Service Charge Check: findings, 31 Oct 2026

Figures computed from the database on 31 Oct 2026; the data was last updated from GOV.UK on 31 Oct 2026. Each figure carries its n (the number of decisions it was counted from) and how it is counted.

Source: First-tier Tribunal (Property Chamber) decisions in the 'Leasehold disputes (management)' category on GOV.UK,
read automatically (regular expressions, then a language model whose every value must quote the decision verbatim),
then checked against two sets of 40 decisions read by hand (https://rentrisecheck.uk/leasehold/accuracy). Open Government Licence v3.0.

1. Decisions read: 7,007 of 7,287 pages in the category.
   Counted: one page per decision; where a case was published more than once the main decision counts and a later costs
   ruling is linked to it; image-only PDFs and anonymised cases are left out.
2. Service charge decisions (case codes LSC and LIS): 2,595; with a readable outcome: 1,395.
   Counted: a readable outcome is payable as demanded, reduced, substantially reduced, not payable or dismissed. A decision
   whose two readings disagree on the outcome, the section 20C result or a total is listed but not counted (1,387 held back).
3. Charges reduced at all: 833 of 1,395 = about 60%.
   Counted: reduced, substantially reduced or not payable, as a share of the decisions in 2 with a readable outcome.
   "About": on 40 decisions read independently (30 Oct 2026) the automatic outcome matched the reader's in 82%, so this share is approximate until a later check clears it (https://rentrisecheck.uk/leasehold/accuracy).
4. Substantially reduced or not payable at all: 146 of 1,395 = about 10%; not payable at all: 140.
   Counted: more than a third off, as the decision's own figures show, or nothing payable; a share of the decisions in 2 ("about" as in 3).
5. Typical reduction where a charge was cut: median 35% off (n=111); across all 243 decisions
   stating both totals, including those payable in full, the median cut is 0%.
   Counted: only decisions that state both the total demanded and the total found payable as single figures; nothing is
   added up from items or estimated.
6. Section 20C orders (the landlord's costs of the case kept out of the service charge): made in 634 of
   1,016 decisions where the order was asked for and decided = 62%; a further 423 asked for but not decided in the document.
   Counted: decisions of any kind that say the order was made or refused; where the main decision deferred the question and
   a later ruling on the same case decided it, the ruling's result counts. On 40 decisions read independently the automatic result matched in 92%.
7. Dispensation applications (case code LDC, s.20ZA): 2,839 decisions; granted 2,684, refused 49, the rest
   not readable as either. A consultation failure (the statutory cap applies) was found in 196 decisions of any kind.
   Counted: by case code. Dispensation applications are landlord-driven and the largest part of the category; they are never in
   the share-reduced figures in 3 to 5, which count service charge challenges (LSC and LIS) only.
8. Items most often disallowed or reduced: repairs and maintenance 361, other 345, management fees 313, administration charges 279, insurance 252, legal and professional fees 236.
   Counted: items the decision disallowed or reduced, from a fixed list of kinds of charge; a decision can count under several.
9. Managing agents named in two or more decisions: 288 (of 868 named at all); freeholders or landlords:
   497 (of 2,604); management, RTM and residents' companies: 121 (of 1,210).
   Counted: a company is stored only when the regex and the model read the same name, as the decision names it; where they
   disagreed the decision counts as 'not stated' (371 decisions). People are never named.
10. Time from application to decision: median 126 days (n=2,016).
   Counted: days between the application date the decision states and the decision date; decisions that don't state the
   application date are left out, so cases withdrawn or settled before a decision cannot be in it.

A share counted from fewer than 10 decisions is given as a count, not a percentage. Figures are as the decisions state them.
Past decisions are not a prediction: each tribunal decides on the evidence before it. This is information, not legal advice.

How to cite: "Service Charge Check, from First-tier Tribunal (Property Chamber) decisions published on GOV.UK", 31 Oct 2026, https://rentrisecheck.uk/leasehold/stats
The decisions are Crown copyright, published under the Open Government Licence v3.0.
Method and accuracy: https://rentrisecheck.uk/leasehold/accuracy