Product preferences and willingness to pay for potable water delivery: Experimental evidence from rural Bihar, India

Abstract

Despite dramatic reductions in global risk exposures to unsafe water sources, lack of access to clean water remains a persistent problem in many rural and last-mile communities. A great deal is known about demand for household water treatment systems; however, similar evidence for fully treated water products is limited. This study evaluates an NGO-based potable water delivery service in rural Bihar, India, meant to stand-in for more robust municipal treated water supply systems that have yet to reach the area. We use a random price auction and discrete choice experiment to examine willingness to pay (WTP) and stated product preferences, respectively, for this service among 162 households in the region. We seek to determine the impact of short-term price subsidies on demand for water delivery and the extent to which participation in the delivery program leads to changes in stated preferences for service characteristics. We find that mean WTP for the first week of service is roughly 51% of market price and represents only 1.7% of median household income, providing evidence of untapped demand for fully treated water. We also find mixed evidence on the effect of small price subsidies for various parts of the delivery service, and that one week of initial participation leads to significant changes in stated preferences for the taste of the treated water as well as the convenience of the delivery service. While more evidence is needed on the effect of subsidies, our findings suggest that marketing on taste and convenience could help increase uptake of clean water delivery services in rural and last-mile communities that have yet to receive piped water. However, we caution that these services should be seen as a stopgap, not a substitute for piped municipal water systems.

Fig 1

CONSORT diagram–survey households allocated into Auction 1 for combined purchase of bottle, dispenser and deliveries (left) or Auction 2 for purchase of bottle and dispenser only (right); number of households in parentheses.

Fig 2

Treated water demonstration–Drinking water 2-minutes after crushed guava leaves are added to untreated ground water (left) versus treated ground water (right).

Fig 3. Decision tree for willingness to pay (WTP) under each auction.

Fig 4

Example DCE choice set, status quo (left) random alternative (center) with scripted description of the choice set faced by respondents (right).

Fig 5

WTP for hardware and 7 deliveries (top) and hardware only (bottom) among all bidders (left) and only auction “winners” (right).

Fig 6

Willingness to pay for hardware and seven water deliveries among all bidders (left) and among winners who purchased (right)–Auction 2 participants scaled to include delivery cost drawn.