A fifth of solicitors won't state their sex and a third won't state their ethnicity: what a hole in the data does to a regulator
The Law Society recently published its annual statistics report, and the growth figures are unremarkable. The roll stands at 208,077, up nearly 3%. Practising certificate-holders reached 172,966, up 3.2%. Admissions were steady at 9,365. Departures fell 17% to 4,930, which is either better retention or a frozen job market, depending on what you think is happening out there.
The number that should stop you is 36,094.
That is how many PC-holders did not tell the regulator their sex in the year to July 2025. In 2022 the equivalent figure was 12,358, which was 8% of the profession. It is now 21%. Ethnicity looks worse in absolute terms: 61,360 people did not disclose it, up from 38,947, so 36% of the profession against 25% three years ago.
Sex non-disclosure has nearly tripled. Ethnicity non-disclosure is up by about 58%. Both trends run in the same direction over the same three years, and neither shows any sign of levelling off.

What a percentage of respondents is not
Here is the sentence in the report that deserves more attention than it will get. Female PC-holder counts fell in each of the last three years, even while the female share among those who did reveal their sex rose. The report's own conclusion: "The apparent decline can therefore be a measurement effect rather than a reversal in representation."
That is the whole problem in one line. Publish a percentage of the people who answered and it will be read as a percentage of the profession. If disclosure rates differ between groups, and they almost certainly do, then the answer moves as the response rate moves, with nothing at all happening in the real world. A count that falls because people stopped replying looks identical on a chart to a count that falls because women left the profession. Those are opposite findings. They produce the same line.
The regulator and the Society are working from this data. So are the committees that ask whether diversity initiatives are working, and so is anyone who wants to argue that they are or are not. With more than a third of the ethnicity data missing, the honest description of most of those conclusions is that they are estimates wearing the clothes of counts.
Why the answers stopped coming
Chancery Lane does not speculate, so anything I offer here is my reading rather than the report's.
One candidate is accumulation. Data is renewed in cycles, and a person who skips one cycle has little reason to re-engage at the next, so a slow leak becomes a flood over three years. A second is that silence costs nothing. There is no sanction for a blank field, the questions are personal, and a regulator is not the obvious recipient for personal information in an era when everyone has become more careful about what they hand over.
The asymmetry is the part I find genuinely odd. If the cause were a general reluctance to disclose anything to a regulator, sex and ethnicity should have drifted together. Sex non-disclosure moved nearly three times as fast. That points at something specific to the sex question rather than at data collection generally, and I don't think anyone has established what it is yet. A monitoring form asks people to settle a live public argument in a tick box, and some of them would rather not.
There is a third explanation, and it is not about method at all. It is about consent, and it needs its own section.
The compulsion problem
I want to be careful here, because the aim is not the argument. Nobody reasonable objects to a profession being open to people who can do the work, and the profession is better for the change of the last thirty years.
What I object to is the mechanism. EDI arrives at a small firm as an obligation rather than a choice: a policy expected at assessment, an objectives plan, a narrative on progress, a named person accountable for all of it, and a set of questions on a form that has to be chased to a number. Very little of that is asked for by clients. It arrives because a framework says it should, and the cost of not doing it is a finding at assessment or a question in a tender that then has to be answered in the vocabulary the questioner expects.
That is compulsion, and it is worth naming as compulsion rather than describing it as good practice. A firm of nine people does not sit down and decide it needs an EDI strategy. It receives one.
The second objection is competence, and this is where I part company with the framing altogether. Whether men and women disclose at different rates, whether disclosure varies by region, by practice area or by the age of the firm, whether the composition of the sector reflects entry, retention or something further upstream — these are societal questions. They are research questions, with methods, datasets and people trained to read them. They are not questions a solicitor with a full caseload is equipped to answer, and the honest response to "why did this percentage move" is usually "I don't know, and neither does anyone else on this evidence".
We are not social scientists. We are asked to behave as though we were: to write a policy that states a position, measure it, explain the movement, and stand behind the explanation. The output is then aggregated into a national statistic that is more than a third empty, and used to shape decisions about the sector we work in. The evidence is thin, the interpretation is contested, and the obligation to interpret it lands on the people least equipped to do so.
There is a practical consequence, which is why this belongs in a post about blank fields. Compulsion does not produce better information. A question put to someone who did not choose to answer it gets a compliant answer, an invented answer, or nothing. A policy written to satisfy a framework is a document. Ask a firm to demonstrate a culture and it will produce a document asserting that the culture exists, an assessor will read the document, and both parties will understand what has happened.
Some of those 36,094 blank fields are the sound of people declining to take part in something they never asked to join. That number can be chased down with process, and the process will produce a figure. Whether it produces a fact is a different question.
None of this is an argument for doing nothing. The alternative to an imposed policy is not indifference. It is a firm that knows its own people, recruits on ability, deals with behaviour when it goes wrong, and can describe what it does without reaching for a framework's vocabulary to do it.
What this means for a small firm
The aggregate is other people's blank fields until you check your own.
Your return. The diversity data the regulator holds is built from what firms and individuals supply. If completion at your firm is a spreadsheet chased by the office manager in the week it falls due, you are part of the 36%, and the fix is an afternoon's work.
Tendering. Panel reviews and public sector work increasingly ask for workforce data and, more awkwardly, for a sentence explaining it. A blank return is noticed at the point where it costs you instructions rather than at the point where it costs you nothing.
Lexcel. The Standard expects you to demonstrate that equality of opportunity is monitored, not just asserted. Monitoring data that is a third empty is not much of a demonstration, and an assessor is entitled to say so.
The direction of policy. Decisions about where supervision effort and support go are being taken on a picture with a sizeable hole in it. Small firms are usually the group most affected by those decisions. When the next consultation arrives, it is fair to ask what the evidence base actually contains.
My view
The striking thing is not that solicitors would rather not say. It is that a profession which spends its working week telling clients that contemporaneous records decide outcomes has let its own records go this thin, and has done so quietly, over three consecutive years, without anyone really noticing until the totals were added up.
There is also a commercial point hiding in the methodology. A firm that cannot answer "how many of your fee-earners are women" without adding a caveat has a governance gap rather than a diversity gap. The same is true of "how many of our clients are on fixed fees" or "how many files did we close late last quarter". If the denominator is unreliable, the discussion is about the data and never about the practice, and the practice is where the risk actually sits.
The remedy is not another policy. It is unglamorous, and it belongs to the firm rather than to the framework: ask your people directly, explain in two lines why you are asking and where the answers go, make it take two minutes, keep the answers confidential, and then do it again next cycle. That produces a number you can trust because someone chose to give it to you. The thing that produced a 36% gap was not one decision. It was three years of nobody owning the task, because the task belonged to a framework rather than to anyone in the building.
I would rather answer for the quality of the work and the fairness of the decisions than for a policy I did not write, measuring a concept I cannot define, in service of a question that properly belongs to sociologists. That is not a refusal to be decent to people. It is a refusal to pretend that a document produced under obligation is the same thing as a practice.
No one is going to be sanctioned for a blank ethnicity field. That is precisely why the field stays blank.
If you would like an independent view of where your firm stands on compliance, including the evidence you would need to show an assessor, I offer a complimentary health check with no obligation. Get in touch and we can arrange a time to talk it through.
The contact page is an enquiry form, so please put "free health check" in the message subject and I will pick it up straight away.
Ashley Barwick, Fraxinus Business Associates. This post is general information and does not constitute legal advice. Fraxinus Business Associates is not authorised to provide legal advice.
Sources
Law Society Gazette, "Huge rise in solicitors refusing to reveal sex and race", Paul Rogerson, 13 November 2026.https://www.lawgazette.co.uk/news/huge-rise-in-solicitors-refusing-to-reveal-sex-and-race/5128527.article
Law Society, Annual Statistics Report 2026 (SRA data, August 2024 to July 2025).
Further reading on this blog:"Lexcel Accredited Firms: Hundreds Utilising Incorrect Logo and Quality Mark Details""The FCA takeover of AML supervision reaches the Commons: what the Law Society's warning means for a small firm"





Comments