中文摘要
超过20亿人缺乏清洁饮用水。现有解决方案要么面临高成本（管道供水），要么面临低需求（终端加氯）。我们利用一项覆盖6万户家庭的整群随机试验，检验一种替代方案：针对农村贫困人口实行分散式处理与清洁用水配送到户。在低价条件下，采用率超过90%，且贯穿整个实验期保持稳定。高价会降低采用率，但可实现私人盈利。我们通过实验还原了显示性偏好的价值评估指标。支付意愿比先前的间接估计高出数倍；接受意愿更高，且超过边际成本。自我报告的健康指标相应改善。按每DALY成本计算，免费送水制度显得极具成本效益。（JEL I12, O12, O13, O18, Q25, Q51, Q53）
Abstract
Over 2 billion people lack clean drinking water. Existing solutions face high costs (piped water) or low demand (point-of-use chlorine). Using a 60,000 household cluster-randomized experiment, we test an alternative approach: decentralized treatment and home delivery of clean water to the rural poor. At low prices, take-up exceeds 90 percent, sustained throughout the experiment. High prices reduce take-up but are privately profitable. We experimentally recover revealed-preference measures of valuation. Willingness-to-pay is several times higher than prior indirect estimates; willingness-to-accept is larger and exceeds marginal cost. Self-reported health measures improve accordingly. On a cost-per-DALY basis, free water delivery regimes appear highly cost effective. (JEL I12, O12, O13, O18, Q25, Q51, Q53)