FCA’s 97.9% Score Hides Five Weak Authorisation Areas

The Financial Conduct Authority’s headline is accurate, but only after reading its denominator correctly. In the three months to June, 5,886 of 6,014 unique cases were determined within the applicable deadline. That is 97.87%, which rounds to the reported 97.9%.

The result does not mean that every part of the authorisation system ran close to 98%. Five of the 12 published category rows were below the FCA’s green band. A simple average of the 12 percentages is 93.1%, but that gives a nine-case money-laundering category the same weight as a 1,652-case payment-services-agent category.

The case-weighted figure is the better aggregate. It is also dominated by high-volume work that performed well, while smaller and more consequential entry routes carried most of the visible weakness.

The denominator needs one adjustment

Adding all 12 rows in the FCA’s Q1 table would produce 6,066 cases and a 97.8% on-time rate. That would double-count 52 adjacent variations of permission. The adjacent row is a subset of the broader 416-case variation-of-permission row, not a separate case population.

Removing that overlap leaves 6,014 cases and 128 deadline misses. The scale effect is large: approved-person applications and payment-services-agent notifications together supplied 4,335 cases, or 72.1% of the denominator. Their three rows scored 98.3%, 97.7% and 100%.

The five below-green rows comprised 454 cases, 7.5% of the unique total, but recorded 55 deadline misses. They therefore generated 43% of all misses despite their small share of volume.

A threshold change moved two rows to green

The current traffic lights count seven green, two amber and three red rows. Since the third quarter of 2025/26, green has started at 95%, down from 98%. Amber covers 90% to below 95%, while results below 90% are red. Change in Control remains an exception and must score 100% to be green.

Applying the former 98% line to the current results would produce five green, four amber and three red. Approved Person applications linked to appointed representatives, at 97.7%, and the main Variation of Permission category, at 97.4%, are the only rows whose colour changes. Change in Control stays green because all 277 cases met its 60-working-day deadline.

That makes the presentation more favourable, but it does not change a single applicant’s wait. Nor does it explain the three red categories, all of which would have been red under either threshold.

The AML correction is small in aggregate and large within the categories

The more substantive revision concerns the clock. The FCA said it had previously calculated completion dates for 3/4MLD and 5MLD registrations using working days rather than calendar days. Its 2025/26 annual metrics page removed both rows pending correction. The Q1 release then restated data back to the second quarter of 2025/26.

For 3/4MLD, the originally published results for those three quarters were 100%, 98.4% and 98.3%, with two misses among 192 cases. The corrected record is 85.7%, 87.5% and 84.5%, with 27 misses. That reclassifies 25 cases as late.

For 5MLD, three successive 100% results became 66.7%, 66.7% and 71.4%. Four of 13 cases were reclassified as late. Together, the correction adds 29 misses among 205 AML decisions. That is only 0.14% of the 20,305 authorisation cases closed across the same three quarters, but it materially changes what an AML applicant would infer from the category history.

The current elapsed times add applicant context, but they are not a direct test of statutory compliance. The 45-day 3/4MLD and three-month 5MLD clocks apply to complete applications. By contrast, the published quartiles run from receipt to determination and include an undisclosed mix of complete and incomplete files. They show that 3/4MLD applicants waited 75 days at the median and 91 at the upper quartile, while 5MLD applicants waited 236 and 297 days. Those are long end-to-end waits, but the table cannot show whether the typical complete application missed its statutory clock. The 5MLD sample is also only nine cases, so it cannot establish system-wide friction.

Three weak rows use faster voluntary standards

The other weak rows need a different reading. The FCA began working voluntarily towards faster targets ahead of legislation. New firm applications are now measured at four months when complete and 10 months when incomplete, rather than the previous six and 12 months. Adjacent variations use an expressly voluntary three or six-month target. The payment and e-money row uses three months for complete applications and 10 months for incomplete ones, tightening the previous 12-month outer limit for incomplete cases.

The published medians combine complete and incomplete applications, so they cannot show whether a typical complete applicant breached the shorter limb. Adjacent variations look least alarming: their 44-day median and 69-day upper quartile both sat inside the tighter three-month target. New firms took 131 days at the median and 190 at the upper quartile. Payment and e-money applicants took 219 and 268 days respectively, but the mix of complete and incomplete files is not disclosed.

The scorecard therefore identifies concentrated friction, not a blocked UK entry system. Its strongest warning is in the corrected AML categories, where compliance rates are weak and end-to-end waits are long, although the quartiles cannot identify the typical complete case. Elsewhere, the traffic lights partly record the FCA choosing a harder voluntary benchmark. For applicants, the useful questions are not whether the aggregate is green, but which clock applies, whether the file is complete and how long comparable cases actually took.

Sources

• FCA Authorisations operating service metrics 2026/27 Q1
• FCA Q1 2026/27 metrics table
• FCA Authorisations operating service metrics 2025/26 Q4
• FCA Q4 2025/26 metrics table
• FCA operating service metrics 2025/26