The Value Equation

The Problem

Most insurance firms can describe what their products cost. Few can defend why that cost represents fair value. The FCA's review of fair value frameworks found firms relying on claims ratios alone, with no clear definition of what fair value actually means for the customer who never claims. The result is a value statement the firm cannot evidence and a regulator who keeps coming back. Fair value has been treated as a number to be calculated, not a definition to be agreed — and a number without a definition will not survive challenge.

The Pattern

The structural move is to redesign the framework so it can produce a credible ‘not fair value’ verdict — and so that verdict carries operational consequence. This means anchoring the value definition in the target market’s actual experience, evidencing benefits with data the firm can defend, and committing in advance to what would trigger withdrawal or redesign rather than renewal:

The structural moves

Defensible definition

Value is defined for this product, this target market, against a stated benchmark — not asserted by reference to the firm’s ethos or market position. The definition specifies the benefits the customer can reasonably expect (in claims terms, contingent protection terms, or service terms as relevant), the limitations that affect those benefits, the total price across the distribution chain, and the comparators against which the answer is tested. The design test: could a reviewer reading the framework predict, before seeing the evidence, what kind of finding would constitute ‘not fair value’?

Granular target market

The target market is specified at a level granular enough that some plausible customer types fall outside it. Drivers of older, lower-value vehicles may be outside the GAP target market; customers with specific pre-existing conditions may be outside a standard travel or income protection target market; customers with low predicted claims frequency may be outside an add-on target market. Where differential outcomes exist between target-market subgroups — channel, age, vulnerability, claims history — the assessment evidences fair value separately for each, not at the aggregate. The design test: does the framework name groups for whom the product would not provide fair value, and does the distribution strategy reflect that?

Capable of failing

The framework specifies, in advance, the conditions under which the product would be assessed as not delivering fair value — commission ratios above a threshold, claims experience persistently below expected, complaint themes that indicate systematic harm, comparative pricing materially out of line with market without justification. When those thresholds are crossed, named owners must escalate to a defined governance forum within a defined timeframe, and the action options must include withdrawal or redesign, not only renewal with monitoring. The design test: in the firm’s history of using this framework, has it ever produced a finding that triggered withdrawal or material redesign? If not, what evidence is there that it could?

Indicators It's Working

✓The framework has, within the last three years, produced at least one ‘not fair value’ finding that resulted in product withdrawal, redesign, or distribution change — evidence the framework can fail.

✓Fair value assessments evidence outcomes separately for at least three target-market subgroups (channel, vulnerability, tenure, or behavioural) for any product where differential outcomes are plausible.

✓Total price is presented in the assessment as a single figure aggregated across the distribution chain, with each component (manufacturer margin, distributor commission, premium finance, sub-distribution) attributed and benchmarked.

✓Pre-committed thresholds (commission ratio, claims frequency, complaint velocity, comparative pricing variance) appear in the framework with named owners and defined escalation routes, and at least one threshold breach has triggered the documented escalation in the past year.

In Practice

Worked example 1

A motor insurer reviewed its add-on portfolio after the FCA’s GAP intervention and found that its existing fair value framework had been signing off all add-ons annually using claims ratios drawn from a five-year aggregate. The firm rebuilt the framework to require, for each add-on, a separately evidenced statement of: the benefit the customer can reasonably expect (with claims frequency and average payout from the most recent twelve months, segmented by channel), the proportion of premium reaching the manufacturer, distributor, and any sub-distributor, and a benchmark from FCA-published value measures data. Two add-on products failed the rebuilt framework on the comparative-rates criterion. One was withdrawn; the other had its commission structure renegotiated and a tighter target market specified, with channels who could not distribute it within the new boundary removed from the panel. The framework had not changed what fair value meant in principle. It had changed whether the answer could be ‘no’.

Worked example 2

A pure protection manufacturer responded to the FCA’s pure protection market study (MS24/1) interim findings by separating its income protection fair value assessment from the pooled assessment that had previously covered all protection products. The granular assessment surfaced that, in the indemnity-commission distribution channels with high clawback, a material proportion of policies were being re-broked within twenty-four months — a pattern the aggregate assessment had not surfaced because new business volumes masked it. The framework was extended to evidence value separately for customers who held the policy for less than twenty-four months, and the firm withdrew the product from two channels where the re-broking pattern persisted. The fair value verdict for the channels that remained was strengthened by being defensible against the cohort that had been failing it.

Watch-Outs

The most common failure mode is the framework that always concludes ‘fair value delivered’ — the FCA explicitly identified this as a tick-box pattern in both the fair value frameworks review and TR24/2. A second is reliance on claims ratio as the proof of value: a high claims ratio does not establish fair value if cover does not match customer need, and a low claims ratio does not refute fair value if the product is genuinely contingent protection. The framework must define value, not delegate it to a single ratio. A third is target-market generality — “anyone who buys a car”, “anyone who travels”, “anyone with a mortgage” — which the FCA called out specifically in GAP and pure protection. If the target market does not exclude any group, the framework has not done its job. A fourth is the ‘peace of mind’ evidence move: claiming customers value a benefit without producing evidence they do, which the FCA flagged in GAP fair value assessments. A fifth is benchmarking against a favourable subset of the market and concluding the product is competitive — a finding that is structurally biased toward passing.

Evidence & AI Lens

E1FCA’s review of fair value frameworks (May 2023) identified four sector-wide gaps: thin evidence of fair value, weak oversight of remedial action, aggregate analysis masking group-level harm, and presentation that prevented robust decision-maker challenge.
E2TR24/2 (Product Oversight and Governance, August 2024) found many manufacturers unable to evidence that products deliver fair value, with shortcomings including not adequately considering total price (including remuneration), insufficient MI to monitor distributor remuneration, and target market statements too high-level to identify customers for whom the product would not provide value.
E3FCA’s GAP intervention (Feb–May 2024) found that in some distribution chains as little as 6% of premium was paid out in claims while up to 70% went to commission — exposing the failure of existing fair value frameworks to surface this. Firms had to demonstrate redesigned fair value assessments before being permitted to recommence sales.
E4FCA Price and Value Outcome update (Sept 2024) cited GAP firms defining the target market as “anyone who buys a car” as poor practice, and identified ‘peace of mind’ as an unevidenced benefit assertion that frameworks had repeatedly relied on without supporting data.

AI & Automation Lens

Natural language processing can audit a portfolio of fair value assessments to identify the linguistic markers of foregone-conclusion reasoning — vague benefit claims, missing benchmark sources, target-market boilerplate that recurs across products with materially different risk profiles, and the absence of any negative or qualifying language. Machine learning models trained on claims, complaints, mid-term cancellation, and renewal-decline data can identify segments of the actual customer base whose experience contradicts the framework’s assumed-value narrative — surfacing the cohorts that the aggregate assessment masks. Automated benchmarking can pull comparable products from value measures data, market panels, and peer pricing to test whether the firm’s comparators are favourable subsets. AI can also generate ‘pre-mortem’ challenge: given the framework conclusion, what evidence would falsify it, and is that evidence in the assessment? The objective is not to remove human judgement but to ensure the judgement is exercised against evidence the framework would otherwise let pass.

The problem, in one line

The move

First signal it's working