SROI on a Rural Sanitation Programme: Missing Baseline and a Fund-Mandated Ratio
A Section 8 not-for-profit has completed a three-year rural sanitation programme across 60 villages in eastern Uttar Pradesh. The programme cost INR 9.6 crore and installed or repaired 4,200 household toilets, with a community triggering and behaviour-change component in every village. A baseline household survey exists for only 12 of the 60 villages; the remaining 48 began after triggering had already started. The endline survey covered 240 households drawn from all 60 villages, and a spot-check of 90 toilets found 71% functional and in use. The client has asked you for a social return on investment study. Their CSR funder has written into the grant agreement that continued funding depends on an SROI ratio of at least 3.0 to 1. The client has suggested using a financial proxy for time saved of INR 45 per hour, taken from a 2019 study in Maharashtra, and does not want a discount rate applied because "the grant is only for three years".
1. How should you decide which outcomes enter the SROI calculation, given that a baseline exists for only 12 of 60 villages?
Read the worked answer
The starting point is that SROI is principle-driven, not formula-driven. Two principles bite here: understand what changes, and only include what is material. An outcome can only enter the calculation if you have credible evidence of the change and evidence that stakeholders consider it material.
With a baseline for 12 villages and none for 48, you have three defensible routes:
- Scope down. Restrict the ratio to the 12 baseline villages and state clearly that it is not representative of the full programme. This is the cleanest option.
- Reconstruct a baseline using respondent recall or secondary data (NFHS/district health data, pre-triggering village records), disclosed as a weakness with a sensitivity test.
- Report change for the whole programme on endline alone - only where the outcome is unambiguous, e.g. toilet functionality verified by spot-check, which is a stock measure rather than a change measure.
What is not defensible is silently extrapolating a ratio computed on 12 villages to 60 and presenting it as programme-level SROI. Materiality should be set with stakeholders - here, diarrhoeal incidence, time saved, and dignity/security are the candidates; groundwater improvement has no plausible link at this scale and should be excluded.
2. The client proposes an INR 45/hour proxy from another state and no discounting. How do you handle the valuation and the adjustments (deadweight, attribution, drop-off, displacement)?
Read the worked answer
The proxy must reflect the local context and the specific outcome. INR 45/hour from a 2019 Maharashtra study is not transferable to eastern UP without justification: use the prevailing local unskilled or agricultural wage from district-level data, disclose the source, and sensitivity-test a range.
Then apply the standard adjustments, each with a stated basis:
- Deadweight - the share of households that would have built toilets anyway under Swachh Bharat Mission incentives. Given the subsidy environment, this could be substantial and must not be set to zero by default.
- Attribution - the programme operated alongside government subsidy and ASHA/anganwadi messaging; credit the programme only with its share.
- Drop-off - behaviour change and toilet durability decay over time. For a three-year grant you should still model drop-off over the outcome duration you claim, or explicitly cap the claim at the funded period.
- Displacement - limited here, but consider whether repaired toilets displaced spending from other household needs.
On discounting: present value is not optional because the grant length is short; it is required whenever benefits accrue beyond the year of spend. Apply a disclosed rate (commonly 3–6% real for social projects) and show the ratio at 0% and at your chosen rate so the reader sees the sensitivity.
3. The funder requires a ratio of at least 3.0. The client asks you to "tune the assumptions" to get there. What do you do?
Read the worked answer
You decline, and you document the request.
The ratio is an output of defensible assumptions, not a target to be reverse-engineered. The SROI principles of do not over-claim, be transparent and verify the result make assumption-tuning a breach of professional integrity - and given the ratio is written into a funding condition, presenting a tuned figure would be misleading to the funder, which is arguably the party most harmed.
The practical response has three parts:
- Deliver the honest number with a sensitivity table showing how the ratio moves across the plausible range of proxy value, deadweight, attribution and discount rate. If the defensible central case is 1.6 and the optimistic case is 2.3, say so.
- Separate the message from the number. Explain that a ratio below 3.0 does not mean the programme failed; it may mean the funder's threshold was set without reference to local wage levels or the government-subsidy baseline.
- Protect the record. Put the assumption changes you refused in the engagement file, and state in the report that stakeholders may reasonably hold different views on the proxies used.
If the client instructs you to publish a tuned ratio, withdraw from the engagement rather than attach your name to it.