Retail participation: what the headline numbers leave out 

A participation rate answers one question about a population, and is routinely asked to answer three more that nobody in Britain has measured. 

Every yes in the total weighs the same 

The FCA’s Financial Lives work for 2024 counted 7.9 million UK adults on an investment platform. The same exercise in 2020 counted 4.4 million. A narrower reading moves the same way: 11 per cent of adults held a non-advised platform account in 2024, against 5.9 per cent in 2020. 

What the instrument records is a yes. A respondent either belongs inside the category or does not, and every member inside counts for one. Somebody who moved a five-figure sum into a portfolio last year and somebody who opened an app in the first lockdown and never signed in again arrive as equals. 

Intensity is the first: how often the account gets used, from a daily habit down to nothing at all. Duration is the second: whether this person turned up last month or has been there since 2016. Amount is the third, and readers assume it must be there somewhere, because a participation question asks whether you hold an account and never what sits inside it. 

None of the three is published beside the rate and none can be recovered from it. That is no criticism of the survey, which measures a headcount competently. It becomes a problem the moment the rate is quoted as evidence that Britain is trading more, a sentence needing all three and handed none. 

A rate can rise while the activity behind it shrinks 

Total activity in a market is participants multiplied by what the average participant does. A participation rate pins down the first term only. Leave the second unmeasured and the product is unknown, so the direction of the rate says nothing reliable about the direction of the whole. 

There is a reason to expect the two terms to move against each other. People enter after a barrier comes down, so the marginal joiner is the one for whom the activity had not been worth the trouble. Adding people at the quiet end lifts the headcount and pulls the average down. 

Which way the product went between the two FCA readings is settled by no published UK figure, because that average is the term nobody publishes. 

Invented numbers, chosen only to show that a participation rate and a total of activity are free to point opposite ways inside the same sample. 

A headcount gap and an exposure gap are different findings 

The same survey reports that 21 per cent of adult men in the UK use an investment platform against 9 per cent of women, a ratio of roughly 2.3 to one. Read the measure before the ratio. It covers investment-platform use across all adults, wider than the non-advised 11 per cent above, so the two can appear in one article but never in one series. Taken at face value, the finding says more men than women answered yes. It says nothing about whether those men hold larger balances, deal more often or stay longer, and each would be a separate gap. Who opens an account and what happens inside one are two questions, and the survey has answered the first. 

The answers are self-classified and nothing checks them 

Underneath every figure here is a person being asked and answering. They must decide whether the thing on their phone counts as an investment platform, whether the help they got amounted to advice, and whether an account untouched since 2021 is still one they use. The answer enters the total unreconciled against any account record. 

If those errors cancelled, this would be a rounding problem. They do not obviously cancel. Advised against non-advised is precisely the line a customer is least equipped to draw, since guidance, a promotion and a personal recommendation are separated by regulatory definitions the customer never sees. Categories flatter or embarrass too, and nobody has estimated the sorting that follows. What is missing is not precision, it is a validation step. 

A validation step being the thing missing, it is fair to ask whether anything here supplies one. Something does. Open a real account, pay the real charges, and set the statement that arrives against what the firm said it would charge. Any discrepancy is visible because two records exist and one was issued by a party obliged to issue it. 

That reconciliation will not turn round and point at a respondent. Ask what a live funded test can establish about how many British adults hold a platform account and the honest answer from The Investors Centre, which opens and funds live accounts with its own money to test UK trading platforms rather than compiling rankings from providers’ published fee schedules, is nothing whatever. A person is not a counterparty and issues nobody a statement. The only instrument here capable of checking an assertion against a record works where the assertion came from a firm obliged to keep one, and that is why a charge can be audited by dealing while a population cannot be audited at all. 

The forty-seven per cent is measured inside the smaller number 

Roughly 1.6 million adults, 3 per cent of the adult population, told the survey they used a trading app, and 47 per cent of that group were aged 18 to 34. The second figure is a composition, describing who is inside the 1.6 million. It does not give the proportion of young adults who use a trading app, a sum needing the young adult population underneath it. 

Composition shares move for two unconnected reasons. That 47 could climb because more people under 35 take up an app, or because people over 35 give one up, and the figure does not say which. Read as a fact about young Britons it does work the arithmetic will not support. Read as a fact about the app population it is exact and more interesting: the category skews young. 

A rising line that two different countries could both draw 

Duration is the quantity with no instrument at all. A participation rate is a photograph, recording who was inside the category while the fieldwork ran. Somebody who joined a fortnight earlier and somebody holding the same account for nine years are one dot each, and somebody who left the month before is absent in the same way as somebody never there. 

Two different countries produce an identical rising line. In one, a stable body of holders grows steadily at the edges. In the other, people pour in, stay briefly and leave, and the level holds up only because arrivals keep coming. 

The published series cannot tell those apart, and neither can this article, since the entry and exit counts are collected by nobody. 

Where the bad writing on this subject starts 

What goes wrong is a sentence shape rather than any particular figure, and it is always the same shape. A participation number goes in at the front and a verb of activity comes out at the back: more Britons hold platform accounts, therefore Britain trades more. The first half was measured. The second was not, and the therefore carries a load no fieldwork authorised. 

None of which argues for ignoring participation rates. They are the only national reading of who is in the market at all, and the move from 4.4 million adults to 7.9 million is a real change in British financial life. It is a change in how many people sit inside a category. How much, how often and how long they stay were never put to anybody, and no figure is responsible for the sentences other people build on it.