Unintended Consequences of Sanitation Investment: Negative Externalities on Water Quality and Health in India

Scientific Impact Assessment

Core Contribution: This paper provides the first causal estimate of the negative water-quality externalities generated by large-scale latrine construction, in the context of India's Swachh Bharat Mission (SBM)—the world's largest sanitation program (100+ million latrines, ~$6B). The central insight is that scaling household latrine construction without complementary fecal-sludge treatment infrastructure displaces fecal contamination from open-defecation sites to rivers, offsetting direct health benefits. The paper finds a 72% increase in river fecal coliform, yet a net 10% reduction in diarrheal child mortality, with the positive health effect fully eliminated where upstream wastewater treatment capacity is low. This reframes a policy widely considered unambiguously beneficial as one with important conditional trade-offs.

Methodological Rigor: The empirical design is credible and thoughtfully constructed. The instrument—Available Water Capacity (soil infiltration proxy) interacted with a post-SBM indicator—is grounded in an official technical guideline that made latrine construction costlier in high-infiltration soils after 2014, yielding a plausible first stage (F-stats 30–79). The upstream–downstream specification is an elegant strengthening of the exclusion restriction: using upstream soil characteristics to instrument upstream latrine construction while examining downstream outcomes plausibly rules out the agricultural-income confound. The author supports identification with event-study pre-trends, falsification tests on nitrate/nitrite, temperature, BOD/DO, and child overweight, plus an alternative DiD design, spatial-spillover checks, urban-exclusion checks, balanced-panel checks, and an alternative NFHS mortality dataset. The heterogeneity analysis by STP treatment capacity convincingly isolates the fecal-sludge-dumping mechanism. Weaknesses: the estimates are LATE-driven by "backward" complier districts (limiting external validity); administrative latrine counts are likely overreported (though this biases toward zero); and several upstream–downstream health estimates are imprecise or hover near zero, making the offsetting-externality decomposition partly reliant on strong assumptions (Appendix D.1 acknowledges this).

Potential Impact: The topic is highly policy-relevant. Sanitation is a core SDG target, and the finding that latrine expansion needs coordinated wastewater infrastructure has direct implications for India (SBM Phase 2 already emphasizes fecal-sludge management) and comparable programs in Bangladesh, Nepal, Pakistan, and China. The cost–benefit analysis (net mortality benefit ~1/3 of subsidy cost without treatment; complementary STP investment yielding benefits exceeding costs) offers concrete guidance. The broader conceptual message—incentivizing private goods without complementary public goods can generate offsetting externalities—connects to a growing "unintended consequences of development interventions" literature (deforestation from CCTs, air pollution from rural roads).

Timeliness & Relevance: Addresses a genuine bottleneck: much prior sanitation evidence comes from village-level RCTs that cannot capture externalities from nationwide scale-up. The paper directly speaks to the gap between local RCT effects (e.g., Geruso & Spears' 48% mortality reduction) and the more modest scaled-policy effect documented here.

Strengths: (1) Novel, important, and clearly-framed question; (2) credible and multi-layered identification strategy; (3) rich integration of administrative water-quality, mortality, soil, and GIS datasets; (4) policy-actionable cost–benefit framing; (5) a conceptual model tying predictions to empirics. The paper won the 2022 Moriguchi Prize.

Limitations: (1) External validity limited by LATE and India-specific setting; (2) latrine data quality and lack of usage data; (3) water-quality monitoring restricted to districts with stations along rivers (sample is more populous/rural than average); (4) imprecision in key upstream–downstream health estimates; (5) no released replication code (though data sources are meticulously documented).

Other observations: The paper corrects/qualifies the widely-held assumption that sanitation subsidies uniformly improve child health, and offers a plausible reconciliation of Greenstone & Hanna's (2014) limited water-regulation effects. It is a self-contained empirical contribution rather than a reusable methodological primitive, though the upstream–downstream IV template is transferable to other pollution-externality settings.

This is a strong, well-executed applied-micro paper likely to be published in a good field or general-interest economics journal and cited across development, environmental, and health economics.