When Better Customer Outcomes Don’t Mean Better Distribution

On 18 September 2026 Japan’s Financial Services Agency published its latest analysis of the common performance indicators that sellers of foreign-currency insurance report¹. The headline is easy to quote. Across the 140 institutions reporting for March 2026, the average share of customers whose holdings showed a return of zero or better was 84.7%, up from 65.0% across 155 institutions a year earlier¹.

It is a better number. It is not yet an explanation.

What changed

Three facts about the figure matter before anything else. It is a simple average across institutions, not a share of all customers. The two years cover different populations — 155 institutions, then 140. And the FSA itself says why it rose: the increase was affected by movements in exchange-rate levels, among other factors¹.

The measure is also specific. Return since purchase is the surrender value plus any amounts already paid, less the single premium, with everything converted into yen at the base-date exchange rate. The FSA notes that this differs from the result of holding to maturity or of taking the proceeds in foreign currency, and that early surrender often returns less than the premium paid¹.

Nothing here is a new obligation. The FSA expects sellers to publish these indicators rather than requiring it³, the underlying principles operate on a comply-or-explain basis, and the FSA states that it does not guarantee the accuracy of the data institutions submit².

Japan’s direction as part of a broader shift

The indicators belong to a sequence the FSA has built over nine years. Its customer-oriented principles came in 2017. Firms then published their own performance measures, which varied too much for customers to compare, so the FSA introduced common indicators for investment trusts in 2018 and extended them to foreign-currency insurance in January 2022³. In September 2024 it added supplementary principles on product governance⁴.

The sequence that follows is Eveil’s synthesis, not an FSA roadmap. Read together, those steps run from transparency to comparability to product governance to measured customer outcomes. The step still missing is attribution.

The direction is not unique to Japan. In the United Kingdom, the Consumer Duty requires firms to monitor whether customers actually receive good outcomes⁵. In the European Union, insurers must check whether their products remain consistent with the needs of the target market and whether they are reaching customers outside it⁷. That is regulatory convergence in intent, not a single global standard — and not a story of Japan catching up.

The next challenge: attribution

Measuring an outcome is one thing. Understanding what produced it is another.

The indicator cannot separate what a firm did from what the market did. A higher figure may reflect product design, target-market selection, distribution quality, cost or surrender behaviour — or simply a weaker yen. The FSA’s own attribution of this year’s rise to exchange-rate movements makes the point for it. So the 84.7% does not show that distribution improved, and a lower figure at any one institution does not show that its distribution was poor.

The difficulty is not peculiar to Japan. In its review of insurers’ Consumer Duty monitoring, the UK Financial Conduct Authority set the expectation that firms identify where customers are not getting good outcomes, understand why, and act⁵. Reviewing monitoring across sectors in July 2026, it observed that it is not always clear what caused an issue in the first place, or whether the actions taken actually fixed it⁶.

Why distributor dispersion matters

Averages conceal a wide range. Among the 90 regional banks reporting, the share of customers at zero or better ran from 100% down to 49.4%; across business types the averages sat broadly between 80% and 90%¹.

This section is Eveil’s interpretation. Dispersion on that scale is exactly where attribution matters most, because the obvious reading is probably wrong. Institutions operated within the same broad currency environment, but their customers entered at different times, in different currencies and through different products. An institution that sold heavily when the yen was weak would tend to show worse yen-converted results than one that sold when it was strong, whatever the quality of its advice. Product mix, customer mix and surrender patterns pull in the same direction. Without decomposing those effects, dispersion cannot be read as a ranking of distribution quality.

From disclosure to management information

The public indicator is disclosure. Managing outcomes needs something more granular, and Japan’s own framework already asks for it. The 2024 supplementary principles expect manufacturers to define a target customer — including customers who should not buy — to verify continuously that a product performs as designed, and to check whether the customers who actually bought it match that target, using information from distributors such as purchaser characteristics, complaints and sales activity⁴.

The UK experience shows how easily that ambition reduces to box-ticking. In its review of 20 larger insurance firms, the FCA found that some reported metrics showing whether a process had been completed rather than insight into the outcomes customers received⁵. An outcome figure that no one can decompose is the same problem, one level up.

Implications for international insurers and distributors

For a global insurer or distributor in Japan, an existing product-governance framework is an advantage. It is not sufficient. A dashboard built for another market will report outcomes; it will not show how Japanese distribution, customer behaviour and a yen-denominated measure produced them.

What to validate next Whether your Japan evidence can trace a single chain — target market, product design, distribution, customer behaviour, cost, outcome — and where the exchange rate sits within it. If a result cannot be traced back through that chain, it cannot be managed; it can only be reported.

Eveil view

Japan has done the difficult first part: comparable outcome data, published by the institutions that sell the product, built deliberately one step at a time. But the September figures show the limit of that first part as clearly as its value. The number rose, and the regulator itself points to the exchange rate.

The more meaningful test under closer scrutiny will not be which firm reports the highest figure, but which can explain its figure — which part came from its own design and distribution choices, and which part came from the market. That is the difference between measuring outcomes and managing them.

What to watch next

The next round of reporting closes on 8 January 2027². Three things are worth watching as it lands: whether outcomes hold if the exchange rate moves the other way, whether manufacturers’ product-governance work begins to show up in how firms explain their results, and whether dispersion between institutions narrows or simply moves.

Assessing how a product or distribution model should be governed for customer outcomes in Japan? Eveil Intelligence is structured work on one organisation’s question.