2ème Conférence annuelle de la FAERE – 10-11 Septembre 2015 – Toulouse The value of forest elephant conservation for indigenous and local communities in the Tridom landscape - Combining Open-Ended and Close-Ended Contingent Valuation methods (work in progress) Jonas Ngouhouo 1 , Jens Abildtrup 2 , Dénis Sonwa 3 , Philippe Delacote 2 1 The Laboratory of Forest Economics – INRA - CIFOR, PhD Candidate, [email protected]2 The Laboratory of Forest Economics – INRA 3 CIFOR - Central Africa Regionl Office Abstract The purpose of this paper is to determine and characterize the social and cultural preferences for forest elephant conservation in the Congo Basin’s Tridom landscape. Using data from a contingent valuation survey with 1035 households conducted in 2014 in 108 villages of landscape, the paper runs five different models comparing the Open-ended and the Closed-ended elicitation format, and finds that combining both formats would lead to approximating the true WTP. According to the hypothesis tested, we found the following outcomes. On average, local and indigenous households will to pay a CFA1139,4 (€1,74) monthly amount for forest elephant conservation. Unlike the second hypothesis, the estimates shows that the existence of human/elephant conflict, as well as the distance of the households‘ location to the nearest protected area do not influence their preferences for elephant conservation. Finally, as expected, the indigenousness has a positive and significant influence on the household’s preferences. This information is all the more important to increasing incentives for elephant conservation as the extinction of forest elephant would lead to a severe threat on spiritual enrichment, cultural identity as well as the way of life of this ethnic minority group. Furthermore, we found that improving household’s education level as well as an increase in their income lead to greater preferences for elephant conservation. Keywords: Forest Elephant Extinction, Local and indigenous people, Combining Open-Ended and Close- Ended formats, Contingent Valuation, Willingness-to-pay Highlights Combining Open-Ended and Close-ended elicitation formats to improve the Stated Preference Valuation Technique Original data from a recent (January - July 2014) face-to-face interview with 1035 households Issue of securing the minority indigenous Baka’s spiritual and cultural way of life through Forest Elephant conservation This work increases awareness of the forthcoming forest elephant loss as well as the incentives for conservation JEL Classification: Q 57, 29. 1. Introduction 1.1. Background and importance of forest elephants Forest Elephant (Loxodonta Africana cyclotis) poaching in Tropical Africa is a big threat for the dynamic of this iconic species. In 2011, The Congo Basin’s forest elephant population was less than 10% of its potential size, occupying less than 25% of its potential range (Maisels and Al, 2013; Blake et al., 2007; Martin and Stiles, 2000). The Tri-national Dja-Odzala-Minkebe’s cross-border landscape (Tridom), spanning Cameroon, Congo (R) and Gabon, deemed to have ecological and biodiversity uniqueness and hosting the most important population of elephant in the world, with the highest density in the Minkébé National Park
18
Embed
Background and importance of forest elephantsfaere.fr/...value-of-forest-elephant-conservation... · 1.1. Background and importance of forest elephants Forest Elephant (Loxodonta
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
2ème Conférence annuelle de la FAERE – 10-11 Septembre 2015 – Toulouse
The value of forest elephant conservation for indigenous and local communities in the Tridom landscape - Combining Open-Ended and Close-Ended Contingent
Valuation methods (work in progress)
Jonas Ngouhouo1, Jens Abildtrup
2, Dénis Sonwa
3, Philippe Delacote
2
1The Laboratory of Forest Economics – INRA - CIFOR, PhD Candidate, [email protected] 2The Laboratory of Forest Economics – INRA 3CIFOR - Central Africa Regionl Office
Abstract
The purpose of this paper is to determine and characterize the social and cultural preferences for forest
elephant conservation in the Congo Basin’s Tridom landscape. Using data from a contingent valuation
survey with 1035 households conducted in 2014 in 108 villages of landscape, the paper runs five different
models comparing the Open-ended and the Closed-ended elicitation format, and finds that combining both
formats would lead to approximating the true WTP. According to the hypothesis tested, we found the
following outcomes. On average, local and indigenous households will to pay a CFA1139,4 (€1,74) monthly
amount for forest elephant conservation. Unlike the second hypothesis, the estimates shows that the
existence of human/elephant conflict, as well as the distance of the households‘ location to the nearest
protected area do not influence their preferences for elephant conservation. Finally, as expected, the
indigenousness has a positive and significant influence on the household’s preferences. This information
is all the more important to increasing incentives for elephant conservation as the extinction of forest
elephant would lead to a severe threat on spiritual enrichment, cultural identity as well as the way of life of
this ethnic minority group. Furthermore, we found that improving household’s education level as well as an
increase in their income lead to greater preferences for elephant conservation.
Keywords: Forest Elephant Extinction, Local and indigenous people, Combining Open-Ended and Close-Ended formats, Contingent Valuation, Willingness-to-pay
Highlights
Combining Open-Ended and Close-ended elicitation formats to improve the Stated Preference Valuation
Technique
Original data from a recent (January - July 2014) face-to-face interview with 1035 households
Issue of securing the minority indigenous Baka’s spiritual and cultural way of life through Forest
Elephant conservation
This work increases awareness of the forthcoming forest elephant loss as well as the incentives for
conservation JEL Classification: Q 57, 29.
1. Introduction
1.1. Background and importance of forest elephants
Forest Elephant (Loxodonta Africana cyclotis) poaching in Tropical Africa is a big threat for the dynamic of
this iconic species. In 2011, The Congo Basin’s forest elephant population was less than 10% of its potential
size, occupying less than 25% of its potential range (Maisels and Al, 2013; Blake et al., 2007; Martin and
Stiles, 2000). The Tri-national Dja-Odzala-Minkebe’s cross-border landscape (Tridom), spanning Cameroon,
Congo (R) and Gabon, deemed to have ecological and biodiversity uniqueness and hosting the most
important population of elephant in the world, with the highest density in the Minkébé National Park
- Traditional knowledge - Ritual component (JENGI, YELI, TUMA, ME, ...) - Cultural identity of indigenous people. - Influencing the social structure of indigenous
ECO
NO
MIC
TOT
AL V
ALU
E
USE V
ALU
E N
ON
USE V
ALU
E
Direct
Use
Value
Indirect
Use
Value
Existence
Value
Option Value
Bequest Value
CONSUMPTIVE - Elephant meat (trade and consumption) - Tusk and Trophies value NON CONSUMPTIVE -Loxodontacyclotis viewing - Loxodontacyclotis riding - Spiritual/ cultural
- Loxodontacyclotis has value for individuals even though not using - Iconic species and unique forest
elephant’s species
- Potential use by future generation
- Possible use in the future
SUP
PO
RTIN
G SER
VIC
ES
Considering the importance for the ecological, socio-cultural and economic equilibria for the local
populations, the extinction of elephants will severely affect human welfare and would lead to irreplaceable
costs to the Tridom society. This makes forest elephants’ preservation a priority for biodiversity conservation
decisions. Thereby, assessing the non-use value and the cultural services taking into account the damage cost
of elephant for the local populations can provide relevant information to highlight the significant costs of
extinction. This could increase incentives for biodiversity conservation. Therefore, to the main objective of
this paper is to reply to the question “what is the willingness to pay of local households to avoid a total loss
of elephant?”
1.2. Objective, interest and hypothesis
This paper aims to determine and characterize the social and cultural preferences for forest elephant
conservation in the Tridom landscape. More precisely, it calculates and compares the willingness-to-pay of
the households to the economic damage caused by elephants and it analyses the factors that influence the
calculated WTP. This will provide important knowledge for conservation managers and design for
conservation policies.
The paper tests the following hypothesis:
1. The extinction of forest elephant can lead to a significant net lost in the household’s wellbeing.
2. The WTP for elephant’s preservation increases with the distance of the household’s location to the nearest
protected areas and decreases with the human-elephant conflicts presence. Indeed, the distance to the
protected area can be considered as an indicator of elephant scarcity. Assuming a decreasing marginal
utility from forest elephant presence leads to the first part of the hypothesis. Hence, households closed to
protected area would be likely to express lower preferences for elephants conservation
3. The WTP is significantly influenced by the indigenousness status of households. Knowledge about the
spatial and ethnic differences in WTP may be used for the design of optimal spatially explicit
conservation policies.
2. Literature review
A large body of research has contributed to enriching the literature on the economics of endangered species
conservation (Bishop, 1978; Barbier et al, 1990; Kremer and Morcom, 2000; Lindsey et al, 2007; Tisdell,
2002; Bulte and Kooten, 2002). However, only a few research papers have been assessing the indirect use
value, the bequest value and the existence value of elephant (Bandara and Tisdel, 2003, 2005), while this
iconic species plays an important role in terms of socio cultural and ecological integrity (Lewis, 2002; Blake
et al, 2009).
Bandara and Tisdell (2002) used data from a face-to-face Contingent Valuation study of a stratified sample
of 300 urban residents in Colombo to assess the willingness to pay for elephant conservation in Sri Lanka.
Their assessment allowed distinguishing between users and non-users values of Asian elephant. The
respondents who have ever used elephant facilities were willing-to-pay Rs.137,38, while the non-user were
willing-to-pay Rs. 82,96 for elephant conservation with an average of pay Rs. 110,17 (€1,65).The results
reveal that urban residents are willing-to-pay for elephant conservation because they want to secure the
existence of elephant (non-use value) and because they have an economic use of the presence of elephants,
i.e. their importance for recreational and tourism. Applying another methodology, using the same data, they
found that the probability of being willing-to-pay an amount among the bids values is significantly and
positively influenced by pro-conservation attitude, higher income and negatively influenced by the scale of
the bid value. They found that the total present discounted annual value (Rs. 40248 million) can yield a
perpetual return of Rs. 2012,43 million if invested, that is sufficient to compensate an annual crop damage
value that is Rs. 1121,42 million if the farmers continue allowing elephants some access to their crops for
food.
Even though a few number of research papers investigate the farmers' values of ecosystem services (Smith
and Sullivan, 2014) as well as the option value and the non-use value of African and Asian elephant as an
incentive to improve its conservation (Vredin, 1997; Bandara and Tisdell, 2003, 2004; Muchapondwa et al,
2009), no research has addressed forest elephants. These are currently severely threatened by poaching and
land acquisition for non-forested activities. Besides being the first paper to point out this non negligible
value of forest elephants, this paper considers the role of landscape factors like the distance between the
households and areas nearest elephant concentrations zones (Protected Areas), the relative density of the
protected areas and the households’ plantation size for WTP. Furthermore, this paper is based on a recent
and not previously exploited rich dataset obtained by face-to-face interviews.
3. Methodology In this paper, we use a Contingent Valuation survey to measure households’ WTP for avoiding reduction and
loss of forest elephant population in the Tridom. The application of CV in developing countries is
increasingly growing; however, the present study is the first to measure the impact the loss of an iconic
biological species in the Tridom landscape.
A main distinction between different CV approaches is OE and closed-ended mechanisms (Cameron and
James, 1987). In OE, or continuous CV questions, the respondents are asked to specify their WTP, while in
close-ended (CE), or discrete CV, respondents are asked to choose whether or not to pay a specified amount
(Kealy and Turner, 1993).
3.1. Valuation Technique As CE and OE mechanism may yield significantly different results for public goods due to the differences in
incentives for strategic behaviour (up., Kealy and Turner, 1993), we use both mechanisms to approximating
the true WTP. Indeed, Kealy and Turner (op. cit.) suggest using either one or the other format. The main
types of CE approaches consist of the single and the double (multiple) bounded dichotomous choice (DBDC)
(Hanemann, 1985 ; Carson, 1985 ; Hanemann and Kanninen, 1998).
We first use a DBDC contingent valuation model proposed by Hanemann (1985) and Carson (1985) to
estimate the WTP of the indigenous and local households in the Tridom. This method reduces the variance of
estimated WTP and therefore, is asymptotically more efficient than the single bounded method developed by
Bishop and Heberlein in 1979 (Hanemann et al., 1991; Bateman et al, 1994; Alberini, 1995; Haab and
McConnel, 2002). Indeed, yet the single bounded approach is quick and easy to administer and analyse, but
it is more likely to result in strategic bias (Harris and Roach, 2013). Nevertheless, the answer to the first
question when using the DBDC may sometimes be inconsistent with the response to second bid and
contribute to lowering the WTP (Hanemann et al, 1991; Herriges and Shogren, 1996).
The DBDC elicitation format can produce more consistent values than the OE format when the anchor is
well defined. Indeed, good starting points have the benefit of preparing and encouraging the respondents to
state their maximum WTP (Frew, 2010-a; Fischhoff and Furby, 1988; Seenprachawong, 2004; Carmona-
Tores and Calatrava-Requena, 2006). Nevertheless, we find that, the real problem when using the DBDC, is
that, as each respondent identifies two amounts that bound their maximum WTP (i.e. one amount greater
than and one less than their maximum WTP (Bateman et al, 2002), the respondents who would have been
willing to pay an amount lower than the lower bid are registered as “no/no” respondents. In the same way,
bidding the WTP of the respondents who would have been willing to pay an amount greater than the higher
bid plays down the importance their true WTP. Therefore, using the DBDC format provides only with partial
information about the WTP of all the respondents, indicating only whether their WTP is below or above a
specified bid, or whether it is within a specified interval. Finally, we propose to combine both closed-ended
DBCD format and OE format. Once the starting points are well defined, the use of the DBDC in a first stage
is expected to yield the best bases for an OE question. Indeed, using the DBDC may reduce unrealistically
large stated WTP. It can also help reducing the non-response rates as the closed-ended questions are
normally considered easier. Hence, combining both closed-ended and OE formats helps to offset the limits
associated with identification of distribution of yes-yes or no-no answers, as well as to reducing the limits
associated with the cognitive limits of OE.
3.2. Survey design The CV questions were imbedded in a questionnaire addressing the characteristics of the households, their
behavior and motivations. The CV first described the socio-economic and ecological attributes of elephant,
as well as the potential change in sociocultural services associated with a possible extinction of elephant. The
following components of the elephant economic value were presented: existence, viewing, disponibility for
future generation, cultural and spiritual enriches, baillonella toxysperma seeds dissemination. Even if we
propose a detailed presentation of the various components of the Non Use Value of elephant to the
households, they were asked to provide only with a unique willingness to pay given the various components.
This is because these components are often overlapping. For instance, the existence value willinsure the
bequest value, as well as the option value (Loomis and Larson, 1994; Bondara and Tisdell, 2003). Next, we
proposed the following hypothetical scenario. “Considering the trend to extinction of Loxodonta cyclotis, if
action is not taken quickly, this multiple use iconic species will disappear in the next few years. To stop this
tendency to extinction and make the species abundant, the Tridom Regional Project Management Unit can
develop a 10 years elephant protections program that aims to seize weapons currently used by poachers, to
effectively fight against cross-border poaching by (1) creating joined checkpoints at the landscape scale, (2)
recruiting more young people in the villages, involving them in a communication network to improve the
anti-poaching control strategy and prevent human-elephant conflicts”. Then each respondent was asked
whether he was willing to contribute to the program by paying some initial CFA monthly bid amount if the
Regional Management Unit demands financial support of all the inhabitants of the village? The payment
vehicle presented was the direct cash payment in secure funds and lodged at the Tridom program. This mean
of payment is the most familiar, credible and feasible according to the economic situation in the landscape It
generated a realistic reaction of the respondents as far as any financial engagement is concerned. To
minimizing the “Yea saying”, the “Nay saying” bias2 and the starting point bias, (1) each respondent was
asked to consider his monthly income, his sources of income, and the usual monthly expenditure; (2) each
respondent was asked to be realistic, making assured that he could actually pay that monthly amount for the
next 10 years before he answered; (3) the respondents were randomly assigned to one of six starting points
developed and validated during two pretests steps with 40 households in four villages (Meyomessi, Oveng,
Kongo and Mbieleme) of two subdivisions of the study area. A lower bid was presented to those who gave a
negative answer, and a higher bid to those who gave a positive answer. The bid cards structure is presented
in Appendix 1. The WTP of each household belongs to one of the four interval formed by his income, the higher bid, the
initial bid, the lower bid and 0 (Haab and McConnel, 2002). To the respondents who gave yes/yes and no/no
answers to bids, an OE question was propose, asking the maximum amount they would be willing to pay for
forest elephant conservation. We finally introduce follow-up questions that examine reasons for zero WTP to
be able to remove protest refusal bias from the data base before estimation (Arrow et al, 1993).
3.3. Study Area Sampling and Data
The Tridom is a cross-border conservation landscape covering a geographical area of 191 541 km2,
representing 7,5% of the total area of the Congo Basin Tropical Forests in central Africa. It was created in
2005 by an agreement between Cameroon, Gabon and Congo governments, as one of The 12 Congo Basin
Forest Partnership (CBFP) priority landscapes, targeting to promote long-term biodiversity and protected
area conservation, a rational use of natural resources and a sustainable development, as well as to contribute
to the poverty reduction. It encompasses of 9 protected area representing 37 498 km2, including four into the
Cameroon segment (BoumbaBek, Kom-Mengame, Nki and Dja Biosphere reserve), three in the Gabonese
segment (Minkébé, Mwagne and Ivindo National Parks) and two in the Congolese segment (Odzala and
Lossi National Parks). Between the protected area, there is a livable inter zone representing 111000 km2.
It includes a rich and diversified flora and fauna. It abounds commercial high value timber species and
houses the largest population of forest elephants in the world with a high concentration in the Minkébé
National Park, as well as other large mammals such as buffalo, bongos, giant pangolin and gorillas. The map
(Appendix 2) shows some of the glades visited by elephant populations in the landscape. Currently, there is a
high inflow movement of human population within the landscape for resource exploitation. With an average
density comprise between 1-4 inhabitants / km², the Tridom buffer zone is subject of numerous economic
stakes, including forest management, rural agriculture, hunting, traditional and industrial mining as well as
fishing and gathering non timber forest products. The field work was held in the Cameroon and Gabon
segment. Both segments are inhabited by more than 43 tribes, dominated by Bantu, while there is a
minority group of indigenous Baka (see Appendix 3.)
The paper uses data from a representative face-to-face survey by questionnaires with a random and stratified
sample of 10353 from approximately 65141 households, conducted between December 2013 and July 2014,
in 108 villages chosen in all the 26 administrative units of the Cameroonian and the Gabonese part of
landscape (Appendix 4). The villages visited are spread over nearly 27,000 km2, which is 2/3 of the
landscape livable inter-zone.
2 The Yea-sayers and Nay-sayers are the respondents who trie to please or to counter the interviewer without
considering the specific amount they are asked about (Carson and Hanemann, 2005 ; Frew, 2010-b). 3 The sample size required at a confidence level of 95% (typical value of 1.96) is 384.
The interviews lasted between 1 to 3 h. In addition, there were evening visits in the various households
surveyed to quantifying and measuring daily production. The survey was supervised by the first author. Ten
Masters Students selected after 5 training seminars participated as surveyors. Every village provided us with
at least two local translators for the situation the head of household doesn’t communication in French. Every
household was geo-localized with a GPS
3.4. Theoretical model specification Household preferences for forest elephant conservation in the Tridom can be described by a random utility
model developed by McFadden's (1973) and formalized by (Manski, 1977) and Hanemann et al (1991).
Therefore, the utility function is given by :
𝑈 𝑖𝑗 = 𝑉𝑖𝑗(Y𝑖 − 𝑎, 𝑋𝑖 ) + 𝜀𝑖𝑗
Where 𝑗 = 1 𝑎𝑛𝑑 𝑎 > 0 if the household 𝑖 accept to pay 𝑎 amount or , 𝑗 = 0 𝑎𝑛𝑑 𝑎 = 0 otherwise.
3.5. Econometric model specification Following Lopez-Feldman (2013) and Barrena et al (2014), the WTP can be modelled as the following
function:
𝑊𝑇𝑃𝑖(𝑋𝑖, 𝜇𝑖) = 𝑋′𝑖β + 𝜇𝑖
Where 𝑊𝑇𝑃𝑖 represents the willingness to pay vector of the ith respondent, 𝑋𝑖 is a vector of explanatory
variables, β is a parameter vector and 𝜇𝑖 a normally distributed error terms.
Taking into account the valuation technique developed above, we specify five econometric models, using the
two elicitation procedures (the DBDC and open ended) to.
The first two models to be estimated are the DBDC with and without protesters. Moving from the first
to the second using the motives behind the respondent answers helps improving the results with respect
to the protest refusal bias (Arrow et al, 1993).
Afterwards, we estimate a Tobit model based on the OE elicitation procedure. Even if DBDC is
supposed to yield systematically higher stated preferences than the CE (Brown et al. 1996; Boyle et al.
1996, Kealy and Turner 1993; Halvorsen and Soelensminde, 1998), both technics reduces the average
WTP. Indeed, the DBDC left-censors all the no/no respondents at the lower bid, while the Tobit model
left-censors from 0, considering the entire “zero” respondent as those who faced negative preferences.
The design used allows us combining both closed-ended and OE format as the third elicitation procedure
applying the interval censored regression models (ICRM). Indeed, we hypothesize that, combining the
DBDC with a good starting point to the OE yields result closer to the true WTP. The ICRM takes into
account the point data and offers the greatest increase in efficiency with the least ambiguity compared to the
DBDC (Haab and Mcconnel, 2000). In addition, it helps account for negative, zero and positive preferences.
The ICRM (1) left-censors from zero (as upper limit), only the households who expressed a loss in utility
caused by elephants and who effectively faced a human/elephant conflict. These individuals were
identified by using the motives behind the answers captured with the follow-up questions given
theirprevious experiences with elephants.
The ICRM (2) lef-censors the households who expressed a possible lost in utility with the presence of
elephants.
3.5.1. DBDC model and DBDC model without protesters Under a closed ended format, an individual 𝑖 will either be willing to pay a given bid amount 𝑎 for elephant
conservation if 𝑊𝑇𝑃 𝑖 ≥ 𝑎 or not be willing to pay if 𝑊𝑇𝑃 𝑖 < 𝑎 with the following probabilities.
MEAN WTP 338,532 722,349 543,033 1 246,332 1 139,402
Legend : * p<0,1 ; ** p<0,05 ; *** p<0,01
The log likelihood can help comparing the models, but, according to the variability of the types of
data, we cannot directly compare the log likelihoods for the DBDC, the ICRM Model. Indeed, the
DBDC use only interval data, while the ICRM uses a mixture of discrete and interval data. We can
see that the ICRM that left-censors expected negative preference fits better than the first (lower
loglikelihood).
All the models present almost similar significant variables, we consider the ICRM that left-Censors
expected negative Preference. This model shows that, improving household’s education level
increases their preferences for elephant conservation by CFA247. A CFA1000 increase in individual
income would increase the individual WTP by CFA6. The WTP for forest elephant conservation
decreases with the head of household age. Among the heads of household, females are more likely to
state a greater preference. Indeed, an additional female will be willing to pay CFA17 more than an
additional male. The coefficient associated to this variable “autochbaka” is positive and significant.
Indeed, an indigenous baka would be likely to pay CFA559 more than the non-indigenous group. The
variable “proteria” has a negative and significant influence on the household’s preferences for
elephant conservation. Indeed, the households who are closed to the protected areas with less elephant
density are willing to pay more than those closed to higher density protected areas. Among the various
land use and activities, one more traditional gold-miner as well as one more forest management or
conservation related worker are likely to increase the WTP by CFA1223 and CFA 725 respectively.
The existence of conflict as well as the distance between the household and the nearest protected area
(dist_narea) don’t significantly drive the household preferences for elephant conservation.
4.3. Aggregate and Net Benefits To estimate the expected preferences for the population living in both Cameroonian and Gabonese
segment of the Tridom for elephant conservation, the simple transferring point estimate produces
robust aggregate with fewer bias (Bandara and Tisdell, 2004; Loomis et al 2000; Rodriguez et al,
2011). The aggregate population size in both segments of the Tridom is 418855 inhabitants
(Cameroonian Central Bureau of Census and Population Studies, 2010; Population Census Report –
Gabon, 2010). Considering the mean household size of the sample (6,43), the number of households is
around 65141. The protest refusal observation weighed 9,57% in the initial sample, the paper finally
considers 58910 households willing to pay CFA1139,4 (€1,74) per month for forest elephants
conservation. the monthly willingness of the overall population is CFA62,8 million (€95778), that is
annually equal to CFA753,9 million (€1,15 million). Considering the current 2.4% population growth
rate (World Bank) and a 3% discount rate, the Tridom local and indigenous households will to pay
CFA7, 4 billion (€11,3 million) for the 10 years elephant protections’ program to avoid losing the
services provided by elephant in cas of extinction. According to the high density population and the
increasing urban sprawl, the Haut-Nyong, the “Dja et Lobo” and the “Boumba et Ngoko” subdivision’s households in Cameroon expressed a greater aggregrae WTP for elephant conservation. among the gabonese subdivions, the “Ivindo”, the “Woleu” and the “Haut-Ntem” divisions expressed greater WTPs (see the figure below).
Figure 2 : Agregate WTP of Tridom local population for elephant conservation
Considering the CFA 28140 (€43) annual mean damage incurred by the 27,7% of the sampled
households, the net preferences of the overall population in both the Tridom segment is CFA3,383
billion (€5,165 millions). Indeed, the households stated value of the services provided by forest
elephants exceeds the aggregate economic cost incurred by the farmers in terms of crops and property
damage caused by elephants. This result states that forest elephant conservation is socially beneficial.
It may also state the tolerance by the farmers of the presence of elephants in their agricultural fields.
5. Discussion
The above estimates and aggregate provide four major outcomes with respect to the hypothesis stated
above.
1. The extinction of Loxodonta cyclotis can lead to a significant net lost in the household’s wellbeing,
indeed, the per month MWTP stated by local population is CFA1139,4 (€1,74). This value is close
to the results found by Bandara and Tisdell, (2005). Indeed, they found that the respondents in
general were willing to pay Rs. 110,17 (€1,65) per month for elephant conservation.
2. The estimates shows that the existence of human/elephant conflict, as well as the distance of the
location of the households to the nearest protected area do not influence the preferences of the
households for elephants conservation. This result rejects the hypothesis that the WTP for
elephant’s preservation increases with the distance of the location of the households to the nearest
protected areas and decreases with the human-elephant conflicts presence. In fact, as the scenario
presented to the respondents indicated that the program for protection also includes measures to
reducing human elephant conflicts. This result shows that local communities are receptive to such
a program. Furthermore, as the households surrounding the low density protected area are willing
to pay more than the one surrounding high density protected areas, it is clear that, no matter far or
closer is the household to the nearest protected area, the scarcity of elephant increases the
preferences of households for conservation.
3. The indigenousness has a positive and significant influence on the household’s preferences. This
information is all the more important to increasing incentives for elephant conservation as the
extinction of forest elephant would lead to a severe threat on spiritual enrichment, cultural identity
as well as the way of life of this ethnic minority group.
4. As methodological and technical outcome, this paper proposes combining both DBDC and Closed-
Ended elicitation formats to investigate de true WTP when doing a CV. We propose the following
steps: (1) after defining good starting points during pre-test (see Frew, 2010-a), (2) one should first
assign the DBDC, varying randomly the starting points among respondents. (3) as a good starting
points prepares and encourages the respondents to state their true (Seenprachawong, 2004;
Carmona-Tores and Calatrava-Requena, 2006, Fischhoff and Furby, 1988) one should then assign
the OE format to improve the level of information about the “no/no” respondents who would be
1 959,20
158,64
2 038,52
975,63
325,21
658,35
245,89
150,71
452,12
277,62
182,44
7 424,33
892,75
72,29
928,90
444,57
148,19
299,99
112,05
68,67
206,02
126,50
83,13
3 383,07
Dja et Lobo
la Mvila
Haut-Nyong
la Boumba et…
la Zadié
l’Ivindo
la Lopé
la Mvoung
le Woleu
le Haut-Ntem
l’Okano
agregate
Per subdivision agregare WTP and Benefit (million of CFA)
Net Benefit (millions of CFA) agregate ICRM (millions of CFA)
willing to pay a positive amount lower than the lower bid, as well as those willing-to-pay above
the upper bid. (4) Afterwards, one should ask the motives behind the answers. The motive should
be defined a way to help removing protesters, as well as distinguishing between negative, zero and
positive preferences (Arrow et al, 1993). Finally (5) one should run an Interval Regression Model
as it takes into account the point data and offers the greatest increase in efficiency with the least
ambiguity compared to the DBDC (Haab and Mcconnel, 2000).
Acknowledgements
This research is part of the CIFOR-GCS [Center for International Forestry Research’s (CIFOR) global
comparative study on REDD+ (GCS)} project with funds provided by the Norwegian Agency for
Development Cooperation (NORAD)
References
Alberini, A., (1995), "Efficiency vs Bias of Willingness-to-Pay Estimates: Bivariate and Interval-
Data Models," Journal of Environmental Economics and Management, 29, 169-180; Url: