Registered Predictions of the Federal EcoULL Implementation Pathway (FEIP): A Pre-Analysis Protocol

This deposit registers, before observation, what the Federal EcoULL Implementation Pathway predicts and what would show it to be wrong. The pathway is a five-stage design sequence - Diagnose, Prioritize, Institutional Design, Pilot, Institutionalize - for building the institutional conditions that an ecology-oriented urban living lab requires. It was derived from the Urban Living Lab Legal Enablement instrument and from two empirical analyses, and it has never been tested. This protocol does not argue that it is correct. It states twelve falsifiable predictions, each with the observation that would falsify it and the date by which it resolves, so that a later assessment can be checked against a dated record rather than against a reconstruction. The design uses two test sets. The primary set is the United States: three metropolitan regions scored under the instrument, and fourteen federal programme streams coded for durability class and evidence-integration level. The out-of-sample set is Istanbul. The pathway was built on United States federal instruments, and Turkiye has no equivalent of a court-supervised consent order; if the pathway's claims hold only where that instrument exists, they are claims about American administrative law rather than about institutional design, and that distinction cannot be drawn inside a single legal system. In Istanbul the functional equivalent of the consent order is the basin protection plan for the Omerli, Elmali and Alibey catchments: a territorially bounded instrument that binds spatial plans at every scale, assigns implementation to named agencies, carries monitoring duties and specifies its own revision. Two predictions carry the argument. The first holds that no case records substantial epistemic translation without at least one of three instruments present in a formal document: a designated evidence-translation role held by a named office, a revision trigger tied to a stated monitoring threshold, or a data-sharing agreement routing monitoring findings to a named planning decision-maker. The second holds that any new instrument moving a case to substantial epistemic translation will be accompanied by a documented concentrated avoidable cost, and that no case will acquire it on the basis of technical assistance or voluntary adoption alone. That second prediction discriminates between three accounts of the observed absence that present evidence cannot separate. A third prediction resolves early and in the legal system the pathway was not built for. The implementation tables of the Omerli and Alibey basin protection plans schedule revision studies for 2027 and assign them to ISKI, so the observation window is fixed independently of this protocol. The prediction is that those revisions will not introduce a provision tying plan revision to monitoring findings, and that revision will remain calendar-scheduled. The protocol states plainly which of its predictions are cheap. Six of the twelve assert that something will not happen where the base rate of it happening is close to zero, and they are registered because a change would be informative, not because their holding would be evidence for the pathway. It also states the confound that bears on the discriminating prediction: an observation window in which nothing changes anywhere is consistent with political resistance and equally consistent with a window that is simply too short, and this protocol commits in advance to not reporting that outcome as though it settled the question. Files: FEIP_Registered_Predictions_v1.docx, the protocol; feip_predictions_register.csv, the twelve predictions in machine-readable form with falsifiers, test sets, resolution dates and a note on the base rate and weight of each. Scoring follows the ULE codebook and scoring protocol at its concept DOI, which resolves to the current version. The register is not revised after deposit; where a prediction proves ambiguous, a new version is deposited stating what changed, and the original remains citable at its own version DOI.