LAND USE AND COMMUNITY
PARTICIPATION
UTILISATION DES TERRES
ET PARTICIPATION COMMUNAUTAIRE
Evaluating farmers’ willingness to adopt integrated
packages for trypanosomosis control in The
Application du mécanisme
de demande revelée pour évaluer la volonté des éleveurs
d´adopter des méthodes intégrées de lutte contre la trypanosomose en Gambie
Jacques
Somda13 , Mulumba Kamuanga2 And Eric Tollens3
1 International Trypanotolerance Centre. PMB 14,
Résumé
Les résultats des recherches conduites par les
économistes pour évaluer la participation volontaire (main d’œuvre et/ou
argent) des agriculteurs pour la fourniture de biens publics a mis en évidence
l’applicabilité de l’analyse contingente dans le cas des méthodes de lutte
contre les maladies animales. Le problème avec cette technique est que les
agriculteurs peuvent ne pas être suffisamment motivés pour révéler leur vraie
capacité de contribution, parce que le bien considéré est soit semi-public soit
public, mais ayant un prix de marché. Utilisant des données collectées auprès
des éleveurs dans des zones à risque
moyen à élevé de pression glossinaire, cette communication applique le
mécanisme de demande révélée afin d’améliorer les résultats de l’analyse
contingente. Elle détermine ensuite les facteurs influant sur la volonté des
agriculteurs de contribuer, de même que le montant de la contribution. Les
données ont été analysées à l’aide de modèles « Probit » et
d’équations simultanées. Les résultats montrent que la combinaison de l’analyse
contingente avec le mécanisme de demande révélée améliore les résultats obtenus
avec la première technique. Les résultats montrent également qu’en plus des
variables socio-économiques souvent utilisées dans pareilles études, il faut
aussi prendre en compte la perception
qu’ont les agriculteurs sur l’efficacité des traitements connus.
Mots clés : Mécanisme de
demande révélée, évaluation contingente, bétail, contrôle de la trypanosomose,
Afrique de l’Ouest.
Summary
Economists
investigating farmers’ willingness to contribute (money and/or labour) to public
goods supply have accumulated evidence showing that the contingent valuation
(CV) technique can be applied to livestock diseases control measures. However,
the trouble with the CV technique is that it may not provide good incentives
for the individual to reveal his true willingness to contribute, because those
goods evaluated are semi-public or public goods with market prices. Using data
collected in The Gambia from livestock owners facing medium to high tsetse
challenge and working closely with the International Trypanotolerance Centre,
this paper applies demand-revealing mechanisms to improve the outcomes of the
CV technique, and tests the factors affecting farmers’ willingness to
contribute labour, money or both forms of payment as well as the amount of
pledged contributions. The analysis, conducted with Probit and simultaneous
equations models, shows that the demand revealing mechanisms can be applied
with contingent valuation techniques to improve the outcome of the latter. The
results also provide a strong case for future studies to expand the range of
variables used away from the broad socio-economic, and demographic factors to
include farmers’ perceptions of the characteristics of the disease control
measures.
Key words:
demand revealing mechanisms, contingent valuation, livestock, trypanosomosis
control,
1. Introduction
Vectors and vector
borne diseases, such as tsetse-transmitted trypanosomosis and tick-born
diseases, constitute a major constraint to increased livestock productivity in
many African countries. These diseases still inflict economic losses to both
governments and farmers. Currently, most farmers control animal trypanosomosis
by avoiding moving herd cattle and other livestock into infested areas and by
using drug therapy. In the absence of a vaccine, only some form of vector
control could result in effective disease control (Brightwell et al., 2001).
In most countries
in
Contingent
valuation surveys have been used to measure the value households place on
tsetse control. Even though farmers were willing to contribute to vector
control measures, the hypothesis that willingness could be influenced by the
attributes of the techniques was never tested. Furthermore, the provision of
tsetse control techniques (semi-public or public goods) would not exclude the
use of trypanocides that are essentially private goods, which farmers are more
used to.
Therefore, when
evaluating farmers’ willingness to contribute to vector control, it is not
known whether farmers perceive this contribution as additional cost (labour or
money) to parasite control or not. Consequently, the contingent valuation
surveys might not provide good incentives for the individuals to reveal their true
willingness to pay, and the effective contribution will not be achieved when
needed. This paper examines a complement to the contingent valuation techniques
usin4g the demand revealing mechanism. The objectives were to assess farmers’
willingness to contribute to trypanosomosis control measures and its
determinants.
2. Research
framework and survey design
2.1. Theoretical concept: the demand revealing
mechanism
The conceptual
framework of this study is based on demand revealing mechanism (Varian, 1992)
that investigates approaches that result in pledging the “right” amount of the
private contribution to the provision of public good. The so-called
Groves-Clarke mechanism (Clarke, 1971; Groves, 1973) for multipart pricing of
public good, induces each agent to truthfully reveal his true value (labour,
money or both) that he is willing to contribute. It works as follows:
(1) Each agent
states his contribution “bid” for the parasite (trypanosome) and vector
(tsetse) control technique, bi. This may or may not be the
true value.
(2) The public
good is provided if 
, and it is not provided if 
(3) Each agent
receives a side payment equal to the sum of the other bids, 
, if the public good is provided.
It is expected
that the reported value would differ from the maximum willingness-to-pay as it
is assumed when conducting traditional contingent valuation surveys. It
represents the net value that each agent attaches to the good being evaluated.
Let us consider a discrete public good, G, that is either 0 if the good is not
provided, or 1 if provided. Let ri be agent i’s
reservation price and si be agent i’s cost share of
the public good. Since the public good costs c to provide, sic
is the total amount of money the agent i must pay if the good is
provided. Let vi = ri-sic
be agent i’s net value for the public good or the private one. Assume
that there are n agents, each with a true value of vi and a
bid value of bi. Applying Groves-Clarke mechanism, we can
show that it is optimal for each agent to report bi = vi
regardless of what the other agents report. That is “truth telling”
is a dominant strategy. Formally, the agent i’s payoff statement is as
follows:
if 
or
if 
Now, suppose that 
, then agent i can ensure that the good is provided by
reporting bi = vi. Suppose, on the other hand,
that 
, then agent i can ensure that the good is not
provided by reporting bi = vi. Either way, it is
optimal for the agent to tell the truth. There is never an incentive to
misrepresent preferences, regardless of what the other agents do. Since the
information gathering mechanism has been modified, each agent faces the social
decision rather than the individual decision problem, and thus each agent has
an incentive to reveal his own preferences correctly.
2.2. Survey design and implementation
The study was
carried out in twenty villages representing two different levels of tsetse
challenge as described in Agyemang et al. (1997). Four villages are situated in
the Kombo South and East districts, which are respectively classified as
regions of low and low to medium tsetse challenge. The remaining villages are
located in Lower River Division, known as medium tsetse challenge region.
The scheme
described in section 2.1 was applied to gather information from 72 farm
households in The Gambia, from October-November 2001 to January-February 2002. A
pre-coded questionnaire was designed to gather information from the head of
households on the households’ socio-economic characteristics, such as age, sex
and educational status (including participation to extension workshops),
resources endowment (human and livestock), the respondent’s perception on
livestock diseases and on the efficacy of trypanocides, and the current
expenditures for veterinary products.
As respondents’
knowledge and information are important inputs of contingent valuation surveys (Mitchell
and Carson, 1989) as well as of the demand revealing mechanism approach, the
following information have been conveyed to the respondents prior to the
survey:
The attributes of
both parasite and vector control measures, as semi-public or public goods
versus private;
Their respective
effects on disease control, particularly for trypanosomose and tsetse fly
control;
And a hypothetical
cost of the tsetse control measure (100 Dalasi/cattle all costs included for
poor-on or spray) and trypanosome control (6 Dalasi/cattle for trypanocide) was
given to the interviewees.
In a latter stage,
farmers were then asked about the kind of contribution (money, labour, money
and labour) they are willing to provide for each disease control technique.
Respondents who volunteered cash were asked, “What would be the net value of
their money contribution per cattle?”. Those volunteering labour were asked
“what would be the number of hours per day?”, and then “how many household’s
members would be involved in such a contribution?”. Finally, the respondents
were asked to indicate in which type of organisation frame they were willing to
contribute in disease control measures.
3. Results
3.1 Socio-economic characteristics of the sample
farmers
The ethnic group
composition was as follows: Mandinka (64%), Fula (21%), Jola (13%) and Wolof
(1%). Other ethnic groups include Serere, Manjako accounting for 1% of the
sample. Most of the respondents were illiterate (99%), but 64% of them received
non-formal education and 75% were participating in extension workshops for over
five years.
The households’
human and livestock resources were also evaluated. On the average, the
respondent’s households were relatively well endowed in human resources, with a
ratio of active —i.e. those providing labour in agricultural activities— to
inactive member of 1.25 and household size averaging 18 persons. The
respondents own cattle, goats, sheep and equines. On average, cattle were the
most important livestock resource (77 heads/household).
3.2 Assessing
farmers’ perception on livestock diseases importance
When asked to rank
the three major livestock diseases encountered in their herds from 2001 to
early 2002, farmers’ responses varied. On average, trypanosomosis is still
perceived as the most important disease reported by 24% of the respondents.
Twelve percent of the respondents reported black quarter, 11% for spots. Foot
and mouth diseases were quoted by 7% of the respondents. Only 3% of the
respondents reported tick and tick-borne diseases in their herds.
To appraise the
economic importance of diseases in farmers’ herds, the number of cattle
affected by each disease was recorded. Surprisingly, trypanosomosis perceived
as the most important did not affect an important number of cattle. Tick and tick-borne
diseases recorded the highest number of cattle affected and treated (41 to 65
cattle/household), even though it occurred in few respondents’ herd (3%). Foot
and mouth diseases appeared having important impact on livestock during the
period under evaluation (on average 4 to 19 cattle affected). Then followed
spot (5 to 9 cattle) and trypanosomosis (2 to 9 cattle).
The amount of
money spent for cattle healthcare was also evaluated. Since it was practically
difficult for farmers to recall expenditures for each disease, the overall
expenditure was considered. The results indicate that 242.91 Dalasi/household
($US 15.38) were spent on cattle healthcare and 18 cattle treated during the
period of 12 months covered. Average expenses on veterinary drugs and care per
sick cattle were estimated to 16.35 Dalasi/cattle ($US 1.04). The ratio of
health expenses on the total number of cattle averaged 3.31 Dalasi/cattle ($US
0.21).
3.3 Willingness to contribute to parasite versus
vector control measures
The results on the
willingness to contribute to trypanosomosis control indicate that 81.94% of
respondents volunteered to contribute with any kind of contribution for
trypanosomes control (use of trypanocides), out of which 62.71% volunteered
labour and 37.29% both money and labour. For the tsetse control measures,
90.28% indicated their willingness to contribute to the establishment of these
control measures. Out of this, 75.38% of those volunteered to contribute were
willing to allocate only labour, while 24.62% would provide both labour and
money. No respondent volunteered to contribute only money to both diseases
control measures.
The amount of
labour contribution in terms of households’ active members and time available
for contribution differ across the diseases control measures. Overall, farmers
were willing to provide more persons to work for vector control than for
parasite control. On average, 3 persons per household providing 2.42 hours each
would be available for vector control, while 2 persons per household with 2.13
hours would be allocated for parasite control techniques.
The amount of
labour contribution also differs whether the farmer is willing to contribute
only labour or both labour and money, reflecting the household’s resources
endowment. Farmers volunteering to contribute only labour to parasite control
measures were willing to allocate 3 persons with 2.18 hours each, while those
volunteering for both money and labour would be providing 3 persons with 3.32
hours each. For the vector control, 3 persons and 2.45 hours would be provided
by those volunteering to contribute only labour. Volunteers of both labour and
money were willing to provide 3 persons with 3.60 hours each.
The amount of cash
farmers volunteered to contribute for both money and labour for a parasite
control technique was only 0.50 to 2 Dalasi, with an average of 1.00 (s.e =
0.09). The contribution for vector control in this group averaged 21.05 (s.e =
6.52) and ranged of 0.20 to 72.00 Dalasi.
Farmers were also
asked whether they would like to contribute individually or within an
organisation. There was a particular interest for farmers to contribute in any
kind within a community frame, rather than as individuals. Therefore, only 9.7%
and 13.9% of the respondents declared they were willing to contribute
individually for parasite and vector control measures, respectively. The
remaining either chose community organisation (33,3% and 31.9% for parasite and
vector control, respectively), or the combination of both approaches—individual
and community (34.7% in the case of parasite control, and 42.3% for vector
control).
There were strong
positive relationships between willingness to contribute to parasite and vector
control measures (Pearson correlation=0.67). That is, farmers volunteering to
contribute to parasite control measures would also likely be willing to
contribute to vector control measures. There was also a positive (R = 0.51)
relationship between the amount of labour pledged for parasite control and that
for vector.
3.4 Factors affecting farmers’ willingness to
contribute to trypanosomosis control
The
household-level factors that were hypothesised to affect decision making to
contribute to disease control measures were tested with Probit model using
SHAZAM (White, 1993). A discrete variable for the willingness to contribute (1
if yes and 0 otherwise) was used as dependent variable. The explanatory
variables were:
Type of disease
control techniques: 1=vector control technique; 0=parasite control technique.
It is hypothesised that farmers attach more value to vector than to parasite
control technique, because of its long-term effect. Therefore the likelihood to
contribute to vector control would be higher than for parasite.
Age of the
household head: younger household heads would be willing to invest more in
vector or parasite control than older household heads.
Contacts with
extension (number of years of working
with livestock extension service). It is hypothesised that participants in
extension workshops over a long time would be more aware of disease issues and
would be more cooperative in enacting disease control measures.
The number of
households’ active members: the size of labour force (in terms of numbers of
active members in the household) would increase the probability to contribute
to disease control measures.
The total number
of cattle: farmers owning large cattle herd size would be more willing to
provide any contribution.
Veterinary
expenditures: household with currently large expenditures on healthcare would
be willing to share the cost by deciding to provide any contribution to
diseases control.
Efficiency of
chemotherapy: farmers’ perception about the efficiency of the drug use for
disease control, as a binary variable that takes on 1 if farmers valued
positively the efficiency of drug treatment and 0 otherwise. It is hypothesised
that if farmer perceives the chemotherapy as efficient disease control
technique, he would be less willing to contribute to collective disease
control.
The results of the
Probit model are reported in Table 1. The age of the household head decreases
the probability of contribution (P<0.01). There is also evidence of a
positive relationship between willingness to contribute and the previous
veterinary expenditures at 10% significance level.
Table 1: Probit model of willingness to contribute to
diseases control; n=144
|
Explanatory
variables |
Estimated
coefficients |
Standard error |
|
Type of control
technique |
-0.57E-01 |
0.83E-01 |
|
Age of the
household heads |
-0.42E-01*** |
0.13E-01 |
|
Participation in
workshops |
-4.96 |
1303.2 |
|
Household active
members |
0.53E-01 |
0.43E-01 |
|
Number of Cattle |
0.14E-04 |
0.54E-02 |
|
Previous
veterinary expenditures |
0.30E-02* |
0.16E-02 |
|
Constant |
17.357 |
3909.6 |
*** and *
significant at 1% and 10% level, respectively. The dependent variable is a
discrete indicator of willingness to contribute (1=yes; 0=no).
Log-Likelihood
function: -40.276; Likelihood ratio test: 35.49 with 6 D.F. Percentage of right
predictions: 83%
Source: Survey
data.
A simultaneous
equations model was estimated using the three-stage least squares procedure
(White, 1993) to evaluate factors affecting the
amount of labour and money contribution. The results are reported in Table 2.
Table 2: Simultaneous equations model for factors
affecting the amount of money and labour contribution (N=144): System R-Square:
0.55
|
|
Money
contribution (Dalasi/cattle) |
Labour
contribution (Hour/day) |
||
|
Explanatory
variables |
Estimated
coefficients |
Standard error |
Estimated
coefficients |
Standard Error |
|
Type of control
techniques |
-1.244*** |
0.419 |
0.217*** |
0.073 |
|
Participation in
workshops |
0.677 |
1.335 |
-0.508*** |
0.180 |
|
Household active
members |
- |
- |
0.079** |
0.034 |
|
Number of cattle |
-0.083** |
0.041 |
0.017*** |
0.006 |
|
Previous vet.
expenditures |
0.003 |
0.006 |
-0.001** |
0.0008 |
|
Amount of
Dalasi/animal |
- |
- |
0.094* |
0.037 |
|
No hours per day |
2.544** |
1.026 |
- |
- |
|
Efficacy of
chemotherapy |
3.258** |
1.493 |
- |
- |
|
Constant |
-0.555 |
5.232 |
1.210 |
0.795 |
*, **, and ***
significant at 10%, 5% and 1% level, respectively.
Source: Survey
data
In the money
equation (Table 2), variables such as the types of disease control and the
number of cattle have a significant and negative effect on the amount of
contribution. Farmers are willing to contribute less money per cattle in vector
control than in parasite control. Also, there is a positive correlation between
the number of cattle and the amount of money the farmer is willing to
contribute. Likewise, The amount of money that the farmer is willing to
contribute significantly increases with the number of working hours per day. In
addition, if the farmer positively evaluates the efficacy of the trypanocide,
he will pledge a high amount of money. In the labour equation (Table 2), both
participation in workshops and previous veterinary expenses showed a negative
effect on the amount of labour contribution. The type of control technique, the
number of active members in the household, the number of cattle and the amount
of money spent per cattle have shown significant and positive effect on the
number of hour per day that the farmer pledged.
Tests for equality
of the estimated coefficients of variables in both equations were run to
evaluate whether farmers have shown different behaviour regarding the amount of
money and labour contribution. The coefficients of the type of control
technique and the number of cattle in the money equation were significantly
different from that in labour equation at 1% and 5% level, respectively.
Farmers are willing to contribute more labour than money in vector control, and
more money than labour in parasite control. Likewise, they pledged more money
when they own small cattle herd and more labour for large herd.
4. Discussion
In general, using frequencies distribution trypanosomosis has been ranked as the first disease causing livestock production losses in The Gambia (Mugalla, 2000). S