The Broker's Cut
A connection through “Follow the incentives”: Look at how the way people are paid changes the decisions they make.
Your business insurance quote is not a price. It is a market position, taken by an insurer whose history of getting reserves right or wrong helps determine the number you pay.
You are told the premium reflects your risk. The mechanism is less clean.
A commercial quote is a chain of judgement, not a verdict from a model.
A commercial insurance quote is not produced by one actuarial model, audited by a regulator, then handed down as fact. It is produced by a chain of judgement: the data the insurer has on you, the model it uses to price what could happen, the claims experience it thinks belongs to you, the weight it gives that experience, and the market signal it takes from the quotes it lost.
That last step is where the story starts to bend. The "market" your insurer benchmarks against is not a neutral crowd. It is a moving collection of carriers with different reserving habits, different appetites for growth, different claims-inflation assumptions and different memories of past mistakes. The Prudential Regulation Authority has warned general insurers that claims inflation may require further strengthening of prior-year reserves, and that some firms had applied claims-inflation allowances "arbitrarily and without technical justification". Once that happens, the quote you see is no longer just a view of your business. It is also a view of the insurer's own past.
Premium is not a fact. It is a position, taken at a moment, against a market that does not agree with itself.
You see £42,000 from one carrier and £50,000 from another. You assume the first insurer thinks your fleet is safer. It may simply be carrying a cheaper memory of the claims it wrote two years ago.
Commercial pricing begins with inputs. The insurer takes what it can learn about the risk: company history, County Court Judgments, vehicle files, occupation, trade, postcode, driver age, claims record, fleet composition and any telematics feed. In a haulage fleet, the same customer can become two different risks depending on whether the insurer sees only registration numbers and claims totals, or vehicle-level telematics showing braking, route density, overnight parking and mileage by driver. A gap at this stage does not remain a gap. It becomes a distortion downstream.
| Component | What it does | Where insurers diverge |
|---|---|---|
| Input enrichment | Turns the submission into usable risk data | One carrier may see clean telematics; another sees a thin broker schedule |
| Exposure model | Prices what could happen before your claims are weighted | Postcode, vehicle, trade and fleet size matter, but model design and appetite differ |
| Experience model | Prices what has happened on your book | Three claims mean little in a six-vehicle fleet, much in a 150-vehicle fleet |
| Credibility weighting | Decides how much to trust exposure vs experience | Size threshold and time window change the answer |
| Lost-to credible-brand check | Compares the quote with what the market did last time | Same lost quote read differently if from an incumbent vs a fast-growth entrant |
Five components. Three are shaped less by your risk than by the insurer's own internal practices.
Personal lines hide much of this from you because the model dominates. UK motor and home pricing are now industrial systems: huge volumes, standardised products, rapid price changes and little underwriter intervention. Commercial fleet, liability and property risks are not that tidy. The smaller the book, the more the insurer is forced back towards exposure assumptions and underwriter judgement. The larger the book, the more your own claims history begins to drag the price.
The consequence is simple. The model does not "produce the price". The model produces a starting point. The final premium is the starting point plus a market judgement. That market judgement is contaminated by every insurer whose past claims have not yet finished developing.
Reserving is the insurer's estimate of what claims already reported, or already incurred but not fully reported, will ultimately cost. You do not see it, but you pay for it. If an insurer has historically reserved too low, its old underwriting years look profitable for longer than they should. Loss ratios look better. The pricing team sees a book that appears cheaper to run. The carrier can write today's business aggressively and still look rational on paper.
Then the claims mature. Repair costs rise, injury claims settle later, litigation lasts longer, and the old reserves prove thin. The insurer strengthens reserves. Capital pressure follows. Pricing hardens. The customer who enjoyed the cheap quote in year one is told in year three that the market has changed.
Higher economic inflation could pass through into claims settlement costs more than firms had assumed, creating a significant risk that the market would identify the need for further strengthening of prior-year reserves.
— PRA, Insights from thematic review of general insurance reserving, June 2023
The opposite distortion also matters. If an insurer reserves too conservatively, it sees more pain than reality will ultimately deliver. It prices defensively, loses business to carriers with lighter reserves, and later releases the margin it did not need. Both insurers may be acting in good faith. Both may be using professional actuaries. They are still selling you different versions of the past.
| Reserving stance | What the insurer sees | What you see | What happens later |
|---|---|---|---|
| Under-reserved | Old years look cheaper than they are | Competitive premiums now | Reserve strengthening and price correction later |
| Accurately reserved | Loss ratios mature roughly as expected | Less dramatic pricing | Fewer violent corrections |
| Over-reserved | Old years look worse than they are | Defensive premiums now | Releases and renewed appetite later |
Three reserving stances. The customer experiences each as a different price for the same risk.
Direct Line's 2023 annual report gives the live case study: the group reported a £331.6 million motor insurance service loss and a £138.4 million reserve strengthening in motor, in a market where claims inflation remained elevated and repair-cost inflation was driven by higher labour costs. The cheapest quote on your renewal schedule can be a signal about the insurer, not your risk. It may be sharp underwriting. It may also be a sign that the carrier has not yet paid the bill for the business it already wrote.
Commercial claims do not close neatly inside the policy year. A liability claim opened in 2024 can still be developing in 2028. To price this year's policy, the insurer takes paid and reported claims from previous years and converts them into an estimate of ultimate cost. That conversion is the development factor.
The mechanism is blunt. If your historical claims are £100,000 paid and incurred to date, one insurer may think the ultimate cost is £120,000. Another may think it is £150,000. The same customer, the same history, the same broker presentation. Two different premiums.
| Same fleet history | Insurer A | Insurer B |
|---|---|---|
| Paid and reported claims to date | £100,000 | £100,000 |
| Development factor | 1.20× | 1.50× |
| Estimated ultimate claims | £120,000 | £150,000 |
| Indicative premium effect | Lower | Higher |
One number. One choice of multiplier. £30,000 of price difference before expenses or margin.
You are rarely told which factor has been used. Yet it can move the price more than the items you are trained to argue about: postcode, vehicle mix, excess, or whether one driver had a minor accident two years ago.
The Bank of England's insurance aggregate annual data exposes the lag. Its non-life charts include claims development by accident or underwriting year, and the underlying 2024 year-end data show claims continuing to emerge across multiple development years rather than landing as a single settled number. The insurer is not pricing only what happened. It is pricing what it believes the old claims will become.
When a broker tells an insurer it lost the fleet to another carrier, the feedback looks simple: "Lost to Carrier X at £40,000." It is not simple. The analyst has to decide whether Carrier X's £40,000 is a market signal or noise.
A sophisticated insurer does not treat every lost quote equally. It builds a memory of which competitors are credible reference points. A price from a mature, conservatively reserved incumbent is not the same signal as a price from a specialist carrier growing hard in a volatile line. A quote from Allianz, Aviva, AXA, QBE or Zurich may be treated as evidence that the insurer's own model is too high. A quote from a carrier with a younger book, a narrower specialty, a fast-growth strategy or a history of volatile results may be discounted before it changes the model.
| Lost-to carrier signal | How a disciplined insurer reads it | Consequence for your renewal |
|---|---|---|
| Tier-one incumbent at 10% below | Strong market evidence | The quote may be challenged or reduced |
| Specialist carrier at 25% below | Context-dependent evidence | The underwriter may adjust only partly |
| New entrant at 35% below | Weak or noisy evidence | The underwriter may refuse to follow |
The same lost quote is treated differently depending on who lost it to whom.
There is no public league table that says "this insurer reserves accurately". The tier system is informal, reputational and constantly updated by underwriters, actuaries and brokers. That makes it powerful. It also makes it invisible to you. The customer sees the second number and calls it cheaper. The insurer sees the same number and asks whether it should be learned from.
The fashionable error is to make commercial insurance behave like personal insurance: more automation, more modelling, fewer underwriters, faster quotes. The mechanism fails at the tails. A personal motor book can average out millions of similar risks. A commercial fleet book cannot treat a refrigerated-food distributor, a scaffolding contractor and a last-mile courier fleet as interchangeable just because they all run vans.
Sub-10-vehicle fleets often have too little claims history to be credible. Larger fleets can be heavily experience-rated, but only if the claims have been developed properly. The middle is where judgement earns its keep. Market cycles show the cost of pretending otherwise.
Marsh reported that UK commercial insurance pricing rose 20% in the first quarter of 2022, following 22% in the fourth quarter of 2021; cyber pricing rose 102%, driven by ransomware claims, market deterioration and reduced capacity. Those increases were not just "risk going up". They were the market correcting a prior view of risk that had become indefensible. A price can look technically sophisticated and still be wrong if the historical claims base underneath it was reserved too lightly.
You do not only buy insurance from an insurer. You buy access to a particular slice of the market. That slice is chosen by your broker's panel, wholesaler relationships, speed preferences, facility arrangements and appetite for complexity.
The broker channel is not a detail. In evidence to the House of Lords, the ABI stated that brokers controlled £18.2 billion of the available £20.9 billion across the UK commercial insurance market. Your premium is partly a function of which insurers were asked, which ones responded, which ones were considered credible and which ones your broker did not approach at all.
If your broker's panel is dominated by fast-quoting carriers, you may be benchmarking against speed rather than quality. If it excludes the insurers whose books are most conservatively reserved, you may never see what the disciplined price looks like. If it includes a market that is buying share, the cheapest quote may anchor the entire renewal discussion even if better underwriters would not trust it.
| Question to ask your broker | What it reveals |
|---|---|
| Which insurers were approached, and which declined? | Whether you saw the market or only the convenient panel |
| Were tier-one composite carriers quoted? | Whether the benchmark includes disciplined incumbents |
| Who was cheapest last year, and did their loss pick mature? | Whether the prior price was sharp or temporarily under-reserved |
| What development factor has been assumed on my claims? | Whether the broker understands the invisible multiplier in your price |
| Which lost-to prices do my carriers consider credible? | Whether the renewal comparison is evidence or noise |
Most buyers ask for discount. Better buyers ask for the shape of the market that produced the quote.
Your commercial insurance premium is not the cost of your risk. It is the cost of your risk as seen through a particular insurer's data, claims assumptions, reserving history, appetite for growth and opinion of its competitors.
That does not mean the market is fake. It means the market is plural. Each insurer is using a different claims memory and calling it price. Each broker panel is selecting which memories get to compete. Each underwriter is deciding which lost quote deserves respect. Each reserving team is revising old years that will quietly rewrite new-business appetite.
The honest version is this. Every quote you receive is partly the broker's choice of carrier panel, partly the carrier's choice of reserving philosophy, partly the underwriter's judgement about who else quoted, and only partly the underlying risk on your fleet, property or liability book.
You are not paying for actuarial science alone. You are paying for the gap between what the insurer reserved and what reality served up.
You’ve looked beneath the surface.
A connection through “Follow the incentives”: Look at how the way people are paid changes the decisions they make.
A connection through “Follow the incentives”: Look at how the way people are paid changes the decisions they make.
A connection through “Follow the incentives”: Look at how the way people are paid changes the decisions they make.