The importance of being informed: Experimental evidence on demand for environmental quality

https://doi.org/10.1016/j.jdeveco.2007.10.002Get rights and content

Abstract

To what extent does information affect the demand for environmental quality? A randomly selected group of households in an Indian city were informed whether or not their drinking water had tested positive for fecal contamination using a simple, inexpensive test kit. Households initially not purifying their water and told that their drinking water was possibly contaminated, were 11 percentage points more likely to begin some form of home purification in the next eight weeks than households that received no information. They spent $7.24 (at PPP) more on purification than control households. By way of comparison, an additional year of schooling of the most educated male in the household is associated with a 3 percentage-point rise in the probability of initial purification, while a standard-deviation increase in the wealth index is associated with a 12 percentage-point rise in this probability and an $11.75 rise in expenditure. Initially purifying households that received a “no contamination” result did not react by reducing purification. These results suggest that estimates of the demand for environment quality that assume full information may significantly under-estimate it.

Introduction

Environmental quality is generally thought to be a normal good, possibly a luxury good. However, observed positive correlations between income and environmental quality might overstate the importance of income because the poor are likely to be less informed about health risks than the rich. For example, 70% of those surveyed in a village in rural Bolivia thought that diarrhea was a normal occurrence in childhood (Quick et. al., 1997) while 45% of those surveyed in an Indian city in the study reported here did not include drinking contaminated water among the possible causes of diarrhea. The literature in economics, unlike that in public health, has laid much more emphasis on household wealth than on information as a determinant of willingness to pay to protect oneself from environmental and health hazards.1 The question is: Just how significant is the role of information in the demand for environmental quality?

To address this question we devised an experiment to provide households with exogenous information and measure its impact on averting behavior and expenditure, that is, on actions taken to reduce exposure to an environmental hazard.2 We chose household water quality as the environmental good because it is possible for households to take averting actions and incur expenditures even in the short run, and because it is one of the important environmental and health issues in developing countries. We provided a randomly selected treatment group of households with information about their water quality and available methods of improving it. The information was based on a test for the presence of bacteria of fecal origin. Eight weeks later, we recorded any changes in their in-home water purification behavior that had occurred in the interim and compared these with the same data for a control group.

The experiment was conducted in Gurgaon, a suburb of Delhi with wealth and education levels that are higher than the national average. Between November and December 2003, we surveyed approximately 1000 randomly selected households and tested their drinking water for the presence of fecal bacteria. About 60% of the water (before any home purification) tested “dirty”.3 To put this figure in perspective, in the United States, if even a single sample of tap water tests positive for fecal coliforms, the local water authority is in violation of federal regulations issued under the Safe Drinking Water Act.4 Moreover, local authorities must inform individual households about any violation of the standards in their jurisdiction. In India, on the other hand, tap water is unregulated and the results of any water quality tests conducted by government authorities are generally not made public (McKenzie and Ray, 2004).

In the second round, we informed a randomly selected group of about half the sampled households of their water test results and told them about the cost and likely efficacy of home water purification methods in use in Gurgaon at the time. Returning about eight weeks later, we found that households who had not been purifying their water and who were told that their water was “dirty”, indicating the likelihood, but not certainty, of fecal contamination, were 11 percentage points more likely to have made changes in water purification, handling and/or storage behavior than households who had not been informed of the test result (p = 0.013). Their averting expenditures increased by $7.24 (PPP) more than those of the control group.5 They were also nearly 13 percentage points more likely to have had discussions on issues concerning drinking water. The latter two results are significant at the 1% level. On the other hand, discovering that their water was “clean” (probably not contaminated) did not induce initially purifying households to reduce their purification expenditure.

The mean averting expenditure prior to our intervention was $42, and a simple calculation shows that if all households were given the information that our treatment group got, averting expenditure (a lower bound for WTP for safe water) would be predicted to rise by 6.5%. This effect on purification expenditure is of the same order of magnitude as that of a one-standard deviation increase in the wealth index, estimated from the cross-section. These results show that even a single instance of targeted information to households, despite the qualifier about the accuracy of the test, can have an effect on awareness and disease averting expenditures. More generally, our results suggest that measuring the demand for environmental quality and attempting to draw welfare and policy conclusions from it without addressing the issue of under-provision of information can lead to significant underestimates of demand.

As yet, little attention has been paid to this issue in the literature. The only other experimental study from a developing-country that we are aware of is that by Madajewicz et al. (2007) who studied the impact of informing Bangladeshi households about the arsenic levels in their well water. They find that 60% of households with unsafe wells changed their water source while only 14% of those with safe wells did so. The proportion of those responding to information is much larger than in our study, possibly because their cost of switching was very low, subjects were unambiguously warned that their wells were unsafe, and generally knew that arsenic poisoning is deadly. Moreover, information dissemination was continuous and intense — wells were marked with skulls and crossbones signs if they did not meet the government's prescribed standards on arsenic levels. Finally, households had a much longer response time.

In another study, different types of information booklets on cancer risks from radon gas were given to a sample of US households in the state of New York by the Environmental Protection Agency (Smith et al. (1995)). The researchers found that households were persuaded to take mitigating action on receiving information on the radon levels and on being convinced about the health risks of radon exposure.

The motivation and a simple model underlying the experiment are laid out in Section 2. The experiment and the sampling design are described in Section 3. Section 4 provides details about the data and reports some summary statistics. Section 5 presents the main results of the experiment and Section 6 concludes.

Section snippets

Motivation and a simple model underlying the experiment

In 1998–99, the National Family Health Survey of India (NFHS) found that 16% of urban households with children in the 0–3 age group had a case of child diarrhea in the two weeks preceding the survey. Despite this high incidence of disease we found (Jalan et. al., 2003) that nearly half of all households in the urban sample of the NFHS did not purify their water. Of those who did, only about 20% used methods (i.e. boiling and electronic filters) that are thought to reliably remove pathogens. In

The experiment

Our study area was the city of Gurgaon, a suburb of the national capital, New Delhi.7 We chose Gurgaon because we wanted a residential urban area where the water supply to households was not of uniform quality, where

Data and descriptive statistics

Table 1, Table 2 report several descriptive statistics on information gathered during the first round of the survey.

In Table 1 we report statistics on the quality (as measured by our water test results) of water supply and purification behavior by households in Gurgaon. Contamination of unpurified water is high — 61% of unpurified water tested “dirty” i.e., positive for the presence of fecal bacteria. Only 42% of surveyed households use private purification methods (including boiling) and of

Results

We estimate the following reduced form of Eq. (2) from Section 2:y=1ifD(p)c(w)+u=xβ+u>0,=0otherwise,xβ=α+β1(Treatment×Dirty)+β2(Treatment×Clean)+β3(Dirty)

In this equation, y is a measure of a behavior adopted by households between the third and first rounds, “Dirty” is a dummy for a positive test result indicating the presence of fecal bacteria in the household's water source, “Clean” denotes a negative test result, and “Treatment” is a dummy for a treatment, as opposed to control,

Conclusion

In this paper, we asked to what extent targeted information affects the demand for environmental quality. We focused on the easily measurable component of willingness to pay for safe drinking water given by averting expenditure. We found that Gurgaon households who were told that their drinking water was “dirty” and was, therefore, likely to contain germs, were nearly 11 percentage points (p-value = 0.013) more likely to begin some form of home purification in the next seven weeks than households

References (16)

  • HarringtonWinston et al.

    The economic losses of a waterborne disease outbreak

    Journal of Urban Economics

    (1989)
  • AbdallaC.W. et al.

    Valuing environmental quality changes using averting expenditures: an application to groundwater contamination

    Land Economics

    (1992)
  • BorghesiSimone

    The environmental Kuznets curve: a survey of the literature

  • BresnahanB. et al.

    Averting behavior and urban air pollution

    Land Economics

    (1997)
  • DasguptaPurnamita

    Valuing health damages from water pollution in urban Delhi, India: a health production function approach

    Environment and Development Economics

    (2004)
  • Government of India

    Household consumer expenditures and employment unemployment situation in India

  • JalanJ. et al.

    Awareness and the demand for environmental quality: drinking water in urban India

  • LevinsonA.

    The ups and downs of the environmental Kuznets curve

There are more references available in the full text version of this article.

Cited by (0)

Financial support from the South Asian Network for Development and Environmental Economics (SANDEE) is gratefully acknowledged. We thank the Office of Registrar-General, Census India, for providing enumeration-block level data and copies of census maps, Professor Shahid Jameel of the International Centre for Genetic Engineering and Biotechnology for incubator space, and Mr. S.N. Prasad at TARA for information about the water test kits. Tushi Baul, Saraswata Chaudhuri, Kamaal un Nabi Khan, Sreemoyee Moitra and Arindam Nandi provided excellent research assistance. Khan and Moitra and the survey team successfully implemented the surveys. Abhijit Banerjee, Shreekant Gupta, M.N. Murty, Arindam Nandi, Arijit Sen, Priya Shyamsundar, Rohini Somanathan and Jeff Vincent gave useful comments at various stages of the project. Finally, we thank Esther Duflo and two anonymous referees of the journal for making this a better paper.

View full text