The Assize

The other regulated legal professions of England and Wales, on one footing

Home › Analysis

Two hundred and seventy-five FOI requests to three legal bodies, and what they produced

The register this site cannot collect has already been handed over once: 1,844 regulated immigration advice organisations, 29% of them at a London postcode. Half the requests to that regulator have no recorded outcome at all.

By Rose Traviss · 29 September 2026

Of the six legal regulators whose annual reports this site reads, one is a public authority under the Freedom of Information Act. Knowing that is not the same as knowing what it has been asked.

So every request on the WhatDoTheyKnow page of all three covered bodies was read: the Office of the Immigration Services Commissioner’s 114, the Legal Ombudsman’s 108, and the Legal Services Board’s 53. Two hundred and seventy-five in all, carrying 379 released files between them. The 172 hanging off the 83 requests that touch something this site is short of were downloaded; the rest — printer fleets, net zero spend, immigration statistics — were catalogued and left where they are.

What happened to them

RequestsSuccessfulPartlyLong overdueNo recorded outcome
Immigration Services Commissioner11433131051
Legal Ombudsman10844161721
Legal Services Board531681213

The largest single category at the immigration regulator is no recorded outcome: 51 of 114. WhatDoTheyKnow classifies a request only when somebody says so, and a request nobody came back to stays unclassified for good. So the honest summary is that rather less is known about what these bodies released than the successful column suggests.

Twelve requests to the Legal Services Board and ten to the immigration regulator are recorded as long overdue.

The one that fills a gap

The immigration regulator’s adviser search will not enumerate. There is no register file to read, which is why this site has no register for that profession while it has one for the other five.

It has handed the register over, though. A request in 2020 asked for the register as it stood on 1 January that year, and the regulator released it as a spreadsheet: eleven columns, 1,845 rows.

1,844 regulated organisations, with a postcode against every one of them. 1,534 distinct organisation names. A telephone number against 1,612 and an email address against 1,163 — none of which is reproduced here, or anywhere on this site, for any source.

What can be said from it is where regulated immigration advice comes from, which nobody publishes:

Postcode areaOrganisationsTown as givenOrganisations
E95London496
B83Birmingham64
W82Glasgow46
SE76Manchester41
N71Harrow25
EC57Ilford23
SW53Sheffield21
G53Bradford20
HA52Croydon20
NW49Leicester20

529 of the 1,844 — 29% — are at a London postcode. The town column undercounts that badly: it gives London 496, because Harrow, Ilford and Croydon are London and are typed as themselves.

The regulator’s own annual report gives 1,615 regulated organisations at 31 March 2020, eleven weeks after the snapshot. The released file has 1,844. The difference is not explained in either document, and the earlier reports split the register into “registered” and “exempt” organisations in a way that would account for a gap of about that size — but neither the release nor the report says so, and no reconciliation is offered here.

What else is sitting there

Among those 83 requests:

None of that has been merged into this site’s figures. Knowing it exists, and where, is the useful first step; deciding what can responsibly be published from a third party’s FOI release is a separate question from whether the file can be downloaded.

How it was collected

WhatDoTheyKnow sits behind a challenge that will not clear for an ordinary HTTP client — curl gets 403 and an “attention required” page every time. A real browser clears it in about a second, so the fetching was done by one, and by nothing else: a script that takes a list of addresses, saves what comes back, and stops. Every figure above was then computed from that cache in Python, so re-doing the sums never launches a browser and never touches the network again.

The site’s robots.txt is unusually precise and was treated as the specification. Authority and request pages are allowed and were used. Search, feeds, profiles and bulk-download paths are disallowed and were not touched — which is why the request list was walked through each authority’s own pagination rather than through the search that would have been quicker. Response pages are disallowed except attachments, which are explicitly allowed, so attachments were fetched and the response pages they hang off were not. Three seconds between requests, no concurrency, and the browser refuses to run without a list naming every address, so it cannot follow a link out of scope.

One thing had to be tuned. A long-lived browser session gets challenged and a fresh one does not: the first pages come back in about a second each and then every page starts costing a minute and a half in challenge waits. Rotating the session every three pages fixed it, and is gentler on the site than repeatedly tripping its protection.

Of 275 request pages and 172 attachments, none failed.

The asymmetry, again

The five regulators this site reads that are not subject to the Act all publish enumerable registers. The one that is subject to it does not publish one — and can be made to hand it over, and has been.

Neither half of that was designed. It is the residue of which statute happened to create which body, and it means the profession whose consumers are least able to shop around is the only one where the list of who is allowed to advise them arrives by request rather than by publication.

The figures

Request counts, outcomes and attachment inventory from WhatDoTheyKnow, read 29 September 2026, and published on sources and licences. Outcomes are WhatDoTheyKnow’s own words. Register counts computed from the spreadsheet released in the 2020 request, which is linked there.

A request marked successful was answered, not necessarily answered in full: one of these released its statistics and withheld the policy behind them under sections 31(1)(b) and 31(1)(g), the law-enforcement exemption.