199

CHAPTER-V

EMPIRICAL ANALYSIS OF SOLID WASTE MANAGEMENT BY

USING CONTINGENT VALUATION METHOD

The Contingent Valuation (CV) method is a widely used non-market valuation

method especially in the areas of environmental cost–benefit analysis and

environmental impact assessment (see Mitchell and Carson, 1989; Cummings et al.,

1986). Its application in environmental economics includes estimation of non-use

values (e.g. Walsh et al., 1984; Brookshire et al., 1983), non-market use values (e.g.

Choe et al., 1996; Loomis and duVair, 1993) or both (e.g. Niklitschek and Leon, 1996;

Desvousges et al., 1993) of environmental resources. In recent years, this method is

commonly used in developing countries to elicit the individuals’ preferences for the

basic infrastructural projects such as water supply and sanitation (see Whittington,

1998; Merrett, 2002). Though a popular non-market valuation method, a group of

academicians criticise this method severely for not being a proper method of estimating

the non-market values (see Hausman, 1993). Hence, the major objective of the concept

which is being used to portray the Willingness to Pay to improve the River water

quality in the study area based on empirical aspects of CV method.

The CV method was originally proposed by Ciriacy-Wantrup (1947) who was

of the opinion that the prevention of soil erosion generates some ‘extra market benefits’

that are public goods in nature, and therefore, one possible way of estimating these

benefits is to elicit the individuals’ willingness to pay for these benefits through a

survey method (see Portney, 1994; Hanemann, 1994). However, Davis (1963) was the

first to use the CV method empirically when he estimated the benefits of goose hunting

through a survey among the goose-hunters. This method gained popularity after the two

200

major non-use values, namely, option and existence values have been recognised as

important components of the total economic values in environmental economics

literature, especially during the 1960s. While the conventional revealed preference

methods such as travel cost method are not capable of capturing these non-use values

(Smith, 1993), the only method that is identified for estimating these values is the

contingent valuation method (CVM) (see, Desvousges et al., 1993). Hence, a

considerable amount of studies on CVM - both theoretical and empirical in nature -

have emerged in the economic valuation literature, including a large number of studies

criticising the CV method.

Summaries of CV studies by different authors reveal that the major criticism of

results of CVM revolves mainly around two aspects, namely, (a) validity and (b)

reliability (Smith, 1993; Freeman, 1993; NOAA, 1993). In simple terms, validity refers

to the ‘accuracy’ and reliability refers to ‘consistency’ or ‘reproducibility’ of the CV

results (Kealy et al., 1990). In other words, validity refers to the degree to which the

CV method measures the theoretical construct of interest which is the true economic

value of individuals (Freeman, 1993). The validity is of three types, namely, content

validity, criterion validity and construct validity (Mitchell and Carson, 1989; Bateman

et al., 2002). The content validity in a CV experiment simply refers to the ability of the

instruments included in the scenario to measure the value in an appropriate manner.

Criterion validity of the CV method may be assessed in terms of another measure, say a

‘market price’ for the same commodity which may be considered as a criterion.

Construct validity has two forms: convergent validity and theoretical validity. The

convergent validity refers to the correspondence between two measures of the same

theoretical construct. CV results can be said to be ‘theoretically valid’ if the results

conform to the underlying principles of economic theory. In other words, the

201

theoretical validity involves assessing the willingness to pay (WTP) values of the CV

method by way of regressing the WTP value against standard economic variables

(Mitchell and Carson, 1989). On the other hand, the reliability refers to the extent to

which the variance of the WTP amounts is due to random sources (Mitchell and

Carson, 1989). According to Loomis (1990), ‘Reliability requires that, in repeated

measurements, (a) if the true value of the phenomenon has not changed a reliable

method should result in the same measurement (given the method’s accuracy) and (b) if

the true value has changed a reliable method’s measurement of it should change

accordingly’. The following section is devoted to discuss various errors/biases affecting

the validity and reliability of the CV method along with illustrations from empirical

studies that looked into addressing these two issues.

Hicks (1946) classified the ‘consumer surplus measure’ into two different

categories, namely, the compensating variation and the equivalent variation. For a

‘proposed welfare gain’ due to provision of public good, the compensating variation

refers to the amount of money income that has to be given up by the consumer to attain

increased level of utility (i.e. WTP measure). The equivalent variation refers to the

amount of compensation required to be provided to the individual so that he could

attain an improved utility level in case the provision of the public good does not take

place (i.e. WTA). For a welfare loss, the compensating variation refers to the amount of

money income that is required to compensate the individual for the welfare loss

experienced (i.e. WTA) and equivalent variation refers to the amount of money income

to be sacrificed by the consumer to prevent the loss from occurring in future

(i.e. WTP) (Bateman and Turner, 1993).

202

In rigorous economic terms, the contingent valuation method estimates the

Hicksian consumer surplus—either the compensating variation or the equivalent

variation—due to change in the provision of public goods (Bateman and Turner, 1993).

One of the important conclusions derived from the above paragraph is that, in principle,

either WTP measure or WTA measure could be used interchangeably to elicit

individuals’ preferences for change in the level of environmental goods and services.

Yet, one of the issues that is supposed to affect the validity of the CV results is the

disparity that arises between the WTP value and WTA value for the same good under

consideration (Mitchell and Carson, 1989). It has been demonstrated both theoretically

as well as empirically that the WTA value is always greater than the WTP value if used

for the same issue (Shogren et al., 1994; Hanemann, 1991; Brookshire and Coursey,

1987; Coursey et al., 1987; Knestch and Sinden, 1984; Bishop and Heberlein, 1979;

Willig, 1976). This being the case, the question that has to be addressed is: which

measure (i.e. either WTP or WTA) should be used in a CV survey to elicit the value of

the changes in the provision of public goods? (Mitchell and Carson, 1989). Before

going on to answer this question, let us see the reasons why the disparity between WTP

and WTA measures arises.

The WTP/WTA disparity has been attributed to many different factors. Willig

(1976), having theoretically demonstrated the disparity between WTP and WTA

measures for a priced commodity, concluded that the WTP/WTA disparity could be

attributed to the ‘income effect’. In a strict sense, the income effect in economics refers

to the effect of additional income on the quantity purchased of a particular commodity

(Diamond et al., 1993). The underlying implication of the income effect is that the

WTP for a good is constringed by income, whereas the WTA compensation is not. In

this sense, Willig’s (1976) conclusion may be interpreted as to imply that WTP and

203

WTA would diverge for a commodity which has high income elasticity of demand,

while accompanied by income constraint. Randall and Stoll (1980), extending Willig’s

(1976) analysis from the ‘price space’ to ‘commodity space’, proved that the WTP and

WTA measures are found to be closer for public goods unless they are affected by an

unusual income effect (Hanemann, 1991). In explaining the disparity further,

Hanemann (1991) showed that: (a) for a change in the quantity (unlike change in the

price), there is no such presumption that WTP and WTA measures should be close; and

(b) not only the income effect but also the ‘substitution effect’ explains the larger

disparity between WTP and WTA. It means that the divergence between WTP and

WTA values could range from zero to infinity depending on the degree of substitution

between goods, accompanied by positive income elasticity (Shogren et al., 1994). Some

of the empirical studies have attempted to understand the influence of substitutes on the

WTA/WTP disparity. For example, using two private goods (a candy bar and a coffee

mug) that are close substitutes and a public good (reduction of human health risk),

which has no close substitute, Shogren et al. (1994) demonstrated that the divergence

between WTP and WTA values for the two private goods disappears after repeated

trials, while the divergence is robust and persistent for the public good. The results of

this study not only proved the role played by the substitution effect in explaining the

disparity between WTP and WTA but also showed how the respondents’ ‘familiarity’

with the valuation experiment plays a role in reducing the disparity. In another

empirical attempt, Adamowicz et al. (1993) used two goods—one without substitute

(a feature film in a local theatre) and the other with substitute (hockey match with a

substitute of live radio/TV telecast). Their experiment demonstrated that substitution

factor does explain the disparity between WTP and WTA estimates. However, the

authors argue that even though the substitution factor reduces the disparity it does not

204

eliminate the disparity entirely. This implies that there are other factors that could

provide possible answers to the WTA/WTP disparity. Apart from income and

substitution effects, some of the theoretical developments in economics and psychology

are also extended to explain the WTA/WTP disparity. For example, the ‘prospect

theory’ developed by Kahneman and Tversky (1979) also explains the variation in the

WTP and WTA measures. According to Kahneman and Tversky (1979), the loss of a

commodity for an individual is considered to be greater than the gain derived from

buying the same commodity. This is because, increase in income is weighted by a

relatively small utility compared to decrease in income which is weighted by much

larger utility (Coursey et al., 1987). More precisely, for the ‘loss averse individuals’ the

disparity would be greater (see Brookshire and Coursey, 1987). Thaler (1980)

demonstrated WTA/WTP disparity through ‘endowment effect’ and noted that it might

be related to prospect theory, which is about decision making under uncertainty.

The existence of loss aversion by individuals has been proved by Brookshire

and Coursey (1987) who found in a hypothetical experiment that the respondents’

WTA value for a particular level of reduction in an area with tree cover was 75 times

greater than the WTP value for the same level of expansion in the same area. The

concept of ‘property rights’ as well as ‘transaction cost’ after Coase (1960) also holds

the reason for the disparity between WTP and WTA. For instance, the WTA/WTP

disparity has been attributed to the rejection of the underlying property rights by

individuals (Mitchell and Carson, 1989). According to Mitchell and Carson (1989), in

many cases the individuals consider the WTA property right as illegitimate or

implausible or both. This is evident from generation of large number of ‘protest bids’ in

CV studies that used WTA measure as elicitation format. Apart from property rights,

the transaction costs involved in the process of obtaining goods and services may also

205

result in WTP/ WTA disparity (Brown and Gregory, 1999). During market transaction,

the consumer calculates the total cost of buying and selling the good that includes not

only the price of the good but also the transaction cost. This is being the case, while

buying the good the individual’s willingness to pay would exclude the transaction cost

and while selling the same good the willingness to accept compensation would include

it. However, since market for public goods are either nil or very weak, the transaction

cost theory may not be expected to play a major role in the WTP/WTA disparity issue.

Apart from theoretical aspects, the value elicitation or experimental aspects of CV

method are also found to affect the WTA/WTP disparity. Cummings et al. (1986) bring

the ‘familiarity’ issue (see also, Shogren et al., 1994) for addressing not only the

disparity between WTP and WTA values but also many other biases in CV method.

How the respondents’ familiarity with the valuation experiment (as well as the good)

and their optimization strategy in a CV study leads to provide answer to some of the

serious problems in CVM have been empirically tested by many authors, especially in

laboratory experiments (e.g. Coursey et al., 1987). These experiments show that

repeated trials (that provide adequate time for the respondents to understand the issue

and optimise) lead to convergence between WTP and WTA measures. In many cases,

it was observed that WTA measure comes down gradually over repeated trials, whereas

the WTP measure remains stable for number of trials (Coursey et al., 1987). These

findings provide support for Hoen and Randall’s (1987) argument that the disparity

between WTP and WTA would be greater for those individuals who lack time to

optimise their decision.

206

In general, studies on WTA/WTP disparity conclude that a larger difference

between WTA and WTP can be attributed to the ‘weak’ experimental features of the

CV studies. These features include hypothetical payments; students being used as

objects, using less incentive compatible elicitation questions, etc. (see Horowitz and

McConnell, 2002). The implication is that if the survey is designed ‘as realistic as

possible’ then the WTA/WTP ratio would become smaller. However, results of a meta-

analysis by Horowitz and McConnell (2002) provide different kinds of insights into the

debate. Horowitz and McConnell’s (2002) meta-analysis tested whether the high

WTA/WTP ratios are an experimental artefact, or they truly present a broad-based

picture of preferences. Forty-five studies on WTA/WTP have been taken for the meta-

analysis and these studies dealt with goods such as chocolates, pens, mugs, movie

tickets, hunting licenses, visibility, nuclear waste repositories, nasty-tasting liquids, etc.

Using a random effects regression model, the study regressed the WTA/WTP ratio

against type of good (ordinary vs. otherwise), survey design (hypothetical or real

payment; elicitation technique; and student and non-student), mean WTP and year.

The results of this study are striking. For example, the WTA/WTP ratios in real

experiments do not differ significantly from that of hypothetical experiments. This

finding is contradictory to the conventional notion that the experiments involving real

payment are superior to experiments involving hypothetical payments. Another

important finding of this study is that WTA/WTP ratio is higher for incentive

compatible elicitation techniques, whereas the earlier studies found that incentive

compatible elicitation techniques provide low WTA/WTP ratio. Another finding which

goes against the earlier findings is that the WTA/ WTP ratio is significantly lower for

students than the general public. This implies that moving the laboratory experiment to

the real world situation will not reduce the ratio. Also, respondents’ familiarity with the

207

experiments does not provide lower ratios. All these important findings of Horowitz

and McConnell (2002) suggest that the high observed WTA/WTP ratios do not appear

to be experimental artifacts and indeed, the answer for this disparity comes from the

broad-based preferences of the individuals.

From the above discussion, it is clear that there exists a disparity between WTP

and WTA and this disparity is influenced by many different factors such as income

effect, substitution effect, transaction costs, broad based preferences, etc. Moreover,

the disparity has been an accepted phenomenon in the CVM literature as demonstrated

by both theoretical as well as empirical studies. However, having agreed that there

exists a difference between WTP and WTA, one important question that one needs to

answer is: how much can it differ? According to Hanemann (1991), the WTA value

can be five times greater than the WTP value. But some of the empirical studies show

that the disparity between WTP and WTA for a same commodity ranges from a low of

2.4 times to a maximum of 61 times (e.g. Brookshire and Coursey, 1987). Occurrence

of large-scale disparity between the WTP and WTA in empirical studies leads to the

following conclusions: (i) one or both of the CV measures are wrong or the theory is

wrong (Carson et al., 2001); and (ii) the WTA measure is not a proper measure of

consumer surplus (see Adamowicz et al., 1993), and therefore, WTP measure rather

than the WTA measure is the proper measure of value that should be used in the CV

studies (Cummings et al., 1986; NOAA, 1993).

One of the major sources of error in the CV method is ‘embedding’ that is

frequently reported in many CV studies (Bateman et al., 1997). Embedding refers to a

phenomenon in which a wide range of variation is found to occur in WTP value for the

same good depending on whether the good is valued on its own or valued as a part of a

208

more inclusive package (Kahneman and Knestch, 1992). The embedding is also called

part–whole bias, dis-aggregation bias, sub-additivity effect, scope effect (see

Cummings et al., 1986; Mitchell and Carson, 1989; Hanemann, 1994; Bateman et al.,

1997) even though the meaning between these concepts differs (Boyle et al., 1994).

The embedding effect is defined in different ways by different authors. According to

Kahneman and Knestch (1992), the embedding occurs if ‘the same good is assigned a

lower value if WTP for it is inferred from WTP for a more inclusive good rather than if

the particular good is evaluated on its own’ (Kahneman and Knestch, 1992, p. 58).

According to Harrison (1992), embedding occurs when the WTP value for one good

differs ‘insignificantly’ with the WTP value for a more inclusive good. The existence

of the embedding effect is supposed to result in a situation where ‘different surveys can

obtain wide variable stated willingness-to-pay amounts for the same public good, with

no straightforward way for selecting one particular method as the appropriate one’

(Diamond and Hausman, 1994, p. 46). Embedding affects the validity of the CV

results, and therefore, it is recommended that every CV study should have an inbuilt

mechanism called, ‘internal consistency test’ to assess the validity of results (NOAA,

1993). One way of assessing the internal consistency of the results is to involve the

results for ‘scope test’. To understand the nature of embedding effect and how the

scope test is used for assessing the validity of the results, let us discuss briefly about

some of the studies that have looked into these issues. Earlier CV studies have widely

reported occurrence of embedding in their results. For example, results of a CV study

by Kahneman (1986) on conservation of fish in lakes showed that WTP value for

cleaning up all the lakes in Ontario (larger area) was only slightly higher than the WTP

value for cleaning up lakes in one region (smaller area). Later in a famous study,

Kahneman and Knestch (1992) used three goods, namely, ‘improved rescue equipment

209

personnel’ (say, Good A), which is embedded in ‘improved preparedness for disasters’

(say, Good B), which in turn is embedded in ‘environmental services’ (say, Good C).

These three goods were assigned to three independent sample groups in such a way that

sample group 1 received Good A, sample group 2 received Good B and sample group 3

received Good C. It was observed that the mean WTP values provided by three

sample groups for all the three goods separately were as follows: $135.91 for Good C;

$151.60 for Good B and $122.64 for Good A. The difference between the mean values

was found to be statistically insignificant, implying that the embedding effect has

occurred in the results.

Hence, Kahneman and Knestch (1992) conclude that the values elicited through

CV are not true values but arising out of individuals’ ‘purchase of moral satisfaction’.

However, Smith (1992) claims that none of the conclusions of Kahneman and

Knestch’s (1992) study (KK study hereafter), which suggest that the embedding effect

is a problem in CV results, are correct. To substantiate his claim, Smith (1992)

provides three reasons: (i) the question asked in KK study to test the embedding effect

was not conveyed in such a way that the respondents could properly understand the

intention of the researchers. More precisely, the reference and target levels of the

provision of the inclusive good were not properly conveyed to the respondents; (b) the

general implementation of the survey was found to have been improperly done. For

instance, issues such as how the questionnaire was developed, whether it was pre-

tested, how sample households have been selected, statistical tools used, etc., have not

been properly addressed in the study; and (c) the interpretation of the results in line

with economic theory was also found to be flawed. Therefore, Smith’s (1992)

arguments imply that the results of the KK study should not be used to assess the

ability of the CV method to measure the non-market values. Moving one step forward,

210

Harrison (1992) not only criticises the way in which the goods were described in KK’s

study but also tries to explain the reason for the embedding effect with the help of

‘Good Cause Dump Hypothesis’ which refers to a situation where the respondents have

a WTP value for a basket of good causes and dump the whole value for any good in the

basket when asked for valuing a single good. Moreover, re-analysing KK’s data with an

alternative method, Harrison (1992) concludes that the KK study had employed

inappropriate statistical and data handling procedures. However, Harrison’s (1992)

claim has been challenged by Nickerson (1993). Nickerson (1993) evaluates KK’s

version of hypothesis, data and statistical analysis and Harrison’s version of KK’s

hypothesis, data and statistical analysis independently and concludes that it is

Harrison’s statistical and data-handling procedures that are inappropriate. Some of the

CV studies conducted subsequently did also report the occurrence of ‘embedding’ in

their results. For example, Desvousges et al. (1993) attempted to test the theoretical

validity which involves evaluations of how well contingent valuation estimates

conform to hypotheses derived from a theoretical construct, using the migratory bird

experiment. Three independent samples at two Malls in Atlanta, Georgia, were

assigned to three different scenarios that differ only in terms of number of bird death

prevented from oil spills. That is, three versions of the survey (2000, 20,000 or

200,000) and their respective percentage (i.e. much less than 1%, less than 1% and

above 2%) of the water fowl deaths prevented through protection of 250,000 ponds in

the Central Flyway were presented to the respondents. The WTP values (of the open-

ended results) for all the three treatments showed that there was no significant

difference between the WTP values for these treatments indicating the presence of

embedding effect. In addition to migratory water fowl scenarios, the authors also used

oil spill scenarios (i.e. small spills vs. all spills) to see whether the theoretical validity

211

of CVM for one commodity (i.e. migratory water fowl) is replicable for another

commodity (i.e. oil spills).

In the oil spill case too, the annual mean willingness to pay values for different

levels of oil spill - derived from both open-ended as well as dichotomous choice

techniques - do not differ significantly. From these results, it has been concluded that

the WTP values are infected with embedding effect, and therefore, the results of the

CVM study are not theoretically valid (Desvousges et al., 1993). Boyle et al. (1994),

using the results of migratory water fowl experiment of Desvousges et al. (1993),

derive the following possible conclusions: (a) the marginal utility for preventing bird

deaths (additional) is zero; (b) the marginal utility is greater than zero but is too small

to detect; (c) the marginal utility is positive but not trivial, but CV is not able to

measure the differences; (d) respondents’ perception about the percentage of the birds

prevented must have been subjective; and (e) the respondents WTP value must have

been based on the number of ponds (over 250,000) to be covered which was same in all

three treatment rather than number of bird death prevented (Boyle et al., 1994).

Having observed ‘embedding’ in CV studies that focused on measuring the non-

use values, Mitchell and Carson (1989) conclude that embedding may be a problem

only in the case of non-use values. However, Kahneman and Knestch (1992) argue that

embedding is not restricted to non-use values alone but it also occurs in the case of

estimation of use values for public goods. It should be noted that embedding has been

found to occur in CV studies on private goods. For example, Randall and Hoehn (1996)

used two private goods - (i) a range of food items and (ii) rice [which is embedded in

(i)] and demonstrated that the ‘embedding effects observed with the contingent

valuation of non-market goods are observed here with market demands for private

212

goods’ (p. 370). In another study, Bateman et al. (1997) used a similar kind of

methodology on WTP for ‘full meals’ and WTP for ‘part’ of it and found the existence

of embedding effect in their results. The conclusion of these two private good studies

is that embedding is an ordinary economic phenomenon that not only occurs in public

goods but also in private market goods. However, some of the studies found opposite

results in which the embedding effect has been reported to be nil. For example, Choe

et al. (1996) in their water quality study found that embedding effect did not occur for

two different water quality levels that differed in scope. It should be noted that

different individual studies have come out with different kinds of results. But, what

these studies have to say collectively? Smith and Osborne (1996) conducted a meta-

analysis to assess the internal consistency of the CV results - especially the theoretical

properties of the underlying preferences for the visibility improvements. This study

utilises meta-analysis of five different CV studies that focused on changes in visibility

at the US national parks. These five case studies were selected in such a way that they

cover both use and non-use values and include both on-site surveys and surveys

conducted in far away distance. Using Box_Cox and Feasible Generalised Least Square

(FGLS), the mean WTP (MWTP) has been regressed against the independent variables

such as, eastern vs western parks; decline or improvement in visibility; and the extent

of the change (Single Park or the whole area). Other variables such as type of

elicitation format used, whether interview took place within the park (or far away

distance) and whether the households live in the park itself were also included in the

analysis. Controlling for the sensitivity of the independent variables, the meta-analysis

found that there exists a statistically significant, positive relationship between WTP and

proportionate change in the visible range. Apart from the statistical significance

between the WTP and the visible range, the analysis also investigated the proportionate

213

change in the WTP value to the visible range and found existence of economic

plausibility in the results. Overall, the meta-analysis reveals that the ‘visibility changes

do seem to be responsive to the NOAA panel’s call for both statistical significance and

economic plausibility in the scope test’ (p. 299). Though meta-analyses and individual

studies provide theoretically consistent results, occurrence of scope effect (or

embedding) in many other individual CV studies still causes disturbance. Having

observed scope effect in CV studies, Diamond and Hausman (1994) provide two kinds

of interpretation regarding this phenomenon. The first interpretation is that the CV

studies valuing two commodities rather than one are unreliable. The other

interpretation, as argued by Kahneman and Knestch (1992), is that the ‘warm glow’

might have occurred in the CV results. It is, therefore, argued that the presence of

scope effect in the WTP/ WTA accept a value, which implies that the contingent

valuation method may not be useful for estimating the non-use values (Kahneman and

Knestch, 1992; Diamond and Hausman, 1994). This is because the occurrence of

embedding effect violates the fundamental principle by which a rational consumer is

directed to prefer more of a good or service to less and pay more for larger quantity

compared to lesser quantity (Desvousges et al., 1993). The proponents of the CV study

do not agree with the argument by the opponents that CV method is not theoretically

valid due to occurrence of embedding or scope effect. One of the reasons why the

embedding shows up in some of the CV studies, according to Hanemann (1994), is due

to simple economic theory. The marginal utility theory provides an answer to this

phenomenon in the sense that individuals’ utility at margin declines for the subsequent

bundle of commodity they consume. However, empirical evidence not only shows the

difference in WTP for two different quantity levels is more or less equal but sometimes

the decline in the value for the larger quantity is zero (NOAA, 1993), which is

214

disturbing. But some of the authors argue that this is because of the way in which the

commodity is being described in the CV studies (see Smith, 1992; Hanemann, 1994;

Carson et al., 2001).

For instance, scenario description that does not facilitate the respondents to

make any real difference between different levels of goods described, convenience

sampling used in surveys (i.e. stopping the respondents at shopping malls and asking

WTP questions in the case of Desvousges et al. (1993) study) and self-administered

questionnaires that are prone to unreliable results, etc., are considered to affect the

validity of the CV results (Hanemann, 1994). Do the elicitation techniques used in the

description of the CV scenario reduce the scope effect and improve the validity of the

CV results? A comprehensive study by Hammitt and Graham (1999) used two different

approaches to understand the influence of preference elicitation on the WTP values for

reduction of health risk. The first approach involves meta-analysis of 25 CV studies on

health-related aspects, such as road safety, medical treatment, hazardous waste

management, etc. These studies were selected on the basis of three criteria, namely,

(i): addressing the programme to protect and enhance the safety; (ii) eliciting WTP

from lay respondents; and (iii) linking WTP value with the numerical change in the

probability of death. These studies can be taken as to have used the conventional

method of conveying different levels of health risk to the respondents. The results of

the meta-analysis suggest that the sensitivity of WTP to probability variation in risk is

limited. That is, the WTP values are found to be insensitive to the scope of the goods

provided. In the second approach, the authors used a split sample test - consisting of

‘internal’ (within sample) and ‘external’ (between samples) tests of sensitivity to

magnitude of risk. Two surveys have been used to estimate the WTP for risk reduction

in two risk contexts, namely, transportation and food consumption. In Survey-1, the

215

respondents were asked about only one risk, namely, dying in an automobile accident.

One-half of the respondents were conveyed a baseline risk of dying in a road accident

as 20/100,000 persons and the other half received a scenario with risk level of

25/100,000 person. Both the samples were asked about their WTP for purchasing a

‘safety bag’ that would reduce fatality risk to 15/100,000, and then to 10/100,000, from

the baseline risk. In Survey-2 also, WTP for a safety bag was asked in addition to the

food-risk scenario. But in Survey-2, an ‘indifference-risk’ approach is introduced to

‘alleviate respondents’ limited appreciation for the relative magnitude of numerical risk

changes, a factor that may cause elicited WTP to be insensitive to magnitude’ (p. 47).

Here, innovations that would make the respondents understand the risk and to remove

factors causing distortions between the preferences and the theory were used. The

innovations include: (a) presentation of analogies in the stimulus to understand the

respondents’ understanding of risk; (b) distinguishing respondents with high and lower

confidence levels; (c) information about the ‘increments’ in risk rather than ‘levels’ to

combat the tendency to focus on percentage change; (d) use of one-and-a-half-bounded

DC elicitation format in the food safety decision.

In this survey, the respondents were asked to choose between two lunch stands -

one is outdoor (cheap, more risky) and the other is indoor (more expensive, safer).

In the case of high-risk scenario, risk levels of 1/13,700 and 1/137,000 were given to

two sub samples, while the low-risk scenario with only 1/100 million has been assigned

to both the samples. The results of Hammitt and Graham (1999) study provide certain

useful insights. In the case of automobile survey, it was found that the indifference risk

approach (Survey-2) provided ‘statistically significant difference’ (between WTP

values) than the conventional approach (Survey-1). In Survey-2, it was also found that

the respondents with ‘high confidence’ in their answers provided WTP values that were

216

proportional to the price difference than the respondents with less confidence. Overall,

the authors conclude that the standard conventional CV methods do not provide

theoretically consistent results. Nonetheless, the author’s new approach that involved

indifference-risk elicitation is found to provide theoretically consistent results.

Therefore, the results of the meta-analysis and that of Survey-1 suggest that the

conventional preference elicitation methods are generally weak in making the

respondents to understand the differences in the level of commodity described. This

implies that the scope effect or embedding could be addressed through proper ‘study

design’.

The findings of Hammitt and Graham’s (1999) study are supported by their

follow-up study reported in Corso et al. (2001). Corso et al.’s (2001) study has used

visual aids to communicate two different levels of risk, namely, one in which the

annual risk of dying in a motor-vehicle crash was 2.5/10,000 and the other in which it

was 2.0/10,000. The information regarding the respondent’s willingness to pay for an

automobile safety devise (i.e. a side-impact airbag) that would reduce the annual risk of

dying in a motor crash to 1.5/10,000 was elicited through the ‘mixed-mode phone–

mail–phone survey’ among four sub sample groups of 1,104 final respondents. The

findings suggest that the WTP values for independent sub-samples presented with each

of the three alternative visual aids (a linear scale, a logarithmic scale and an array of

dots) were responsive to the magnitude of risk reduction. However, the WTP values

elicited from the fourth sub-sample that received ‘no visual aid’ was not responsive to

the magnitude. The conclusion of this study supports the notion that theoretically

consistent CV results could be obtained through improved and effective

communication devices. One of the general conclusions regarding occurrence of scope

effect in CV studies (that basically use conventional preference elicitation approach) on

217

health risk is that WTP is not sensitive because, (i) respondents may not be sensitive to

variation in risk magnitude due to their lack of understanding of probabilities and poor

appreciation for numerical differences in magnitude (Kahneman and Tversky, 1973);

(ii) respondents do not treat the probabilities as applicable to them and base their values

on their prior beliefs and on the information contained in the scenario description.

Finally, the respondents may not value the health risk in line with the expected utility

theory (Hammitt and Graham, 1999). To conclude, we have analysed different kinds of

argument, supporting as well as rejecting the existence of embedding or scope effect in

CV results. In the case of studies reporting the existence of embedding, it is observed

that the fundamental flaws in the design of a survey instrument, improper

implementation of the survey, improper sampling procedure, inability of respondents’

understanding of the survey questions, the properties attributed to standard value theory

to substantiate the claims for embedding, etc., are some of the factors that are found to

cause this problem (Hanemann, 1994; Smith, 1992; Harrison, 1992). Mitchell and

Carson (1989) suggest survey design features to minimise the potential for part–whole

biases. Describing the larger and smaller commodities, and then asking respondents to

focus their attention on the smaller commodity, using maps and photographs to

describe the scenario, debriefing, providing opportunity to respondents to revise the

bids, etc., are some of the measures suggested to minimise the embedding in CV

studies (NOAA, 1993; Mitchell and Carson, 1989).

Dichotomous choice contingent valuation questions have gained popularity over

the last several years. This is due primarily to their purported advantages in avoiding

many of the biases known to be inherent in other formats used in the contingent

valuation (CV) method. Two standard references which discuss different CV

techniques are Cummings, Brookshire, and Schulze (1986) and Mitchell and Carson

218

(1989). Whereas several varieties of bias may be minimized by dichotomous choice

valuation questions, this elicitation method can be highly statistically inefficient in that

vastly larger numbers of observations are required to identify the underlying

distribution of resource values with any given degree of accuracy. An alternative

questioning strategy, intended to reduce this inefficiency, was first proposed and

implemented by Carson, Hanemann, and Mitchell (1986). They advocate introducing a

second offered threshold in a "follow-up" dichotomous choice CV question which

elicits a second discrete response. In practice, if a respondent indicates a willingness to

pay the first offered amount, the new threshold is about double the first one. If the

respondent is unwilling to pay the first offered amount, the second threshold is reduced

to about half the original amount. This questioning strategy has also been called a

"double-bounded referendum" approach.

Arrow et al. (1992) advocate discrete choice contingent valuation questions

over other formats in their assessment of the reliability of CV techniques for

quantifying passive use values in the context of oil spills. However, they note

parenthetically that "If a double-bounded dichotomous choice or some other question

form is used in order to obtain more information per respondent, experiments should be

developed to investigate biases that may be introduced”. This research addresses these

possible biases for this double-bounded case. Carson and Mitchell (1987) employ

survival analysis statistical techniques to analyze dichotomous choice with follow-up

data. These methods were originally conceived to handle product failure data collected

at irregular intervals. Much of this literature has emphasized Weibull distributions for

the variable in question. Hanemann, Loomis, and Kanninen (1991) use maximum

likelihood models to analyze double-bounded referendum contingent valuation data

under an assumption of normality. Both of these papers, however, maintain the

219

hypothesis that a single implicit true valuation drives respondents' answers to both of

the questions in this survey format. This paper proposes a more-general maintained

hypothesis. Our estimation method allows the valuation information elicited at each of

the two stages to be the same, or different, as the data dictate.

Production is always accompanied by waste according to the Material Balance

Principle. To meet the need of rapidly growing population, it is obvious that production

has to be increased by at least the population growth rate which leads to waste creation

that is beyond the absorptive capacity of the environment. The management of solid

waste is one of the challenges facing urban areas in the world. An aggregation of

human settlement has the potential to produce a large amount of solid waste. Municipal

governments in the developed world have generally assumed the collection, transfer

and disposal of waste. Municipal Solid Waste (MSW) management has been a major

issue of concern for many under-developed nations, however, as population increases

(Zerbock, 2003). The problem of municipal solid waste management is compounded

as many nations continue to urbanize rapidly; 30-50% of the population in many

developing countries is urban (Thomas-hope, 1998) and in many African countries the

growth rate of urban areas exceeds 4% per annum (Senkoro, 2003).

Although developing nations do spend between 20 and 40% of municipal

revenues on waste management this is often unable to keep pace with the scope of the

problem. In fact, when African countries were asked by the World Health Organization

(WHO) to prioritize their environmental health concerns, the results revealed that while

solid waste was identified as the second most important problem (after water quality),

less than 30% of the urban population have access to proper and regular garbage

removals (Zerbock, 2003). Most attempts to improve solid waste management in cities

220

in developing countries, like India, have focused on the technical aspects of different

means of collection and disposal. Recently much attention has been given to

investigating the demand side aspects related to solid waste management (Hartwick,

1998). In India also the provision of solid waste management services had for long

been from municipality which had focused on the technical aspects of collecting and

disposal of solid waste. These arrangements to solve the problem of solid waste in

urban areas of the country have not been achieving the desired goal. This resulted in

searching for another alternative that better minimizes the problem of urban solid

waste. It was this trend that pushed / forced the Indian government to form an agency

that assumed the full responsibility of administering the management of municipal solid

waste since 2003. Since then there is an inclination towards the inclusion of the

demand side aspects of solid waste management in addressing the problem of

municipal solid waste management. Thus it is important to find efficient ways to treat

solid waste. The improvement in environmental quality these treatments produce may

lead to increase in individual and social welfare for which, in principle, there should be

a positive willingness to pay (WTP).

Developing countries have a range of solid waste problems, including:

inadequate waste collection systems, open dumping and other forms of improper final

disposal and the resulting environmental pollution, scavenging at landfill sites by waste

pickers, and illegal dumping. These problems are being aggravated by growing waste

generation rates associated with economic growth, increases in consumption levels, and

the transition to mass consumption lifestyles in developing countries. There is concern

that these problems, if left unaddressed, will become a serious challenge for

generations to come. This concern has been shared by the international community

since the 1990s. Agenda 21, a global action plan for sustainable development adopted

221

at the UN Conference on Environment Development in Rio de Janeiro (the Earth

Summit) in 1992, called for the environmentally sound management of solid wastes,

among other priority issues (Toru, 2004). A major implication of the pattern of

urbanization and rapid population growth in developing countries, like India, is

expanding adequate infrastructures, waste management service being one of them.

However, it seems that this has not been the case in developing countries in general and

in India in particular.

For instance, waste generation every day in Addis Ababa was estimated to be

1386m3 out of which 750m3 was left uncollected. That means out of the total amount

of waste generated about 46% remained uncollected and the rest 54% was collected

(ENDA, 1996/7). In another study, the amount of waste generated in Addis Ababa in

2003 was estimated to be 1841.8m3 every day (0.22kg per household per day on

average). Out of this, about 50 - 60 % is collected and the rest is unattended

(Walelegne, 2003). The above examples may imply that ongoing SWM has become a

major problem both for the concerned authority and the population in Addis Ababa in

terms of the environment, economy and health. It is common observing the sanitation

workers busy removing garbage from roads, drainage systems, and streets in Addis

Ababa, the implication of which is extra cost is being incurred. Thus it is reasonable to

say that the amount of solid waste left uncollected in the city of Addis Ababa is

significant implying that the service is not being given satisfactorily for the population

living (residing) in it. If the accumulation and diffusion of solid wastes in this city

continues the way it is now, property values, scenery and aesthetic values,

environmental and health problems will get worse and in general quality of life

deteriorates. To reduce the adverse effects of the solid waste in terms of economic,

222

health and environmental impacts and provide efficient service for the population of

concern there has to be a way out of it, and this is partly addressed in this study.

Currently one of the major problems facing urban cities of developing

countries, like Ethiopia, is lack of efficient infrastructure such as solid waste

management service for urban households which was assumed by the municipality and

efforts made to improve it emphasized to a large extent on the service providers only,

ignoring or giving less emphasis to the demand side. Therefore, this study, as it focuses

on the demand side of the SW management service, is expected to provide the

following valuable information both for the public and private policy makers

concerning solid waste management services in designing of municipal SW

management service and service charge rates, and in understanding the role that

households can play in solid waste management. Another important contribution of

this study is that it can be used as the starting point in applying Choice Modelling (CM)

which to our knowledge has not been applied in any of the environmental valuation

studies conducted so far on SW management service either in Addis Ababa or in other

towns in the country. In general, knowledge obtained from this study may help to

reduce the gap between the service that will be provided and the public demand for SW

services.

In Ethiopia, particularly in Addis Ababa, a lot of contingent valuation method

(CVM) based studies have been made on improvement of solid waste management

services in the last decade. To begin with, Yitbarek1 (1998) undertook a CV survey on

solid waste collection in A.A by selecting 210 households in which a face to face

interview was conducted and open ended question method was applied to elicit

respondent’s willingness to pay (WTP). In his research, linear and log-linear models

223

(OLS) were used to study the significance of the independent variables- respondent’s

interest in environmental problems, awareness of solid waste problems, perceived level

of the existing solid waste collection system, responsibility or various solid waste

problem, age, sex, family size, level of education and house hold income. Of all these

variables, awareness of solid waste problem, household income and interest in

environmental problems were important factors determining respondents’ WTP values

(Yitbarek 2003). Aklilu2 (2002) used CVM, based on closed ended with a follow-up

format, to elicit willingness to pay for an improved solid waste management of

households’ of Addis Ababa, drawing 430 households at random. The model used in

the study, the Tobit model, showed that the income of households, time spent in the

area, quantity of waste generated, responsibility of solid waste management, education

dummies, ownership of the house and the number of children had a positive and

significant effect on WTP. Hagos (2003) also used CVM in his study to elicit

individual willingness to pay for improved solid waste collection and disposal services

for Mekele town. He employed an open ended with the iterative bidding game format

and selected a total of 164 households using stratified sampling based on the smallest

administrative unit ‘Kebele’ thereby applying systematic random sampling for selecting

households from each stratum. He employed Ordinary Least Square (OLS) in

estimating the bid function where the Willingness To Pay (WTP), is function of sex,

age, education, household size, household income, house ownership, household

awareness about SW problem, household satisfaction with the existing level of SW

service. Of these variables, household’s income, awareness about SW problem, age,

size of the household, were found to significantly influence the dependent variable

(WTP).The remaining explanatory variables were found insignificant. Walelegne

(2003) conducted a CVM survey on valuation of improved solid waste management in

224

Addis Ababa. He selected a sample of 500 households using two stage stratified

sampling followed by a random sampling applied on each stratum to ensure

representative ness. The information on different characteristics of households obtained

from the survey was used to test the model that explains WTP for improved solid waste

services as a function of income age, sex, education, and number of children, wealth,

and interest in environmental issues, service provider dummies and starting bids. The

results using Ordinary Least Square (OLS) and the result showed that income,

education, wealth, and types of service (private versus public) were significant while

the remaining repressors were found insignificant in influencing households’ WTP for

improved solid waste management service. The literature review shows that almost all

the studies made to evaluate the non-market environmental goods in Addis Ababa and

other towns in the country, particularly on solid waste management services, employed

the contingent valuation technique (CVM) to elicit consumer’s willingness to pay for

an improvement in the non marketed environmental goods. The potential socio-

economic, demographic and other factors that influence household’s willingness to pay

for improved solid waste management options attribute can be identified from the

literature review. It also shows that majority of the studies used the Ordinary Least

Square (OLS) for their estimation. It is clear from the literature review that the WTP

was used to estimate only the total economic value to a society of a marginal increase

in the non-marketed environmental good as a whole as CVM cannot value the

environmental improvement brought about by an increase in a solid waste management

attribute. CVM as a technique to reveal individual’s preference for non marketed

environmental commodities also suffers from several potential biases such as strategic

bias, information bias, hypothetical bias, constant budget bias and interviewer or non-

respondent bias.

225

Given the above background and the flaws with respect to CVM as a technique

of revealing individual preference for non-marketed environmental commodities and

others, it is reasonable to find a technique that takes into account the pitfalls in CVM as

a technique in revealing individual preferences for valuing the marginal as well as the

aggregate benefits brought about by a change in an attribute in solid waste management

option. The technique thought appropriate for the proposed study, is the Choice

Modelling (CM). CM has never been used, as to the best of our knowledge, in valuing

the benefits to a society of improvement in environmental quality as a result of a

marginal change in an attribute of a SW management option in Addis Ababa in

particular, in Ethiopia.

Let us go into the research area where CV can be used as a tool to measure the

Willingness to Pay to improve Solid Waste Management (SWM) in Tirunelveli

Corporation which is being highly environmentally deteriorated due to various

improper management.

Today, the municipal solid waste affects environment in several ways and

challenges faced by Tirunelveli Corporation is one of the most serious. Therefore, this

study would bring about the existing practice of solid waste disposal is ineffective

which cause unhygienic environment and thereby to frame a new strategy to improve

the solid waste management from the threshold level. However, the chronic problems

of effective municipal solid waste disposal throughout Tirunelveli city are daunting

task. The waste collection in India is much unorganized. The collection bins used in

various cities are neither properly designed nor properly located and maintained. This

has resulted in the poor collection efficiency. The average collection efficiency for

MSW in Indian cities and states is about 70% (Khan, 1994; Maudgal, 1995; Gupta et

226

al., 1998; Nema, 2004; Rathi, 2006; and Siddiqui et al., 2006). The Central Pollution

Control Board (CPCB) has collected data for the299 Class-I cities to determine the

mode of collection of MSW. It has been observed that manual collection comprises

50%, while collection using trucks comprises only 49% (CPCB, 2000b).

Under this circumstance, it is every individual’s responsibility to preserve the

urban environment very sound which preserves the environment in a sustainable

manner for future generation. Thus the question must be raised – where does the

money come from? Inevitably, its the responsibility of the urban dwellers who have

been resorting good environment for their healthy livelihoods. At the end, however,

everyone in the city-dwellers and business alike- have to bear the operating and

maintaining cost for clean environment, since all stand to gain in terms of better

environment. In carrying out a survey about their Willingness to Pay, the people who

pay must be convinced that the end result is worth the expense.

The analysis was divided into two parts. The first explored the respondents’

willingness and ability to pay for the improving Solid Waste Management. The second

part formulated econometric models to estimate the respondents’ WTP.

Here, the researcher has used regression models in which the dependent or

response variable itself could be dichotomous in nature. Basically, a ‘Yes’ or ‘No’ type

answer was received from the respondents with regard to the MSW. The researcher has

used 1 or 0 value to measure this. Some of the respondents were willing to pay and

some were not. To estimate and infer problems, Logit Model is used. The researcher

has classified all categories according to their actual contribution, in terms of rupees, to

avoid foul smell or odours nearby his/her residence. It founds that nearly 33.73 per cent

of the respondents were not willing to pay to improve MSW. To measure the actual

227

contribution of the respondents for MSW management for better management, the

researcher has used the Tobit Model.

WILLINGNESS TO PAY

The respondents’ willingness and ability to pay for improving MSW

management for better environmental quality provides useful information not only in

understanding the basic characteristics of the respondents but also in properly

distinguishing them when formulating WTP estimates.

Table:-5.1. WTP responses towards Municipal Solid Waste

S. No Opinion Percent

1 Willing and able to pay 66.27

2 Willing but unable to pay Nil

3 Able but unwilling to pay 33.73

4 Unable and unwilling to pay Nil

5 Others Nil

Source: Computed data

The above table 5.1 shows that majority of the respondents (66.276%) are both

able and willing to pay for the service. Nearly 33.73 percent of the respondents are

unwilling to pay to improve solid waste management for diverse reasons. For example,

they said they have already paid taxes to the Corporation, such a facility should be

provided by the Corporation or most of the solid waste generates from the industrial

sector and some extent affordable people are highly responsible for more solid waste

generation.

228

WILLINGES TO PAY ESTIMATES

The analysis on WTP estimates was based on qualitative approach, and it

attempts to explore the factors governing the decision whether a person was willing to

pay for improve MSW in Tirunelveli Corporation for better environmental quality. The

exploration of whether a person is willing to pay for the improving MSW management

was done using a logit model. The model was chosen because of its ability to deal with

a dichotomous dependent variable and a well-established theoretical background.

Consider, the specification of the Logit equation is,

WTP=a+b1AGE+b2SEX+b3MS+b4INCOME+b5DIS+b6RUPWTP+b7HCOST+

b8WLOSS+b9PRIEDU+b10HEDU+b11DEDU+b12PRI+b13GOVT+b14BUSI+

b15PQUAL+b16MQUAL+b17FAIRLY+b17HIGHLY+Ui

Where,

Dependent variable WTP=1; if willing to pay for solid waste improvement=yes

= 0 otherwise

Dummy Independent variables (Description)

b1 = Age of the respondent (years)

b2 = 1 if Sex = Male, 0= otherwise

b3 = 1 if Married, 0=otherwise

b4 = 1 if Income earner, 0=otherwise

b5 = 1 if Distance is closer to Dustbin, 0=otherwise

b6 = 1 if Rupee Willingness to Pay, 0=otherwise

b7 = 1 if Health Cost is high, 0=otherwise

b8 = 1 if Wage loss, 0=otherwise

b9 = 1 if primary educated, 0=otherwise

229

b10 = 1 if high school educated, 0=otherwise

b11 = 1 if graduated, 0=otherwise

*Base Category – Illiterates (for education)

b12 = 1 if private employee, 0=otherwise

b13 = if govt. employee, 0=otherwise

b14 = 1 if business, 0 =otherwise

**Base category- Unemployed

b15 = 1 if SWM is poor, 0=otherwise

b16 = 1 if SWM is fair, 0 =otherwise

*** Base category- Very Poor

b13 = 1 if the person is fairly accepted to improve Solid Waste, 0=otherwise

b14 = 1 if the person is highly accepted to improve Solid Waste, 0=otherwise

****Base category – not at all interested to pay WTP for SWM

230

Table : 5.2

Logit Estimates of WTP for Improvement of Solid Waste Management

Dependent variable: WTP (Willingness to Pay to improve SWM)

Independent Variable Co-efficient Marginal Effects

CONSTANT 0.39776 (2.51) 0.0119

AGE -0.00109 (-0.83)*** 0.4067

SEX 0.03530 (0.89)** 0.3734

MS -0.00453 (-0.08)*** 0.9361

INCOME -9.71238 (-3.59)* 0.0003

DISTANCE 7.20779 (0.73)** 0.4642

RUPWTP 0.00557 (19.51)* 2.8865

HCOST 1.42612 (0.12)*** 0.9033

WAGE LOSS 1.02141 (1.36)** 0.1721

PRIMARY 0.12401 (1.93)** 0.0524

HIGH 0.09747 (1.55)** 0.1193

DEGREE 0.18956 (3.08)* 0.0020

PRIVATE -0.20014 (-1.75)*** 0.0784

GOVT -0.15817 (-1.44)*** 0.1471

BUSINESS -0.16333 (-1.43)*** 0.1511

POOR -0.00644 (-0.15)*** 0.8793

MIDDLE 0.04714 (1.04)** 0.2980

FAIRLY 0.05496 (1.04)** 0.2945

GREATLY 0.11306 (2.32)* 0.0199

Log likelihood 113. 0455

Restricted log likelihood 129. 8231

Chi-square 46.10672

Pseudo R2 0.37

Source: Computed from primary data

Note: figures in parenthesis show the t-values

*Statistically significant at the 1% level;

** Statistically significant at the 5% level;

*** Statistically significant at the 10 % level

231

The actual estimation in the logit model is to formulate to capture a simple

Yes/No answer on whether a respondent’s would pay to improve municipal solid waste

management in Tirunelveli city. All the data were employed at this stage in order to

understand a broad perspective of the factors underlying a respondent’s decision. This

survey report of WTP for the improvement of solid waste management has been

included; the data from various stretches were pooled to get a WTP group. Out of 510

respondents that were questioned on their willingness to pay to improve solid waste

management, 322 that is, 63.14 percent gave positive answers while the rest gave

negative answers. These results have several noteworthy features. First, the model has

a good fit. The chi-square value is 46.10, which is highly significant at 1 percent.

Pseudo R2 value is 0.37, which means that about 37 percent of the variables in WTP are

explained by the included independent variables. Almost all the independent variables

have positive influence on WTP except the variables of occupational classifications.

Age variable is positively related to WTP. As the age goes up, the probability of WTP

also increases. The variables – sex, income, education and distance – have a higher

probability of influencing WTP for improving municipal solid waste management in

Tirunelveli city. In the case of sex variable, there was a higher probability of positive

response from female respondents towards improving solid waste management

compared to male respondents. Awareness and income level was also higher among

the female respondents compared to the males. When the respondent’s income rises by

1 percent, the probability WTP for better municipal solid waste quality also rises by

0.0219 percent. Distance has a positive significance. Respondents residing closer to

street dust bins showed higher probability for WTP for improving solid waste

management. WTP dropped when the street dust bins little away from the dwellings.

Those dwellings near the landfills or disposal areas are important factors to decide

positive influence on WTP since the coefficients of the variable is positive, it implies

232

that a respondent, who has a higher education, knows about the importance of

improving solid waste management in urban city and who has a higher probability of

paying for it. Age and income are having a positive significance at 1 and 5 percent and

sex has 5 percent level of significance.

The variable health cost also play a major role in determining WTP. If the

health cost was high, the probability of WTP for improving the solid waste

management increased at 10 percent level of significance. Wage loss was less deciding

criterion for WTP among the respondents. It directly influenced WTP and indicated a

positive sign and significance at 5 percent. It was a good sign among the respondents

that deterioration of health was due to improper management of solid waste disposal

and something has to be done to improve the surrounding environment.

Education at Primary and High school education had a negative sign and

significance at 5 percent. But education at degree level had a positive sign at 1 percent

level of significance. Thus, education may also be interpreted as a proxy for the

knowledge about the poor quality of solid waste management practices taking place in

Tirunelveli city and it clearly highlighted the importance of education at graduation

level. As the level of education goes up, the probability of WTP for improving solid

waste management goes up. This was evident with respondents who had degree level

education and those who had primary and high school level education do not have

much aware of municipal solid waste improvement. It had a positive sign and

significance at 5 percent confirming the earlier results.

Occupation with private, government and business groups had a negative

impact on WTP. The respondents who have been working in the private sector are

getting lower wages causing poor response towards WTP for improving municipal

solid waste management.

233

Table:-5.3. Tobit Estimates of WTP for Solid Waste Management Improvement

Dependent variable: Willingness to Pay (WTP)

Independent Variables Co-efficient Marginal Effects

CONSTNAT 0.18581 (0.79) 0.4266

AGE -0.00168 (-0.86)*** 0.3865

SEX 0.05621 (0.96)** 0.3332

MS -0.00250 (-0.02)*** 0.9763

INCOME -1.4628 (-3.65)* 0.0002

DISTANCE 9.93958 (0.68)** 0.4957

RUPWTP 0.00764 (17.09)* 2.8865

HCOST 1.07757 (0.06)*** 0.9515

WAGE LOSS 1.41199 (1.29)** 0.1947

PRIMARY 0.19387 (1.97)** 0.0483

HIGH SCHOOL 0.14875 (1.54)** 0.1227

DEGREE 0.2856 (3.05)* 0.0022

PRIVATE -0.29345 (-1.75)*** 0.0786

GOVERNMENT -0.22826 (-1.42)** 0.1528

BUSINESS -0.23514 (-1.40)*** 0.1592

POOR -0.00925 (-0.14)*** 0.8834

MIDDLE 0.07359 (1.09)** 0.2720

FAIRLY 0.08049 (1.02)** 0.3056

GREATELY 0.16324 (2.23)* 0.0251

T.NALLUR 0.00904 (0.13)*** 0.8932

P.KOTTAI -0.06933 (-0.92)*** 0.3570

M.PALAYAM -0.02614 (-0.35)*** 0.7257

Sigma 0.478

Likelihood function 2029.1032

N 338

Source: Computed primary data

Note: figures in parenthesis show the t-values

*Statistically significant at the 1% level;

** Statistically significant at the 5% level;

*** Statistically significant at the 10 % level

234

Under this Tobit model, each variable has its own significance. In this model,

the actual value of WTP is directly linked with the respondent’s Willingness to Pay for

improving municipal solid waste management. If the coefficient sign were positive, one

unit increase in age when other things remain constant would increase the WTP amount

by about 0.14 percent. Sigma value (0.478) is highly significant. Because of this OLS

is an unbiased estimate, which is highly significant, and it shows that leaving the

sample would lead to selection bias. It is the case with that of the variable sex. It

would increase the probability of WTP by about 0.3332 percentages for 5 percent. If

the distance decreases, the probability of WTP increases by 0.4957 percent. Health

cost and wage loss have negative influence on WTP and some extent variable health

has got positive influence at 5 percent level of significance. It would increase the

probability by 0.9515 percent and 0.1947 percent respectively. Education has much

influence. As expected, it had improved the WTP amount at each level of education,

namely, primary, high school and degree level, by about 0.0483 percent, 0.1227 percent

and 0.0022 percent respectively. Occupation doesn’t care about the poor maintenance

of solid waste management and less concerned about improving it. Therefore, as

predicted in theory, so many reasons influence them to deviate. It has negative sign

and insignificant too. However, the variables government employees and business

group have a positive effect on WTP. In contrast, one unit increase in the business

group variable will influence WTP by 0.1592 percent as against the government

employee, which is just 0.1528 percent. The respondents have responded to the

personal health loss due to solid waste as a subjective variable that influences an

increase in the WTP amount by 0.3056 and 0.0251 percent, respectively, for fairly

affected and highly affected variables. Thus, the Tobit Model is a good model to

explain the results.

235

The variable of different zones also influences WTP for improving municipal

solid waste management. One unit rise in all three zones would increase WTP by about

0.8932, 0.3570 and 0.7257 percent respectively. Melapalayam (M. Palayam) zone is

being generated more solid waste might have positively influence towards WTP since

New Bus Stand comes under this zone and it also has huge Muslim population who

have more mutton and chicken stalls to generate huge volume of solid waste in this

region. Another significant factor needs to be noted that, the lifestyle of the people in

this zone is also one of the influence factor to generate more solid waste. The

percentage of respondents who said ‘yes’ for WTP is less compared to other zones,

whereas the actual mean WTP is high for Palayamkottai and Thatchanallur zones

showing their high income status. However, WTP is not much attracted as we expected

in all the zones towards solid waste management. It has the negative sign but positive

insignificant value. .

Above all, it might be possible and correct to directly formulate a linear

regression model using Ordinary Least Square (OLS) method using maximum WTP

figures as a dependent variable. Similar to WTP, the data employed also exclude that

of protest bid. In addition, outliers are also identified and excluded from the estimation

process. The outcome of the OLS model is given in Table 5.2 and the outcomes of

both linear models are mostly as expected, not only in terms of variables that affect the

fees but also their signs and levels of significance. For WTP, coefficients of

referendum fee, level of education, income, need of the study, severity of existing solid

waste management practices and whether a respondent is living near street dust bins are

observed to be significant. Their positive relationships are also consistent with what

we have predicted. It should be noted that sex seems to play a significant role in

determining WTP in this equation. However, a relatively low R-squared is a bit of

236

concern. The explanation might lie on the fact that the actual fee (value of WTP) that

respondents are willing to sacrifice has a very wide dispersion and their increments are

very small, making it difficult for a model to precisely determine each individual

figure. Results indicate a significant positive relationship between the referendum fee

and the stated fee.

Regression results indicate that 46. 10 percent of the variation in WTP was

explained by the hypothesized household characteristics R2 0.37. Only age and marital

status were not significant in explaining household WTP, the latter variable perhaps

due to multi-colinearity. All coefficients for the income and education variables were

highly significant and negative as expected, suggesting that respondents who were

degree holders (17.25%) with household’s incomes of Rs. 15,001 to Rs. 20,000 or more

were willing to pay significantly more than those in other income and education

categories. Hypothesis testing indicates that respondents who were degree holders with

household incomes between Rs.15,001 to Rs.20,000 were also willing to pay

significantly more than most respondents with less education and equal or less income.

This finding, in conjunction with the lack of significance of most other differences in

the above table suggests that income and education may not significantly influence

WTP unless a respondent has a degree holder. The significance of the coefficient for

income and education suggests that degree holders who were in the highest income

class were willing to pay about Rs.300 more than degree holders with household

incomes between Rs.20,001 and Rs.25,000. Hence, for degree holders, the level of

income appears to be important.

The coefficients for sex, 0.96 had their hypothesized signs and were highly

significant. Female respondents (SEX) were willing to pay about Rs. 175 more than

237

male respondents, while respondents who depended on piped water or bottled water for

drinking were willing to pay about Rs.60 less than those who relied on bore well or

river water. Those who said they were very concerned about health risks from the

proposed landfill were willing to pay Rs.300 more than those who said they were

unconcerned. Another major finding of the results show that household size

significantly reduced WTP only for households with more than four members. It seems

possible that the lack of difference in WTP between the one-to-three person households

may be due to likely presence of children in the latter offsetting the effect of the lower

per capita income on ability to pay. The year of residence in Tirunelveli is also

significantly determines WTP. However, certain anomalies existed in its parameter

estimates. Because of the residence loyalty, one might expect that the respondents who

had lived in Tirunelveli City longer to be willing to pay more than those who had

moved in more recently. Another interesting point to be noted in the field survey, the

household characteristic that exhibited an unusual patter in its coefficients was that the

household’s distance from the dustbins. Respondents who lived closer to the dust bin

were expected to be willing to pay more than those who lived farther away. This

pattern holds for distance of up to 500 meters, with households who lived within 100

meter or 200 meter willing to pay between Rs.200 and Rs. 150 more than those who

lived between 300 to 500 meters. However, those who lived more than 300 meters

from the dust bin were not willing to pay significantly less than those who living within

200 meters. They were, in fact, willing to pay significantly more than those households

located between 300 meter to 500 meters.

These findings are valuable to policymakers for several reasons. First,

Tirunelveli Corporation is to develop comprehensive waste management plans could

use a similar approach in evaluating the external costs or benefits of all wastes disposal

238

alternatives including using of street dust bins, landfill disposal, incineration and

recycling. If such expenditures are to be made in a cost-effective manner, more

complete analysis is needed to compare the total costs of all solid waste disposal

alternatives. Second, if minimizing overall costs were the only objective and if similar

results were found to hold for other areas, one might conclude that landfills should be

sited in areas with fewer degree holders in higher income classes. However, equity

considerations would likely limit acknowledgement of such a strategy, at least

explicitly or officially.

Conclusion

The increasing threat posed to human health as a consequence of improper way

of MSW dumping in Tirunelveli Corporation has become burning issue in recent years.

This research work has attempted to introduce environmental tax in order to improve

solid waste collection by the way of WTP. The results of the WTP indicate that

unhygienic conditions and mosquito menace due to unplanned dumping of MSW. The

survey result shows that the respondents are well aware of the present situation of

MSW collection and management by the Corporation and the necessity of their

participation to maintain the city clean and tidy. The study results also give a positive

scope for introducing environmental tax (user fee). Most of the WTP studies carried

out in developing countries in the past have been mainly limited to the estimation of

user’s mean WTP and sometimes been controversial. This research has pioneering

attempt to extend the use of WTP survey results indicating that charging for improving

the solid waste management in Tirunelveli Corporation may not have negative impact

amongst the city dwellers. In this case of observed behaviour method, the assumptions

made about the use of WTP for MSW quality improvement services may be far from

239

true in the developing countries. Also, there could be difficulties in charging excessive

amounts as WTP because of low levels of education and incomplete perception about

environmental values due to lack of awareness. The paucity of adequate data on the

degree of solid waste and its effects on people’s health cannot screen the fact.

Generally, the community and nature in general can only be speculated but the fast

deteriorating trend of the urban civic environment can never been denied. It has been

discussed (See Chapter IV) that the rate of vector borne diseases has been increasing at

an alarming rate in recent days due to improper maintenance of MSWM. If the present

quantum of MSW generation continues to exist in future without taking preventive

steps by the Tirunelveli Corporation to control it, the society’s foregone health

expenditure would become many folds which menace huge burden among the people.

The most important requirement is to do the proper urban MSWM planning and

execution. The District administration is needed to help with the corporation hand in

hand for better implementation of solid waste regulation among the public in order to

achieve sustainable civic environment.

240

References

1. Baker, Brian (1982). Land Values Surrounding Waste Disposal Facilities.

Research Report, Department of Agricultural Economics, Cornell University.

Ithaca, New York.

2. Freemen, A. Myrick, III (1986). “On Assessing the State of the Art of the

Contingent Valuation Method of Valuing Environmental Changes.” Valuing

Environmental Goods: A State of the Arts Assessment of the Contingent

Valuation Method, Ronald G. Comings, David S. Brookshire, and William

D. Schulze, eds., Totawa, New Jersey: Rowman and Allanheld.

3. Havlicek, Joseph Jr., Robert Richardson, and Lloyd Davies (1971) “Measuring

the Impacts of Solid Waste Disposal Site Location on Property Values”

American Journal of Agricultural Economics, 53:869.

4. Arrow, Kenneth (1969) “The organization of economic activity: issues pertinent

to the choice of market vs. non-market allocation”, The Analysis and Evaluation

of Public Expenditures: The PPB System. Washington DC: Joint Economic

Committee, 91st

Congress, 47-64.

5. Carson, Richard T. y Robert C. Mitchell, (1993) “The value of clean water: the

public’s willingness to pay for boatable, fishable, and swimmable quality

water” Water Resources Research, Vol. 29, No. 7, Pp.2445-2454.

6. Diamond, Peter A. y Jerry A. Hausman, “Contingent valuation: is some number

better than no number” Journal of Economic Perspectives. Vol. 8, No. 4,

pp. 45-64.

241

7. Duston, Thomas E. (1993) Recycling Solid Waste: The First Choice for Private

and Public Sector Management Edit. Quorum USA.

8. Environmental Protection Agency (2000), The Municipal Solid Waste Fact

book, Internet version. Office of Solid Waste, United States. Last update May

15, 2000. http://www.epa.gov/epaoswer/non-hw/muncpl/factbook/

9. Hanemann, Michael W. (1994) “Valuing the environment through contingent

valuation”, Journal of Economic Perspectives, Vol. 8 , No. 4, Pp. 19-43.

10. Hanley, Nick, Jason F. Shogran, y Ben White (1997) Environmental Economics

in Theory and Practice. Oxford University Press.

11. Holmes, John R. (1984) Managing Solid Wastes in Developing Countries Edit.

Wiley, USA.

12. Platteau, J.P. (1994) “Behind the Market Stage: Where Real Societies Exist”,

Journal of Development Studies.

13. Portney, R. Paul (1994) “The contingent valuation debate: why economist

should care” Journal of Economic Perspectives. Vol. 8, No. 4, Pp. 1-17.

14. Arlosoroff, Saul 1982. “WB/UNDP Integrated resource recovery project:

Recycling of wastes in developing countries.” In: Appropriate Waste

Management for Developing Countries, edited by Kriton Curi. New York:

Plenum Press, 1985.

15. Arrow, J.K., R. Solow, P.R. Portney, E.E. Leamer, R. Radner, and

H. Schumand (1993) Report of the NOAA Panel on Contingent Valuation to the

General Council of the US National Oceanic and Atmospheric Administration,

Resources for the Future, Washington, DC .

242

16. Bartone, C.L & Bernstein, J.D (1993). “Improving Municipal Solid Waste

Management in Third World Countries", Resources, Conservation and

Recycling; 8; 43-45.

17. Bateman I. J. and Turner R. (2000) Evaluation of the Environment: The

Contingent Valuation Method: The Centre for Social and Economic Research

on the Global Environment (CSERGE) Working Paper GEC 92-18.

18. Cameron, T.A., and M.D James (1987) ‘Efficient Estimation Methods for Use

with Closed-Ended Contingent Valuation Surveys Data’, the Review of

Economics and Statistics, 69: 269–276.

19. Carson R.T., Flores, N., and Meade N. (2001), Contingent Valuation:

Controversies and Evidence, Environmental and Resource Economics 19(2):

173-210.

20. Carson, R.T. (2002) Contingent Valuation: A Comprehensive Bibliography and

History, Cheltenham: Edward Edgar.

21. Christian Zurbrügg Urban Solid Waste Management in Low-Income Countries

of Asia How to Cope with the Garbage Crisis.

22. Cointreau, S. J. (1982), Environmental management of urban solid wastes in

developing countries: a project guide, Urban Development Dept: The World

Bank.

23. Cornwell, D. (1998) Introduction to environmental Engineering 3rd

ed.

McGraw-Hill Company: USA Department of Economics, University of

California (2005) ‘Recent Literature on Contingent Valuation Method for

Valuing, Environmental Goods’, University of California, Los Angeles

243

24. Economic Research Unit Indian Statistical Institute, (2003) Economics of Solid

Waste Management: A Survey of Existing Literature.

25. Egner, Eldon D, and Smith, Bradley F (2006) Environmental Science: A study

of interrelationship 10th

ed McGraw-Hill company: New York USA.

26. El-Fadel, M., Findikakis, A.N., and Leckie, J.O. (1997), “Environmental

impacts of Solid Waste Land filling”, Journal of Environmental Management

50(1): 1-25.

27. Haab, C.T., and K.E. McConnell (2002) Valuing Environmental and Natural

Resources: The Econometrics of Non-Market Valuation, Cheltenham: Edward

Elgar.

28. Hanemann, W.M. (1984) ‘Welfare Evaluations in Contingent Valuation

Experiments with Discrete Responses’, American Journal of Agricultural

Economics, 66: 332–341.

29. Hanemann, M.W. (1994) Valuing the Environment Through contingent

valuation method: Journal of economic perspectives, Vol. 8 No.4 PP 19-43.

30. IDB (2003), “Economic Instruments for Solid Waste Management: Global

Review and Applications for Latin America and the Caribbean”, Environment

Network regional policy dialogue study series, Inter-American Development

Bank, Washington DC.

31. Jamal, O. (2002), “Household preference for solid waste management”, the

economy and environment program for South East Asia (EEPSEA) publication,

Malaysia.

244

32. Seleshi Fantu (2007) Household heads’ Willingness to Pay for improved Solid

Waste Management: The case of common building residence in Addis Ababa,

Master thesis, Department of economics, Addis Ababa.

33. Solomon Tarfasa (2007) Households Willingness To Pay for Solid Waste

Management Options, the case of Yeka sub-city, Addis Ababa, Master’s thesis,

Addis Ababa University, Department of economics, Addis Ababa, Ethiopia.

34. Teklay Tesfay (2004). An integrated approach to solid waste management in

Mekelle City, The 86th

International course on housing and urban development

(ICHUD) Rotterdam, the Netherlands.

35. UNEP (2004). The Use of Economic Instruments in Environmental Policy:

Opportunities and Challenges, Geneva.

36. UNEP, (2005). Selection, Design and Implementation of Economic

Instruments in the Solid Waste Management Sector in Kenya, the Case of

Plastic Bags.

37. Zerbock, O. (2003). Urban Solid Waste Management: Waste Reduction in

Developing Nations, School of Forest Resources & Environmental Science,

Michigan Technological University.