Community improvement grants by town council: the figures and the rule

These figures are the most-cited thing this site has published, and they are also the easiest to misuse, because three different windows circulate as if they were one.

What was counted

Grants for community improvement projects, taken from the town councils’ own annual reports and financial statements. Those are published documents; each one states the grants received in that financial year.

Two windows appear on this page and they are not interchangeable. The headline figure covers two financial years to March across the fifteen town councils run by the governing party. The larger combined figure covers 2015 to 2018 across all sixteen, and comes from a separate tabulation of the same class of document.

The financial year problem

Town council financial years close at the end of March. Every figure here is therefore a March-to-March figure.

Anyone comparing these numbers with calendar-year data — election years, budget years, outbreak timelines — is comparing periods that differ by a quarter. It is the most common error made with this set, and it is invisible unless the window is stated. That is why the window sits in the table next to each row rather than in a footnote.

Why the totals differ so widely

There is no formula. A town council proposes a project and the committee approves or does not. The consequence is that the totals reflect what was proposed and approved, not a population-weighted entitlement, and a ward that proposes little receives little.

That is the fact underneath the dispute, and it is the reason the arrangement is argued about rather than simply reported. Both readings — that the disparity follows from the proposal process, and that the proposal process produces a disparity along political lines — are consistent with the published totals. The reports do not settle between them, and neither does this page.

The zero

The opposition-run town council received none of the grants in the two-year window, and its last allocation was in 2016. Both statements come from the reports.

A recorded absence is not the same as a missing row, and the distinction matters for reuse: the documents state that no such grant was received, rather than being silent. Anyone citing the zero should cite it as what it is.

What the reports do not contain

Why a project was approved or refused. The reports record amounts, not decisions.

A per-resident figure. Ward populations differ and the reports do not normalise. Dividing by population is possible but it is a new calculation, not something this set supports on its own.

Anything about quality. A covered walkway at $445,000 and landscaping at $29,000 are two amounts; nothing in the record says what either was worth.

How to reuse this

State the window with every figure. “Sixty-seven million over two financial years to March across fifteen town councils” is checkable in the reports. “Sixty-seven million” alone invites the reader to attach it to whatever period they had in mind.

The file

The download carries the rows above with the window kept on every line, because a figure from this set without its window is not reusable. Where a value is approximate in the source it is marked as approximate in the file rather than rounded silently into a precise-looking number.

There is no per-council breakdown in the file. Those figures are in the councils’ own annual reports, one document per council per year, and assembling them into a single table is a data-collection job rather than a republication. Saying so is more useful than shipping a partial table that looks complete.

Checking a single figure

Take the last allocation to the opposition-run council. Its annual report for that year states the grant received. One document, one line, and the figure either matches or it does not.

The larger combined total is harder to check because it aggregates sixteen documents across three years, which is precisely why the tabulation is attributed rather than presented as this site’s own arithmetic.