How would we know assisted access worked?

The question

Does assisted welfare access at labour nakas increase successful benefit receipt among informal interstate migrants in three Mumbai wards?

Randomising help desks across nakas is possible in principle and politically expensive in practice. The strategy paper already wants a comparison ward and quarterly funnel re-surveys. That points to a difference-in-differences with a matched comparison ward, plus a process evaluation. An RCT is a later instrument, if a partner can actually randomise.

Treatment: naka help desks and case management, as in Option A. Outcome: successful receipt (rations, BOCW, insurance, pension) among a defined segment, not registrations. Process indicators: time-to-case-close, refusal grievances filed, dealer exception-handling use if ePoS can be pulled. Identification: parallel trends is an assumption, and a short pre-period from ePoS is the check. Threats: spillover if migrants move between wards; contamination if FCS reforms hit both; selection if desks attract the already-informed. Measurement: household survey plus administrative traces. If the diagnostic in months 0–3 shows the break is at the counter, this design is the wrong one, and the evaluation should follow Option B instead.