London, filed short.

City Hall’s budget cycle is the spine of the London year. Most of what happens to a bus route, a library or a pedestrian crossing is settled long before anyone stands at a lectern. We read the draft consolidated budget, the assembly’s scrutiny letters, and the written answers that arrive in batches late on a Thursday. The current draft runs to 214 pages, with a capital programme that assumes receipts the boroughs have not yet confirmed.

The assembly filed 386 written questions last month. Most will be answered with a paragraph and a link; a handful will contain the sentence that changes a story. Our method is dull and reliable: read the papers before the meeting, note the discrepancy between the paper and the answer, and ask again in writing. When a figure moves between drafts, we publish both versions and the date each was signed.

We do not treat a mayoral press notice as a decision. A decision has a signature, a date and usually a legal reference; until we have seen one, the line stays “proposed”.

Transport for London publishes more operational data than most city desks can digest: asset condition registers, planned closure notices, bus consultation outcomes, and a monthly performance report that is longer than some novels. We read the appendices. The difference between a “planned” and an “unplanned” closure matters, because it decides who pays for the replacement buses and which contractor carries the cost.

Our standing questions are simple. Which assets are past their design life? Which renewals slipped a year, and by how much? What does a consultation actually change, as opposed to acknowledge? On the Bakerloo file, the answers run to a £240m gap and a four-year difference between the cheapest and the most expensive option. On the buses, they run to a diversion order extended twice, and a route that now spends eleven minutes longer on the road than it did in the spring.

We ride the routes we write about, at the times the documents claim are busiest. It is not a scientific sample. It is a check on whether the timetable describes the service.

Housing is where London’s numbers stop being abstract. The county courts publish possession lists; the planning portal publishes applications and appeals; the boroughs publish temporary accommodation spend, usually late and usually in a different format each time. We put the three together, because a possession claim, a planning consent and a budget line are often the same story told by three offices that do not read each other.

Last week’s tribunal determinations ran to 41 cases. Two-thirds were adjourned or withdrawn; of the rest, most turned on paperwork rather than rent arrears. On the planning side, an appeal for 38 homes on a former depot site in Southwark was allowed with eleven pages of conditions, three of which concern a bus stop that does not yet exist. We publish the reference, the stage and the deciding body, so the file can be checked without taking our word for it.

We do not name tenants. Possession is a public process and a private disaster, and the record does not require both.

The cause lists are the most under-read documents in London. Magistrates’ lists, the Royal Courts of Justice daily list, licensing reviews and planning statutory challenges all appear in public before anyone writes them up. We read them each morning and note what moved: a judicial review permission granted, a licensing decision reserved, a case that has been listed four times without a hearing.

Our rule on names is narrow. We name public bodies, companies and offices as a matter of course. We name individuals only where the case is already in open court and the identification is material to the public interest; we do not name private tenants, witnesses or family members. Where a judgment is reserved we say so, and we do not report the mood of the room as if it were a finding.

Court copy is checked twice: once against the list, once against the judgment or the clerk’s note. If neither is available, the item is held.

The Square Mile produces more paper than any other square mile in the country. Consultation papers, listing notices, syndicate results and enforcement decisions arrive daily, most of them written to be skimmed and filed. We read the parts that change behaviour: commencement dates, fee thresholds, the small print on reporting deadlines.

The desk covers London’s financial services as a local beat, not a market report. That means the effect on employment in the boroughs, on office occupancy in the City, and on the firms that supply both. We do not give investment advice, we do not report share tips, and we do not treat a company announcement as a fact until the filing confirms it.

A consultation closing date is a diary item, not a story. The story is what changes between the consultation and the policy statement, which usually arrives nine months later with a different title.

London’s arts coverage is mostly preview and applause. We cover the money and the dates: Arts Council decisions, borough culture budgets, venue closures and the programme announcements that follow a funding review. A season is a schedule and a balance sheet before it is a set of reviews, and the schedule is usually the first thing to change.

We keep a running list of venues with a capacity under 300, because those are the rooms that disappear quietly and take a borough’s night economy with them. When a South Bank institution confirms a spring programme after six months of review, the useful detail is not the headline show but the number of weeks the building will be dark.

We do not review openings. We do report what a programme costs, who paid for it, and what happens to the building between seasons.