Payment reports connector

The deduction calculator has been a stub since genesis and 21 [open] claims across four corpus nodes turn on it. The recorded blocker was dew-foundation, which is now wired with three sources, and the block did not lift.

It did not lift because the blocker was wrong about where the deduction is. That has now been established exhaustively rather than inferred:

Summary_SFPR, 30 columns The transfer channel is exactly two lines, S - Educational Service Center and T - Other Adjustments. Neither is named for a scholarship or a community school. Detail_SFPR, 58 columns The full formula decomposition, [a] Enrolled ADM through [N] Total Formula Funding. No deduction line. Profile Report, 61 columns None either.

And the one place a deduction could have been hiding is measured. T - Other Adjustments is the only unlabeled line in the report; its negative half is $95.6m across 577 districts, 1.12% of total state support, against Ohio scholarship spending on the order of a billion a year. A deduction channel would have to be ten times the size of the residual it would have to sit in.

So the deduction is not missing from the current model. It is absent from it by design: under the Fair School Funding Plan community and STEM students are funded directly rather than deducted from a resident district. The calculator is right and the corpus’s expectation was wrong.

What is genuinely missing is the era before — the years the deduction governed, roughly FY1998 through FY2021. Those amounts live in the department’s School Foundation Payment Reports, which are a different artifact from the funding calculator: the calculator models a year, the payment report records what was paid. They are also the only source that could settle whether ECOT recovery money returned to the districts it had been deducted from, which the deduction skill names as the sharpest open question in its scope.

Declare dew-payment-reports as a twelfth connector, with status Declared and a recorded blocker, feeding program, education-agency and revenue-stream.

Declared rather than retrievable, because the reports are posted per fiscal year with no index and no stable path, and the years before roughly 2015 are not on the current host at all. That is a retrieval problem and not a parsing one — the workbooks, where they exist, are the same format the wired connectors already read.

The deduction calculator stays a stub, but for a stated reason that is now correct rather than for one that was wrong. Its blocker moves from dew-foundation to this connector.

The [open] claim “could a deduction be inside the negative transfers” is closed by the measurement above and does not wait on this. What waits on it is narrower and clearly bounded: deduct-era amounts per resident district, and the ECOT recovery question.

A twelfth connector also means the registry’s approval list grows again, which is the third time. The list is derived from these decision records precisely so that growth is deliberate; a connector appearing in the registry without one should keep failing the test that caught this.

Leave it undeclared. Rejected: the corpus’s rule is that a gap nobody has written down is indistinguishable from a gap nobody has noticed, and this one is now well enough understood to name. It has a publisher, a scope, and a specific reason it cannot be retrieved.

Fold it into dew-foundation. Rejected: that connector retrieves the department’s model, and the payment reports are a record. The corpus keeps that distinction everywhere else and the two would have different blockers, different vintages, and different failure modes.