Appendix
Summaries of selected papers

Paper no. 1 [valuation of health endpoints]

Authors:	Alberini, Anna, Maureen Cropper, Tsu-Tan Fu, Alan Krupnick, Jin-Tan Liu, Daigee. Shaw, and Winston Harrington
Title:	Valuing Health Effects of Air Pollution in Developing Countries: The Case of Taiwan
Published in:	Journal of Environmental Economics and Management, 34, 107-126, 1997
Commodity valued:	Avoiding the recurrence of an episode of minor respiratory illness identical to that most recently experienced by the respondent
Location of the study:	Three cities of Taiwan: Taipei, Kaohsiung, and Hualien
When was the survey conducted?	September 1992
Is the questionnaire reported in the paper?	No, but it can be found in Alberini et al. "What is the Value of Reduced Morbidity in Taiwan?" in The Economics of Pollution Control in the Asian Pacific, Mendelsohn and Shaw (eds.), London: Edgar Elgar Publishing Ltd., 1996.
Payment vehicle:	This is a private risk reduction. The payment question asks respondents how much they would be willing to pay to avoid the illness altogether.
Type of Survey:	In-person
Elicitation method:	Dichotomous choice question, plus two dichotomous choice follow-up questions.
Sampling Frame:	Respondents were adults who had previously participated in a prospective cohort study aimed at establishing the acute respiratory illness associated with pollution exposures. The prospective cohort study participants were recruited by random sampling after stratification by age among the residents of the areas near five air quality monitors in Taipei, Kaohsiung, and Hualien.
Econometric modeling	Interval-data model, estimated by maximum likelihood
Methodological aspects and experimental design	The approach taken by these authors is somewhat different than in the previous literature: rather than asking people to report their WTP for a specified type of illness defined for the respondent, the surveys asks people to define their own illness and report their WTP to avoid it. The advantage of this approach is that people are familiar with the commodity to be valued. In addition, this study focuses on the episode of illness, instead of individual symptom-days.
Difficulties:	The econometric model might be affected by errors-in-variables if the questionnaire fails to capture all of the relevant attributes of the illness.
Lessons learned:	WTP increased with the duration of the illness and with the number of symptoms, but at a decreasing rate, confirming findings from US studies. WTP increases with education and income, its income elasticity being 0.4. WTP reported by persons with chronic illness or serious respiratory illness is systematically higher.
General summary:	The paper performs a benefit transfer from US studies to the Taiwan, adjusting for income differentials, and from Taiwan to the US. The results show that one can neither confirm nor rule out the appropriateness of a benefit transfer procedure that adjusts for income.

Paper no. 2 [valuation of health endpoints]

Authors:	Chestnut, L.G., B. Ostro, N. Vichit-Vadakan, K.R. Smith and Feng C. Tsai
Title:	Final Report. Health Effects of Particulate Matter Air Pollution in Bangkok
Published in:	prepared for the Air Quality and Noise Management, Pollution Control Department, Royal Thai Government, Bangkok, Thailand, under grant from the World Bank (1998). Available from Stratus Consulting, Boulder, CO.
Commodity valued:	Three minor respiratory health endpoints characterized by different severity: (i) one symptom-day; (ii) one restricted-activity day; and (iii) one work-loss day.
Location of the study:	Bangkok
When was the survey conducted?	April 1996
Is the questionnaire reported?	Yes
Payment vehicle:	This is a private good. Respondents are asked to report how much they would pay to avoid the specified illness.
Type of Survey:	In-person
Elicitation method:	Payment card
Sampling Frame:	Respondents had previously participated in a prospective cohort study recording their minor respiratory symptoms on a daily basis. The participants of the prospective cohort study included nurses at specified hospitals with air pollution monitors, school children (who were not given the CV questionnaire) and adults living near Odean Circle.
Econometric modeling	OLS and tobit regressions
Difficulties:	Some persons reported a WTP equal to zero, but comments offered during the survey and other debriefs suggest that did place a value on avoiding illness. They simply rejected the tradeoff between income and health.

Paper no. 3 [Sanitation]

Authors:	Mir Anjum Altaf and Jeffrey A. Hughes
Title:	Measuring the Demand for Improved Urban Sanitation Services: Results of a Contingent Valuation Study in Ouagadougou, Burkina Faso.
Published in:	Urban Studies, 31(4), 1763-1776 (1994)
Commodity valued:	Improved sanitation, described in terms of attributes rather technologies: (a) wastewater disposal; (b) on-site sanitation; and (c) off-site sanitation. Not all households were asked to report WTP for each of these.
Location of the study:	Ouagadougou, Burkina Faso, city with population 688,000.
When was the survey conducted?	February and March 1991
Is the questionnaire reported in the paper?	Parts are reported in the Appendix.
Payment vehicle:	If the respondent is a renter, higher rent. If the respondent is a homeowner, he or she would have to pay the service charges to the financing agency. Owners would be able to recover the cost of (a) and (b) by passing it on to their tenants in the form of higher rent, but for (c) they would have to arrange it separately.
Type of Survey:	In person
Elicitation method:	Dichotomous choice, two dichotomous-choice follow-ups, and a final open-ended question.
Sampling Frame:	Absent a list of households, a simplified two-stage stratified sample was drawn to arrive at a sample of 600 households. The city was divided into grids, grids were selected at random, dwelling units in the selected grid were counted, and every tenth unit in each block was interviewed. 605 interviews were completed for a total of 593 usable surveys.
Econometric modeling	The ordered probit model fitted by the authors finds several individual characteristics and wealth indicators significant predictors of WTP.
Worthwhile findings:	Households with improved pit had WTP for on-site sanitation almost twice as large as those without, which suggests that WTP for marginal improvement is not decreasing. The authors conclude that although almost all households has a positive WTP for off-site sanitation, from a financial perspective this type of plan still remains unfeasible. By contrast, on-site sanitation appears to be feasible.
General summary and background:	Only 38% of the households in Ouagadougou are connected to the piped water supply system. All connections are metered, and all water used must be paid for. There is no sewer system, and the most frequently used form of sanitation at the time of the study was private pit latrine. Respondents were generally dissatisfied about the available sanitation. The researchers decided not to use descriptions of technologies (e.g., WC) because the residents would not be familiar with them. Instead, they decided to describe attributes, arguing that the latter would be better understood by respondents and more useful to policy makers.

Paper no. 4 [water supply]

Authors:	Mir Anjum Altaf, Dale Whittington, Haroon Jamal, and V. Kerry Smith
Title:	Rethinking Rural Water Supply Policy in the Punjab, Pakistan
Published in:	Water Resources Research, 29(7), 1943-1954, 1993
Commodity valued:	Connection to piped water supply system with (a) standard and (b) improved reliability
Location of the study:	Villages in the agricultural area of the Punjab, Pakistan
When was the survey conducted?	Not reported
Is the questionnaire reported in the paper?	No
Payment vehicle:	Monthly tariff
Type of Survey:	In person
Elicitation method:	Dichotomous choice, one dichotomous choice question, plus one final open-ended follow-up
Methodological aspects and experimental design	(a) in villages without connections to public water system, households were asked to report WTP for private connection to a system with standard reliability, and to a system with improved reliability. (b) in villages with connections, households were asked to report their WTP for connections to a system with improved reliability.
Worthwhile findings:	Results of the valuation effort are used to compute the percentage of connecting households and the resulting revenues for various prices. Revenues are compared to the cost of the investments necessary to provide the specified number of connections.
General summary and background:	Water for domestic use is obtained by individual households using pumps (community wells provided by the government having been abandoned) in the sweet water zone. In the brackish water zone, groundwater is too salty for drinking and cooking, but is adequate for other uses such as bathing and washing. Household keep pumps even when they have connections to the piped water system (where available), because the latter is so unreliable. The government policy is to provide water to villages with at least 5000 residents. In smaller villages, the government provides public taps. No public water systems have been or are being built in the sweet water zone, because the government feels that the pump system is adequate. Private connections to the public water system require that households incur the cost of connecting. A flat fee, independent of actual water usage, is charged per month.

Paper no. 5 [sanitation and surface water quality]

Authors:	KeyongAe Choe, Dale Whittington, and Donald T. Lauria
Title:	The Economic Benefits of Surface Water Quality Improvements in Developing Countries: A Case Study of Davao, Philippines
Published in:	Land Economics, 72(4), 519-537, 1996
Commodity valued:	Sanitation and Improvement in surface water quality
Location of the study:	Davao, Philippines (urban area, with population of approximately 850,000, based on the 1985 Census)
When was the survey conducted?	Not reported
Is the questionnaire reported in the paper?	No
Payment vehicle:	Monthly fee (amounts: 25, 50, 100, 150 and 200 pesos, or about 1, 2, 4, 6, and 8 US dollars)
Type of Survey:	In-person
Elicitation method:	Elicitation starts with offering price (listed above, and varied to the respondents), then a "series of follow-ups" is implemented, but the actual number of follow-up questions is not reported. If the respondent answers all "nos" to the payment question, he or she is asked a final open-ended question.
Sampling Frame:	Two-stage stratified sampling to draw 1200 households. Overall response rate is about 65 percent, because about 32 percent of households could not be located by enumerators, and the remaining 3 percent refused. There were a total of 777 completed interviews. Households were recruited from 3 groups: (1) households who owned dwellings and had a water closet (WC) for their exclusive use; (2) renters (n=87) and (3) all others (with house but no WC, and squatters) (n=326). Eighty-five percent of the respondents had a WC for their household's exclusive use.
Econometric modeling	Interval-data estimation based on maximum likelihood
Methodological aspects and experimental design	A split-sample methodological investigation is presented in this paper (see experimental design below).
Worthwhile findings:	The WTP figures were compared with results from a single-site travel cost model. The trip data for this model (trips to Times Beach and alternative sites before and after the health advisory issued for this beach) came from the same survey questionnaire. WTP and the welfare estimates from the travel cost model were found to be in the ballpark of one another, at least for beach users.
Lessons learned:	The WTP from non-users of the beaches were found to be rather low. In general, WTP for improved sanitation and surface water quality was found to be rather low (both in absolute terms and as a percentage of income), suggesting that environmental quality is not a high priority for the residents of Davao. While the responses to the payment questions were generally consistent with economic theory (e.g., the higher the initial bid, the lower the percentage of respondents agreeing to pay), there was little sensitivity to scope.
Methodological aspects and experimental design:
Who	Scenario
Respondents in group (2) and (3), and half of the respondents in group (1)	City-wide plan (with details intentionally left unspecified) to clean up rivers and sea, and make Times Beach suitable for swimming and recreation (cleanup of surface water). Monthly fee required of the household to implement this plan. Respondents were told that industry would also pay or reduce their discharges of wastewater into rivers.
One half of respondents in group (1).	Cleanup of surface water as above, plus construction of sewer lines in the neighborhood and treatment of wastewater. Respondents were described how sewers and wastewater treatment work.
Clearly, this implementation allows the researchers to perform a scope test. Only one half of the respondents could be given the second version of the questionnaire, because they were already familiar with a minimum of sanitation. Scope test is based on respondents in group 1.

Paper no. 6 [recreational amenities; conservation of resources]

Authors:	Nadini Hadker, Sudhir Sharma, Ashish David, T.R. Muraleed Haran
Title:	Willingness to Pay for Borivli National Park: Evidence from a Contingent Valuation
Published in:	Ecological Economics, 21, 105-122 (1997)
Commodity valued:	Management program for the park that would stop the deterioration of the park itself
Location of the study:	Borivli National Park, India. This national park is located near the city of Bombay, and receives about 2.5 million visits every year.
When was the survey conducted?	Not reported
Is the questionnaire reported in the paper?	Yes, in the Appendix
Payment vehicle:	Monthly payments over the next 5 years. Due to the lack of credibility of the government, the researchers had to come up with a scenario that calls for "a completely honest and trustworthy agency" that would be delivering the plan. People were told they would pay or could volunteer labor services, in which case they were asked to state how many days of labor they would volunteer.
Type of Survey:	In person. Respondents were given brochures with park information, containing pictures plus descriptions. If the respondent could not read English, he would still be able to look at the pictures while the enumerator was translating the text into local languages.
Elicitation method:	Dichotomous-choice with follow-up
Sampling Frame:	Residents of Bombay intercepted at various locations and times of the day. [Note that this leaves the sampling frame completely unclear.]
Econometric modeling	Double-bounded interval data models estimated by maximum likelihood
Methodological aspects and experimental design	"Attitude questions" were asked in various parts of the survey.
Worthwhile findings:	No significant association was found between WTP and respondent income.
Difficulties:	The researchers were concerned with minimizing all possible sources of bias, and especially strategic behaviors on the part of the respondents. Despite all precautions, about 25 percent of the respondents provided responses identified as "protest zeros," stating that "the government should pay for environmental preservation."
Lessons learned:	Although India is a poor country, people are willing to pay for conservation. The authors calculate that the WTP bids reported in the survey imply a total WTP of 20 million rupias per month for the next 5 years, or a present value of over 1 billion rupias.

Paper no. 7 [water supply]

Authors:	Alexander A. McPhail
Title:	Why Don't Households Connect to the Piped Water System? Observations from Tunis, Tunisia
Published in:	Land Economics, 70(2), 189-196, 1994
Commodity valued:	Individual water connection and subsequent monthly charges (respondents were told that they would be charged monthly by the water used, and that the initial down payment for connecting would be spread over five years).
Location of the study:	Tunis, Tunisia (urban setting, population 1.5 million). Two areas within the city were studied: Medina (the oldest part of the city) and Melassine (poor area built in the 1970s).
When was the survey conducted?	Late Spring of 1990 - surveys took place over three days
Is the questionnaire reported in the paper?	No
Payment vehicle:	Monthly fees
Type of Survey:	In-person
Elicitation method:	Dichotomous choice question, one follow-up, plus one final open-ended question (two different starting points used to check for starting point bias)
Sampling Frame:	A comprehensive list of unconnected households was provided by SONEDE, the local water utility. A sample of 82 households were surveyed.
Econometric modeling	Simple averages and frequencies
Difficulties:	The respondents were told to consider an usage of water that would take care of all their needs. The author notes that households did find it hard to estimate how much water they would use if connected.
General summary and background:	The author wishes to examine why households in Tunis do connect to the existing water system. The study reveals that the down payment required by the water utility is what turns away households, which are otherwise able to cover the monthly charges for water and sewer.

Paper no. 8 [recreational amenities; conservation of resources]

Authors:	Stale Navrud and E.D. Mungatana
Title:	Environmental Valuation in Developing Countries: The Recreational Value of Wildlife Viewing.
Published in:	Ecological Economics, 11, 135-151 (1994)
Commodity valued:	Recreational Value of Wildlife Viewing, with specific reference to flamingos. The authors argue that this is a conservative estimate of the total economic value of the species of wildlife in the park.
Location of the study:	Lake Nakuru National Park, Kenya
When was the survey conducted?	1991
Is the questionnaire reported in the paper?	No
Payment vehicle:	N/A - respondent's own expenses (the survey asks how much higher the cost of the trip could be before the respondent gives us the trip altogether)
Type of Survey:	In person
Elicitation method:	Open-ended
Sampling Frame:	Random sample of visitors (both Kenya nationals and non-nationals) for a total of 185 persons
Methodological aspects and experimental design	The survey asked for the expenses sustained during this trip, and then queried the respondent about how much higher the cost could be before the respondent would give up the trip altogether. The questionnaire thus elicited the respondent's surplus. Two measures were asked of the respondent: WTP for a visit to the national park, and both WTA and WTP for the flamingos alone.
Worthwhile findings:	WTP figures were compared with surplus estimates from a travel cost analysis. The researchers conclude that the two alternative approaches provide values with each other's ballpark.
Difficulties:
Lessons learned:	The researchers worried that embedding effects could influence the results of the study, and chose to ask first the question about the national park and then the question about the flamingos for this reason. They found that the item response rate to the valuation question was lower when the WTA format was used.

Paper no. 9 [drinking water supply]

Authors:	Marcia Rosado
Title:	Willingness to Pay for Drinking Water in Urban Areas of Developing Countries
Published in:	Unpublished. Presented at the American Agricultural Economics Association meetings, Salt Lake, August 1998.
Commodity valued:	This is NOT a CV study, and is included in this survey simply because it represents an alternative approach for valuing safe drinking water. The paper uses the defensive behavior approach, which provides a lower bound for WTP for safe water. It uses revealed preference data, collected though a survey of households in Brazil, and fits a nested logit model of the choice of action undertaken by households prior to drinking water (boiling it, filtering it, etc.). It then uses the estimated coefficients of the nested logit model to compute the compensating variation of eliminating that option. The attributes of the various actions used to predict the probability that that action is selected are the time the action takes and out-of-pocket costs, interacted with individual characteristics of the respondents.
Location of the study:	Urban agglomeration in the state of Espirito Santo, Brazil.
When was the survey conducted?	Not reported
Is the questionnaire reported in the paper?	No
Payment vehicle:	Not Applicable - out of pocket costs sustained by the respondents.
Type of Survey:	Not reported
Elicitation method:	Not applicable
Sampling Frame:	Unclear
Econometric modeling	Nested logit model of revealed preference data

Paper no. 10 [recreational amenities; conservation of resources]

Authors:	Steven Shultz, Jorge Pinazzo, and Miguel Cifuentes
Title:	Opportunities and Limitations of Contingent Valuation Surveys to Determine National Park Entrance Fees: Evidence from Costa Rica
Published in:	Environment and Development Economics, 3, 131-149 (1998)
Commodity valued:	Improvements in infrastructure and services (the exact nature of the improvements was not specified to the respondents)
Location of the study:	Poas National Park and Miguel Antonio National Park in Costa Rica
When was the survey conducted?	1995
Is the questionnaire reported in the paper?	No
Payment vehicle:	Entrance fee for a future visit
Type of Survey:	In person
Elicitation method:	Single-bounded dichotomous choice
Sampling Frame:	Random sample of visitors to the two parks, for a total of 124 usable surveys
Econometric modeling	Logit analysis.
Worthwhile findings:	Mean WTP to Poas National Park is $11 for residents and $23 for foreigners; mean WTP for entrance to Miguel Antonio National Park is $13 for residents and $14 for foreigners. Separate analyses were performed for the two parks and for foreign nationals and residents. Of the various covariates of WTP that were attempted, hardly any was found significant, a result probably to be blamed on the small sample and on the focus on participants (visitors), as opposed to the general population.
Lessons learned:	The authors explain that Costaricans had a high WTP relative to their income, perhaps as a result of "cultural and strategic bias related to their unfamiliarity with personal surveys and with providing truthful negative responses to interviewers." Enumerators noted that while foreign respondents asked to participate in the survey treated the event very informally and quickly went about reading and answering the survey, Costa Rican visitors were generally apprehensive and started an extensive dialogue before beginning the survey. The authors also recognize that a proper study should address the issue of non-participation (e.g., predict whether participation rates would change a result of changing the entrance fee).
General summary and background:	The authors explain that it is important to evaluate the effect of entrance fees carefully: raising the entrance fee can raise revenues but decrease visitation rates. Until 1994, fees to national parks in Costa Rica were $1 for foreigners and $0.65 for residents. In July 1995, the entrance fees were raised to much higher levels, with different fees for more and less highly visited parks. In April 1996, fees were reduced somewhat. This study took place in 1995, right after the fee hike. The two national parks here studied boast significant ecological/diversity riches: Poas offers a highland tropical habitat and volcanic views, whereas Miguel Antonio offers jungle and beach flora and fauna.

Paper no. 11 [conservation of natural resources]

Authors:	Priya Shyamsundar and Randall A. Kramer
Title:	Tropical Forest Protection: An Empirical Analysis of the Costs Borne by Local People.
Published in:	Journal of Environmental Economics and Management, 31, 129-144 (1996)
Commodity valued:	Loss of access to large area of forest lands (caused by the establishment of a national park)
Location of the study:	Near Mantadia National Park, eastern Madagascar, Africa.
When was the survey conducted?	July 1994
Is the questionnaire reported in the paper?	No
Payment vehicle:	Baskets of rice (and not cash payment) given to households.
Type of Survey:	In person
Elicitation method:	The survey elicits willingness to accept (WTA) and not WTP, using a single-bounded approach with 7 different bid levels, plus one follow-up question.
Sampling Frame:	Villages sampled to reflect regional differences
Econometric modeling	It is not clear which response (the one to the initial WTA question, or that combined with the follow-up?) is modeled in the paper. It is assumed that WTA is a function of household size, preference for forests over buffer zones around the forests; annual household coffee production; location dummies and gender dummies.
Worthwhile findings:	Twenty percent of the respondents did not answer the WTA question, but the authors did not find any relationship between failure to answer this question and individual or household characteristics. No biases were created by the missing responses. Mean WTA is 6.15 vata of rice, or about $50 per household per year). Over a lifetime and using discount rate of 10%, this leads to a net present WTA value of $312,331.
General summary and background:	The paper attempts to answer the question whether re-forestation and protection of ecological systems, carried out through the establishment of national parks, are good conservation strategies in developing countries. Madagascar is an example of a poor country which is very rich in biological resources and diversity. Poor people rely extensively on agriculture and on foraging forest products. The authors recognize that Willig (1976) shows that for price changes there should be little difference between WTP and WTA, but that the psychology, economic-theoretical and empirical literature has found reasons for and actual differences between WTP and WTA. They mention Hanemann's result (1991) that WTA can be much larger than WTP is the elasticity of substitution between public and private good is close to zero. During the development of the questionnaire, the authors evaluated carefully whether it was OK to ask people how much they would pay to avoid loss of access to forest lands, concluding that the few respondents that had reported a positive, non-zero WTP had done so out of a sense of coercion, and not because they were truly willing to pay. Since residents considered loss of access as a decrease in welfare, WTA was chosen. The use of photos was also ruled out, because they attracted too much attention to the detriment of the remainder of the survey scenario.

Paper no. 12 [water supply]

Authors:	Singh, Bhanwar; Radhika Ramasubban, Ramesh Bhatia, John Briscoe, Charles C. Griffin, and ChongChum Kim
Title:	Rural Water Supply in Kerala, India: How to Emerge from a Low Level Equilibrium Trap
Published in:	Water Resources Research, 29(7), 1931-1942, 1993
Commodity valued:	Yard taps; house connections to the piped water supply for both current service level and improved service level
Location of the study:	Kerala, India (rural area).
When was the survey conducted?	Not reported
Is the questionnaire reported in the paper?	No
Payment vehicle:	Down payment for connection, plus monthly service charges
Type of Survey:	In person
Elicitation method:	One-way down double-bounded
Sampling Frame:	At each location, two types of household were selected: (a) households from areas where improved water supply has been available for years, and (b) households from areas with no connections.
Econometric modeling	Instead of using the proper interval-data estimator, the researchers form a panel dataset (with the responses provided by the respondents to the payment questions for various commodities) and estimate a random-effects probit model.
Worthwhile findings:	WTP figures are plugged into simulations to trace out revenues and number of connections as a function of the monthly tariff and the cost of the connection. The paper recommends that the connection charges be folded into the monthly payments. The authors report that improved water supply does not elicit higher WTP, and that females (who bear the burden of getting water for domestic consumption) bid significant lower WTP amounts.
General summary and background:	The current problem in the Kerala area is that connections to the piped water system (which has low pressure and reliability problems to start with) further reduce pressure and reliability, a problem aggravated by the fact that water usage is unmetered. People have low WTP for connecting to such an unreliable water system, and the revenue are insufficient to cover improvements to the system.

Paper no. 13 [public health programs]

Authors:	B.M. Swallow and M. Woudyalew
Title:	Evaluating Willingness to Contribute to a Local Public Good: Application of Contingent Valuation to Tse-Tse Control in Ethiopia
Published in:	Ecological Economics, 11, 153-161 (1994)
Commodity valued:	Monitoring of baits for tse-tse flies to avoid theft (and hence disruption of the fly control program)
Location of the study:	Rural area in Ethiopia
When was the survey conducted?	January 1993
Is the questionnaire reported in the paper?	No
Payment vehicle:	People could volunteer cash or labor (an option allowed because labor markets in Ethiopia are very restricted and there is limited cash available to households), or both
Type of Survey:	In person
Elicitation method:	Open-ended
Sampling Frame:	The population from which the sample was drawn was the households within the Ghibe Valley tse-tse control area. 850 households were listed. These were stratified by cattle ownership and residential location to form 9 strata. Random sampling from these strata resulted in a final sample of 180 households.
Econometric modeling	Two variables, time the respondent wishes to donate to the program, and cash, are treated as simultaneous dependent variables and estimated as a system using three-stages least squares (3SLS).
Methodological aspects and experimental design	The paper accepts the findings about giving respondents time to think about the proposed commodity or program (Whittington et al., 1992), and the respondents were provided with materials and information about two to four weeks prior to being interviewed.
Worthwhile findings:	The regression analyses find that gender, information, household size and ownership of cattle are significant predictors of cash and labor donations. Interestingly, women have lower WTP or labor offers.
General summary and background:	Tse-tse fly control is investigated as an alternative to the treatment of trypanosomiasis in Southwestern Ethiopia, due to increased resistance to available forms of medication. A tse-tse control trial using 500 baited targets was instituted in the area of the current study. The baited targets are pieces of black cloth covered in pesticide and odor-baited to attract flies. The system proved effective in its first year, but during the second year, the success of the program was undermined by theft (people found ways to use the cloth to make clothes, bags and bedding, and recycled the small containers containing the bait substances to make coffee cups. This made it necessary to secure monitoring of the targets and baits and cooperation of the local populace.

Paper no. 14 [water supply]

Authors:	Dale Whittington, Donald T. Lauria, and XinMing Mu
Title:	A Study of Water Vending and Willingness to pay for Water in Onitsha, Nigeria.
Published in:	World Development, 19 (2/3), 179-198, 1991
Commodity valued:	Connection to the New Onitsha Water Scheme, with meter, for specified water price
Location of the study:	Onitsha, Nigeria (urban setting, with population of about 700,000; the residents of this city are comparatively wealthy)
When was the survey conducted?	July and August 1987
Is the questionnaire reported in the paper?	In the appendix (along with formal statistical analysis of the data)
Payment vehicle:	Monthly charges
Type of Survey:	In person (n=235)
Elicitation method:	Bidding game
Sampling Frame:	There was no available list of households from which to draw a sample. Enumerators were dropped in various districts of the city and instructed to interview every other person.
Worthwhile findings:	The percentage of income spent on water is at least 5%, and is higher in the dry season. For comparison, the authors note that in Haiti poor people sometime spend 20 percent of their income on water, and in Addis Ababa and in Kenya the fraction of income spent of water is about 9 percent. They conclude that the percentage here found is high, but not inconsistent with other places.
Lessons learned:	The paper uses the WTP results to predict the number of connecting households and the projected revenue to the water authority at various prices. The authors conclude that WTP for improved water connection is high in Onitsha, and that households can afford to pay for the full economic cost of connecting, which in turn generates revenue sufficient to cover the costs. Even the households who elect not to connect are predicted to benefit from the lower prices vendors would charge.
General summary or background:	Onitsha's water supply infrastructure was hard hit during the 1967-70 civil war, and has been inadequate ever since. Any residents believe that piped water should be provided for free by the government. Tension exists between local agencies, which believe water should be provided to the populace at low price, and donors, who would prefer efficiency pricing in the interest of promoting conservation. Water for domestic use is provided primarily by the private sector, using tanker trucks. Some households re-sell the water they purchase from vending tanker trucks. Extensive primary data collection was necessary to find out about water vending practices. The authors found that the prices charged by the vendors are very high, and that vendors must capture significant rents.

Paper no. 15 [sanitation and surface water quality]

Authors:	Dale Whittington, Kyeongae Choe, and Donald T. Lauria
Title:	The Effect of Giving Respondents "Time to Think" on Tests of Scope: An Experiment in Calamba, Philippines
Published in:	"Determining the Value of Non-marketed Goods," edited by Raymond J. Kopp, Walter W. Pommerehne and Norbert Schwartz; Kluwer, 1997.
Commodity valued:	(a) connection to sewer system only; (b) sewer system plus wastewater treatment plant; (c) sewer system, wastewater treatment plant, plus regional plan to preserve surface water quality.
Location of the study:	Calamba, Philippines (urban environment, population 175,000, situated on large in-land lake used for commercial fishing)
When was the survey conducted?	November and December 1992
Is the questionnaire reported in the paper?	No
Payment vehicle:	Monthly fee
Type of Survey:	In person
Elicitation method:	Initial bid (5 different amounts), followed by a series of follow-up questions ("short bidding game")
Sampling Frame:	Stratified random sample of 1500 households
Econometric modeling	OLS, tobit and probit
Methodological aspects and experimental design	The study incorporated the test of the hypothesis that giving respondents time to think lowers their WTP, based on an earlier study by Whittington et al. Three different scenario were created: (A) sewer system only; (B) sewer system plus wastewater treatment plant; and (C) sewer system, wastewater treatment plant, plus regional plan to preserve surface water quality. For each of these commodities, two different elicitation techniques were used: (a) a "long questionnaire," completed in one sitting (of the average duration of 45 minutes); and (b) the "time to think" variant, whereby the interview was completed over 2 or 3 days. The resulting 2×3 cells can be used to test whether allowing respondents time to think makes them appreciate the differences between the "sizes" of the different commodities, resulting in scope effects, and whether effects earlier noted by Whittington et al are also seen here.
Worthwhile findings:	People were able to appreciate the difference in the size of the plan between plan A and B, whether or nor they had been given time to think. People who had had time to think liked plan C less than plan B (even if the latter is a less comprehensive plan than the former). The possible reasons are that they did not like government intervention, or that the time to think had given them the opportunity to realize the potential for free riding.

Paper no. 16 [water supply]

Authors:	Dale Whittington, V. Kerry Smith, Apia Okorafor, Augustine Okore, Jim Long Liu, and Alexander McPhail
Title:	Giving Respondents Time to Think in Contingent Valuation Studies: A Developing Country Application
Published in:	Journal of Environmental Economics and Management, 22, 205-225, 1992
Commodity valued:	(i) Public taps; (ii) private connections to the public water supply system. Study also contains a methodological investigation, for which it is well known.
Location of the study:	Agricultural communities in Anambra State, Nigeria. Three large Igbo villages, with population ranging between 10,000 and 25,000.
When was the survey conducted?	April and May 1989
Is the questionnaire reported in the paper?	Yes, in the Appendix.
Payment vehicle:	Fee to water authority
Type of Survey:	In person. Complete data for n=395.
Elicitation method:	Dichotomous choice question with one dichotomous choice follow-up question, and one final open-ended question.
Sampling Frame:	Unclear
Econometric modeling	The authors fit four alternative models to the WTP responses: (a) a regression model that uses the responses to the last, open-ended question and treats them as continuous. (b) double-bounded; (c) ordered probit; (d) single-bounded probit, using only the responses to the initial payment questions.
Methodological aspects and experimental design	Experimental design first divides the respondents into two subsamples. One was asked to report information about WTP immediately, but was also interviewed after some time. The other was asked to go home, and interviewed about WTP after a day or two. Subsample. WTP immed. WTP after time to think I yes yes II yes
Worthwhile findings:	WTP of people that were given time to think is systematically lower for that of people that had to report WTP immediately. The size of the difference is larger for public taps than for private taps, and when the interval-data estimator for double-bounded data is used, as opposed to OLS or probit. The authors also examine the revised WTP bids by respondents in subsample II. They use two approaches, OLS estimation, and two-stage estimation to account for endogeneity of initial WTP and revised WTP. Once potential endogeneity is explicitly accounted for, the downward revision of the bids is significant only for public taps. The authors consider strategizing on the part of the respondent as a likely explanation for the results, but also note that there was no evidence of respondents getting together in small groups to discuss ways to free ride or influence the provision of the good. They also point out that had the respondents strategized in this fashion, they would have probably done so for both types of commodities, and yet the difference in WTP is significant only for public taps.
General summary and background:	The authors report that the reason why they implemented the "time to think" was their fear that unfamiliarity with the commodity or the tradeoff required to produce WTP amounts would result in many missing responses or strategic considerations on the part of the respondent. In the area of the study, households must use vendors equipped with water tank trucks, and household expenditure on water can be in excess of 10 percent of household income.The authors report that their result that WTP is lower in the group of respondents who were given time to think is in sharp contrast with the predictions of the theoretical result by Hoehn and Randall (1987) that WTP is non-decreasing in time. The authors further comment that WTP for improved water must be constrained by cash restrictions (resulting from restricted labor markets) or by divergences between the individual's and the household's WTP.

Paper no. 17 [sanitation]

Authors:	Dale Whittington, Jennifer Davis, Harry Miarsono and Richard Pollard
Title:	Urban Sewer Planning in Developing Countries and The "Neighborhood Deal." A Case Study of Semarang, Indonesia.
Published in:	UES Working Paper. World Bank Water and Sanitation Program.
Commodity valued:	Sewage system plus wastewater treatment plant.
Location of the study:	Semarang, provincial capital of central Java (population 1.3 million)
When was the survey conducted?	August 1995
Is the questionnaire reported in the paper?	No
Payment vehicle:	Monthly service fee (flat rate for sewerage plus use-based water fee), plus fixed fee for connection to the sewage system.
Type of Survey:	In person
Elicitation method:	Single-bounded? With 50000, 150000, 300000, and 500000 rp as the prices for the full-service option.
Sampling Frame:	Random sampling could not be done. Instead, three kelurahans were selected (one with low-income residents; one with middle to high-income residents, and one with middle to low-income residents). Households were selected at random within each kelurahan.
Worthwhile findings:	Willingness to pay for sanitation is low and calls into question the appropriateness of undertaking investments at this time. The percent of households willing and not willing to connect is shown below: FULL SELF-HELP NO PARTICIPA SERVICE TION Yes, would Connect 17% 12% 5% No, would 6% 9% 50% Not connect
Difficulties:	Authors conclude that trying to get households without water connections to get both water connection and sewage connection is not likely to work.
Lessons learned:	Although water quality was deemed "good" for odor (89%), color (96%) and taste (78%), 100% of the respondents boil water. At the time of the study, there were no sewers in Semarang, and wastewater flows in open combined sewers/storm drains to the Java Sea, without receiving any treatment. Most people have toilets (with septic tanks). The authors noted that "sanitation is not viewed as a high priority by many residents," but concern over drainage of flood waters and safe water supply seen to be stronger than that over roads.
General summary and background:	Instead of entirely entrusting the government with the provision of the plan for improved sanitation, the authors chose to introduce (following the local traditions and institutions) a grass-roots neighborhood organization who could choose to have an engineering contractor ("full service") or just an engineering consultant ("self-help"), with members of the community performing simple tasks and helping out with the construction of the sewer lines. Households would be connected only if they want the connection and are willing and able to pay for it.

Paper no. 18 [sanitation]

Authors:	Dale Whittington, Donald T. Lauria, Albert M. Wright, Kyeongae Choe, Jeffrey A. Hughes and Venkateswarlu Swarna
Title:	Household Demand for Improved Sanitation Services in Kumasi, Ghana: A Contingent Valuation Study.
Published in:	Water Resources Research, 29(6), 1539-1560 (1993)
Commodity valued:	Improved sanitation systems: (a) ventilated improved pit latrines, and (b) water closets (WCs) connected to a piped sewerage systems. These were offered to respondents in various combinations (see below).
Location of the study:	Kumasi, Ghana
When was the survey conducted?	Fall 1989
Is the questionnaire reported in the paper?	Some parts of the questionnaire are reported in the Appendix.
Payment vehicle:	Monthly fees
Type of Survey:	In person
Elicitation method:	Dichotomous choice, one dichotomous choice follow-up question, plus one final open-ended question
Sampling Frame:	Households were selected using a two-stage stratified procedure, giving a sample of 1633 households. Usable interviews were completed for 1524 households.
Econometric modeling	Interval-data maximum likelihood estimation. Covariates include (a) questionnaire characteristics, such as a dummy for the time-to-think variant or high starting point for the bidding game; (b) socioeconomic characteristics of the households; (c) characteristics of the individual respondent; (d) existing water and sanitation conditions. Separate regressions were run for each of the commodities being valued.
Methodological aspects and experimental design	Different versions of the questionnaire were developed, including one that allows to test whether giving respondents one day to think about their WTP resulted in different WTP. Four different commodities were created: (i) Kumasi ventilated improved pits; (ii) sewer connections (for households that already had a WC); (iii) WCs with sewer connections (for households that did not have a WC); and (iv) private water connections, plus WC with sewer connections (for households that had neither). To sum, survey only queried respondents for their WTP for an improvement over the current situation. The survey questionnaire was subject to extensive development work and was pre-tested on a sample of 100 households.
Worthwhile findings:	The most important predictors of WTP are the current water usage and sanitation available to the household. The evidence about the possible presence of starting point bias is mixed (there seems to be no evidence of starting point bias for water, and some evidence of starting point bias for sanitation). Perhaps the most striking result is that time to think does not seem to influence WTP, in contrast with the evidence from the Whittington et al 1992 study. The presence of other people at the interview also does not seem to influence WTP.
Lessons learned:	The authors make the point that reported WTP bids reflect the respondents' private valuation - had the respondents been aware of the public health externalities of improved sanitation, their valuation would have been much higher. The results of the valuation effort are input into the calculation of the number of households connected and the resulting revenue for the water and sanitation authority.

Paper no. 19 [water supply]

Authors:	Dale Whittington, John Briscoe, Xinming Mu, and William Barron
Title:	Estimating the Willingness to Pay for Water Services in Developing Countries: A Case Study of the Use of Contingent Valuation Surveys in Southern Haiti
Published in:	Economic Development and Cultural Change, 293-311, 1990
Commodity valued:	Public water posts and private connections
Location of the study:	Laurent, a village in a rural area in Southern Haiti, pop. 1500
When was the survey conducted?	August 1986
Is the questionnaire reported in the paper?	Parts of the commodity description and the bidding game are reported in the article.
Payment vehicle:	Fee per month
Type of Survey:	In person
Elicitation method:	Bidding game (but authors find no evidence of starting point bias)
Sampling Frame:	"The majority of the households in Laurent were interviewed. Enumerators were instructed to try to interview someone in every house."
Econometric modeling	Interval-data model estimated by maximum likelihood
Methodological aspects and experimental design	The researchers conducted a split-sample methodological investigation, testing whether changes in the wording of the payment question (which were argued to contain potential for strategic responses) resulted in different WTP figures. No significant differences were found in mean WTP across the two independent subsamples given the two different versions of the survey. Our reading of the payment questions is that the potential for strategic behavior is obscure at best.
Worthwhile findings:	WTP is positively related to household income, an occupation index, education and distance to alternative water sources. This suggest that WTP does depend - as economic theory suggests - on the opportunity cost of alternative sources of water.

Paper no. 20 [water supply]

Authors:	Whittington, Dale, Apia Okorafor, Augustine Okore, and Alexander McPhail
Title:	Strategy for Cost Recovery in the Rural Water Sector: A Case Study of Nsukka District, Anambra State, Nigeria
Published in:	Water Resources Research, 26(9), 1899-1913
Commodity valued:	Public tap or private connection to improved water supply.
Location of the study:	Three large Igbo villages in the Nsukka district of Anambra State, Nigeria
When was the survey conducted?	May 1989
Is the questionnaire reported in the paper?	No
Payment vehicle:	Monthly payment in perpetuity with no seasonal variation.
Type of Survey:	In-person
Elicitation method:	Dichotomous choice question, one follow-up, plus one open-ended question.
Sampling Frame:	Since there was no adequate list of household from which to draw random samples, the enumerators were instructed to randomly interview people in the community.
Econometric modeling	Unspecified - presumably, sample averages of the open-ended WTP amounts.
General summary:	The government in the past had drilled boreholes for wells, but these had been subsequently capped, were no longer operational, or no longer used. The abundant rainwater during the rainy season provides water for domestic use, with the problem of insufficient water supply arising during the dry season. Based on the information gathered in the survey, it is estimated that during the dry season 5-7 hours a day must be spent by household members to collect water from local springs. On an annual basis, households spend 6-10 percent of their income on water.
Lessons learned:	The most striking result of this study is the low WTP for water, despite the fact that much time and money is spent during the dry season to get water. Since this finding has important implications, the households were interviewed. Reasons indicated by the respondents for their low WTP for improved water supply include: the respondents did not trust the government to deliver or maintain an improved water system, or thought it was the government's responsibility to provide free service. Alternatively, respondents had already paid their taxes and did not want to pay additional taxes for services that were not expected to work; cash flow restrictions; respondents did not want to commit to a uniform flow of monthly payment when they felt the problem was of a seasonal nature.