Most English learners (ELs) eventually gain sufficient English proficiency to be reclassified and receive instruction without linguistic supports. Though well-identified, prior regression discontinuity estimates for the effect of reclassification are estimated too imprecisely to detect policy-relevant effects. Applying a student fixed-effect design to data from Indiana, we show that ELs experience a substantial negative shock immediately following reclassification, and this negative impact remains over time. We apply insights from a stylized model for changes in ELs outcomes as they gain facility with English to bound the causal effect of reclassification. A difference-in-differences approach making across-student comparisons between treated students and not-yet reclassified ELs yields similar results. Our analysis illustrates the potential for panel data strategies with transparent assumptions to produce valuable information about the effect of important but understudied services in cases where experimental and quasi-experimental strategies are unavailable and/or uninformative.