Five year forecast connector

Decision proposals approved nine connectors and report-card-connector a tenth. Every one of the ten retrieves either a modeled figure — what a formula computes, what a district spent per pupil on the department’s own definitions — or an outcome. None retrieves a record of money arriving in a bank account, and none says what a district holds.

That gap surfaced from a direct question: can this corpus show carry-over funding and cash on hand? It could not, and neither could it show a district’s actual state aid in any year. The FY2027 funding calculator is a single-year model carrying no prior-year amounts, and the department replaces rather than archives it, so a year-over-year comparison of published funding was unavailable from the sources already wired.

The five-year forecast is where both live. Ohio districts file under R.C. 5705.391 and the department republishes every filing at https://public.education.ohio.gov/School District Five-year Forecasts/ — an open directory, no key, no form, 70 files back to 2008. Each filing carries three years of audited actuals alongside its five forecast years, so two filings three years apart tile six consecutive closed years without a gap.

What it reaches that nothing else here does:

7.010 / 7.020 Beginning and ending cash balance — carry-over and cash on hand, the only measure in this corpus of what a district holds rather than what it is given. 1.035 Unrestricted grants-in-aid: state foundation money as the district books it, per district per year. FY2020 is the year the temporary transitional aid guarantee holds districts at, and it had until now only been inferrable from the guarantee itself, which made every check of the guarantee circular. 1.010 / 1.030 Property and income tax actually collected — the local side as revenue. tax-abstract would have supplied valuation and remains unbuilt.

Approved. An eleventh connector, dew-five-year-forecast, publisher Ohio Department of Education and Workforce, feeding education-agency, revenue-stream, metric and fiscal-period, status Wired, with two sources: the FY2023 and FY2026 required spring update submissions. The registry test that pins the connector list is updated from ten names to eleven.

It is a separate connector rather than more sources under dew-foundation, on the same reasoning that separated dew-report-card: different infrastructure (public.education.ohio.gov rather than the CMS), a different release calendar, a different filing obligation, and a different epistemic status. Merging them would put a treasurer’s submission and the department’s own calculator behind one status flag.

Two constraints are approved with it, and both are enforced in code rather than left to discipline.

Only the actuals are extracted. Each row carries the filing year’s own forecast and four future ones. Those are a district treasurer’s projection made under incentives — a board arguing for a levy and a board defending a balance want different numbers on the same line — and this repository already has a projection engine whose assumptions it can state. connect::forecast::Line::current is parsed and named so that a caller who wants a submitted forecast has to ask for it, and the committed fixture contains none.

The parser is header-driven and refuses a headerless file. The department has published this file under three layouts: headered from FY2022, headerless with a different column order from FY2013 to FY2021, and zero-padded fixed-width before that. A positional parser would read a district name as a line code on the second era and return an empty result, which looks exactly like a clean parse. Extending the panel before FY2020 is a per-era reader and is not approved here.

The cost is the same one report-card-connector weighed and it is now larger: the connector count is a rough proxy for maintained surface, and the nine at genesis were chosen to bound it. Eleven is a second deliberate expansion.

What justifies it is that the corpus’s central subject is a guarantee, and until now the corpus could not observe the year the guarantee is defined against. Every statement it made about whether the guarantee holds districts whole was derived from the guarantee’s own output. Two filings and an afternoon turn that into a measurement, and the measurement disagrees with the plain reading: the median guaranteed district receives 89.9% of what it booked in FY2020, in nominal dollars, while the median formula-funded district receives 118.6%. That comparison is only available because both groups can now be measured the same way against the same source — which is precisely the argument the provenance rule makes for holding data rather than quoting it.

The narrower option — pinning only the FY2026 filing, which alone answers the cash-on-hand question — was rejected. It would carry FY2023 to FY2025 and miss FY2020 entirely, and FY2020 is the year that makes the guarantee checkable.

The broader option — parsing all three eras back to 2008, giving a twenty-year panel — was deferred rather than rejected. It is a per-era reader against layouts with no header to validate against, and the pre-2020 years would arrive without the FY2027 funding model to join them to. Recorded in the catalog entry as what a later phase would take up.

A note against a future reader’s assumption: approving this connector does not make the general fund the district’s budget. Capital, food service, and most federal programs sit in other funds, so an expenditure figure from this source is not total spending and must never be compared to a per-pupil expenditure figure from the profile report without saying so.