.1 (handbook.fca.org.uk). Figures are the published minimums at the date shown on each source and are not advice; check your own body’s current rules.
The trade-offs, mechanically
Every step to the right trades certainty for expected saving. Retaining risk removes the insurer’s costs and margin from the retained layer — that is the saving — but exposes the business to its own claims volatility — that is the price. Three tests tell you whether the trade is sound:
- Stability. Retention works on losses that are frequent enough to be predictable. A stable pattern of small claims is ideal retention material; a thin record with occasional large losses is not.
- Capacity. The retained layer must be absorbable in a bad year — not an average year — without threatening the business. The balance sheet, not optimism, sets the ceiling.
- Discipline. Retained losses still need managing: claims handling, reserving, and honest accounting for the cost. Self-insurance done casually is just being uninsured with extra steps.
There is also a cash-flow dimension: premium is a known outflow on day one, while retained losses fall when they fall. And there is a market-cycle dimension — in a hard market, moving right along the spectrum lets a good risk stop paying for the market’s bad ones; in a soft market, transfer gets cheap and the case for retention weakens. The position deserves revisiting at each renewal, not fixing forever.
Who should think about which
For most small businesses, conventional transfer is the right answer and will stay that way: premiums are modest relative to the volatility being transferred, and management time is better spent elsewhere. Mid-sized businesses with a few years of clean claims data are usually the ones leaving money on the table — typically by carrying a default excess that no longer reflects their record. Larger organisations and groups with substantial, stable premium spend across several lines are the natural candidates for the formal end — structured retentions across a programme, and eventually a captive feasibility conversation, particularly where a group programme already exists.
The broker’s role: structuring the conversation
None of this requires exotic products. It requires the decision to be framed properly: what the claims record actually shows; what each retention level does to premium, tested against real market appetite rather than theory; what a bad year at each retention would look like against the balance sheet; and what the wording must do above the retention — because a policy that sits over a large retention needs its conditions, notification provisions and claims handling to fit that structure. That framing is broker work. We have no product to sell at any point on the spectrum, which is precisely what makes the conversation useful.
Frequently asked questions
What is the difference between self-insurance and just not buying insurance?
Going uninsured is passive: losses land wherever they land. Self-insurance is deliberate risk financing — choosing which losses to retain, sizing the retention against the balance sheet, budgeting for expected claims and keeping insurance for the losses that could genuinely hurt. The retention is a decision, not an accident.
What is a self-insured retention and how does it differ from an excess?
Both leave the first part of each loss with the insured. Under a conventional excess (or deductible) the insurer typically manages the claim from the ground up and recovers the excess; under a self-insured retention the insured is usually responsible for handling and funding claims within the retention itself, with the policy sitting above it. The practical difference is who runs, and funds, the small claims.
Is a captive a form of self-insurance?
Essentially yes — it is self-insurance made formal. A captive puts the retained risk inside a licensed insurance company owned by the group, with capital, reserves, and reinsurance behind it. That formality brings discipline, data and access to the reinsurance market, at the cost of running a regulated entity.
Who should consider moving along the spectrum?
Organisations whose claims record consistently outperforms their premiums, who have stable spend, a balance sheet that can absorb a bad year, and the appetite to manage risk actively. If premiums feel expensive but the claims record is thin, unstable or poor, the priority is risk management and presentation — not retention.
What does a broker add to this decision?
Independence from the answer. A broker can model what different excess levels do to premium, test the market's appetite for each structure, benchmark how the programme performs across the cycle, and structure the conversation about retention against your actual claims data — rather than against a product someone needs to sell.
Apex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority (FRN 724952). This page is general information, not advice on a specific policy.