Imagine you're running a food truck. Each meal costs you $8 to make, so charging $10 seems rational — $2 profit per plate. But now imagine you're running a cafeteria where the kitchen, tables, and building cost $100,000 whether you serve 10 people or 1,000. Suddenly, charging $10 per meal to cover costs backfires: the high price drives away 900 people, and your fixed costs are now spread over 10 diners at $10,000 each. Dropping the price to zero, funded externally, fills the room and drops per-diner cost to $100. That's the core mechanism of this paper, applied to piped water networks. The committed claim: for piped water infrastructure in rural Bangladesh, the price that minimizes cost per connected household is zero. Not low — zero. The authors construct an actual pipeline network and use an incentive-compatible willingness-to-pay elicitation mechanism (a Becker-DeGroot-Marschak variant where stated WTP determines real connection offers) to measure demand. They then combine this demand curve with actual project cost data showing strong economies of scale in the network. The result is that the revenue gained from charging even modest user fees is overwhelmed by the cost increase from serving fewer households over the same fixed infrastructure. This directly attacks a standard assumption in development economics and infrastructure policy: the equity-efficiency trade-off. The conventional wisdom says subsidies help the poor (equity) but sacrifice cost recovery (efficiency). The paper argues this framing breaks down when marginal costs are dwarfed by fixed costs that decline per-unit with scale. In such regimes, universal free access IS the efficient outcome, and equity comes for free. The demand-side finding is itself noteworthy: willingness to pay is highly price-sensitive near zero. This is well-documented across developing countries for water, sanitation, and electricity connections. Many households value the service but drop out at even small positive prices. Combined with the cost-side economies of scale, this creates a regime where the demand elasticity near zero is the fulcrum — small price increases cause disproportionate disconnection, which in turn causes disproportionate per-unit cost increases. The validation is grounded in a real infrastructure project with real cost data and incentive-compatible preference elicitation — not a simulation or a thought experiment. That said, the generalizability hinges on two conditions the authors themselves identify: (1) significant economies of scale, and (2) high demand sensitivity to price near zero. Both hold for many network infrastructure settings in developing countries, but the quantitative thresholds will vary. The practical implication is stark for any policymaker setting connection fees for piped water, electricity grids, or sanitation networks in contexts with high fixed costs: user fees may be actively counterproductive at the margin, raising per-household costs rather than defraying them. The paper doesn't say all infrastructure should be free — it says the assumption that positive prices always improve cost-effectiveness is wrong when economies of scale are strong enough. The obvious next step is testing this framework on other network goods — rural electrification, broadband, sanitation — where economies of scale and price sensitivity near zero may similarly hold. The authors don't run those comparisons, likely because each requires its own field experiment with real cost data, which is expensive and slow.
