Impact of nature-based tourism on earnings of local people: evidence from Keenjhar lake in Pakistan.
Mangan, Tehmina ; Lohano, Heman D.
Keenjhar lake is one of the most important freshwater lake of
Pakistan. The lake has a vital wetland area of great ecological
significance and has been recognised as a wetland of international
importance under Ramsar Convention. The lake provides numerous goods and
services including recreational services for tourists. The main
objective of this study is to investigate the contribution of tourism at
Keenjhar lake towards the earnings of local people living in the
adjoining areas of the lake. For this study, primary data were collected
from 264 households selected by stratified random sampling method. This
study applies endogenous dummy variable model to evaluate the impact of
households' participation in tourism related work on their
earnings. Results show that the households who participate in the
tourism related work enjoy 7.6 percent higher level of earnings relative
to the conditions they would have been in had there been no tourism
activities. Furthermore, the estimates of marginal effect of productive
assets indicate that any additional Rs 100 investment in productive
assets would result in higher earnings by Rs 20. Education level of
earning members also increases the earnings of the household. This study
finds that tourism at Keenjhar lake improves the standard of living of
local people by raising their earnings, and that the nature-based
tourism can be an effective poverty alleviation strategy.
JEL Classification: Q56, R20
Keywords: Nature-based Tourism, Earnings, Endogenous Dummy
Variable, Impact Evaluation, Wetland, Keenjhar Lake, Pakistan
1. INTRODUCTION
Nature-based tourism is the fastest growing part of tourism [Kuenzi
and McNeely (2008)]. Wetland areas including lakes are an important
source of nature-based tourism as tourists like their scenic views and
enjoy doing different activities including swimming, canoeing, diving
and learning about nature [van der Duim and Henkens (2007)]. Wetlands
are amongst the most important ecosystems on Earth and provide numerous
goods and services including recreational services [Mitsch and Gosselink
(2007)]. Increasing demand for nature-based tourism has raised the
importance of wetlands.
In the developing countries, millions of people depend on wetlands
for their livelihoods. However, due to population growth and lack of
alternative livelihood resources, wetlands have been threatened due to
over-exploitation of their resources, which, in turn, would affect the
livelihood of poor people and lead to increased poverty. In order to
break this vicious circle, tourism has increasingly been considered as a
possible solution [van der Duim and Henkens (2007)]. Pro-poor tourism
can be best strategy for both poverty alleviation and wetland
conservation [Ashley, et al. (2001)]. According to United Nations World
Tourism Organisation [UNWTO (2011)] there are many ways by which the
poor can get economic benefits from tourism such as by getting
employment, supplying of goods and services to tourism enterprises,
direct sales of goods and services to tourists, revenue generation,
voluntary support and investment in infrastructure. Poor households have
surplus labour that is well suited to tourism activities. Measures can
be taken to increase the level of employment of poor people within all
kinds of tourism related activities and enterprises including hotels,
resorts, transport companies and tourism services.
Keenjhar lake is one of the largest natural freshwater lake of
Pakistan. Keenjhar Lake, also known as Kalri Lake, is located in Thatta
district. It is 24 km long and 6 km wide and has an area of 14,000
haectares [WWF-Pakistan (2007)]. The lake has a vital wetland area of
great ecological significance and provides habitat for internationally
important water birds. Due to its ecological functions and economic,
cultural, scientific and recreational value, the lake has been declared
as one of the Ramsar sites recognised as the wetlands of international
importance under Ramsar Convention in 1971. The lake has also been
declared as wildlife sanctuary under Sindh Wildlife Protection
Ordinance, 1972. The lake is located 122 km from Karachi city and 19 km
from Thatta city in Sindh province. The lake has great scenic beauty and
attracts national and international tourists. In the year 2010, the
official annual visitor count at Keenjhar lake was 385,000. Tourists pay
an entrance fee varying between 2 Pakistan Rupees (Rs) for students and
children under five years and Rs 5 for every adult (and additionally Rs
5 for a scooter and Rs 20 for a bus). The revenues from entrance fees
are US$ 38,000 [STDC (2010)]. With proper sustainable management of the
recreational facilities at Keenjhar Lake, the number of tourists could
be increased and the tourism could become an even more important source
of revenues for lake conservation and improvement of the livelihoods of
the poor living around the lake.
According to WWF-Pakistan (2007), about 50,000 people from
surrounding villages depend on the lake for their livelihood, especially
on fishing and tourism. Most of the local people who depend on this lake
for their livelihoods are landless and earn marginal incomes for their
families. Keenjhar lake and its aquatic ecosystem are seriously
threatened by over-exploitation and poor management of the lake. Due to
illegal fishing, improper fishing methods, and poor management, the fish
stock in Keenjhar Lake is depleting and fishing cannot sustain
livelihood of poor people due to reduced catch rates [WWF-Pakistan
(2007)]. Thus, these poor people need alternative earning opportunities.
Keenjhar lake has a great potential for nature-based tourism,
largely because of its location near Karachi, the most populated city of
Pakistan with population over 13 million and among top ten mega-cities
of the world [Pakistan (2010)]. Tourism can potentially be an effective
strategy that can provide income generating opportunities for local poor
people and generate revenue for wetland management and conservation.
Thus, for effective sustainable planning and policy-making, there is
need to evaluate the contribution of tourism on livelihood of local
people. Knowing the economic value of this contribution provides an
important indicator of the social desirability of maintaining and
further improving the site [Carrier and Macleod (2005)]. Previous
studies on contribution of Keenjhar lake have focused on the valuation
of various goods and services, especially recreational services [e.g.,
Mangan, et al. (2013); Dehlavi and Adil (2011); Amjad and Kidwai
(2003)]. Although these studies have highlighted the importance of
tourism by providing recreational value of Keenjhar lake, there is a
lack of information on the contribution of tourism towards the
livelihood of local people who live in the adjoining areas of the lake
and participate in the tourism related work.
The main objective of this study is to investigate the contribution
of tourism at Keenjhar lake to local livelihoods. This study attempts to
answer the question, do the households who participate in the tourism
related work enjoy higher levels of welfare relative to the conditions
they would have been in had there been no tourism activities? This study
uses econometric model with endogenous dummy variable to investigate the
impact of tourism participation on the household earnings of local
people. In the econometric modeling, we account for self selectivity of
household's decision whether to participate or not in tourism
activities. To our knowledge, this is the first application of
endogenous dummy variable model to estimate the impact of nature-based
tourism on local livelihoods in Pakistan.
The remainder of this paper is organised as follows. The next
section presents a brief literature review. Section 3 specifies the
model of this study and estimation methods. Section 4 describes the data
used in the study. Section 5 presents the empirical results of the
study. Finally, Section 6 draws conclusions and offers their policy
implication.
2. LITERATURE REVIEW
Poverty has been one of the most complex social challenges facing
the world today. A review of literature indicates that poverty and
wetland degradation are interlinked [van der Duim and Henkens (2007);
Goodwin (2006); Jamieson, et al. (20040; Holland, et al. (2003); Ashley,
et al. (2001); Bennett, et al. (1999)]. There are many strategies that
can be followed for poverty reduction and to improve wetland management
and conservation. Tourism can potentially be one of the most important
strategies that can provide income generating opportunities for local
poor people and can generate revenue for wetland management and
conservation. This section provides a brief review of previous studies
on the contribution of nature-based tourism towards local livelihoods.
Bennett, et al. (1999) highlighted the importance of tourism as a
tool for ensuring minimum environmental damage (green tourism),
conservation of resources through community-based tourism, and enhancing
welfare and wellbeing of poor people.
Guha and Ghosh (2007) examined the contribution of tourism in
providing livelihood of the local people in Indian Sundarbans. In this
study, household expenditure was compared between tourism participants
and non-participants using regression analysis in order to control for
other factors. The results of their study showed that the households who
participate in tourism activities were found to spend 19 percent more on
food items per capita and 38 percent more on non food items per capita
as compared to non-participants households.
Leon (2007) evaluated the impact of tourism on rural livelihoods of
the Dominican Republic's coastal areas. This study conducted survey
of 23 coastal communities covering a range of tourism levels and types
and followed the Dominican Republic's Central Bank's
methodology to estimate household income. This study used household
income as a measure of the standard of living. Results of this study
also highlighted that tourism play a very important role in improving
the standard of living of people involved in tourism related income
generating activities.
Bandyopadhyay and Tembo (2010) in their study on "Household
consumption and natural resource management around National Parks in
Zambia" investigated the impact of community-based wildlife
management and participation in related community institutions on
household welfare. They used household and community level survey data
from Game Management Areas (GMAs) and other areas near national parks
(non-GMAs) and employed Maddala's treatment regression techniques.
Their study found significant welfare gains, measured as consumption per
capita, in some GMAs but these gains were unevenly distributed. The
welfare gains accrued mainly to the relatively well off, while the poor
did not gain. Bandyopadhyay, et al. (2004) evaluated the benefits of
community conservancies in Namibia based on a survey covering seven
conservancies and 1192 households. They divide their study in two parts
i.e. evaluation of conservancy impact and evaluation of economic impact
of participation in conservancies. They used multivariate analysis
method to evaluate the impact of household participation in
conservancies. They found a positive impact of conservancies on standard
of living of local poor people.
3. MODEL AND ESTIMATION METHODS
Keenjhar lake is an important source of livelihood for the poor
people living in the adjoining areas of the lake. About 50,000 people
from the surrounding villages depend on the lake for their livelihood
[WWF-Pakistan (2007)]. Majority of people depend on fishing for their
livelihood. Other professions of these local people include providing
tourism services, agriculture labour, farming, livestock rearing, stone
mining, shop-keeping, business, mat making, transport, teaching,
government service, tailoring and nursing.
Local people working in tourism at Keenjhar lake are involved in
different income generating activities and provide services such as
boating, huts for resting, vending services, swimming dresses, and tour
guidance. Local people also work on part-time basis on the restaurants
and furnished huts established by the Sindh Tourism Department.
Households working in tourism at Keenjhar lake do not entirely depend on
tourism-based earnings due to seasonal variation in tourism activities.
Annual visitor count at Keenjhar lake is 385,000. During the peak season
of summer from May to August, more than 15,000 tourists visit the lake
weekly. During the off-peak season of winter from October to January,
the number of tourists decline significantly and reach up to 50 tourists
per week during very cold days.
Households living in the adjoining areas of Keenjhar lake make a
choice whether or not to participate in the tourism related work. This
study examines the impact of tourism on the income of households who
participate in tourism related work.
This section specifies the model and estimation methods to measure
the impact of tourism on the income of households who participate in
tourism related work. We first specify a model where participation in
tourism related work is assumed to be an exogenous variable. Next we
relax this assumption because it is the household's choice whether
to participate in the tourism related work or involve in other income
generating activities. We then specify an endogenous dummy variable
model, where participation in tourism related work is assumed to be
endogenous variable.
An early work on self-selection of professions is discussed in Roy
(1951) who studied the problem of individual's choice between two
professions, hunting and fishing, based on their productivity (earnings)
in each. The issue of self-selectivity has also been addressed in the
studies on the behaviour of females' labour supply in Gronau (1974)
and Heckman (1974).
Endogenous dummy variable model used in the present study has been
used in a variety of application. This model has been used for
evaluating the impact of participating in natural resource management in
Game Management Areas in Zambia on the household welfare [Bandyopadhyay
and Tembo (2010)]. This model has also been applied for measurement of
treatment effects and programme effectiveness when there are
cross-sectional data. The model presented in this section is based on
the conceptual framework for evaluating treatment effects as given in
Greene (2012) and Stata (2011).
3.1. Model with Exogenous Dummy Variable
To evaluate the impact of tourism, the econometric model is
specified as:
[y.sub.i] = [x.sub.i][beta] + [delta][z.sub.i] + [[epsilon].sub.i]
(1)
where [y.sub.i] denotes annual income of household; [x.sub.i] is
the vector of explanatory variables including number of earning members
of household, value of household's productive asset, average years
of schooling of earning members, and average age of earning members of
the household; [beta] is the vector of unknown parameters; [delta] is
unknown parameter; [[epsilon].sub.i] is the error term representing the
unobserved other factors; and [z.sub.i] is a dummy variable indicating
whether or not the household participates in tourism related work:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
If [z.sub.i] is an exogenous dummy variable, then the expected
earnings of household who participates in tourism related work are given
by:
E[[y.sub.i]| [z.sub.i] = 1,[x.sub.i] = [x.sub.i] [beta] + [delta]
(3)
In this case, the impact of participating in tourism related work
on household earnings is:
E[y.sub.i]|[z.sub.i] = 1, [x.sub.i]] - E[[y.sub.i]|[Z.sub.i] = 0,
[x.sub.i]] = [delta] (4)
3.2. Endogenous Dummy Variable Model
In the above model, the dummy variable indicating whether or not
the household participates in tourism related work, [z.sub.i] is assumed
to be exogenous variable. However, is an endogenous dummy variable and
is selected by the household as the household makes a decision whether
to participate in tourism related work or involve in any other income
generating activities. In this case, household's earnings
([y.sub.i)] and decision to participate in tourism related work
([z.sub.i]) are jointly determined by two equations:
[y.sub.i] = [x.sub.i][beta] + [delta]][z.sub.i] + [[epsilon].sub.i]
(1)
Prob([z.sub.i] = 1 |[w.sub.i]) = [PHI] ([w.sub.i][gamma]) (5)
where Equation (5) represents a probit model; [PHI](x) is the
standard normal cumulative distribution function; and wt denotes the
vector of exogenous covariates that may affect household's decision
to participate in tourism related work. In this study, [w.sub.i]
includes a variable defined as distance from household's village to
the recreational site of Keenjhar lake. The probit model is represented
based on an underlying latent variable model. Let [z.sup.*.sub.i] be a
latent variable that determines whether or not the household participate
in tourism related work:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)
We do not directly observe [z.sup.*.sub.i] but instead we observe a
binary outcome [z.sub.i] that depends on [z.sup.*.sub.i], as given in
Equation (6). It is assumed that [z.sup.*.sub.i] is a linear function of
wt and a random error term [u.sub.i].
[z.sup.*.sub.i] =[w.sub.i] + [u.sub.i] (7)
The two error terms [[epsilon].sub.i]; and [u.sub.i] have bivariate
normal distribution with mean zero and the following covariance matrix:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)
where [rho] is the correlation between the two error terms
[[epsilon].sub.i] and [u.sub.i], and [sigma] is the standard deviation
of [[epsilon.sub.i]. The expected earnings of household participating in
tourism related work are given by:
E[[y.sub.i]|[z.sub.i] = 1, [x.sub.i], [w.sub.i]] = [x.sub.i][beta]
+ [delta] + [rho][sigma][[phi]([w.sub.i][gamma]/
[PHI]([w.sub.i][gamma])] (9)
where [phi](x) is the standard normal density function, and
[PHI](x) is the standard normal cumulative distribution function. The
expected earnings of household not participating in tourism related work
are given by:
E[[y.sub.i]|[z.sub.i] = 0, [x.sub.i], [w.sub.i]] = [x.sub.i][beta]
+ [rho][sigma][-[phi]([w.sub.i][gamma]/1 - [PHI]([w.sub.i][gamma])] (9)
(10)
In this case, the impact of participating in tourism related work
on household earnings is given by:
E[[y.sub.i]|[z.sub.i] = 1, [x.sub.i], [w.sub.i]] -
E[y.sub.i]|[z.sub.i] = 0, [x.sub.i], [w.sub.i] = [delta] + [rho][sigma]
[[phi]([w.sub.i][gamma])/[PHI]([w.sub.i][gamma])[1 - [PHI]
([w.sub.i][gamma])]] (11)
In this study, the above model is estimated by the maximum
likelihood estimation method using 'treatreg' command in Stata
11.2.
The last term in Equation (9),
[[phi]([w.sub.i][gamma])/[phi]([w.sub.i][gamma]] is referred to as
selectivity correction variable. Comparing Equations (3) and (9) shows
that the selectivity correction term is an omitted variable in Equation
(3) where the self selectivity of [z.sub.i] is not accounted for. If the
correlation between the error terms is zero, [rho] = 0, then the
Equations (4) and (11) will yield the same results for estimating the
impact of working in tourism sector on household earnings. However, if
[rho] [not equal to 0] and the selectivity correction term is omitted,
then the least squares estimates through Equation (3) would be biased
and the impact of working in tourism sector on household earnings given
by Equation (4) may be overestimated or underestimated.
4. DATA
To examine the impact of participating in tourism related work on
earnings of households, we collected data from two types of households:
participants and nonparticipants in tourism related work. Tourism
participant household has been defined as the household with at least
one of its family members earns from the activities directly related to
the tourism sector while households having none of its family members
engaged with tourism related income generating activities are defined as
the nonparticipants.
Through a preliminary survey conducted on recreational area, it was
identified that tourism related households come only from some of the
villages in two union councils, namely Sonda and Ongar. Total number of
villages in these two union councils is 44 villages (27 in Sonda and 17
in Ongar17). Social mapping of these villages was done to identify
villages where both tourism and non tourism households are living. Based
on social mapping, we selected six villages: Abdullah Gandhro, Wadero
Adam Manchri, Haji Khameso Khaskheli, Yousuf Hilayo, Sonehri, and Jafar
Hilayo. These villages are located within 10 kilometers from
recreational site in north-east to south of Keenjhar lake. Total
population of these six villages is 1345 households. Figure 1 shows the
map indicating the location of Keenjhar lake while Figure 2 presents map
of the study area where household data were collected.
Stratified random sampling method was used to select 264 households
from the selected six villages. From each of these six villages, 44
households were selected with 22 tourism participants and 22
non-participants. In each village, starting at a certain location,
surveyors were asked to knock at every third house on their left,
alternating between left and right at every turn. In case of
non-response, they were asked to knock on the next door.
Face to face interviews of head of the households were conducted
using a structured questionnaire pre-tested through a pilot survey of 25
households. The data were collected for twelve months of year. The
survey was conducted two times for ensuring the accuracy of data. The
first survey was conducted to collect data for six months (March to
August 2010) which included peak season of tourism. The second survey
was conducted to collect data from the same households for six months
(September 2010 to February 2011) which included off-peak season of
tourism.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
5. EMPIRICAL RESULTS
5.1. Descriptive Statistics
Table 1 presents the summary statistics to compare the average
values of variables between the households who participate in tourism
related work and households who do not participate in tourism related
work. Simple test of means between both types of households are also
included (last column). The results indicate that the average annual
earnings of tourism participants are higher than the nonparticipants by
Rs 16,021. However, this difference in earnings cannot be attributed as
the impact of the participation in tourism because of the difference in
other household characteristics. Partial effect of participation in
tourism related work on household earnings can be statistically
identified using regression analysis, presented in the next subsection.
Results in Table 1 show that the average household size is
statistically not different in both types of households but the number
of earning members in tourism participant households is higher than the
non-participant households. Average education of earning members is
statistically not different while earning members of participant
households are younger (28 years) than non-participants households (33
years). However, the average value of assets owned by households is
statistically different. On average, the distance from participant
households' villages to the recreational site of Keenjhar lake is
2.2 km while it is 3.45 km from non-participant households'
villages. Summary statistics in Table 1 also indicate that both groups
of households have overall very low earnings, low education level, low
value of assets, and large family size.
Tourism related income generating activities are presented in
Figure 3. Providing boating and hotel facilities to the tourists are the
highest income earning activities with 18 and 17 percent contribution in
the earnings of households, respectively. Providing tent and tubes to
the visitors are the subsequent highest incomes earning tourism
activities with 15 and 13 percent contribution, respectively. Renting
productive assets in tourism business is also a profitable business and
makes 10 percent contribution in the earnings. Providing transport and
labour services at the recreational site of the lake make 5 percent
contribution each. Vendor services, shop keeping and government services
account for 4 percent of their earning, while car washing is the lowest
earning activity at the lake (3 percent).
5.2. Regression Results
In our model, household earnings depend on the number of earning
members, value of productive assets, average age and average education
level of earning members. Participation dummy is equal to one if the
household participates in tourism related work, otherwise zero. As
discussed in Section 3, the household makes a decision whether to
participate in tourism related work or involve in any other income
generating activities. This makes the decision to participate as an
endogenous dummy variable.
Table 2 presents the regression results of three models. Third
column of the table presents results of a model where participation
dummy variable is assumed to be exogenous. In this case, the model is
represented by only Equation (1) with earnings as a dependent variable.
Fourth and fifth columns present results of the endogenous dummy
variable model. In this model, participation dummy variable is assumed
to be endogenous. In this case, the model is represented by two
equations: Equation (1) with earnings as a dependent variable and
Equation (5) with participation dummy as a dependent variable. For this
model, two specifications are presented. In the fourth column, the
exogenous variables affecting the participation include distance as well
as other variables which also affect household earnings. Following the
exclusion restriction, in the last column of the table, the exogenous
variable affecting the participation is distance only.
The estimate of the correlation between the error terms ([rho]) is
reported in Table 2. The Chi-squared test results show that this
correlation estimate is statistically significant at 1 percent
significance level. The test indicates that we have [rho] [not equal to]
0 and supports the endogenous dummy variable model. Endogenous dummy
variable model is also supported by the Jarque-Bera statistic for
normality test for normality of the error term. In this test, the null
hypothesis is that the error term is normally distributed. As p-value is
much greater than 0.05, the test does not reject the null hypothesis.
Thus, the diagnostic tests support the endogenous dummy variable model.
As Specification 2 of this model (in the last column of Table 2)
satisfies the exclusion restriction, we will discuss and interpret the
result of this model.
The results of participation equation in the last column show a
negative coefficient estimate for distance variable which is
statistically significant at 1 percent significance level. These results
show that the likelihood of household's participation in tourism
related work decreases when distance from household's village to
the recreational site of Keenjhar lake is higher. Results of earning
equation in the same column show that the explanatory variables earning
members, assets, and education are statistically significant at 1
percent significance level. The estimates indicate that the marginal
effect of an additional earning member on household's average
annual earning is Rs 13,987.
The marginal effect of productive assets is 0.2, which indicates
that any additional Rs 100 investment in productive assets would result
in higher earnings by Rs 20. The marginal effect of an additional year
of education level is Rs 5,258 on household's average earnings.
As explained in Section 3, the impact of participating in tourism
related work on household earnings is given by Equation (11). Results in
Table 2 show that the impact of the participation on household annual
earning Rs 9,251, which is 7.6 percent of the earnings. These results
show that the households who participate in the tourism related work
enjoy 7.6 percent higher level of earnings relative to the conditions
they would have been in had there been no tourism activities. When the
participation dummy variable is assumed to be exogenous, the impact of
the tourism participation on household annual earnings is Rs 12,003 (9.8
percent of the earnings), which is overestimated as the correlation
between the error terms ([rho]) is statistically significant. In
endogenous dummy variable model, the results of two specifications are
similar. The impact of the tourism participation on household annual
earnings is 7.4 and 7.6 percent, respectively. Results of this study
show that tourism at Keenjhar lake makes a positive contribution in the
earnings of the poor local people and in sustaining their livelihoods.
Results of the present study are similar to those in the study by
Bandyopadhyay and Tembo (2010), which also shows that tourism has
positive impact on overall welfare of households. Findings by Ashley
(2000) are also supported in the results of this study. Ashley (2000)
found that tourism has positive impact on livelihoods of rural people
and generally generates various types of cash income for rural
households.
6. CONCLUSIONS AND POLICY IMPLICATIONS
This study examines the impact of nature-based tourism on the
livelihood of local people at Keenjhar lake in Pakistan. For this study,
primary data were collected from 264 households selected by stratified
random sampling method. This study applies endogenous dummy variable
model to evaluate the impact of households' participation in
tourism related work on their earnings.
Results of this study show that the households who participate in
the tourism related work enjoy 7.6 percent higher level of earnings
relative to the conditions they would have been in had there been no
tourism activities. Study finds that tourism at Keenjhar lake makes a
positive contribution in the earnings of the poor local people and in
sustaining their livelihoods. Furthermore, the estimates of marginal
effect of productive assets indicate that any additional Rs 100
investment in productive assets would result in higher earnings by Rs
20. Education level of earning members also increases the earnings of
the household.
This study finds that tourism at Keenjhar lake improves the
standard of living of local people by raising their earnings, and that
the nature-based tourism can be an effective poverty alleviation
strategy.
REFERENCES
Amjad, S. and S. Kidwai (2003) Fresh-water, Brackish-water and
Coastal Wetlands of Sindh. Report prepared by National Institute of
Oceanography, Karachi.
Ashley, C. (2000) The Impacts of Tourism on Rural Livelihoods:
Namibia's Experience. Working Paper, Overseas Development Institute
(ODI), Portland House, Stag Place, London.
Ashley, C., H. Goodwin, and D. Roe (2001) Pro-poor Tourism
Strategies. Making Tourism Work for the Poor. A Review of Experience.
Overseas Development Institute (ODI) and International Institute for
Environment and Development (IIED), United Kingdom. (Pro-poor Tourism
Report No. 1).
Bandyopadhyay, S. and G. Tembo (2010) Household Consumption and
Natural Resource Management Around National Parks in Zambia. Journal of
Natural Resources and Policy Research 2:1, 39-55.
Bandyopandhyay, S., P. Shyamsundar, N. Humvindu, and L. Wang (2004)
Do Households Gain from Community-Based Natural Resource Management? An
Evaluation of Community Conservancies in Namibia. World Bank
Environment, Washington, DC, USA. (Policy Research Working Paper No.
3337).
Bennett, O., D. Roe, and C. Ashley, (1999) Sustainable Tourism and
Poverty Elimination Study. A Report to the Department for International
Development. London, UK: Overseas Development Institute.
Carrier, J. G. and D. V. L. Macleod (2005) Bursting the Bubble: The
Socio-Cultural Context of Ecotourism. Journal of the Royal
Anthropological Institute 11:2, 315-334.
Dehlavi, A. and I. H. Adil (2011) Valuing the Recreational Uses of
Pakistan's Wetlands: An Application of the Travel Cost Method.
Kathmandu, Nepal. South Asian Network for Development and Environmental
Economics (SANDEE) (Working Paper No. 58).
Goodwin, H. (2006) Measuring and Reporting: The Impact of Tourism
on Poverty. Cutting Edge Research in Tourism--New Directions, Challenges
and Applications. School of Management, University of Surrey, UK.
Greene, W. H. (2012) Econometric Analysis. (7th Ed.). New York:
Prentice Hall.
Gronau, R. (1974) Wage Comparisons: A Selectivity Bias. Journal of
Political Economy 82, 119-143.
Guha, I. and S. Ghosh (2007) Does Tourism Contribute to Local
Livelihoods: A Case Study of Tourism, Poverty and Conservation in the
Indian Sundarbans. South Asian Network for Development and Environmental
Economics (SANDEE). Kathmandu, Nepal. (Working Paper No. 26).
Heckman, J. J. (1974) Shadow Prices, Market Wages, and Labour
Supply. Econometrica 42, 679-694.
Holland, J., M. Burian, and L. Dixey (2003) Tourism in Poor Rural
Areas. Pro-poor Tourism, Department for International Development
(DFID), UK. (Working Paper No. 12).
Jamieson, W., H. Goodwin, and C. Edmunds (2004) Contribution of
Tourism to Poverty Alleviation Pro-Poor Tourism and Challenges of
Measuring Impacts Transport Policy Section, Transport Division, The
United Nations Economic and Social Commission for Asia and the Pacific
(UNESCAP), Bangkok.
Kuenzi, C. and J. McNeely (2008) Nature-Based Tourism. In Global
Risk Governance: Concept and Practice Using the IRGC Framework edited by
Renn O. and Walker K. D. International Risk Governance Council Book
series Volume 1, 155-178.
Leon, Y. (2007) The Impact of Tourism on Rural Livelihoods in the
Dominican Republic's Coastal Areas. Journal of Development Studies
43:2, 340-59.
Mangan, T., R. Brouwer, H. D. Lohano, and M. Nangraj (2013)
Estimating the Recreational Value of Pakistan's Largest Freshwater
Lake to Support Sustainable Tourism Management Using a Travel Cost
Model. Journal of Sustainable Tourism 21:3, 473-486.
Mitsch, W. J. and J. G. Gosselink (2007) Wetlands (4th Ed). New
York: John Wiley and Sons.
Pakistan, Government of (2009) Pakistan Economic Survey 2008-09.
Islamabad: Finance Division, Economic Advisor's Wing.
Pakistan, Government of (2010) Pakistan Economic Survey 2009-10.
Islamabad: Finance Division, Economic Advisor's Wing.
Roy, A. (1951) Some Thoughts on the Distribution of Earnings.
Oxford Economic Papers 3, 135-146.
Stata (2011) PDF Documentation, Stata Statistical Software: Release
11.2.
UNWTO (2011) UN WTO Tourism Highlights. United Nations World
Tourism Organisation.
van der Duim, R. and R. Henkens (2007) Wetlands, Poverty Reduction
and Sustainable Tourism Development, Opportunities and Constraints.
Wetlands International, Wageningen, The Netherlands.
WWF-Pakistan (2007) Preliminary Socio-economic Baseline Study
Report. Indus for all Programme. World Wide Fund for Nature-Pakistan,
Karachi.
Tehmina Mangan <t.economist@hotmail.com> is Assistant
Professor of Agricultural Economics at Sindh Agriculture University,
Tando Jam, Sindh. Heman D. Lohano is Senior Economist with SANDEE in
Kathmandu, Nepal and Associate Professor of Economics at IBA, Karachi.
Authors' Note: This study was conducted with financial and
technical support from the South Asian Network for Development and
Environmental Economics (SANDEE).
Table 1
Summary Statistics
Mean
Tourism Non-tourism Comparison
Household Household Test
Variables Definition (mean) (mean) (p-value)
Earnings Annual earnings of 138,412 122,391 0.034
household in Rupees
Household Number of family 7.77 7.60 0.664
Size members in household
Earning Number of earning 2.61 2.30 0.047
Members members in household
Education Average years of 4.51 4.15 0.610
schooling of earning
members
Age Average age of 28.15 33.80 0.000
earning members in
years
Assets Value of productive 23,440 28,748 0.354
assets owned by the
household in Rupees
Distance Distance in 2.22 3.45 0.001
kilometers from
household's village
to the recreational
site of Keenjhar
lake
Table 2
Regression Results
Model with Model with Endogenous Dummy
Exogenous
Dependent Explanatory Specification Specification
Variable Variables Dummy 1 2
Earnings
Constant 47,365 *** 1,091 24,276 *
(3.680) (0.0622) (1.800)
Earning 15,619 *** 15,133 *** 13,987 ***
members (7.219) (6.184) (6.546)
Assets 0.236 *** 0.268 *** 0.204 ***
(4.248) (4.243) (3.718)
Education 5,523 *** 5,487 *** 5,258 ***
(1161) (10.21) (1142)
Age 279.5 1,046 *** 400.4
(0.937) (2.794) (1.371)
Participation 12,003 ** 58,073 *** 62,677 ***
dummy (2.317) (5.095) (5.802)
Participation Dummy
Constant -- 1.839 *** 0.360 ***
(4.455) (3.586)
Distance -- -0.127 *** -0.114 ***
(-4.516) (-5.126)
Earning -- -0.0546 --
members (-0.757)
Assets -- -3.87e-06 * --
(-1.953)
Education -- -0.0114 --
(-0.742)
Age -- -0.0381 *** --
(-3.759)
Observations 264 264 264
R-squared 0.586
Goodness of 73.16 *** -- --
fit
F-statistics
Goodness of -- 308.2 *** 367.3 ***
fit Chi-Square
Correlation -- -0.655 *** -0.704 ***
between error
terms ([rho])
Jarque-Bera 11.88 0.298 0.434
statistic for
normality test
p-value of 0.003 0.861 0.805
above
Impact of Participation in Tourism on Earnings
Impact in 12,003 9,051 9,251
Rupees
Impact in 9.8 7.4 7.6
percentage
t-statistics in parentheses.
*** p < 0.01, ** p < 0.05, * p < 0.1.
Fig. 3. Income Earned from Tourism Related Economic Activities
Providing swimming dress 2%
Boat 18%
Hotel 17%
Tube 13%
Tent 15%
Transport 5%
Rent asset 10%
Labour 5%
Government services 4%
Shop 4%
Vendoring 4%
Car washing 3%
Note: Table made from pie chart.
