Hilsa’s

non-consumptive
value in Bangladesh
Estimating the non-consumptive
value of the hilsa fishery in
Bangladesh using the contingent
valuation method
Essam Yassin Mohammed, Liaquat Ali, Shahjat Ali,
Belayet Hussein, Md Abdul Wahab, Nathan Sage

Working Paper Fish

Keywords:
November 2016 hilsa fishery; economic valuation;
contingent valuation
About the author Partner Organisation
Essam Yassin Mohammed is a Senior Researcher in the The Bangladesh Centre for Advanced Studies (BCAS) is an
Shaping Sustainable Markets Group at the International independent, non-profit, non-government, policy, research and
Institute for Environment and Development, UK. implementation institute working on Sustainable Development
Email: [email protected] (SD) at local, national, regional and global levels. It was
Liaquat Ali, Shahjat Ali and Belayet Hussein, of the Bangladesh established in 1986 and has grown to become a leading
Centre for Advanced Studies, Bangladesh research institute in the non-government sector in Bangladesh
and South Asia.
Md Abdul Wahab of WorldFish Centre, Bangladesh
WorldFish is an international, nonprofit research organization
Nathan Sage of USAID, Bangladesh that harnesses the potential of fisheries and aquaculture to
reduce hunger and poverty. WorldFish is a member of CGIAR, a
Acknowledgements global agriculture research partnership for a food secure future.

The authors acknowledge the financial support of USAID.

Special thanks are given to the EcoFishBD project team of the Published by IIED, November 2016
WorldFish for their active support in the implementation of the
Mohammed, E.Y., Ali, L., Ali, S., Hussein, B., Wahab, M.A.,
research and to Syed Arif Azad, Director General, Department
Sage, N. 2016. Hilsa’s non-consumptive value in Bangladesh:
of Fisheries, Bangladesh for his continuous inspiration, sincere
Estimating the non-consumptive value of the hilsa fishery
interest and support towards the implementation of the
in Bangladesh using the contingent valuation method. IIED
research activities. Finally, the enthusiasm and sincere support
Working Paper. IIED, London.
from the officials of the Department of Fisheries at the field
level, as well as at headquarters, are gratefully recognised. http://pubs.iied.org/16626IIED
ISBN 978-1-78431-425-5
Produced by IIED’s Shaping Sustainable Printed on recycled paper with vegetable-based inks.
Markets Group
The Shaping Sustainable Markets group works to make sure
that local and global markets are fair and can help poor people
and nature to thrive. Our research focuses on the mechanisms,
structures and policies that lead to sustainable and inclusive
economies. Our strength is in finding locally appropriate
solutions to complex global and national problems.

International Institute for Environment and Development

80-86 Gray’s Inn Road, London WC1X 8NH, UK
Tel: +44 (0)20 3463 7399
Fax: +44 (0)20 3514 9055
email: [email protected]
www.iied.org
@iied
www.facebook.com/theIIED
Download more publications at www.iied.org/pubs
 IIED Working PAPER

Hilsa is Bangladesh’s most important single-species

fishery: for cultural identity, earnings and employment.
However, overfishing, habitat destruction, siltation,
pollution and climate change have driven catches down,
and management policies have not adequately intervened
— probably because the fishery’s total economic value
is under-appreciated. This study is the first to estimate
the non-consumptive (non-use) value of a well-managed
hilsa fishery. It used the contingent valuation method
and asked 1006 fishing and non-fishing households
how much they would be ‘Willing To Pay’ (WTP) for
an effectively-managed fishery. In Barisal Division, an
improved fishery could be worth BDT 651.8M – 1,384.2M
a year (approximately US$8.3M – 17.7M). Nationally, a
better-managed fishery could be worth BDT 13,128.6M
– 27,882.1M per year (US$167.5M – US$355.7M).
Poorer people are willing to pay the highest proportion
of their income, suggesting fishery restoration would be
pro-poor. However, any interventions must share benefits
equitably and address the systemic constraints facing low
income groups.

Contents
Summary4 4 Results and discussion 12
Estimating non-consumptive value of hilsa fishery  13
1 Introduction 5
Factors explaining willingness to pay 15
Income elasticity of willingness to pay 16
2 The contingent valuation method (CVM) 7
Hypothetical market design 9 5 Conclusion 17

3 Methodology 9 References  18
Sampling design  11
Descriptive statistics 12 Abbreviations19

www.iied.org  3
Hilsa’s non-consumptive value in Bangladesh

Summary
Hilsa is the most important single-species fishery in When calculating value, we used a Kaplan Meier
Bangladesh. Almost all the catch is Tenualosa ilisha, survival estimate to establish that respondents had
and the fishery contributes significantly to cultural considered their budget constraints and were acting
identity, the national economy and employment. rationally. We also asked a follow up ‘how certain
However, overfishing, habitat destruction, siltation, are you that you would be willing to pay?’ question,
pollution and climate change mean catches have which was presented in a 10 point Likert scale. A cut
declined in recent decades. off point of 8 and above was used to calibrate WTP
statements. However, the estimates were not statistically
Management policies have not intervened adequately
distinguishable from the uncalibrated values. This further
to reverse this trend — probably because the total
suggests that the survey protocol was well executed
economic value of the fishery is not well understood.
and respondents were behaving as they would in a real
While citizens and policymakers recognise the cultural
world market scenario.
significance of the fishery, until now no monetary value
has been put on this aspect. Yet valuing the fishery only We then explored the distributional implications of
on landed catches underestimates its true importance benefits from an improved hilsa fishery by estimating
and the benefits of better management. income elasticity of WTP. This was found to be less
than one, suggesting lower income groups are willing to
This study is the first attempt to value the non-
pay proportionately more for hilsa fish restoration than
consumptive (non-use) aspects of a well-managed hilsa
higher income groups will pay. This implies that low
fishery. It surveyed 1006 households in Barisal Division,
income segments of society are most reliant on the hilsa
asking how much they would be Willing To Pay (WTP)
fishery, and therefore investment in hilsa fish restoration
for a hypothetical fisheries restoration programme
is pro-poor.
run over ten years. We used the contingent valuation
method to convert WTP statements to estimates of We believe that investments equivalent to only a
the fishery’s economic value, using the median and fraction (5-10 per cent) of the fishery’s estimated
the mean amounts people were willing to contribute to non-consumptive value would bring about the desired
calculate lower and upper estimates. change. Such investments could: restore crucial fish
habitats; effectively enforce the fishing regulation (as
In Barisal Division, an improved fishery is estimated
stipulated in the Fish Protection and Conservation Act-
to be worth between BDT 651.8M (approximately
1950), and provide incentives to local fishers to stop
US$8.3M) and BDT 1,384.2M (approximately
destructive fishing practices.
US$17.7M) per annum. Extrapolating the analysis to
the national level suggests a better-managed hilsa
fishery would be worth between BDT 13,128.6M
(approximately US$167.5M) and BDT 27,882.1M
(US$355.7M) per annum to Bangladesh. It must be
noted that these estimates do not include the fishery’s
use or consumptive values.

4 www.iied.org
1
IIED Working paper

Introduction
Coastal and marine resources provide a range of The hilsa fishery is the biggest single-species fishery in
ecological functions that directly and indirectly support Bangladesh1 with landings contributing approximately
human lives and economies, often categorised as 10 per cent to annual fish production (FRSS 2014), and
supporting, regulating, provisioning and cultural services 1 per cent to the country’s annual GDP (DoF 2014).
(MEA, 2005). Hilsa spend much of their life in coastal waters but they
migrate upstream to spawn in coastal rivers (Rahman
For many of the world’s countries, fisheries play an
and Naevdal 2000).While hilsa is broadly distributed
important role in meeting food demands, in addition
from Vietnam to the Persian Gulf (Freyhof 2014),
to providing employment and income. Fisheries alone
Bangladesh takes 50–60 per cent of the catch with
support close to 250 million livelihoods around the
relatively smaller proportions taken by Myanmar (20–25
world and produce food worth nearly US$190 billion
per cent), India (15–20 per cent) and other countries
per year (WAVES, 2012). In 2010, fish accounted for
(5–10 per cent) (Rahman et al. 2012). In Bangladesh
16.7 per cent of the global population’s animal protein
alone, an estimated half a million people directly depend
intake (FAO 2014). In 2012, around 58.3 million people
on the fishery and a further 2.5 million are indirectly
were engaged in capture fisheries and aquaculture,
involved in supply-chain activities such as processing,
with 84 per cent located in Asia (FAO 2014). Over the
transportation and marketing (Rahman et al. 2012).
past 50 years, global landings of fish have increased at
an average rate of 3.2 per cent per year (FAO 2014). Hilsa has cultural and religious significance in the South
Despite this upwards trend in landings, coastal fisheries Asian region. The people of Bangladesh and West
are declining due to overfishing, compromising the Bengal in India, as well as Bengali-speaking people
sustainability of this important resource (Pauly 2006). throughout the world, love fish. They like to define
themselves with the phrase ‘mache bhate Bengali’, or
Bangladesh is one of the world’s leading fish-producing
‘fish and rice make the Bengali’. Hilsa holds the highest
nations. According to the latest available data, in
position among the rich biodiversity of the Ganges
2011–2012, fish production contributed 4.4 per cent
river system, and its importance is further emphasised
to the country’s national GDP, 2.5 per cent to foreign
through different dishes and their use in ceremonial
exchange earnings, and 60 per cent of all consumed
festivals (see Box 1). Thus, hilsa is important socially,
animal protein (FRSS 2013). In addition to its economic
culturally, and religiously to the Bengali people and
importance, the fisheries sector is a significant source
people in many Indian states including Odisha, Bihar
of employment, with 11 per cent of the country’s
and Assam (Mohammed and Wahab 2013).
population directly or indirectly involved in this sector
(FRSS 2013).

1
While there are three separate species of Hilsa (Hilsa kelee, H. toli and Tenualosa ilisha), where we use the term ‘hilsa’ throughout this paper we are referring
to T. ilisha as it the most numerous species making up to 99 per cent of the annual fish catch in Bangladesh.

www.iied.org 5
 Hilsa’s non-consumptive value in Bangladesh

However, numerous studies have concluded that a Typically, markets do not capture non-use values,
burgeoning human population and a corresponding or as we call them in this study, non-consumptive3
demand for fish protein has driven mass overfishing values of artisanal and small-scale fisheries. For
of both adults and jatka (juvenile hilsa) in the gill net2 example, fisheries’ socio-cultural values are usually
fishery (Amin et al. 2008 and 2002; Rahman et al. not fully understood, or are poorly accounted for in
2013). Hilsa, which was abundantly available in the national accounts (Lafolley et al. 2009). Such values
100 rivers of Bangladesh until the 1960s and 1970s, are rarely factored into decision making processes,
declined gradually over 30 years to reach a low point which instead typically focus on short-term, produce-
in 2002 when catches were 0.19 million tonnes. This based commodities.
decline was due to a combination of factors: the
Therefore, this study aimed to estimate the non-
closure of migratory routes, river siltation, overfishing,
consumptive values of the hilsa fishery in monetary
indiscriminate harvesting of brood stocks and juveniles,
terms that encompassed cultural, religious and
use of fishing nets with very small mesh sizes,
sentimental values. We used the Contingent Valuation
mechanisation of fishing, increasing numbers of fishers,
Method (CVM) to estimate the ‘willingness to pay’
pollution, and climatic variability.
of residents in Barisal Division for non-consumptive
We argue that such threats from overfishing are (socio-cultural, religious and sentimental) benefits of a
exacerbated by policymakers’ insufficient investments hypothetically-restored fishery.
in restoring the fishery, despite its economic and
Unlike the majority of CV studies, which are conducted
cultural importance. This probably occurs because the
in relatively high income countries, this study aims to
fishery’s true value is poorly understood. An explicit,
estimate the non-consumptive values of hilsa fishery
rather than implicit, understanding of the multiple values
in Bangladesh, a least developed country (LDC). We
that artisanal hilsa fishing communities provide to, and
will also identify the determinants of willingness to
receive from, coastal and marine ecosystems is vital for
pay statements. Subsequently, Section 2 reviews the
well-informed policy.
contingent valuation method and Section 3 describes
the survey design. Section 4 reports and discusses
empirical results, and Section 5 finally presents
concluding remarks.

2
A fishing net that is hung vertically in the water column and typically made of monofilament or multifilament nylon. The mesh sizes are designed to allow fishes
to get only their head through the netting. The gills of the fish then become caught in the net when they try to escape.
3
While non-use value is commonly used in existing literature, the use of ‘non-use’ was creating unwanted confusion during consultation meetings and survey
design phase of the study. This is primarily because all the stakeholders consulted were interpreting the phrase literally and did not feel with the term. Moreover,
we were not able to find a close equivalence of the phrase in Bengali language. Therefore, non-consumptive was used instead.

6 www.iied.org
2
IIED Working paper

The contingent
valuation method
(CVM)
The contingent valuation method (CVM) puts a For example, a person may want to buy some fruit and
monetary value on something for which there is no vegetables, school uniforms for their children, and also
market and therefore no price (or compensation restore a fish stock that they rely on.
payment). Non-use or non-consumptive values of small-
If that person faces some budget constraint, s/he
scale fisheries fall into this category. The contingent
will be forced to forgo something in order to afford
valuation is a survey-based method where people
something else. Let us assume that the consumer has
are asked how much they are willing to pay for an
an exogenous disposable income y, which is to be spent
improvement and/or how much compensation they
on some or all n commodities. These can be bought for
would accept for the deterioration of a given ecosystem
an overall price P in non-negative quantities at given,
quality (Mohammed 2009). The improvement or
fixed, strictly positive component prices:
deterioration is the contingency; i.e., the hypothetical
state, which the survey respondents are asked to P : P = {p1, p2, …, pk} (3)
imagine (Blore, 1996).
The best choice lies on, rather than below, any budget
The method was first mathematically articulated by constraint (Johansson 1991). Hence, one can write the
Maler (1974), who sought to extend standard welfare utility function of the consumer as:
theory of price changes to changes in the supply of a
U(X,Q) (4)
public good (Willis and Corkindale, 1995, pp. 84–85).
A consumer has preferences over n conventional market Put simply, utility (or happiness) can be attained
commodities, such as groceries, subscription to internet by consuming a certain marketable good (x) and
service, a mobile telephone, and so on, denoted here ecosystem service (q). As consumers intend to
as X: maximise their utility with respect to x marketable goods
(subject to the usual budget constraints), the problem of
X : X = {x1, x2, …, xn} (1)
utility maximisation, can be written as:
Since the basic premise of neoclassical economic
max xU(X,Q); such that p∗x ≤ y and Q = Q0  (5)
theory is that people have preference over both
marketable and non-marketable goods, it is vital to Champ et al. (2003) explain that there are two
incorporate the consumer’s preference over a set of k constraints that people face in maximising utility. First,
other items (such as public goods), here called Q: the total expenditure on market goods cannot exceed
income. Second, the levels of the non-market goods
Q : Q = {q1, q2, …, qk} (2)

www.iied.org 7
Hilsa’s non-consumptive value in Bangladesh

are fixed. This is because the consumer cannot control Accordingly, the individual’s utility increases from:
the level of ecosystem service provided. The (x) that
U0 ≡ v(P, Q0, y) to U1 ≡ v(P, Q1, y) (6)
solves this problem then depends on the level of income
(y), the prices of all the market goods (P), and the Where U0 is the person’s initial utility before
level of the rationed, non-market goods (Q). For each improvement, and U1 is utility or happiness after
market good, we have an optimal demand function improvement. Based on the welfare theory of price
that depends on these three elements, xi∗ = xi(P,Q,Y); changes, Maler (1974) defined compensating and
the vector of optimal demands can be written similarly, equivalent variation measures for this utility change (see
x∗ = x(P,Q,Y) where the vector now lists the demand also Willis and Corkindale, 1995). The compensating
function for each market good. variation, which is denoted as ω, corresponds to
the individual’s willingness to pay (WTP) for the
If we are inserting the set of optimal demands into the
improvement, and satisfies the following equation:
utility function, we obtain the indirect utility function,
U(X∗, Q) = v(P, Q, y). Because the demands depend v(P, Q0, y) = v(P, Q1, y − ω)(7)
on prices of the goods that a consumer wants to buy,
The basic idea behind ω is that if a person gives up ω
the levels of the non-market goods such as restored
for the improvement (e.g. a restored fish stock), then he/
fish stock, and the consumer’s income, the highest
she is back to the original utility. Champ et al. (2003)
obtainable level of utility or satisfaction also depends
further argue that ω could be positive or negative
on these elements. Now, suppose that Q (the level of
depending upon how much benefit is gained or lost.
environmental quality, the level of river water quality for
instance) increases from Q0 (initial quality or quantity)
to Q1 (improved quality or quantity) while prices and
income remain constant at (P, Y). As long as prices and
income are kept constant, then we expect the person to
be happier when the level of the environmental quality is
improved.

8 www.iied.org
3
IIED Working paper

Methodology
Hypothetical market design Our participatory approach developed a brief
description of hilsa’s non-consumptive value. This
Initial focus group discussions with 146 households and identified the most important cultural values and the
31 fisheries managers and experts were conducted to most significant threats the fishery currently faces. We
define the use and non-use (or non-consumptive) values explored the investments needed to reverse the trend
of the hilsa fishery. and restore the fishery, and developed a hypothetical
scenario for achieving this, summarised in Box 1.

Box 1. Hypothetical market scenario

Hilsa is the most preferred fish of the people of pre-1970s situation where hilsa was available in all
Bangladesh and West Bengal in India, and is of major rivers, the average weight of caught fish was
religious and cultural importance, forming part of back up to 800g (from around 300g now), and most
Bengali festivals. Hilsa has been recognised as the people could afford to buy hilsa. We asked fisheries
‘national fish’ of Bangladesh. In some Hindu Bengali managers and experts what would be needed for
families, large hilsa fish are bought for engagements a national hilsa fish restoration programme. They
and pre-wedding ceremonies. One such important suggested a ten year programme to:
occasion is the Jamai Sashti, when the son-in-law
1. Dredge river beds;
visits his prospective parents-in-law. A Jamai Sashti
meal is never complete without at least one dish of 2. Control pollution;
hilsa, and it is often expected that the bridegroom will
3. Compensate fishers for the adhering to a fishing
bring a pair of hilsa for the occasion.
ban during spawning season; and
Pohela Boishakh, the first day of the Bengali New
4. Boost capabilities to enforce the closed season
Year, is ceremonially observed in both Bangladesh
and ban on harmful fishing gear (such as the jal).
and the Indian state of West Bengal as a national
day. Bengali communities celebrate Pohela Boishakh Such a national programme requires a large amount
with a special menu of Panta-Ilish (fermented rice and of money. Each household needs to pay a monthly
fried hilsa). contribution towards the programme over the 10
years. The payment would be an additional fee on
Recent and significant declines in catches have
Union Parishad taxes (on buildings and land holding
pushed up prices, meaning most low income
tax) and would be directed to a National Hilsa
groups can no longer afford to buy hilsa. Decline
Conservation Foundation – which would administer
in hilsa fish stock also poses a major threat to the
the fund and work closely with the government and
socio-cultural benefits of the fishery. Significant
fisher communities to implement restoration activities.
investment is required to reverse the trend to the

www.iied.org 9
Hilsa’s non-consumptive value in Bangladesh

There are many ways to ask people their willingness to programme, and not necessarily because they do not
pay, and these are called ‘elicitation methods’. Widely value the ecosystem service. However, since there
used elicitation methods include: open-ended, single or was no statistically-observable relationship between
double bound dichotomous choices, discrete choices, people’s support for a programme and their willingness
or offering a ‘payment card’ from which respondents to pay statements, no interviews were discarded at the
can pick a value. Every method has its own advantages data analysis stage.
and disadvantages and none is immune to criticism.
We also addressed ‘hypothetical bias’ – divergence
For example, open-ended questions such as ‘how much
between actual and hypothetical willingness to pay
would you pay to restore the fishery?’ may provide a
(Neil et al. 1994; List and Gallet 2001; Murphy et al.
straightforward valuation on the ecosystem service in
2005; Blumenschein et al. 2008), which leads to
question (Ahmed and Gotoh, 2007), but the method is
over-valuation of the ecosystem service in question.
susceptible to ‘hypothetical bias’ in that when the idea
There are several approaches for mitigating (and
is hypothetical people say they are willing to pay, but
where possible eliminating), hypothetical bias. These
may be less willing when the payment becomes a reality
range from the ‘cheap talk method’, which reminds
(Mohammed, 2012). Similarly, bidding games suffer
respondents to consider their budget constraints and
from starting point bias.
not to ‘overstate’ their willingness to pay, to using a
Therefore, selecting the right elicitation method is follow up ‘certainty question’ that asks respondents to
crucial to gathering valid and reliable data. Our focus state how certain they are about their choice.
groups discussed and pre-tested a number of elicitation
The effectiveness of ‘cheap talk’ in mitigating
techniques. A payment card method was found to
hypothetical bias is still debated, but studies conclude
be the most easily-understood and straightforward.
that a follow up certainty question can be used to
Therefore, following Carson and Michell (1993), the
effectively calibrate willingness to pay statements
study developed a payment card designed to cover the
(Murphy et al. 2004). In this study, following Champ
likely range of responses, as identified in the pre-test
et al. (2004) and Mohammed (2012), a 10 point Likert
surveys (see Box 2).
scale was used, where 0 was ‘very uncertain’ and
Our full survey covered 1006 households (see section 10 was ‘very certain’. Those respondents who are
on sample design). Before asking people about ‘sufficiently’ certain are considered as giving a ‘true’
their willingness to pay, they were asked whether WTP statement (Blomquist et al. 2008). Following
they supported the hypothetical National Hilsa Fish Champ et al. (2004), a cut-off point of 8 and above was
Restoration Programme. We used a Likert scale used to calibrate WTP statements. What this means is
between 0 (strongly disagree) to 10 (strongly agree) that, those respondents who stated certainty level 8 and
and asked respondents to choose their point on the above were considered to be ‘sufficiently’ certain.
scale. This was done to check for ‘protest bidders’,
The questionnaire survey also included a question on
who would understate their willingness to pay simply
the respondent’s socioeconomic characteristics, and a
because they do not agree with the proposed
set of attitudinal and behavioural questions.

Box 2. Willingness to pay elicitation question

What is the highest amount of money in Bangladeshi Taka, if anything, that your household would pay each
month for the next 10 years to make a National Hilsa Fish Restoration Programme possible?
(Circle the highest amount at which your household would still vote for the programme).’

   0   20   40 60  80  100  130

 150  180  210 240 300  350  400
 500  600  700 800 900 1000 1500
2000 3000 4000 > 5000

If more than BDT 5000, then how much? ____________________

10 www.iied.org
 IIED Working paper

Sampling design Household respondents were selected from each sub

district through stratified random sampling. Each sub
The study surveyed 1006 households in five districts district sample was allocated proportionately to the
(Barisal, Bhola, Patuakhali, Pirojpur and Jhalokathi) selected villages, according to population size of the
of Barisal Division. See Figure 1. Within each of the village. To achieve this, all households in the selected
districts, two sub-districts (upazila) were identified village were serially numbered and every nth household
through consultations with local fisheries officers. Since was selected for interview, where n is the total number
it is believed that hilsa is important to both fishers and of residents in the village divided by the total sub
non-fishers alike, sub-districts with sufficient variation district level sample size. Household respondents were
in households’ main source of livelihood and in income questioned in person by interviewers using a structured
levels were selected. interview (available on request).
The site selection process also ensured sufficient
variety in distance to major rivers and fish markets. This
was done to investigate ‘distance decay’ in willingness
to pay statements.

Figure 1. Map showing location of the study site and villages surveyed

www.iied.org 11
4
Hilsa’s non-consumptive value in Bangladesh

Results and discussion

Descriptive statistics
The questionnaire included questions on household
income. Since most respondents do not have a
steady income level, and if they do, they may not feel
Since interviews were conducted ‘in person’, all the
comfortable to state the exact figure, a range of values
1006 households provided answers to all questions.
was presented (e.g. below BDT 1000; BDT 15000 –
As mentioned in the previous section, the issue of BDT 20,000; BDT 1000 – BDT 2000; BDT 20,000 –
‘distance decay’ has been discussed in a number BDT 30,000 and so on). A midway value of the intervals
of CVM studies. This is the phenomenon in which was used to estimate average income. The average
respondents’ willingness to pay falls with distance from household income was estimated at BDT 12,190.85
the main area of interest (e.g. a major river or market). per month.
According to Hanley et al. (2003), such relationships
Questions on whether respondents are concerned with
are common where people are being asked to bid for
the current state of hilsa stock; whether or not they had
the use-value (as opposed to non-use value) of an
participated in environmental awareness programmes;
ecosystem service or good. However, when estimating
and whether or not their jobs involve environmental
non-consumptive values (as is the case in this study),
conservation were used as proxies for attitudinal and
such an inverse relationship is not necessarily expected
behavioural influences. Forty four per cent of the
to be observed.
respondents stated that they are very concerned about
Male and female representation was balanced (51 per the current state of hilsa fishery. Sixteen per cent of the
cent and 49 per cent respectively). The average age respondents had participated in programmes to raise
of the respondents was about 41. The sample also awareness about fisheries management at least once.
had a good mix of household representatives ranging A staggeringly high 98 per cent of respondents stated
from aratdars (middlemen), labourers and fishers that overfishing is the major threat to hilsa fishery.
to merchants, farmers and professionals (including
teachers, government officers, etc.).

12 www.iied.org
 IIED Working paper

Table 1. Descriptive statistics

Variable Definition Mean/ Std.

per Dev.
cent
Gender 1 = male; 0 = otherwise 51% 0.500
Age In years 40.821 14.03
Education Years of schooling 6.130 4.151
EconActiveHH Total number of persons who are economically active or 1.352 0.753
earn income.
Aratdar 1 = respondent is aratdar; 0 = otherwise 0.002 0.045
Fisherman 1 = if respondent is fisherman; 0 = otherwise 0.083 0.275
Labourer 1 = respondent is labourer; 0 = otherwise 0.219 0.414
Merchant 1 = respondent is merchant; 0 = otherwise 0.266 0.442
Professional 1 = employed in government and/or non-government 0.182 0.386
sectors
Farmer 1 = respondent is farmer (including animal husbandry) 0.240 0.427
DistanceWholesale Distance from major wholesale fish market in Km. 4.457 5.100
AverageIncome Family’s average income (midpoint of interval). 12190.85 8278.27
VeryConcerned 1 = respondent is very concerned about the state of 0.439 0.497
hilsa fish; 0 = otherwise
PartAwarness 1 = respondent has participated in awarness raising 0.158 0.365
programmes; 0 = otherwise
WorkEnvNGOs 1 = respondent works for environmental NGO; 0.116 0.321
0 = otherwise
OverFishing 1 = respondent thinks main threat to hilsa fishery is 0.978 0.146
overfishing; 0 = otherwise
WTP Payment card bid value in BDT (monthly) 63.713 170.04
CertaintyQ Likert scale between 1 and 10 7.417 3.944

Estimating non-
However, the median value may be a more realistic
measure of the central tendency of WTP in a world

consumptive value of where decisions are based on voting and where people
have concerns about the distribution of a programme’s

hilsa fishery benefits and costs.

It must be also noted that mean values are higher than
Once the respondents’ WTP statements had been median values. Therefore, in some studies they have
elicited, estimating the aggregate non-consumptive been used as upper estimates of WTP, and median
value of a hypothetically improved hilsa fishery is values are used as an indication of the lower limit. We
straightforward. However, the value estimates vary adopt this approach here. Tables 2 and 3 estimate
depending on whether mean or median values are used. value of the fishery based on the mean and median
respectively. Total estimated value is computed by
Using the mean value may reflect the Kaldor–Hicks
multiplying mean or media WTP statements by the
potential compensation criterion, which says that there
number of households at division and national levels.
will be a net gain in social welfare if those who have
Average monthly WTP statements were multiplied by 12
welfare gains can both compensate losers and still
to obtain annual estimates.
have a net gain for themselves (as argued by Cameron
and Huppert 1989 and echoed by Mohammed 2009).

www.iied.org 13
Hilsa’s non-consumptive value in Bangladesh

Table 2. Value estimates using mean WTP

Geographical Mean WTP Number of Total annual

scope statement (BDT) households* estimated value
(BDT)
Barisal Division 63.71  1,810,444.44 1,384,186,162.67
Bangladesh (national) 63.71 36,468,330.11 27,882,080,597.39
* Population statistics were obtained from National Census 2011 (latest available data) and is available at: http://203.112.218.69/binbgd/RpWebEngine.exe/
Portal?BASE= HPC2011_short&lang= ENG

Table 3. Value estimates using median WTP

Geographical Median WTP Number of Total annual

scope statement (BDT) households estimated value
(BDT)
Barisal Division 30  1,810,444.44 651,760,000.00
Bangladesh (national) 30 36,468,330.11 13,128,598,840.45

At Barisal division level, the non-consumptive value in Table 2 and 3, we estimate lower and upper
of the hilsa fishery was estimated between BDT national values for the fishery at BDT 13,128,598,840
651,760,000.00 (approximately US$8.3M) and (approximately US$167.5M) and BDT 27,882,080,597
BDT 1,384,186,162.67 (US$17.7M) per annum. (US$355.7M) respectively.
While extrapolating value estimates by multiplying To examine whether respondents have seriously
average or median WTP statements is a common considered their budget constraints in answering the
practice when estimating values at sub-national level, WTP question we ran a Kaplan-Meier survival estimate.
the validity of employing the same principle to estimate This shows the probability that a given payment card bid
national level estimate is often disputed. value is selected. In the real world we would expect a
decreasing curve showing that higher bids are less likely
One of the main criticisms of such a general approach
to be selected — and this is what we found. Figure 2
is that some determining factors can be significantly
demonstrates that respondents have indeed considered
different across regions. For example, and as previously
their budget constraint and were behaving rationally.
alluded to, distance from an environmental good or
This can be used to demonstrate the reliability of the
service may make WTP vary significantly. But in this
data collected.
study WTP did not diminish with distance from rivers.
On the contrary, the further the distance the higher We tested for statistical variability between calibrated
the WTP (even though only slightly higher). Therefore, (using certainty question cut-off point of 8) and
assuming all other factors are constant, we believe uncalibrated WTP values, and it was found that they
that it is reasonable to extrapolate to a national value, were not statistically distinguishable. We conclude, with
particularly since hilsa is revered as part of national a word of caution, that there was no evidence found to
identity across the country. Therefore (and as presented suggest that the survey suffered from hypothetical bias.

14 www.iied.org
 IIED Working paper

Figure 2. Kaplan-Meier survival estimate of WTP statements

1.00

0.75

0.50

0.25

0.00

0 500 1000 1500 2000 2500

Payment card bid value

Factors explaining It was also found that if a household has one more
economically active member, then its WTP is likely
willingness to pay to increase by 5.1 per cent. This is mainly because
there is a positive correlation between the number
Beyond value estimates, we are also interested in of earning household members and its income level.
examining the factors that affect WTP statements. To do Interestingly, but not unexpectedly, aratdars (middlemen
so, we regressed the natural logarithm of WTP against in the hilsa supply chain) are willing to pay nearly
socioeconomic, attitudinal and behavioural variables. double than non-aratdars. This is chiefly because their
The value function can be mathematically denoted as: livelihood relies heavily on the status of hilsa fish stock.
As mentioned earlier, distance from major fish market
LnWTP(ω) = α + β1X1 + β2 X2 + β3X3 + … + βnXn(8)
was also positively related to WTP (even though by
The regression analysis is presented in Table 4. While a less remarkable magnitude). This is presumably
the coefficients of respondents’ age and gender are because hilsa fish are regarded as a national treasure
not statistically significant, educational level, average and are highly valued by Bangladeshis regardless of
monthly income, number of economically active their geographic location. According to Hanley et al.
household members, whether the respondent is an (2003) and Mohammed (2009), this is not uncommon
aratdar (middleman), and distance from wholesale fish in studies that estimate non-use or non-consumptive
market were found to be statistically significant. A one values of ecosystem services.
per cent increase in income yielded a 0.13 per cent
The study also started with a hypothesis that attitudinal
increase in WTP.
and behavioural characteristics affect respondents’
WTP. While attitudes are primarily about beliefs and
opinions, behaviour involves actions. If a respondent
believes that overfishing poses the main threat to
the hilsa stock, our analysis suggests WTP will be
31.1 per cent higher.

www.iied.org 15
Hilsa’s non-consumptive value in Bangladesh

Those who have participated in sustainable fisheries Perhaps one way of assessing whether communities
management awareness programmes, and those who see investment in fish stock restoration as a luxury or as
work for environmental groups and non-governmental essential is to look at how demand varies with income.
organisations had higher WTP by 22.5 per cent and But since there is no market where environmental goods
14.4 per cent respectively. and services are traded, estimating income elasticity
of demand for environmental goods is difficult. Hokby
Finally, we found that if a respondent is more certain
and Soderqvist (2003) argue that “independent of the
about his/her WTP statement by one unit on the one
issue whether environmental services are luxuries or
to ten scale, then s/he is very likely to state 13.2 per
not, there are distributional reasons to be concerned
cent higher WTP than those who are not. This further
about what income groups in society are relatively more
demonstrates that the respondents were considering
willing to pay for an increased provision of environmental
a real-world scenario where they considered making a
services”, i.e., a measure of willingness to pay (WTP) is
payment only when they thought they were certain.
affected by changes in income.
Table 4. Regression result
Ebert (2003) emphasised that income elasticity of
demand for environmental improvement does not
Variables β Std. Err. enable us to assess distributional impacts; therefore,
Age 0.001 0.002 an assessment of the income elasticity of WTP for
environmental good is decisive.
Gender 0.024 0.043
Recall equation (7) from the CVM section. We
Education 0.023*** 0.005 established that a consumer would have the same utility
by giving up some of his income to attain a desired
LnIncome 0.133*** 0.047
improvement in ecosystem service (from q0 to q1)
EconActiveHH 0.051*** 0.028 so long as the price of other marketable goods and
services remains constant.
Aratdar 0.947* 0.447
Therefore, solving for WTP (ω) from equation 7 we
Distance from major 0.007*** 0.004 obtain:
fish market q1 ∂v/∂q q1
wtp = ∫ q0 ∂v/∂q ∂q = ∫ q   0 Mwtp ∂q(8)
Affordable 0.259*** 0.048

Since this study included income as an explanatory
VeryConcerned –0.041 0.043
variable, it is possible to use the value function to
OverFishing 0.311** 0.136 compute income elasticity of willingness to pay (εwtp ) as;
PartAwareness 0.225*** 0.059 y ∂W ∂(ln W)
εwtp = wtp
= (9)
∂y ∂(ln y)
WorkEnvNGO 0.144** 0.066
If εwtp is less than one the benefits of environmental
CertaintyQ 0.132*** 0.005 improvement are distributed regressively. If εwtp equals
(Constant) 0.515 0.417 one the distribution is proportional. If εwtp is greater than
one, the benefits are distributed progressively.
R2 0.5118
This study found that the income elasticity of WTP
Prob > F 0.000 to restore hilsa fish in Bangladesh is 0.133, which is
Dependent variable = LnWTP, n = 1006. Single (*), double (**), and triple (***)
less than 1 (see coefficient for LnIncome in table 4).
asterisks indicate significance at 90%, 95%, and 99%, respectively. We interpret this to mean lower income groups are
willing to pay a higher proportion of their income for

Income elasticity of
hilsa fish restoration than are better-off groups. This
clearly suggests that the low income segments of the

willingness to pay society are more reliant on hilsa fishery, and therefore
investment in hilsa fish restoration is pro-poor. However,
one must note that such investments should be carefully
It has been long debated whether sustainable
designed so that they address some of the systemic
management of ecosystem services is a luxury or a
constraints faced by low income groups and so that
necessity. Kristrom and Riera (1995) note there is a
benefits are shared equitably.
widely held notion that the environment is a luxury good
and therefore receives little attention from policymakers.

16 www.iied.org
5
IIED Working paper

Conclusion
The hilsa fish is the most important single-species Using a Kaplan Meier survival estimate, we established
fishery in Bangladesh. Its contribution to national that respondents have considered their budget
economy (export earnings) and employment constraints and were acting rationally. To further test the
opportunities is widely recognised. However, reliability of the data, we also used a follow up ‘certainty’
overfishing, habitat destruction, siltation, pollution question which was presented in a 10 point Likert scale.
and climate change pose significant threats to the A cut off point of 8 and above was used to calibrate
fishery. These threats are compounded by insufficient WTP statements. However, the estimates were not
investment to restore the fishery, making the future of statistically distinguishable from the uncalibrated values.
the fishery and the people who rely on the resource This further suggests that the survey protocol was well
uncertain. We argue that this underinvestment is chiefly executed and respondents were behaving as they would
due to the fact that the total economic value of the in a real world market scenario.
fishery is not well understood.
We also explored the distributional implications of
This study is the first attempt to estimate the non- benefits from an improved hilsa fishery by estimating
consumptive value of the hilsa fishery. While citizens income elasticity of WTP. This was found to be less
generally recognise the cultural significance of than one, suggesting lower income groups are willing to
the fishery, no monetary value has been put on pay proportionately higher for hilsa fish restoration than
this. We surveyed people’s ‘Willingness To Pay’ higher income groups will pay. This suggests that low
(WTP) for a hypothetical fisheries restoration to income segments of the society are more reliant on hilsa
estimate the fishery’s value. The study estimated fishery, and therefore investment in hilsa fish restoration
the lower and upper non-consumptive values of an is pro-poor. However, such investments should be
improved hilsa fishery in Barisal Division to be BDT carefully designed to ensure that benefits are shared
651,760,000.00 (approximately US$8.3M) and BDT equitably and that interventions address some of the
1,384,186,162.67 (approximately US$17.7M) per systemic constraints faced by low income groups.
annum respectively. Extrapolating the estimate to
We believe that investments equivalent to only a fraction
national level gives lower and upper estimates of BDT
of the estimated value of non-consumptive benefits
13,128,598,840.45 (approximately US$167.5M) and
of hilsa fishery would bring about a desired change
BDT 27,882,080,597.39 (approximately US$355.7M)
including: restoring critical habitats of the fishery,
per annum respectively for a better-managed fishery. It
effective enforcement of the fishing regulation (as
must be noted that these estimates do not include the
stipulated in the Fish Protection and Conservation Act-
fishery’s use or consumptive values.
1950), and providing incentives to local fishers to stop
destructive fishing practices.

www.iied.org 17
Hilsa’s non-consumptive value in Bangladesh

References
Ahmed SU and Gotoh K (2007) The choice of elicitation Ebert, U (2003) Environmental goods and the
methods in CVM and their impact on willingness to pay distribution of income Environmental and Resource
in environmental assessment Reports of the Faculty of Economics 25: 435–459
Engineering, Nagasaki University 37(68): 47–52
FAO (2014) The State of World Fisheries and
Amin SMN, Rahman MA, Haldar GC, Mazid MA Aquaculture Food and Agriculture Organization of the
and Milton D (2002) Population dynamics and United Nations, Rome
stock assessment of hilsa shad, Tenualosa ilisha in
Freyhof J (2014) Tenualosa ilisha IUCN Red
Bangladesh. Asian Fishery Science 15: 123–128
List of Threatened Species 2014. See: http://
Amin SMN, Rahman MA, Haldar GC, Mazid, MA and dx.doi.org/10.2305/IUCN.UK.2014-1.RLTS.
Milton, DA (2008) Catch per unit effort, exploitation T166442A1132697.en
level and production of hilsa shad in Bangladesh. Asian
FRSS (2013) Fisheries Statistical Yearbook of
Fishery Science 21: 175–187
Bangladesh 2011–2012 Fisheries Resources Survey
Blore, I (1996).How useful to decision-makers is System 29: 44 Department of Fisheries, Bangladesh
contingent valuation of the environment? Journal of
FRSS (2014) Fisheries Statistical Yearbook of
Public Administration and Development 16: 215–232
Bangladesh 2012–2013 Fisheries Resources Survey
Blomquist, Glenn C.C., Blumenschein, Karenand System 30: 52 Department of Fisheries, Bangladesh
Johannesson, Magnus (2008). Eliciting Willingness
Hanley N, Schlapfer F and Spurgeon J (2003).
to Pay without Bias Using Follow-Up Certainty
Aggregating the benefits of environmental
Statements: Comparisons between Probably/Definitely
improvements: distance-decay functions for use and
and a 10-point Certainty Scale, iHEA 20076th World
non-use values. Journal of Environmental Management
Congress: Explorations in Health Economics Paper.
68: 297–304
Available at SSRN: htpp://ssrn.com/abstract=993509.
Hokby S and Soderqvist T (2003) Elasticities of
Blumenschein K, Glenn C, Blomquist M, Johannesson
demand and willingness to pay for environmental
N, Horn, and Freeman P (2008) Eliciting willingness
services in Sweden Environmental and Resource
to pay without bias: evidence from a field experiment
Economics 26: 361–383
The Economic Journal 118(525): 114–137
Johansson PO (1991) An introduction to modern
Cameron, TA and DD Huppert (1989). OLS versus ML
welfare economics. In Champ PA, Boyle KJ and Brown
estimation of non-market resource values with payment
TC eds., A Primer on Nonmarket Valuation Dordrecht:
card interval data. Journal of Environmental Economics
Kluwer Academic Publishers
and Management, 17, 230–246.
Kristrom B and Riera P (1996) Is the income elasticity
Carson RT, and Michell RC (1993) The value of clean
of willigness to pay less than one? Environmental
water: the public’s willingness to pay for boatable,
Resource Economics 7: 45–55
fishable, and swimmable quality water Water Resource
Research 29: 2445–2454 Lafolley, Dd’A and Grimsditch G. (eds) (2009) The
Management of Natural Coastal Carbon Sinks IUCN,
Champ PA, Boyle KJ and Brown TC eds. (2003). A
Gland, Switzerland. 53 pp
Primer on Nonmarket Valuation Dordrecht: Kluwer
Academic Publishers List JA and Gallet C (2001) What experimental protocol
influence disparities between actual and hypothetical
Champ PA, and Bishop RC (2004) Hypothetical bias:
stated values? Environmental and Resource Economics
the mitigating effects of certainty questions and cheap
20: 241–254
talk Selected paper prepared for presentation at the
American Agricultural Economics Association Annual Maler KG (1974) Environmental economics. In Champ
Meeting, Denver, Colorado PA, Boyle KJ and Brown TC, eds., A Primer on
Nonmarket Valuation. Dordrecht: Kluwer Academic
DoF (2014) Fishery Statistical Yearbook of Bangladesh
Publishers
Department of Fisheries, Dhaka, Bangladesh

18 www.iied.org
 IIED Working paper

MEA (2005) Ecosystems and human well-being: Neil HR, Cummings RG, Ganderton P, Harrison GW
biodiversity synthesis Millennium Ecosystem and McGuckin T (1994). Hypothetical surveys and real
Assessment, World Resources Institute, economic commitments Land Economics 70: 145–154
Washington, DC
Pauly D (2006) Major trends in small-scale fisheries,
Mohammed EY (2009) Measuring benefits of river with emphasis on developing countries, and some
quality improvement using contingent valuation method: implications for the social sciences. MAST 4: 7–22
the case of Ping River, Chiang Mai, Thailand. Journal
Rahman M and Naevdal G (2000) Population genetic
of Environmental Assessment Policy and Management
studies of hilsa shad, Tenualosa ilisha (Hamilton),
11(3): 349–367
in Bangladesh waters: evidence for the existence
Mohammed EY 2012. Contingent valuation responses of separate gene pools Fisheries Management and
and hypothetical bias: mitigation effects of certainty Ecology 7: 401–411
question, cheap talk, and pledging Environmental
Rahman M, Rahman B, Hasan S, Flura A,T and Haidar
Economics 3(3): 62–71
M (2013) Impact of eleven days fishing ban in the major
Mohammed EY and Wahab MA (2013) Direct spawning grounds of hilsa (Tenualosa ilisha) on its
economic incentives for sustainable fisheries breeding success Bangladesh Research Publications
management: the case of hilsa conservation in Journal 9: 116–122.
Bangladesh. International Institute for Environment and
Rahman MJ, Rahman MA and Bhaumik U (2012)
Development, London. http://pubs.iied.org/16527IIED.
Biology and ecology of hilsa shad, Tenualosa ilisha
html.
(Ham.) In: Hilsa: Status of Fishery and Potential for
Murphy JJ, Allen PG, Stevens TH, Weatherhead D Aquaculture. Proceedings of the regional workshop
(2005). A meta-analysis of hypothetical bias in stated held in Dhaka 16–17 September 2012. The WorldFish
preference valuation Environmental and Resource Centre, Bangladesh and South Asia Office, Dhaka
Economics 30: 313–325
WAVES (2012) Moving beyond GDP: how to factor
Murphy JJ and Stevens TH (2004) Contingent valuation, natural capital into economic decision making The
hypothetical bias, and experimental economics World Bank, Washington DC, USA.
Agricultural and Resource Economics Review 33(2)
Willis KG and Corkindale JT eds. (1995) Environmental
Northeastern Agricultural and Resource Economics
Valuation: New Perspectives. Wallingford, United
Association
Kingdom: CAB International.

Abbreviations
BDT Bangladeshi Taka
US$ United States dollar
WTP willingness to pay
GDP Gross Domestic Product

www.iied.org 19
Hilsa is Bangladesh’s most important single-species fishery: for cultural
identity, earnings and employment. However, overfishing, habitat
destruction, siltation, pollution and climate change have driven catches
down, and management policies have not adequately intervened —
probably because the fishery’s total economic value is under-appreciated.
This study is the first to estimate the non-consumptive (non-use) value
of a well-managed hilsa fishery. It used the contingent valuation method
and asked 1006 fishing and non-fishing households how much they
would be ‘Willing To Pay’ (WTP) for an effectively-managed fishery.
In Barisal Division, an improved fishery could be worth BDT 651.8M
– 1,384.2M a year (approximately US$8.3M – 17.7M). Nationally, a
better-managed fishery could be worth BDT 13,128.6M – 27,882.1M per
year (US$167.5M – US$355.7M). Poorer people are willing to pay the
highest proportion of their income, suggesting fishery restoration would
be pro-poor. However, any interventions must share benefits equitably
and address the systemic constraints facing low income groups.

IIED is a policy and action research

organisation. We promote sustainable
development to improve livelihoods
and protect the environments on which
these livelihoods are built. We specialise
in linking local priorities to global
challenges. IIED is based in London and
works in Africa, Asia, Latin America, the
Middle East and the Pacific, with some
of the world’s most vulnerable people.
We work with them to strengthen their
voice in the decision-making arenas that
affect them — from village councils to
international conventions.

International Institute for Environment and Development

80-86 Gray’s Inn Road, London WC1X 8NH, UK
Tel: +44 (0)20 3463 7399
Fax: +44 (0)20 3514 9055
email: [email protected] Knowledge
www.iied.org Products
Funded by: