Suhani Bhatia
← Work

Independent analysis · 2026

On the lists, not at the counter

Where a ₹5 crore grant can change migrant welfare outcomes inside 24 months.

Mumbai · Diagnostic · options · stage-gate · cost and M&E

The decision

A philanthropy has ₹5 crore and 24 months. Government is a partner, not the grantee. How to improve access to existing welfare among low-income informal workers in Kurla L, Govandi M-East and Andheri K-East, with success defined as benefits received. Ten percent for independent M&E. Figures below are my estimates.

Current-state journey

Eligibility in principle is not receipt. Friction concentrates at documents and at the counter. Awareness campaigns do not open those gates.

EligibilityThe rule exists. Knowing it applies to you is a separate fact.
DocumentationRation cards, employer proof, an address a counter will accept.
ApplicationA claim filed at destination, often through an intermediary.
VerificationThe file is checked. Exception handling is rare. Biometrics fail.
ApprovalA name on e-Shram, BOCW, NFSA. Lists can accumulate without a benefit.
ReceiptGrain, cash, or cover in hand. The only stage that should count as success.

Cited evidence · not a Mumbai baseline

Awareness of portability is 71 percent. Receipt is 30 percent. The grant has to buy the drop between those two numbers.
On a scheme list100%Knows portability exists71%Tried to claim51%Actually served30%

CLRA, One Nation Unequal Access (December 2024), n = 1,012 seasonal migrants. Registration at ~100% is an estimate relative to source-state lists. Mumbai’s own baseline is a diagnostic output, not assumed.

CLRA’s 2024 survey of 1,012 seasonal migrants is the closest published funnel, not Mumbai’s baseline. Two breaks do most of the work: documents at destination (68% in that study do not carry a ration card), and the counter even when eligibility exists (interstate PDS portability 0.8%; dealer refusal and unused exception handling in the 2022 Dalberg / Kantar / Omidyar ONORC study).

Two strategic hypotheses

Hypothesis to cost

A. Assisted access

Fifteen naka help desks, WhatsApp and IVR, case-managed claims, a grievance on every refusal. Starts without waiting for a government reform. Stops when the grant stops.

Reach 60,000 and benefits 25,000 are estimates, not observed outcomes.

Hypothesis to cost

B. Systems / counter reform

Dealer incentive for portability, flexible stock, notified exception handling, BMC residence attestation. Cheaper if it works. Depends on an FCS memorandum within six months.

Institutional dependence is the binding constraint, not the idea.

Criteria

CriterionWeightAB
Impact inside 24 months0.2052
Reach / equity for interstate women0.1043
Cost per successful receipt0.1524
Feasibility, government dependence0.1551
Scalability and systems change0.1515
Technology requirement (lower is better)0.0534
Adoption at the frontline0.1042
Measurability0.1054

Scores 1–5, author judgement before the diagnostic. Weighted totals: A 3.75, B 2.95. The matrix is a prior. The gate at month 3 can reverse it if H1 and H2 both favour B.

Recommendation

Start with assisted access

Use the first 90 days to determine whether the binding constraint warrants a shift toward systems reform. Default to assisted access (Option A) in the three wards. If the survey shows the break at served and the MoU is signed, switch to fixing the counter (Option B). Otherwise keep ₹50 lakh to pilot B’s dealer incentive in one ward so the systems option does not die.

Who: a CSO with naka presence, FCS and BMC as partners. Where: Kurla L, Govandi M-East, Andheri K-East. When: diagnostic immediately; operations from month 3. How much: ₹5 crore, 10% independent M&E. Dependency: for B, the MoU. Metric: successful receipts and a rebuilt funnel, quarterly.

The 90-day diagnostic

Estimate · hypothesis

₹30 lakhFirst ninety days, before locking the remaining grant.

90 daysRecruit a CSO, open the FCS conversation, test H1–H4.

  1. H1Does Mumbai’s funnel break at served, or at attempt? Survey n≈1,200 plus ePoS pull.
  2. H2Will Maharashtra FCS sign an MoU on dealer incentives within six months?
  3. H3Can assisted access unlock a benefit for under ₹2,000? Benchmark Haqdarshak, Aajeevika Bureau, MRC.
  4. H4Are interstate migrants the majority of the segment, or do intrastate workers already draw rations in the state?

Twenty-four months

Months 0–3₹30 lakh. Diagnostic on H1–H4. Recruit a CSO. Open the FCS conversation. Survey n≈1,200 plus ePoS pull.
Gate, month 3If the break is at served and an MoU is signed, switch toward Option B. Otherwise default to A, keeping ₹50 lakh to pilot a dealer incentive in one ward.
Months 3–12Fifteen naka desks, case management, grievance on every refusal. Comparison ward running.
Months 12–24Scale or stop. Reset unit-cost target after H3. Independent M&E reports quarterly. North star: rupees of benefits received per rupee spent.

What ₹5 crore buys under Option A

Option A, as costed: ₹3.5 crore for operations, ₹1 crore for technology, ₹50 lakh ring-fenced for independent monitoring.
Operations (desks, casework)₹3.5 crTechnology (WhatsApp, IVR, tracker)₹1 crIndependent M&E₹0.5 cr

Author estimates for an independent strategy model. Independent analysis.

What would be measured

  • North starRupees of benefits received per rupee of grant. Target ≥ 3 after the unit-cost test.
  • OutcomeSuccessful receipts at the frontline: rations, BOCW, insurance or pension.
  • CheckComparison ward and a quarterly funnel re-survey. Help desks that vanish with the grant, or a low interstate share in the diagnostic, would reverse this.

Sources

  1. Dalberg, Kantar, Omidyar Network India, ONORC frontline study (2022).
  2. IMPDS via IndiaSpend (2023), interstate portability 0.8% of PDS.
  3. CLRA, One Nation Unequal Access (Dec 2024).
  4. MPRA 114603 on document carry-rates.
  5. PRS on NFSA list vintage (2024).

12-slide strategy deckCost model CSVDecision matrix CSVWorkbook narrativeEvaluation design