Aquaculture xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Aquaculture
journal homepage: www.elsevier.com/locate/aquaculture

Production economics of striped catfish (Pangasianodon hypophthalmus,

Sauvage, 1878) farming under polyculture system in Bangladesh
⁎
Hazrat Alia, , Muhammad Meezanur Rahmana, Khondker Murshed-e-Jahanb,
Goutam Chandra Dhara
a
WorldFish, Bangladesh and South Asia Office, Bangladesh
b
International Maize and Wheat Improvement Center (CIMMYT), Bangladesh Country Office, Bangladesh

A R T I C L E I N F O A B S T R A C T

Keywords: The present study assesses the production costs, profitability and the importance of inputs (seed, feed, fertilizer and
Pangasianodon hypophthalmus labor) in striped catfish (Pangasianodon hypophthalmus) production under polyculture system in Bangladesh. A
Production cost survey was conducted by structured questionnaire during November 2011–June 2012 that included 201 pangasius
Profitability farmers of four locations, namely Bogra, Jessore, Mymensingh and Narshingdi districts. Based on the aquaculture
Return to scale
operation area and management characteristics, these farms were categorized into small, medium and large scale
Allocative efficiency
Cobb-Douglas production function
farms. The average annual production costs, gross income, net income and benefit cost ratios were higher
(P ≥ 0.05) in medium farms, followed by small and large farms and these parameters varied significantly
(P ≤ 0.05) between farm locations. The Cobb-Douglas production function model was applied to evaluate several
independent variables (inputs) related with fish production. The results showed that feeding and stocking density
were the most significant factors influencing fish production. The analysis also showed an increasing return to
scale, implying an increase in inputs will more than proportionately increase production. The estimated marginal
physical productivity of the inputs suggests that stocking density, feeding and labor should be increased in order to
increase farm profitability. However, some inputs were inefficiently used by different farm categories. Based on the
findings of this study, it is recommended that farmers should consider more attention on the use of these inputs
more efficiently in order to facilitate a sustainable increase in fish production.

1. Introduction The high capital requirements for commercial production have

Striped catfish (Pangasianodon hypophthalmus) also known as pan-
gasius or locally as pangas or Thai pangas, is an important fast-growing
aquaculture commodity in the Asian region, particularly Vietnam,
Thailand, India, Myanmar, Indonesia and Bangladesh (Phan et al.,
2009; Ali et al., 2013). Pangasius farming is a significant component of
aquaculture production in Bangladesh with a total production of
494,357 t in the fiscal year 2015–2016, accounting for 29% of the total
farmed fish supply in the country (DoF, 2017a). Commercial pangasius
production was first started in 1993 (Ali and Haque, 2011) and it has
rapidly developed into an economically significant activity (Ali et al.,
2013). The species possesses several characteristics that have helped
drive this growth: good survival rates, fast growth, large size and a high
market demand (Sarker, 2000). Additionally, it can be stocked at a
much higher density in ponds compared with other aquaculture species
(Alam, 2011). As a result, pangasius farming has rapidly established
itself as an important aquaculture industry in Bangladesh.
usually been thought to limit the potential for direct involvement by the
resource-poor (Belton et al., 2014a). However, the development of in-
tensive pangasius farming has benefited landless laborers by creating
employment opportunities, particularly involving the provision of
supporting goods and services in associated value chains (Ali, 2009;
Belton and Azad, 2012; Ali et al., 2013). Pangasius farming generated a
higher intensity of on-farm wage employment, resulting in higher
wages and shorter working hours compared to agricultural labor
(Belton et al., 2017). It has also created off-farm employment oppor-
tunities in areas such as feed mills, inputs trading (e.g. feed and che-
micals), fish harvesting, transportation and marketing (Ali, 2009;
Belton et al., 2017). This industry provides many livelihood opportu-
nities with long backward and forward linkages for a wide range of
value chain actors (Ali, 2009; Belton and Azad, 2012).
Fish is the main animal-source food consumed, accounting for 60%
of animal protein intake (DoF, 2017b) and is the most frequently con-
sumed nutrient rich food (Toufique and Belton, 2014) in Bangladesh.

⁎
Corresponding author at: WorldFish, Bangladesh and South Asia Office, House #22/B, Road #7, Block #F, Banani, Dhaka, Bangladesh.
E-mail address: ali_03ku@yahoo.com (H. Ali).

https://doi.org/10.1016/j.aquaculture.2017.12.004
Received 15 April 2017; Received in revised form 27 November 2017; Accepted 3 December 2017
0044-8486/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Ali, H., Aquaculture (2017), https://doi.org/10.1016/j.aquaculture.2017.12.004
H. Ali et al.
Pangasius is now the most consumed fish in rural areas of Bangladesh
(Hernandez et al., 2017). It has a significant local market demand and
almost all of the production is consumed domestically, providing a year
round supply of animal protein to consumers in both rural and urban
areas (Ali et al., 2013; Belton et al., 2014b; Hernandez et al., 2017).
Pangasius aquaculture has improved access to fish among poor con-
sumers in Bangladesh by increasing supply and reducing relative prices
(Toufique and Belton, 2014; Belton et al., 2014a). Moreover, it has
proven particularly popular among poor consumers in urban areas due
to its low market value, arguably making it one of the most important
cultured species for food security in Bangladesh (Belton et al., 2011).
There appears to be good potential for further development of this
industry in Bangladesh. Considering the potential of pangasius as its
vital role in food security for poorer consumers, its production needs to
be increased through the targeted application of inputs that are readily
available to farmers. Whilst several studies have been conducted in
Bangladesh related to pangasius productivity (Ali et al., 2013; Jahan
et al., 2015) only one study to date has focused on an economic analysis
of different pangasius production systems (Ahmed et al., 2010).
2
Aquaculture xxx (xxxx) xxx–xxx
Fig. 1. Map showing the geographical distribution of the
surveyed farms.
Furthermore, the data on which the study of Ahmed et al. (2010) is
based was collected in only one sub-district in 2006, since which pro-
duction systems in Bangladesh have intensified further (Ali et al.,
2013). It is therefore vital that the economics of pangasius production
are reassessed in order to facilitate informed farmer decision making.
The objectives of this study were to assess production costs, profitability
and the role of inputs in pangasius production.
2. Materials and methods
2.1. Study area
This study was performed as part of the Cereal Systems Initiative for
South Asia (CSISA) project funded by the United States Agency for
International Development (USAID). The survey was conducted in Bogra,
Jessore, Mymensingh and Narshingdi districts, which constitute the major
pangasius production areas of Bangladesh (Fig. 1). The geographical dis-
tribution of Pangasius farms approximately corresponds with the major
locations of pangasius hatcheries listed by Ali et al. (2013).
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

Table 1 2.3. Data analyses

Characteristics and classification of small, medium and large scale farms.
Characteristics Farm scale
Small Medium
Total farm area (ha) < 0.60 0.60–1.2
Full time labor (non- No No or yes
family)
Management Household or Household or
extended family salaried manager
2.2. Data collection procedure
A stratified random sampling strategy was adopted to select pan-
gasius farms for the survey. The survey procedure followed was in line
with that described earlier by Jahan et al. (2015). In brief, the villages
(lowest level administrative unit) with high concentrations of house-
holds practicing pangasius aquaculture were identified through key
informant interviews and subsequent follow-up visits. A census of
households practicing aquaculture was conducted in each of the se-
lected villages, and farm households were selected at random from this
list for interview. A total of 201 pangasius farms (Mymensingh-98,
Jessore-59, Narshingdi-27 and Bogra-17) were selected from 914 farms.
These farms were categorized into three size categories; small, medium
and large scale depending on the total farm area, employed labor and
farm management procedures (Table 1). A structured questionnaire was
prepared and piloted in the study area. The questionnaire captured
information on farm management practices, including production in-
puts, costs and profitability. The survey was conducted between No-
vember 2011 and June 2012. It was very difficult to obtain reliable
financial data (i.e., costs and returns) because most farmers did not
keep any financial records. Therefore, extra attention was paid and
great care had to be taken in compiling financial information using
different data collection methods (both qualitative and quantitative).
Twenty semi-structured interviews were conducted to obtain qualita-
tive information from key informants possessing particularly detailed
knowledge of certain aspects of the industry. The key informants in-
cluded district and sub-district fisheries officers, a pangasius hatchery
owner, fingerling traders, feed dealers, commercial pelleted feed com-
pany managers and food fish traders. In addition, 10 focus group dis-
cussion (FGD) meetings were organized at village level with eight to ten
farmers/farm operators per meeting to gather information on farming
practices, production costs and returns and to crosscheck the informa-
tion obtained in different data collection methods.
Questionnaire responses were coded and entered into an MS Access
(Microsoft Corporation, Redmond, WA, USA) database before exporting
Large the data into MS Excel (Microsoft Corporation) and SPSS 16.0 (SPSS,
Chicago, IL, USA). SPSS was used for data analysis and production of
> 1.2
descriptive statistics. Results from the data analyses, in combination
Yes
with qualitative information were used to describe pangasius produc-
Salaried tion in polyculture systems. An economic analysis was also conducted
manager to determine net returns of pangasius production in polyculture system
(Ali et al., 2016a). The economic analysis was based on fish sales prices
received by the farmers and the prices paid by farmers for inputs,
converting these to USD post hoc (USD 1 = BDT 80.86). The value of
fish consumed and gifted by the farmers was also included in the
analysis by imputed average sales prices. Table 2 shows the equations
used to calculate costs and returns.
2.4. Production function model
The Cobb-Douglas production function model was applied to assess
the production economics of pangasius in polyculture system. Several
studies have used this model to analyse fish culture systems. For ex-
ample, Phuong et al. (2007), Boonchuwong et al. (2007) and Ahmed
et al. (2010) for pangasius farming in Vietnam, Thailand and Bangla-
desh, respectively; and Asamoah et al. (2012) for tilapia farming in
Ghana. Moreover, Ali et al. (2016a) and Karim et al. (2017) used the
Cobb-Douglas production function model to analyse the carp-mola
polyculture system in Bangladesh.
Four inputs or explanatory variables (i.e., seed, feed, fertilizer and
labor) were assumed to explain pangasius production in polyculture
system by the Cobb-Douglas production function model. It was hy-
pothesized that changes in the amounts of these inputs used would
affect production levels and farm income. Regression analysis (ordinary
least squares method) was used to assess the effect of these inputs. The
Cobb-Douglas production function model can be expressed in the fol-
lowing general form:
Log Y = log a + Σbi log(Xi) + log Ui (1)
However, the empirical Cobb-Douglas production function models
used for this study, one with feed types, culture system and location
dummies (3) and the other without (2) are expressed as follows:
Log Yi = log a + b1 log X1i + b2 log X2i + b3 log X3i + b4 log X 4i + b5 log X5i
+ log Ui (2)

Table 2
The equations to calculate production costs and returns of pangasius farming in polyculture system (USD ha− 1).

Total cost = Variable costs + fixed costs

Variable costs
Fixed costs
Depreciation costs
Production cost per kg fish
Gross revenue
Gross revenue per kg fish
Net margin
Net margin per kg fish
Income above variable cost
Benefit-cost ratio (BCR)
Rate of farm income
Break-even production
= Cost of fingerling, feed, fertilizer, chemical inputs, labor (familya and hired), harvesting & marketing and miscellaneous (electricity, fuel, water
supply etc.)
= Cost of depreciation + land lease + others (land tax, interest on loan etc.)
=(Purchase price-salvage value)/economic life
= Total costs/total production
= Total production × actual sold value
= Gross revenue/total production
= Gross revenue − total cost
= Net margin/total production
= Gross revenue − variable costs
= Gross revenue/total cost
= Net margin/gross revenue × 100%
= Total costs/actual sold value

a
Family labor was treated as an opportunity cost for the farmer. Quantification of family labor was challenging because farmers could not accurately estimate the amounts of family
labor used for different purposes.

3
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

Table 3 3. Results and discussion

Demographic information of studied farmers by location (percentages in the parenthesis).
3.1. Demographic information
Variables Farm locations

Mymensingh Jessore Narshingdi Bogra Overall All of the 201 farmers surveyed were male (Table 3) implying that
n = 98 n = 59 n = 27 n = 17 n = 201 pangasius farming is principally a male-oriented activity. Traditionally
men are deemed to be the heads of the household unit in Bangladesh
Gender
Male 98 (100) 59 (100) 27 (100) 17 201 and farms owned and operated by a family are likely to be in the name
(100) (100) of the head of the family. The mean age was 41 years (ranging from 18
Female to 75) and the majority (64%) were below 45 years of age (Table 3).
Age (years) Thirteen percent of farmers had no formal educational qualification.
≥ 30 32 (33) 7 (12) 4 (15) 5 (29) 48 (24)
More than 22% of respondents had completed only primary education,
31–45 34 (35) 24 (41) 14 (52) 9 (53) 81 (40)
46–60 23 (23) 22 (37) 7 (26) 3 (18) 55 (27) 40% secondary, 14% higher secondary certificate and 10% a bachelor's
≤ 61 9 (9.2) 6 (10) 2 (7.4) 0 (0.00) 17 (8.5) degree or higher. There was no significant difference in mean schooling
Education years between farmers in different locations and farm categories. These
Illiterate 13 (13) 11 (19) 2 (7.4) 0 (0.00) 26 (13)
levels of educational attainment were considerably higher than the
Primary 21 (21) 12 (20) 7 (26) 4 (24) 44 (22)
Secondary 30 (31) 29 (49) 14 (52) 8 (47) 81 (40) national average (BBS, 2013), even higher than other aquaculture
school farmers in Bangladesh (Jahan et al., 2015; Ali et al., 2016b). This result
certificate indicates the relatively high socio-economic status of most pangasius
Higher 20 (20) 5 (8.5) 1 (3.7) 3 (18) 29 (14) farmers prior to their entry into pangasius farming. The mean years of
secondary
aquaculture experience were 7.0 ± 4.0, 8.4 ± 4.1, and 9.8 ± 4.6 in
certificate
Bachelor's 14 (14) 2 (3.4) 3 (11) 2 (12) 21 (10) small, medium and large farm respondents respectively (Table 3). The
degree or large farm operators had significantly longer (P ≤ 0.05) experience
higher than small farm operators. Aquaculture experience was counted as the
Aquaculture experience (years)
number of years that an operator was involved in any form of aqua-
≥5 23 (23) 36 (61) 10 (37) 2 (12) 71 (35)
6–10 45 (46) 16(27) 9 (33) 6 (35) 76(38)
culture production (not exclusively pangasius). Only 9.5% of farmers
11–15 25 (26) 6 (10) 6 (22) 5 (29) 42(21) reportedly received formal training on pangasius aquaculture from
≤ 16 5 (5.1) 1 (1.7) 2 (7.4) 4 (24) 12 (6.0) Department of Fisheries (DoF) officials and the remaining gained
Received aquaculture training knowledge through experience and from neighboring farmers.
Yes 11 (11) 3 (5.1) 2 (7.4) 3 (18) 19 (9.5)
No 87 (89) 56 (95) 25 (93) 14 (82) 182 (91)
3.2. Farming practices

3.2.1. Farm general information

Log Yi = log a + b1 log X1i + b2 log X2i + b3 log X3i + b4 log X 4i + b5 log X5i The average size of operated landholding (i.e. including all land
+ d1D1 + d2 D2 + d3D3 + d 4 D4 + d5D5 + d6 D6 + d7 D7 + d8D8 + d9 D9 owned, leased in, and shared, for agricultural and other uses) by
farmers was 1.0 ± 1.1 ha (Table 4), which was larger than the national
+ log Ui
average operated area of farm holdings, 0.60 ha (BBS, 2010). These
(3) results indicated that pangasius farmers possess higher than average
resources. Similar observations were made by Belton and Azad (2012)
where Y is the total fish production (kg ha− 1) including pangasius, carp
and Belton et al. (2014c). The average farm size and water surface area
and tilapia; a is the constant parameter in the equation, mathematically
was 1.0 ± 0.88 ha and 0.79 ± 0.72 ha, respectively (Table 4). The
interpreted as the intercept; X1 the stocking density (kg ha− 1) including
number of ponds per farm ranged from 1 to 24 and it was significantly
pangasius, carp and tilapia fingerlings; X2 the supplementary feed
different between farm categories. The mean pond size was 0.24 ha,
(kg ha− 1); X3 the fertilizer (kg ha− 1); X4 the labor (man-day ha− 1); D1
ranging from 0.19 ha to 0.39 ha between farm locations. The mean
(1: if farm is used commercially produced pelleted floating feed; 0:
pond size at Bogra was significantly (P ≤ 0.05) larger than at My-
otherwise); D2 (1: if farm is used commercially produced pelleted
mensingh and Narshingdi. The individual pond size (i.e. including
sinking feed; 0: otherwise); D3 (1: if farm is used farm-made pelleted
dykes) and water surface area were also significantly different
feed; 0: otherwise); D4 (1: if farm is used farm-made mash feed; 0:
(P ≤ 0.05) between farm categories. Pond water depth varied from
otherwise); D5 (1: if farm is stocked carp; 0: otherwise); D6 (1: if farm is
0.91 to 2.7 m (mean 1.5 ± 0.29) across all farms. For approximately
stocked tilapia; 0: otherwise); D7 (1: if farm is located in Mymensingh
90% of farms, pond water depth ranged from 1.1–1.8 m during the
district; 0: otherwise); D8 (1: if farm is located in Jessore district; 0:
culture season (Table 4 and Table 5). A similar range of pond water
otherwise); D9 (1: if farm is located in Bogra district; 0: otherwise);
depths has previously been reported by Ali et al. (2013) for pangasius
b1–b5 is the coefficient of the relevant variables; Ui the indexes of ob-
aquaculture in Bangladesh. A positive correlation was found between
servations (1, 2, 3……………n).
pond depth and yield (R2 = 0.08). A similar trend was also observed for
The four explanatory variables specified in the model are within the
pangasius production in Vietnam (Phan et al., 2009).
control of farmers. Other independent variables like water quality,
water exchange and soil condition that might have affected production
3.2.2. Water supply and pond preparation
were excluded from the model on the basis of some preliminary as-
All surveyed farmers' pumped groundwater by shallow tube-wells or
sessments: 10 independent variables were initially included in the
submersible pumps for water supply. However, during monsoon season
analysis but based on preliminary estimations, further analysis was
(June–September) farmers were mainly reliant on rainwater. Fifty five
restricted to those explanatory variables that were likely to have the
percent of the farms exchanged pond water when water quality dete-
greatest impact on production. Moreover, farm management factors
riorated or following a disease outbreak. A wide number of chemical
were not included in the model because specification and measurement
and biological compounds were applied to treat water quality, pollution
of these are complicated by the prevalence of farm operators who are
and diseases (Table 5). All farmers dried their ponds after the final
both the main laborer and the manager.
harvest of fish, with pumps being used to drain the ponds. Almost all
farms treated pond bottoms with lime prior to filling and stocking.

4
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

Table 4
Characteristics of pangasius polyculture farming system in the study area by farm location.

Characteristics Location (mean ± standard deviation)

Mymensingh Jessore Narshingdi Bogra

Farm size including agriculture land (ha) 1.0 ± 1.3a 0.92 ± 0.80a 1.0 ± 1.1a 1.4 ± 1.5a
Total fish farm size (ha) 1.0 ± 0.92a 0.92 ± 0.70a 1.0 ± 0.89a 1.4 ± 1.2a
Total farm water surface area (ha) 0.78 ± 0.72a 0.67 ± 0.52a 0.82 ± 0.74a 1.2 ± 1.0a
No. of ponds farm− 1 4.5 ± 4.0a 2.6 ± 1.6b 5.1 ± 3.1a 5.2 ± 4.1a
Sample pond size (ha) 0.19 ± 0.10ac 0.30 ± 0.22bcd 0.21 ± 0.11abc 0.39 ± 0.29bd
Sample pond water surface area (ha) 0.16 ± 0.09ac 0.24 ± 0.19bcd 0.18 ± 0.09abc 0.30 ± 0.21bd
Pond water depth (m) 1.5 ± 0.34a 1.6 ± 0.17a 1.6 ± 0.34a 1.5 ± 0.14a
Stocking density (kg ha− 1)
Pangasius 1176 ± 751a 1652 ± 768b 772 ± 410a 846 ± 590a
Carp 113 ± 98a 560 ± 246b 171 ± 174a 613 ± 489b
Tilapia 18 ± 29a 37 ± 75a 27 ± 42a 102 ± 116b
Feed (t ha− 1) 78 ± 23a 39 ± 13b 77 ± 28a 29 ± 14b
Crop duration (days) 302 ± 30ac 316 ± 26bc 308 ± 29abc 267 ± 61d
Production (t ha− 1)
Pangasius 41 ± 10a 21 ± 6.9b 41 ± 13a 18 ± 10b
Carp 2.8 ± 1.7abc 2.7 ± 1.3abc 2.2 ± 1.6ab 3.5 ± 1.5ac
Tilapia 1.1 ± 1.2a 1.4 ± 1.2a 1.4 ± 1.3a 0.74 ± 0.68a

Mean values followed by different superscript letters indicate significant difference (P ≤ 0.05).

Fertilizers were applied to improve natural productivity (e.g. phyto-
plankton) that enhances the production of additional crops (carp and
tilapia). This is consistent with earlier reports (Ali et al., 2013) on
pangasius farming in Bangladesh.
3.2.3. Seed stocking
Fifty nine percent of farmers (across all farm scales) purchased fry
from hatcheries and reared them in their own nursery ponds (Table 5).
This tendency was found to be higher in large (90%) and medium
(80%) farms than small (27%) farms. The remaining farmers (41%)
who had no nursery ponds purchased fingerlings from nursery farmers.
Furthermore, 13% of farmers both reared their own fish from fry and
purchased fingerlings from nursery farmers. Farmers started nursing fry
in early March and continued until the end of April, followed by

Table 5
Characteristics and management by farm size category.

Characteristics Farm size

Small (n = 87) Medium (n = 66) Large (n = 48) Overall (n = 201)

a b c
Pond size (ha) 0.16 ± 0.08 0.24 ± 0.12 0.38 ± 0.27 0.24 ± 0.18
Pond water surface area (ha) 0.13 ± 0.07a 0.20 ± 0.10b 0.31 ± 0.21c 0.20 ± 0.15
No. of ponds farm− 1 2.3 ± 1.1a 4.0 ± 1.6b 7.5 ± 5.2c 4.1 ± 3.5
Average water depth (m) 1.4 ± 0.27ab 1.5 ± 0.27abc 1.6 ± 0.33bc 1.5 ± 0.29
% of farms included nursery pond for pangasius 27 80 90 59
Stocking density of pangasius (kg ha− 1) 1297 ± 746a 1242 ± 769a 1107 ± 805a 1234 ± 768
% of farms stock carp 90 80 88 86
Stocking density of carps (kg ha− 1) 314 ± 312a 315 ± 371a 315 ± 240a 315 ± 315
% of farms stock tilapia 54 42 65 53
Stocking density of tilapia (kg ha− 1) 34 ± 67a 25 ± 33a 36 ± 72a 32 ± 61
Farm exchange pond water (% of farms) 56 65 38 55
Frequency of water exchange (times per cycle) 2.3 ± 1.1 2.3 ± 1.4 1.5 ± 0.52 2.1 ± 1.0
Water exchange % (% pond water volume) 55 52 23 51
Application of organic fertilizer (% of farms) 29 17 19 22
Dosage of organic fertilizer (kg ha− 1) 1681 ± 972a 1545 ± 1085a 1691 ± 1077a 1650 ± 999
Application of inorganic fertilizer (% of farms) 38 32 52 39
Dosage of inorganic fertilizers (kg ha− 1)
Urea 192 ± 105a 158 ± 89a 159 ± 137a 172 ± 112
Triple super phosphate 128 ± 70a 105 ± 59a 106 ± 91a 115 ± 75
% of farms applied various health products
Water and soil treatment compounds 99 100 98 99
Disinfectant 39 44 46 42
Antibiotics 9.2 12 17 12
Pesticide 18 17 29 20
Feed additives 16 29 25 22
Probiotics 6.9 6.1 6.3 6.5
Amount of feed (t ha− 1) 61 ± 26a 66 ± 26a 59 ± 37a 62 ± 29
FCR 1.70 ± 0.31a 1.73 ± 0.32a 1.72 ± 0.36a 1.71 ± 0.32
Labor (man-day ha− 1) 597 ± 252a 542 ± 231a 411 ± 262b 535 ± 257
Crop duration (days) 307 ± 30a 304 ± 33a 298 ± 44a 304 ± 35
Pangasius production (t ha− 1) 32 ± 12abc 37 ± 13ab 30 ± 16ac 33 ± 14
Carp production (t ha− 1) 2.7 ± 1.1a 2.7 ± 2.0a 2.9 ± 1.7a 2.7 ± 1.6a
Tilapia production (t ha− 1) 1.2 ± 0.87a 1.4 ± 1.5a 1.0 ± 1.3a 1.2 ± 1.2a

Mean values followed by different superscript letters indicate significant difference (P ≤ 0.05).

5
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

grading and transfer to grow-out ponds in several stockings. In com-
parison with Ali et al. (2013), farmers in the present study stocked
larger size of pangasius fingerlings in their grow-out ponds and the
mean size was 41 ± 30 g in weight and 15 ± 5.9 cm in length. The
stocking of larger size pangasius fingerlings is associated with high
productivity (Alam, 2011).
The main grow-out period is from March to December. Pangasius
production includes polyculture with Indian major carps (Rohu, Labeo
rohita; Catla, Catla catla; Mrigal, Cirrhinus cirrhosus), Chinese carps
(Silver carp, Hypophthalmichthys molitrix; Grass carp, Ctenopharyngodon
idella; Common carp, Cyprinus carpio) and Tilapia (Oreochromis nilo-
ticus). At grow out, stocking density of pangasius fingerlings ranged
from 69 to 4117 kg ha− 1 (Table 5). The mean stocking density was
significantly higher (P ≤ 0.05) in Jessore compared with Mymensingh,
Narshingdi and Bogra, but it did not differ significantly (P ≥ 0.05)
between farm categories. The mean stocking density of carp was sig-
nificantly higher (P ≤ 0.05) in Bogra (613 ± 489) and Jessore
(560 ± 246) than Narshingdi (171 ± 174) and Mymensingh
(113 ± 98) (Table 5). Fifty three percent of farms stocked tilapia at a
density between 1.2 and 411 kg ha− 1 and no significant difference was
observed in stocking density between farm categories.
3.2.4. Feeds and feed management
The pangasius polyculture system in Bangladesh is intensive
(Edwards, 2009; Jahan et al., 2015), with all farms applying pelleted
feeds, of which 69% used commercially produced pelleted feeds (the
remainder used farm-made pelleted feeds). The details of typical farm-
made feeds can be found in Ali et al. (2013). Of those that used com-
mercially produced feeds, 62% used sinking pelleted feeds, 25% used
floating pelleted feeds and 18% used a combination of both sinking and
floating pelleted feeds. Moreover, 21% of farms also used farm-made
mash feeds (non-pelleted) including rice bran/polish, rice products,
wheat products, mustard oilcake, and maize meal. Sixteen percent of
small scale farmers applied both commercial and farm-made pelleted
feeds, usually opting for commercial feeds during rainy and cloudy
weather due to the increased difficulty of preparing and drying farm-
made pelleted feeds. Many national and international feed companies
are established in Bangladesh and > 100 mills produce commercial
pelleted fish feeds (Mamun-Ur-Rashid et al., 2013). These companies
are linked with a range of dealers for marketing their feeds. Farmers
were able to obtain feed from local market at a competitive price. Most
of the farmers purchased feed by cash payment; however 35% of
farmers reported receiving feeds from dealers using in-kind loan as a
form of credit.
The amount of feed used ranged between 9.1 and 148 t ha− 1
(Table 5), with no significant difference (P ≥ 0.05) observed between
completed within a single day, however for a large pond it could take
two to three days, netting several times. Fish harvesters work year
round, with the peak harvest season between August–December. Seine
nets were used and fish were then kept alive in plastic containers with
water until they were sold.
3.2.6. Productivity
Pangasius yields ranged from 9.7 to 69 t ha− 1 (Table 4) with sig-
nificantly higher (P ≤ 0.05) yields in Mymensingh and Narshingdi
compared with Bogra and Jessore. This may be due to Mymensingh and
Narshingdi having medium to high quality agricultural land with sui-
table soil pH (Shamsuddin et al., 2007); easy access to industrially
manufactured pelleted feed due to better transport linkages to the ca-
pital city Dhaka (Ahmed et al., 2010); and the presence of fisheries
research institutions (Bangladesh Agricultural University and Bangla-
desh Fisheries Research Institute) (Ali et al., 2013). Moreover, farmers
in this area received training on pangasius farming from Mymensingh
Aquaculture Extension Project which resulted in a dramatic increase in
pangasius production (Ahmed, 2007). The mean annual production was
highest in medium farms (37 ± 13 t ha− 1) followed by small
(32 ± 12 t ha− 1) and large (30 ± 16 t ha− 1) farms (Table 5). The
pangasius yields reported in this study are consistent with earlier re-
ports (Ali et al., 2013). Carp yields ranged from 0.17–9.6 t ha− 1 and
were higher in large farms than in small and medium farms. Carp
productivity was higher (P ≥ 0.05) in Bogra than in other districts. This
may be due to the practice in Bogra of stocking carp fingerlings when
they are larger and at higher densities when compared with other
districts (Jahan et al., 2015). Lastly, tilapia yields ranged from
0.11–5.9 t ha− 1 (Table 5), with a similar range being found in previous
studies on pangasius aquaculture (Ali et al., 2013).
3.3. Production costs structure
The annual production costs of pangasius farms under polyculture
system ranged from USD 8127–50,336 ha− 1 (Table 6). The mean an-
nual cost per hectare was the highest in the medium farms, followed by
small and then large farms. The higher costs in medium farms were due
to higher variable costs including fish fingerlings, feed, fertilizer, che-
micals, labor, harvesting and marketing. The annual variable costs were
USD 25,975 ha− 1, USD 24,334 ha− 1, and USD 22,707 ha− 1 in
Table 6
Average production costs (USD ha− 1) for pangasius production in polyculture system by
farm categories (percentages in the parenthesis).
Cost items Farm category

farm categories. However, it was found to be significantly higher Small Medium Large Average
(P ≤ 0.05) in Mymensingh and Narshindi compared to Bogra and Jes-
sore. All farmers applied feed twice per day throughout the production Variable cost (VC)
cycle. The feed conversion ratios (FCRs) were 1.70 ± 0.31, Fish fingerlings 2775 (11) 3128 (12) 2853 (12) 2910 (11)
Feed 18,564 20,233 (75) 17,805 18,931 (75)
1.73 ± 0.32 and 1.72 ± 0.36 in small, medium and large farm cate- (74) (75)
gories, respectively. Similar FCRs were also reported by Ahmed et al. Fertilizer 49 (0.19) 31 (0.11) 50 (0.21) 44 (0.17)
(2010) for pangasius aquaculture in Bangladesh. Chemical 161 (0.64) 156 (0.58) 151 (0.64) 157 (0.62)
Labor (family and hired) 2361 (9.4) 2149 (8.0) 1582 (6.7) 2105 (8.3)
Harvesting and 221 (0.88) 188 (0.70) 232 (1.0) 213 (0.84)
3.2.5. Labor
marketing
The labor input (man-day ha− 1) was significantly higher (P ≤ 0.05) Miscellaneous 202 (0.80) 92 (0.34) 35 (0.15) 126 (0.50)
in small (597 ± 252) and medium (542 ± 231) farms compared to Sub-total 24,334 25,976 (96) 22,707 24,485 (96)
large farms (411 ± 262) (Table 5). Labor input was disaggregated for (97) (96)
men and women. Men were the main source of labor which provided Fixed cost (FC)
98% of labor requirement in pangasius farming. It was also observed Depreciation 483 (1.9) 512 (1.9) 387 (1.6) 470 (1.8)
that wages for men (USD 5.2/day) were 15% higher compared to Land leased 220 (0.87) 297 (1.1) 441 (1.9) 298 (1.2)
Other (land tax, interest 162 (0.64) 176 (0.65) 126 (0.53) 158 (0.62)
women (USD 4.4/day). Labor classifications included full time, part-
on loan etc.)
time and family members. Part-time workers were generally hired for Sub-total 866 (3.4) 985 (3.7) 954 (4.0) 926 (3.6)
re-excavating, cleaning and repairing ponds, as well as feeding. Fish
Total cost (TC) 25,199 26,961 23,661 25,410
harvesting is a specialized job in the studied areas for which farmers = VC + FC (100) (100) (100) (100)
depend on commercial fish harvest teams. Harvest was generally

6
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

medium, small and large farms, respectively. Variable costs accounted
for about 96% of total costs across the three farm categories. This is in
line with previous studies (Ara, 2005; Ahmed et al., 2010) which found
variable costs of 91–95% of the total costs. Feed costs were the most
important variable cost in the pangasius polyculture system, accounting
for 74%, 75% and 75% of total costs in small, medium and large farms,
respectively and with no significant difference (P ≥ 0.05) between
farm categories. This is in line with Ara (2005) and Alam (2011), who
found similar feed cost structures in pangasius production: 75% and
72%, respectively. However, a higher level of feed costs were reported
by Phuong et al. (2007), accounting for 93% of total costs for more
intensive pangasius production in Vietnam.
The annual fingerling cost in small, medium and large farms was
estimated to be USD 2775 ha− 1, USD 3128 ha− 1 and USD 2853 ha− 1,
respectively. This cost was much lower (at 11% of total costs) than feed
cost, despite fingerlings being a key input in pangasius polyculture
system. A significantly higher (P ≤ 0.05) labor cost was found in small
(USD 2361 ha− 1) and medium (USD 2149 ha− 1) farms than large (USD
1582 ha− 1) farms (Table 6). Whilst labor costs account for the third
highest variable cost, they constitutes a smaller proportion of total
production costs at 9.4% in small farms, 8.0% in medium farms and
6.7% in large farms. This is in line with the findings of Boonchuwong
et al. (2007) and Alam (2011) for pangasius farming in Thailand and
Bangladesh, respectively. Lastly, farmers in all three size categories
reported similar chemical, fertilizer, harvesting and marketing costs.
The mean annual fixed cost was USD 926 ha− 1, varying from USD
866 ha− 1 for small farms to USD 985 ha− 1 for medium and USD
954 ha− 1 for large farms (Table 6). The fixed costs accounted for 4% of
total costs on large farms, compared with 3.7% for medium farms and
3.4% for small farms. A similar result was also reported by Ahmed et al.
(2010), who found that fixed costs accounted for 5% of total costs for
pangasius production in Bangladesh.
small (USD 9426 ± 6885 ha− 1) and large (USD 9269 ± 7100 ha− 1)
farms. A significantly higher (P ≤ 0.05) net margin was observed in
Mymensingh and Narshingdi than in Jessore and Bogra districts. All
farmers reported decreasing returns over the last few years due to in-
creasing costs of production and stagnant market prices for fish. The
annual income was higher than total variable cost in all farm categories.
It was higher (P ≥ 0.05) in medium farms (USD 12,256 ± 5495 ha− 1)
than small (USD 10,291 ± 6966 ha− 1) and large (USD
10,223 ± 7148 ha− 1) farms. Therefore this suggests that farmers can
continue pangasius production in polyculture system as the returns to
variable costs were positive.
The benefit-cost ratio (BCR) was higher (P ≥ 0.05) in medium
farms followed by large and small farms (Table 7). The BCR was sig-
nificantly different (P ≤ 0.05) between farm locations. This is con-
sistent with earlier studies (Alam and Rahman, 2005; Alam, 2011),
suggesting pangasius polyculture system has been a sustainable and
profitable industry for many years. Accordingly, a higher rate of income
was observed in medium farms (29%) compared with large (27%) and
small (25%) farms. This is in accordance with earlier findings on in-
tensive pangasius production in Bangladesh by Ahmed et al. (2010).
The break-even yields were at 26 ± 9, 28 ± 9 and 24 ± 12 t ha− 1
for small, medium and large farms, respectively. The average yields of
small, medium and large farms exceeded break-even production by
25%, 29% and 27%, respectively. The level of economic efficiency in
the present study is comparable to carp, tilapia, shrimp and prawn
farming systems (Jahan et al., 2015), integrated rice-fish farming
(Ahmed et al., 2011) and the main agricultural crops, rice (Afroz and
Islam, 2012) in Bangladesh. This suggests that fish yields are sufficient
to consider it as a business with the potential to generate profit.
3.5. Production function model

The estimated values of the coefficients and related statistics of the

3.4. Profitability
The mean annual gross revenue ranged between USD
32,930–38,232 ha− 1 and was not significantly different (P ≥ 0.05)
between farm categories (Table 7). However, farms located at My-
mensingh and Narshingdi had significantly higher (P ≤ 0.05) annual
gross revenue compare to Jessore and Bogra. This higher revenue was
in accordance with significantly higher yields at Mymensingh and
Narshingdi (Table 4). Yield, operating costs and fish market price
contribute to the net margin of the farms. Net margin was higher
(P ≥ 0.05) in medium farms (USD 11,271 ± 5412 ha− 1) compared to
Cobb-Douglas (C-D) production function model are presented in
Table 8. Regression analysis showed that the coefficient of multiple
determinations (R2) for Model 1 was 0.74, implying that 74% of the
variation in fish production can be explained by the four explanatory
variables (stocking density, feeding, fertilization and labor) included in
the model. Nine dummy variables for feed types, culture systems and
location (model 2) increased R2 to 0.81. Both models were highly sig-
nificant (ANOVA; P < 0.01). These findings are similar to those of
Ahmed et al. (2010) and Ali et al. (2016a). Results of model 1 indicated
that fish production was significantly influenced by stocking density
(P < 0.01), feeding (P < 0.01) and labor (P < 0.05). The estimated

Table 7
Return and economic indicators of pangasius production in polyculture system by farm categories.

Performance indicators Farm category

Small Medium Large Overall

−1
Variable cost (VC) (USD ha ) 24,334 ± 7578 25,976 ± 8106 22,707 ± 10,951 24,485 ± 8703
Fixed cost (FC) (USD ha− 1) 866 ± 683 985 ± 741 954 ± 827 926 ± 737
Total cost (TC) = VC + FC (USD ha− 1) 25,199 ± 7594 26,961 ± 8134 23,661 ± 11,054 25,410 ± 8750
Total production (TP) (t ha− 1) 35 ± 12 40 ± 13 33 ± 17 36 ± 14
Production cost per kg fish (TC/TP) (USD) 0.74 ± 0.11 0.70 ± 0.15 0.74 ± 0.11 0.73 ± 0.13
Gross revenue from pangasius (USD ha− 1) 30,617 ± 12,281 34,302 ± 11,739 29,270 ± 16,500 31,505 ± 13,332
Gross revenue from carp (USD ha− 1) 3367 ± 2003 3360 ± 2621 3037 ± 1873 3286 ± 2191
Gross revenue from tilapia (USD ha− 1) 641 ± 907 570 ± 1128 623 ± 1012 613 ± 1005
Total gross revenue (GR) (USD ha− 1) 34,625 ± 12,143 38,232 ± 11,100 32,930 ± 16,307 35,405 ± 13,050
Market price per kg fish (MP) (GR/TP) (USD) 0.99 ± 0.11 0.98 ± 0.13 1.0 ± 0.10 0.99 ± 0.11
Net margin (NM = GM − TC) (USD ha− 1) 9426 ± 6885 11,271 ± 5412 9269 ± 7100 9994 ± 6523
Net margin per kg fish (NM/TP) (USD) 0.25 ± 0.13 0.28 ± 0.09 0.27 ± 0.12 0.27 ± 0.11
Income above variable cost (GR − VC) (USD ha− 1) 10,291 ± 6966 12,256 ± 5495 10,223 ± 7148 10,920 ± 6600
Benefit-cost ratio (BCR = GR/TC) 1.37 ± 0.22 1.44 ± 0.22 1.39 ± 0.22 1.40 ± 0.22
Rate of farm income (NM/GR × 100%) 25 ± 12 29 ± 10 27 ± 11 27 ± 11
Break-even production (TC/MP) (t ha− 1) 26 ± 9 28 ± 9 24 ± 12 26 ± 10

7
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

Table 8 Table 9
Estimated values of coefficients and related statistics of the Cobb-Douglas production Cobb-Douglas production function estimation for small, medium and large size farms.
function.
Explanatory variables Regression coefficient of different farm category
Explanatory variables Model 1 Model 2
Small Medium Large
Y-intercept 2.91 (0.59)c 2.22 (0.68)c
Stocking (X1) 0.29 (0.04)c 0.26 (0.04)c Y-intercept 1.71 (1.08)a 3.32 (1.39)c 2.63 (1.10)b
Feeding (X2) 0.59 (0.06)c 0.57 (0.06)c Stocking (X1) 0.24 (0.08)c 0.30 (0.12)c 0.28 (0.06)c
Fertilization (X3) 0.02 (0.02) 0.01 (0.03) Feeding (X2) 0.67 (0.09)c 0.54 (0.12)c 0.67 (0.10)c
Labor (X4) 0.12 (0.06)b 0.21 (0.07)c Fertilization (X3) 0.10 (0.05)a 0.01 (0.05) − 0.01 (0.04)
D1 0.10 (0.05)a Labor (X4) 0.01 (0.09) 0.22 (0.12)a 0.10 (0.10)a
D2 −0.05 (0.06) R2 0.76 0.75 0.80
D3 0.02 (0.08) Adjusted R2 0.73 0.73 0.76
D4 −0.03 (0.05) F-value 22.17c 8.35c 20.36c
D5 0.11 (0.12) Return to scale Σbi 1.03 1.07 1.04
D6 0.00 (0.04)
D7 0.22 (0.12)a Figures within parentheses indicate standard error.
D8 0.00 (0.03) a
Significant at P < 0.1.
D9 0.07 (0.08) b
Significant at P < 0.05.
R2 0.74 0.81 c
Significant at P < 0.01.
Adjusted R2 0.72 0.77
F-value 44.24c 18.69c
Return to scale Σbi 1.02 1.48 in small farms. The production function model indicated pangasius
production in polyculture system had an elasticity of returns to scale
Figures within parentheses indicate standard error. (Σbi) of 1.02. A value greater than 1 means that fish production exhibits
a
Significant at P < 0.1. disproportionately increasing returns to scale. For example, fish pro-
b
Significant at P < 0.05. duction would increase by 1.02% if all inputs increased by 1%. A si-
c
Significant at P < 0.01.
milar trend in returns to scale was reported by Ahmed et al. (2010) for
the less intensive pangasius production system in Bangladesh. In the
coefficients for stocking density, feeding and labor were 0.29, 0.59 and
present study, the elasticity of returns to scale of small, medium and
0.12 respectively, meaning that (keeping other factors constant) a 10%
large farm categories were 1.03, 1.07 and 1.04, respectively (Table 9).
increase in stocking density, feeding and labor, would increase fish
This implies that if all inputs (except fertilization in large farms) are
production by 2.9%, 5.9% or 1.2%, respectively. The result of a positive
increased by 1%, fish production will increase by 1.03% in small,
coefficient for labor could be explained by the resultant increase in
1.07% in medium and 1.04% in large farms, respectively. This elasticity
monitoring of farm management activities. For model 2, the estimated
of returns to scale indicates that there is potential to increase pangasius
coefficients for stocking density, feeding and labor were 0.26, 0.57 and
production in polyculture systems in Bangladesh by applying more
0.21, respectively. This suggests that more attention should be paid to
specific inputs.
these parameters when trying to increase fish production.
The dummy variables were included in the model to assess the
sensitivity (if any) of fish production to qualitative factors (Kurbis, 3.6. Allocative efficiency of input use
2000). The dummy variable for commercially produced pelleted
floating feeds (D1) was found to have a statistically significant The value of marginal product (VMP) is the resource's marginal
(P < 0.05) effect on fish production. However, there was no sig- physical product (MPP) multiplied by the unit price of the output. The
nificant effect for commercially produced pelleted sinking feeds (D2), MPP of an input is the change in output resulting from employing one
farm-made pelleted feeds (D3) and farm-made mash feeds (D4). There more unit of a particular input, assuming that the quantities of other
was a positive relationship between fish production and the dummy inputs are kept constant (Anthony, 2010). The VMP should be equal to
variables for commercially produced pelleted floating feeds and farm- its price for the most efficient use of any input (Wattanutchariya and
made pelleted feeds; with commercially produced pelleted sinking feeds Panayotou, 1982). If the VMP of an input is greater than its price, profit
and farm-made mash feeds having a negative coefficient. This indicates could be increased by increasing the volume of use of the input. On the
that farm-made pelleted feeds resulted in better fish production than other hand, if the VMP of an input is less than its price, profit could be
commercially produced pelleted feeds which is consistent with Phan increased by decreasing the level of use of that input. Allocative effi-
et al. (2009) and Ali et al. (2013) for pangasius production in Vietnam ciency for inputs used in pangasius farms are presented in Table 10. The
and Bangladesh, respectively. The dummy variables for stocking carp C-D production function model indicated that stocking density, feeding
(D5) and tilapia (D6) were not statistically significant, implying that
there were no significant differences in fish production between farmers Table 10
Marginal physical product (MPP), value of marginal product (VMP) and input price of
who stocked carp or tilapia with pangasius. The dummy variable for
pangasius polyculture farms.
Mymensingh (D7) was significant (P < 0.1) when compared to Nar-
shingdi district, however Jessore (D8) and Bogra (D9) districts were not Regression Stocking density Feeding Fertilization Labor
significantly different from Narshingdi district. Despite this, there was a
Model 1
positive relationship between fish production and the dummies for all
MPP 6.03 0.33 1.26 14.24
locations. VMP 5.98 0.33 1.25 14.13
Three separate C-D production function models were run for the Px 0.05 0.32 0.40 4.02
small, medium and large farm categories and the estimated values of Input use Increase Increase Increase Increase
coefficients and related statistics are presented in Table 9. Models Model 2
showed that explanatory variables in each of the separate models were MPP 6.03 0.33 1.26 14.24
similar to the overall model with the exception of labor, which had no VMP 5.98 0.33 1.25 14.13
Px 0.05 0.32 0.40 4.02
significant effect, and fertilization, which had a significant effect Input use Increase Increase Increase Increase
(P < 0.1) in small farms only. This effect of fertilization could poten-
tially be explained by the higher stocking densities of filter feeding carp MPP = (input coefficient × mean production)/mean of input used; VMP = Py × MPP;
Py is the price of the production, Px = input price.

8
H. Ali et al. Aquaculture xxx (xxxx) xxx–xxx

Table 11 Kingdom for his constructive suggestions and editorial support with this
Marginal physical product (MPP), value marginal product (VMP) and input price by farm paper. We also thank Imran Khan of the WorldFish for his kind assis-
categories.
tance with producing the study area map.
Regression Stocking density Feeding Fertilization Labor
References
Small farm
MPP 5.29 0.39 11.34 0.71
Afroz, S., Islam, S., 2012. Economics of Aus rice (Oryza sativa) and jute
VMP 5.24 0.38 11.24 0.70 (Corchorusolitorius) cultivation in some selected areas of Narsingdi district of
Px 0.05 0.31 0.42 4.07 Bangladesh. Agriculturists 10, 90–97.
Input use Increase Increase Increase Decrease Ahmed, N., 2007. Economics of aquaculture feeding practices: Bangladesh. In: Hasan,
Medium farm M.R. (Ed.), Economics of Aquaculture Feeding Practices in Selected Asian Countries.
FAO Fisheries Technical Paper. No. 505. 2007. FAO, Rome, pp. 33–64.
MPP 7.78 0.33 1.66 15.96
Ahmed, N., Alam, M.F., Hasan, M.R., 2010. The economics of sutchi catfish
VMP 7.64 0.32 1.63 15.66
(Pangasianodon hypophthalmus) aquaculture under three different farming systems in
Px 0.05 0.32 0.39 4.01 rural Bangladesh. Aquac. Res. 41, 1668–1683.
Input use Increase Increase Increase Increase Ahmed, N., Zander, K.K., Garnett, S.T., 2011. Socioeconomic aspects of rice-fish farming
Large farm in Bangladesh: opportunities, challenges and production efficiency. Aust. J. Agric.
MPP 6.21 0.38 − 0.63 8.07 Resour. Econ. 55, 199–219.
Alam, M.F., 2011. Measuring technical, allocative and cost efficiency of pangas (Pangasius
VMP 6.28 0.38 − 0.63 8.16
hypophthalmus: Sauvage 1878) fish farmers of Bangladesh. Aquac. Res. 42, 1–14.
Px 0.06 0.32 0.37 3.92
Alam, M.F., Rahman, O., 2005. Input-output relationship and economics of pangas
Input use Increase Increase Decrease Increase monoculture and carp-pangas polyculture in two districts of Bangladesh. Bangladesh
J. Fish. Res. 9, 105–106.
MPP = (input coefficient × mean production)/mean of input used; VMP = Py × MPP; Ali, H., 2009. Assessment of Stakeholders' Perceptions on Pangasius Aquaculture Dialogue
Py is the price of the production, Px = input price. (PAD) Standards in Two Villages, Mymensingh, Bangladesh. MS Thesis. Department
of Aquaculture, Bangladesh Agricultural University, Mymensingh, Bangladesh.
Ali, H., Haque, M.M., 2011. Impacts of pangasius aquaculture on land use pattern in a
and labor were statistically significant and these inputs were used ef- selected area of Mymensingh district, Bangladesh. J. Bangladesh Agric. Univ. 9,
ficiently (Table 10). It suggests that an increase of these inputs has the 169–178.
potential to increase profitability. Ali, H., Haque, M.M., Belton, B., 2013. Striped catfish (Pangasianodon hypophthalmus,
Sauvage, 1878) aquaculture in Bangladesh: an overview. Aquac. Res. 44, 950–965.
Allocative efficiencies for inputs used by each farm category are Ali, H., Jahan, K.M., Belton, B., Dhar, G.C., Rashid, H.O., 2016a. Factors determining the
presented in Table 11. Regression analysis showed that stocking den- productivity of mola carplet (Amblypharyngodon mola, Hamilton, 1822) in carp
sity, feeding and labor significantly contributed to increasing farm polyculture systems in Barisal district of Bangladesh. Aquaculture 465, 198–208.
Ali, H., Rico, A., Jahan, K.M., Belton, B., 2016b. An assessment of chemical and biological
productivity. However, fertilization in large farms and labor in small
product use in aquaculture in Bangladesh. Aquaculture 454, 199–209.
farms were inefficient. Stocking density and feeding should be in- Anthony, B., 2010. The Making of the Classical Theory of Economic Growth. Taylor &
creased in all farms, whilst fertilization should be decreased in large Francis, pp. 192.
Ara, I., 2005. Profitability of Pangas Monoculture in Trishal Upazila of Mymensingh
farms to obtain more profit. Similarly, labor should be increased in
District. MS Thesis. Bangabandhu Sheikh Mujibar Rahman Agricultural University,
medium and large farms for proper monitoring of farming activities to Gazipur, Bangladesh.
maximize profit. Asamoah, E.K., Nunoo, F.K.E., Osei-Asare, Y.B., Addo, S., Sumaila, U.R., 2012. A pro-
duction function analysis of pond aquaculture in Southern Ghana. Aquac. Econ.
Manag. 16, 183–201.
4. Conclusions Bangladesh Bureau of Statistics (BBS), 2010. Report of the Household Income and
Expenditure Survey (HIES) 2010. Bangladesh Bureau of Statistics, Ministry of
Planning, Government of the People's Republic of Bangladesh, Dhaka.
Pangasius is a major aquaculture commodity in Bangladesh and has
Bangladesh Bureau of Statistics (BBS), 2013. Report of the Household Expenditure and
the potential to increase production further. The results showed that Income Survey 2010. Bangladesh Bureau of Statistics, Ministry of Planning,
production costs, gross incomes, net incomes and BCR were slightly Government of the People's Republic of Bangladesh, Dhaka, Bangladesh.
Belton, B., Azad, A., 2012. The characteristics and status of pond aquaculture in
higher in medium farms than small and large farms. The BCR was > 1
Bangladesh. Aquaculture 358–359, 196–204.
for all categories of farms (1.37–1.44) which indicated pangasius pro- Belton, B., Karim, M., Thilsted, S.H., Jahan, K.M., Collis, W., Phillips, M., 2011. Review of
duction in polyculture system is a profitable business. The regression Aquaculture and Fish Consumption in Bangladesh. Studies and Reviews 2011-53. The
analysis showed that stocking density, feeding and labor were im- WorldFish Center, Penang.
Belton, B., Ahmed, N., Murshed-e-Jahan, K., 2014a. Aquaculture, employment, poverty,
portant factors for increasing farm production suggesting more atten- food security and well-being in Bangladesh: a comparative study. In: Program Report:
tion should be paid to these parameters for increasing fish production. AAS-2014-39. CGIAR Research Program on Aquatic Agricultural Systems, Penang,
The results showed that farmers have the scope to increase fish Malaysia.
Belton, B., van Asseldonk, I.J.M., Thilsted, S.H., 2014b. Faltering fisheries and ascendant
production; however, they need to be economically efficient to sustain aquaculture: implications for food and nutrition security in Bangladesh. Food Policy
pangasius production in polyculture systems in the long run. Allocative 44, 77–87.
efficiencies for stocking density, feeding and labor suggest the potential Belton, B., Ahmed, N., Jahan, K.M., 2014c. Aquaculture, employment, poverty, food se-
curity and wellbeing in Bangladesh: a comparative study. In: Program Report: AAS-
of increasing these inputs to increase farm profitability. However, the 2014-39. CGIAR Research Program on Aquatic Agricultural Systems, Penang,
current use of these inputs was found to be inefficient in some farm Malaysia.
categories (e.g. labor in small and fertilization in large farms, respec- Belton, B., Asseldonk, I.J.M.V., Bush, S., 2017. Domestic crop booms, livelihood pathways
and nested transitions: charting the implications of Bangladesh's Pangasius boom. J.
tively). This suggests that there is some scope to improve production
Agrar. Chang. 17 (4), 694–714.
practices and farmers should be careful with their use of these inputs, in Boonchuwong, P., Boonchuwong, K., Noorti, K., 2007. In: Hasan, M.R. (Ed.), Economics
order to improve production efficiencies. of aquaculture feeding practices: Thailand. In: Economics of Aquaculture Feeding
Practices in Selected Asian Countries, FAO Fisheries Technical Paper 505. Food and
Agriculture Organization (FAO) of the United Nations, Rome, Italy, pp. 159–181.
Acknowledgments DoF, 2017a. Fishery Statistical Yearbook of Bangladesh. Fisheries Resources Survey
System. Department of Fisheries, Ministry of Fisheries and Livestock, Dhaka,
Bangladesh.
This paper is a contribution to the CGIAR Research Program on DoF, 2017b. Magazine National Fish Week 2014. Department of Fisheries, Ministry of
Livestock and Fish (L&F). We gratefully acknowledge financial con- Fisheries and Livestock, Government of the People's Republic of Bangladesh.
tributions from the United States Agency for International Development Edwards, P., 2009. Traditional Asian aquaculture: definition, status and trends. In:
Burnell, G., Allan, G. (Eds.), New Technologies in Aquaculture, Improving Production
(USAID) Feed the Future Cereal Systems Initiative for South Asia pro- Efficiency, Quality and Environmental Management. Woodhead Publishing,
ject (CSISA-BD, Grant no.: EEM-G-00-04-0013). The authors of the Cambridge, pp. 1029–1063.
manuscript would like to express cordial gratitude to Douglas Tenison- Hernandez, R., Belton, B., Reardon, T., Hu, C., Zhang, X., Ahmed, A., 2017. The “quiet
revolution” in the aquaculture value chain in Bangladesh. Aquaculture in press.
Collins, Institute of Aquaculture, University of Stirling, Stirling, United

9
H. Ali et al.
https://doi.org/10.1016/j.aquaculture.2017.06.006.
Jahan, K.M., Belton, B., Ali, H., Dhar, G.C., Ara, I., 2015. Aquaculture technologies in
Bangladesh: an assessment of technical and economic performance and producer
behavior. In: Program Report: 2015-52. WorldFish, Penang, Malaysia.
Karim, M., Ullah, H., Castine, S., Islam, M.M., Keus, H.J., Kunda, M., Thilsted, S.H.,
Phillips, M., 2017. Carp–mola productivity and fish consumption in small scale
homestead aquaculture in Bangladesh. Aquac. Int. 25 (2), 867–879.
Kurbis, G., 2000. An Economic Analysis of Tilapia Production by Small Scale Farmers in
Rural Honduras. M.Sc. Thesis. Department of Agricultural Economics and Farm
Management. University of Manitoba, Winnipeg, Manitoba (102 pp.).
Mamun-Ur-Rashid, M., Belton, B., Phillips, M., Rosentrater, K.A., 2013. Improving
aquaculture feed in Bangladesh: from feed ingredients to farmer profit to safe con-
sumption. In: Working Paper: 2013-34. WorldFish, Penang, Malaysia.
Phan, L.T., Bui, M.B., Nguyen, T.T.T., Gooley, G.J., Ingram, B.A., Nguyen, H.V., Nguyen,
P.T., De Silva, S.S., 2009. Current status of farming practices of striped catfish,
Pangasianodon hypophthalmus in the Mekong Delta, Vietnam. Aquaculture 296,
227–236.
10
Aquaculture xxx (xxxx) xxx–xxx
Phuong, N.T., Sinh, L.X., Thinh, N.Q., Chau, H.H., Anh, C.T., Hau, N.M., 2007. Economics
of aquaculture feeding practices: Vietnam. In: Hasan, M.R. (Ed.), Economics of
Aquaculture Feeding Practices in Selected Asian Countries, FAO Fisheries Technical
Paper 505. Food and Agriculture Organization (FAO) of the United Nations, Rome,
Italy, pp. 183–205.
Sarker, M.T., 2000. Pangus Chash Babastapana (Management of Pangus culture). A
Fisheries Leaflet. Sheuli, Comilla Cantt., Comilla, Bangladesh (in Bengali).
Shamsuddin, M., Alam, M.M., Hossein, M.S., Goodger, W.J., Bari, F.Y., Ahmed, T.U.,
Hossain, M.M., Khan, A.H.M.S.I., 2007. Participatory rural appraisal to identify needs
and prospects of market-oriented dairy industries in Bangladesh. Trop. Anim. Health
Prod. 39, 567–581.
Toufique, K.A., Belton, B., 2014. Is aquaculture pro-poor? Empirical evidence of impacts
on fish consumption from Bangladesh. World Dev. 64, 609–620.
Wattanutchariya, S., Panayotou, T., 1982. The economics of aquaculture: the case of
catfish in Thailand. In: Aquaculture Economics Research in Asia. International
Development Research Center, Ottawa, Canada, pp. 26–34.