Skip to main content
SpringerLink
Log in

Household Demand for Water in Rural Kenya

  • Published:
Environmental and Resource Economics Aims and scope Submit manuscript

Abstract

To expand and maintain water supply infrastructure in rural regions of developing countries, planners and policymakers need better information on the preferences of households who might use the sources. Using data from 387 households in rural Kenya, we model source choice and water demand using a discrete-continuous (linked) demand model. We find that households are sensitive to the price, proximity, taste, and availability in choosing among sources, but are not sensitive to other source qualities including color, health risk, and risk of conflict. Estimates of the value of time implied by our model suggest that households value time spent collecting water at one third of unskilled wages. We use the linked demand framework to estimate own-price elasticities in the rural setting. These estimates range between − 0.13 and − 1.33, with a mean of − 0.56, and are consistent with other elasticity estimates from small and large cities.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1

Similar content being viewed by others

Notes

  1. A basic water service is a source within 30 min roundtrip of the household which, by nature of its design and construction, has the potential to deliver safe water (WHO/UNICEF 2017).

  2. Since our target sample was 400 households and the most recent census indicated a population of 3005 households in our study site, we targeted approximately 20% of the total population, or every fifth household. In 23 sampled households, the respondents in the household were unavailable so that call backs had to be scheduled. In 15 of these 23, an interview was later completed. Six households declined to be interviewed. Therefore, of the 402 households contacted, 387 were interviewed giving a response rate of 96%.

  3. Many households reported collecting from their “neighbor’s” well or private tap. Often these households reported walking significant distances to these “neighbors” and paying financial costs to collect, so we assume that many respondents were referring to the public sources just described.

  4. We do this for two reasons. First, the survey was conducted in the dry season and we have more confidence in respondents’ recall of water collection choices and trips when asked about the prior 7 days, rather than representative numbers for an “average week during the rainy season” several months prior. Second, many households collect rainwater in small containers during the rainy season and rely less on collection outside the home. Ninety-six percent of households collect rain water during the rainy season; the average household uses 230 l of rain water per week (during the rainy season). Although this seasonality is important for water supply planning, during the rainy season it is more difficult to observe households making the sorts of tradeoffs that help us identify their preferences.

  5. Based on the set of household-source pairs where we collected GPS locations as well as self-reported one-way walk times (we did not anticipate so many households to report collecting from their “neighbors”, and did not take GPS measurements at these sources), we estimate an implied walking speed of 1.7 miles per hour (results available on request). This is in line with other estimates (White et al. 1972; Calvo 1994; Tanser et al. 2006), suggesting the household’s self-reported walk times are plausible. We also calculate the difference in a household’s reported walk time to a given public source and their nearest neighbor’s reported walk time to the same source (we thank a referee for this suggestion). Among the 211 observations for which a household and their nearest neighbor reported having access to the same public source, the average difference (absolute value) in their reported walk times is 7.68 min (median: 2.91, SD 19.64). This represents on average a 12% difference in reported walk times between nearest neighbors (median: 5%), suggesting self-reported walk times are fairly precise.

  6. In practice, conditional demand equations have sometimes been misspecified in the water source choice and demand literature. Using either the Lee correction (Lee 1983) or Dubin-McFadden correction (Dubin and McFadden 1984) methods, analysts should estimate one conditional demand equation for each choice alternative (see original papers, or Mannering (1986), or Wu and Babcock (1998) for examples).

  7. Researchers in the recreational demand literature face similar issues. Here the discrete-continuous choices are which site to visit and how many visits to make. Suppose sites were fishing destinations, and the choice set included a salt-water site and three freshwater lakes. Data aggregation would collapse the choice set to a salt-water site and a freshwater lake, and fail to make use of any information on heterogeneity in lake characteristics like fish stocking, boat ramps, or water quality.

  8. We failed to achieve convergence when we allowed coefficients on the source type dummies to be random parameters.

  9. One explanation might be that perceived quality variables like taste, color and health risk are highly correlated, resulting in poor parameter estimates. We estimated several models that controlled for the modest correlation in our data (less than 0.5) between these quality variables. We also estimated models that added correlated quality variables one by one and in combinations. Finally, we estimated a model using a water quality index generated by principal component analysis. Results are generally consistent with our main model and are available upon request. Another concern might be that source-type dummies absorb the effects of quality attributes if there is little variation of quality attributes within source-types, though it is apparent from the raw results in Table 2 that there is substantial variation in source attributes within source-types.

  10. The reported coefficients for the random parameters logit are the mean of each individual parameter estimate. Note that the mean shadow value of walk time is given by \(60\times mean\bigg(\frac{{\hat{\psi }}_i^{walk}}{{\hat{\beta }}_i^{price}}\bigg)\) which is not the same as \(60\times \frac{mean({\hat{\psi }}_i^{walk})}{mean({\hat{\beta }}_i^{price })}\).

  11. The wealth index is calculated following Filmer et al. (2001) and Filmer and Scott (2012). It includes data on durable assets, electricity connection, sanitation, number of rooms, number of buildings, and main cooking fuel. More information on construction of the wealth index is available on request.

  12. \(E[{Pr_{ij}}\times {q_i}]=\hat{Pr_{ij}}\times \hat{q_i}\) if \(Cov({Pr_{ij}},{q_i})=0\). This can be tested empirically: in our sample we find \(Cov({Pr_{ij}},{q_i})=-0.05\), which is nominal compared to \(\hat{Pr_{ij}}\times \hat{q_i}\) (mean: 20.88).

  13. Since changes in attributes of sources not included in the household’s choice set are irrelevant to the household’s source choice and demand decision, these inequalities are not strict.

  14. \(\epsilon _j =\sum _{i=1}^{N}\bigg (\frac{\partial }{\partial P_{j}}\big (\hat{Pr_{ij}}\big )\times \hat{q_i}+\hat{Pr_{ij}}\times \frac{\partial }{\partial P_{j}}\big (\hat{q_i}\big )\bigg )\times \frac{P_{j}}{Q_j}=\sum _{i=1}^{N}\bigg (\beta _{i}^{price}Pr_{ij}\big (\eta Pr_{ij}+q_i(1-Pr_{ij})\big )\bigg )\times \frac{P_{j}}{Q_j}\)

References

  • Acharya G, Barbier E (2002) Using domestic water analysis to value groundwater recharge in the Hadejia–Jama’ are floodplain, Northern Nigeria. Am J Agric Econ 84(2):415–426

    Google Scholar 

  • Adamowicz W, Swait J, Boxall P, Louviere J, Williams M (1997) Perceptions versus objective measures of environmental quality in combined revealed and stated preference models of environmental valuation. J Environ Econ Manag 32(1):65–84

    Google Scholar 

  • Basani M, Isham J, Reilly B (2008) The determinants of water connection and water consumption: empirical evidence from a Cambodian household survey. World Dev 36(5):953–968

    Google Scholar 

  • Boardman AE, Boardman AE, Greenberg DH, Vining AR, Weimer DL (2018) Cost-benefit analysis: concepts and practice. Cambridge University Press, Cambridge

    Google Scholar 

  • Bockstael NE, Hanemann MW, Kling CL (1987) Estimating the value of water quality improvements in a recreational demand framework. Water Resour Res 23(5):951–960

    Google Scholar 

  • Boone C, Glick P, Sahn DE (2011) Household water supply choice and time allocated to water collection: evidence from madagascar. J Dev Stud 47(12):1826–1850

    Google Scholar 

  • Briscoe J, Chakraborty M, Ahmed S (1981) How Bengali villagers choose sources of domestic water. Water Supply Manag 5:165–181

    Google Scholar 

  • Brouwer R, Job FC, van der Kroon B, Johnston R (2015) Comparing willingness to pay for improved drinking-water quality using stated preference methods in rural and Urban Kenya. Appl Health Econ Health Policy 13(1):81–94

    Google Scholar 

  • Calvo CM (1994) Case study on the role of women in rural transport: access of women to domestic facilities (English). Sub-Saharan Africa Transport Program (SSATP) working paper series; no. 11. Washington, DC: World Bank. http://documents.worldbank.org/curated/en/910381468742552233/Case-study-on-the-role-of-women-in-rural-transport-access-of-women-to-domesticfacilities

  • Cheesman J, Bennett J, Hung Son TV (2008) Estimating household water demand using revealed and contingent behaviors: evidence from Vietnam. Water Resour Res 16:470

    Google Scholar 

  • Cook J, Kimuyu P, Blum A, Gatua J (2016) A simple stated preference tool for estimating the value of travel time in rural Africa. J Benefit Cost Anal 7(2):221–247

    Google Scholar 

  • Coulibaly L, Jakus PM, Keith JE (2014) Modeling water demand when households have multiple sources of water. Water Resour Res 50(7):6002–6014

    Google Scholar 

  • Creel M, Loomis J (1992) Recreation value of water to wetlands in the San Joaquin Valley: linked multinomial logit and count data trip frequency models. Water Resour Res 28(10):2597–2606

    Google Scholar 

  • Dalhuisen JM, Florax RJGM, De Groot HLF, Nijkamp P (2003) Price and income elasticities of residential water demand : a meta-analysis. Land Econ 79(2):292–308

    Google Scholar 

  • Deaton A, Muellbauer J (1980) An almost ideal demand system. Am Econ Rev 70(3):312–326

    Google Scholar 

  • Deely J, Hynes S, Curtis J (2019) Are objective data an appropriate replacement for subjective data in site choice analysis? J Environ Econ Policy 8(2):159–178

    Google Scholar 

  • Dubin JA, McFadden D (1984) An econometric analysis of residential electric appliance holdings and consumption. Econometrica 52(2):345–362

    Google Scholar 

  • Espey M, Espey J, Shaw WD (1997) Price elasticity of residential demand for water: a meta-analysis. Water Resour Res 33(6):1369–1374

    Google Scholar 

  • Filmer D, Scott K (2012) Assessing asset indices. Demography 49(April):359–392

    Google Scholar 

  • Filmer D, Scott K, Prichett L (2001) Estimating wealth effects without expenditure data-or Tears: an application to educational enrollments in states of India. Demography 38(1):115–132

    Google Scholar 

  • Gross E, Elshiewy O (2019) Choice and quantity demand for improved and unimproved public water sources in rural areas: evidence from Benin. J Rural Stud 16:688

    Google Scholar 

  • Haab TC, McConnell KE (2002) Valuing environmental and natural resources. Measurement 8:326

    Google Scholar 

  • Hanemann WM (1982) Applied welfare analysis with qualitative response models. Department of Agricultural & Resource Economics, UCB CUDARE Working Papers

  • Heckman JJ (1979) Sample selection bias as a specification error. Econometrica 47(1):153

    Google Scholar 

  • Heien D, Wessells CR (1990) Demand systems estimation with microdata: a censored regression approach. J Bus 8(3):365–371

    Google Scholar 

  • Hensher DA, Greene WH (2003) The mixed logit model: the state of practice. Transportation 30(2):133–176

    Google Scholar 

  • Herriges JA, Kling CL, Phaneuf DJ (1999) Corner solution models recreation demand: a comparison of competing frameworks. In: Herriges JA, Kling CL (eds) Valuing recreation and the environment: revealed preference methods in theory and practice. Aldershot: Edward Elgar

    Google Scholar 

  • Koehler J, Thomson P, Hope R (2015) Pump-priming payments for sustainable water services in Rural Africa. World Dev 2015(74):397–411

    Google Scholar 

  • Kremer M, Leino J, Miguel E, Zwane AP (2011) Spring cleaning: rural water impacts, valuation, and property rights institutions. Q J Econ 126(1):145–205

    Google Scholar 

  • Larson B, Minten B, Razafindralambo R (2006) Unravelling the linkages between the millennium development goals for poverty, education, access to water and household water use in developing countries: Evidence from Madagascar. J Dev Stud 42(1):22–40

    Google Scholar 

  • Lee L-F (1983) Generalized econometric models with selectivity. Econometrica 51(2):507

    Google Scholar 

  • Madanat S, Humplick F (1993) A model of household choice of water supply systems in developing countries. Water Resourc Res 29(29):1353–1358

    Google Scholar 

  • Mannering FL (1986) A note on endogenous variables in household vehicle utilization equations. Transp Res Part B Methodol 20(1):1–6

    Google Scholar 

  • McFadden D, Train K (2000) Mixed MNL models for discrete responses. J Appl Econ 15:447–470

    Google Scholar 

  • Mu X, Whittington D, Briscoe J (1990) Modelling village water demand: a discrete choice approach. Water Resourc Res 26(4):521–529

    Google Scholar 

  • Murphy KM, Topel RH (1985) Estimation and inference in two-step econometric models. J Bus Econ Stat 3(4):370–379

    Google Scholar 

  • Nauges C, Van Den Berg C (2009) Demand for piped and non-piped water supply services: evidence from southwest Sri Lanka. Environ Resour Econ 42(4):535–549

    Google Scholar 

  • Nauges C, Whittington D (2010) Estimation of water demand in developing countries: an overview. World Bank Res Obs 25(2):263–294

    Google Scholar 

  • Nauges C, Strand J (2007) Estimation of non-tap water demand in Central American cities. Resour Energy Econ 29(3):165–182

    Google Scholar 

  • Olea JLM, Pflueger C (2013) A robust test for weak instruments. J Bus Econ Stat 31(3):358–369

    Google Scholar 

  • Onjala J, Ndiritu SW, Stage J (2014) Environment for development risk perception, choice of drinking water, and water treatment evidence from Kenyan Towns. J Water Sanit Hyg Dev 4(2):268–280

    Google Scholar 

  • Persson TH (2002) Household choice of drinking water in the Phillipines. Asian Econ J 16(4):303–316

    Google Scholar 

  • Phaneuf DJ, Requate T (2017) A course in environmental economics: theory, policy, and practice. Cambridge University Press, Cambridge

    Google Scholar 

  • Phaneuf DJ, Requate T, Smith VK (2005) Chapter 15 recreation demand models. In: Karl-Gran M, Jeffrey RBT (eds) Handbook of environmental economics vincent. Valuing environmental changes, vol 2. Elsevier, New York, pp 671–761

    Google Scholar 

  • Revelt D, Train K (1998) Mixed logit with repeated choices: households’ choices of appliance efficiency level. Rev Econ Stat 80(4):647–657

    Google Scholar 

  • Rural Water Supply Network (2013) Myths of the rural water supply sector

  • Shonkwiler JS, Yen ST (1999) Agricultural & applied economics association two-step estimation of a censored system of equations two-step estimation of a censored system of equations”. Am J Agric Econ 81(4):972–982

    Google Scholar 

  • Stelmach RD, Clasen T (2015) Household water quantity and health: a systematic review. Int J Environ Res Public Health 12(6):5954–5974

    Google Scholar 

  • Strand J, Walker I (2005) Water markets and demand in Central American cities. Environ Dev Econ 10(3):313–335

    Google Scholar 

  • Tanser F, Gijsbertsen B, Herbst K (2006) Modelling and understanding primary health care accessibility and utilization in rural South Africa: an exploration using a geographical information system. Soc Sci Med 63(3):691–705

    Google Scholar 

  • Uwera C, Stage J (2015) Water demand by unconnected households in Rwanda. Water Econ Policy 1(1):145002

    Google Scholar 

  • Von Wartburg M, Waters WG (2005) Congestion externalities and the value of travel time savings, chapter 2. In: Zhang A, Boardman AE, Gillen D, Waters WG (eds) Towards estimating the social and environmental costs of transportation in Canada : a report for Transport Canada. Centre for Transportation Studies, University of British Columbia, Vancouver, B.C.

    Google Scholar 

  • White GF, Bradley DJ, White AU (1972) Domestic water use in East Africa, Drawers of water. University of Chicago Press, Chicago

    Google Scholar 

  • Whittington D, Cook J (2019) Valuing changes in time use in low- and middle-income countries. J Benefit Cost Anal 10:51–72

    Google Scholar 

  • Whittington D, Cook J, Xinming M, Roche R (1990) Calculating the value of time spent collecting water: some estimates for Ukunda, Kenya. World Dev 18(2):269–280

    Google Scholar 

  • WHO/UNICEF (2017) Progress on drinking water, sanitation and hygiene. Technical Report, WHO/UNICEF Joint Monitoring Programme for Water Supply and Sanitation

  • World Health Organization (2013) How much water is needed in emergencies. Tech Notes Drink Water Sanit Hyg Emerg 9:1–4

    Google Scholar 

  • Wu JJ, Babcock BA (1998) The choice of tillage, rotation, and soil testing practices: economic and environmental implications. Am J Agric Econ 80(3):494–511

    Google Scholar 

Download references

Acknowledgements

We thank Annalise Blum, Josephine Gatua, Mark Mwiti, and John Wainana for valuable assistance in the field and in data analysis. We also thank Dale Whittington, Celine Nauges, and two anonymous reviewers for helpful comments and suggestions. Funding for the project was provided by https://www.efdinitiative.org/kenya Environment for Development-Kenya with support from the Swedish International Development Cooperation Agency.

Author information

Authors and Affiliations

  1. School of Economic Sciences, Washington State University, Hulbert Hall 101, PO Box 646210, Pullman, WA, 99164-6210, USA

    Jake Wagner & Joseph Cook

  2. Commission for Revenue Allocation, Nairobi, Kenya

    Peter Kimuyu

Authors
  1. Jake Wagner
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Joseph Cook
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Peter Kimuyu
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Jake Wagner.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix: Materials

Appendix: Materials

See Tables 7, 8 and 9.

Table 7 First stage results for household demand model
Full size table
Table 8 Correlation between average source attributes, and estimated own-price elasticity
Full size table
Table 9 Correlation between average source attributes, and estimated own-price elasticity
Full size table

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Wagner, J., Cook, J. & Kimuyu, P. Household Demand for Water in Rural Kenya. Environ Resource Econ 74, 1563–1584 (2019). https://doi.org/10.1007/s10640-019-00380-5

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10640-019-00380-5

Keywords

Search

Navigation