Productivity Growth in the Shrimp Farming
Industry of Bangladesh: A Luenberger
Productivity Indicator Approach
Hasneen Jahan, PhD
Bangladesh Agricultural University, Bangladesh
Tihomir Ancev, PhD
The University of Sydney, Australia
9/7/2014
IIFET 2014, Brisbane, Australia
1
Shrimp Aquaculture




One of the fastest growing aquaculture activities in Bangladesh
Earns foreign exchange
earns 350 million US dollars a year, making up about 3% of total
export earnings
Contributes significantly to the employment and community
development
600,000 people employed directly in shrimp aquaculture, who
support approximately 3.5 million dependents
Shrimp farming has therefore rapidly expanded in the coastal
districts of Bangladesh due to favourable agro-climatic condition and
low production cost. Currently, there are about 250,000 hectares of
land occupied with shrimp farms in these regions, producing about
156 million tonnes of shrimp annually
9/7/2014
IIFET 2014, Brisbane, Australia
2
Shrimp farming
districts
Khulna
Bagerhat
Satkhira
Cox’sbazar
Chittagong
Patuakhali
Pirojpur
Jessore
9/7/2014
IIFET 2014, Brisbane, Australia
3
Typical shrimp farm
9/7/2014
IIFET 2014, Brisbane, Australia
4
Shrimp production and export earning
350
700
Shrimp (V)
Fish(V)
300
Shrimp(Q)
600
Fish(Q)
250
500
200
400
150
300
100
200
50
100
0
Qty in MT
Million $
Linear (Shrimp(Q))
0
Year
9/7/2014
IIFET 2014, Brisbane, Australia
5
Shrimp area trend
Area (ha)
Khulna
Chittagong
Barisal
Dhaka
Total
Year
9/7/2014
IIFET 2014, Brisbane, Australia
6
Shrimp Aquaculture




Environmental degradation is a major concern
Increasing salinity, declining productivity of land, destruction of
mangrove forests, declining inland capture fishery production
The five-fold increase in soil salinity levels in many areas seriously
hampers crop cultivation (Islam et al., 1999). The effects of shrimp
farming-induced salinity on paddy rice production have a significant
negative impact on its productivity (Umamaheswar et al., 2009).
the rapid expansion of shrimp farming has drastically reduced the
stock of indigenous fish varieties and destroyed mangrove flora and
fauna (Gain, 1998).
The ecological problems created by unplanned shrimp cultivation,
particularly in the Chakaria Sundarbans resulted in extinction of
mangroves in this area (Gregow, 1997).
9/7/2014
IIFET 2014, Brisbane, Australia
7
Area affected by soil salinity
Soil Salinity status
80
% of salinity affected area
70
60
50
40
2000
30
2010
20
10
0
Khulna
Bagerhat
Satkhira
Jessore
Cox’s
Bazar
Chittagong Patuakhali
Pirojpur
Districts
9/7/2014
IIFET 2014, Brisbane, Australia
8
Water salinity
Water Salinity status
35
Water salinity in EC
30
25
20
15
2000
2010
10
5
0
Khulna
Bagerhat
Satkhira
Jessore
Cox’s
Bazar
Chittagong Patuakhali
Pirojpur
Districts
9/7/2014
IIFET 2014, Brisbane, Australia
9
Mangrove forest destruction
9/7/2014
IIFET 2014, Brisbane, Australia
10
Research Objectives




9/7/2014
Productivity growth is an important goal for the sustainable growth
of the shrimp industry. The application of productivity growth
measurement with undesirable outputs is a comparatively new
technique which can effectively address the productivity of farms
and industry in the sustainable development perspective.
The present paper aims to assess the dynamics of productivity
growth of shrimp farming industry in Bangladesh over time,
considering both economic and environmental outcomes
associated with this sector
A modified, environmentally adjusted Luenberger productivity
indicator is applied for this purpose. This model can easily
decompose the productivity growth into technological change and
technical efficiency change components
The analysis of productivity growth will help to identify
performance of shrimp farms over time when the production
technology adjusts some inputs and outputs in terms of economic
and environmental variables
IIFET 2014, Brisbane, Australia
11
Research Methods




This paper applies the environmentally adjusted productivity
measurement using directional distance function
The directional distance function plays an important role in
production theory and has the powerful advantage of taking into
account the variation of both input and output bundles
simultaneously.
The Luenberger productivity indicator is constructed in terms of
difference between directional distance functions. One of the major
advantages of using the Luenberger productivity indicator is that it
allows for inclusion of undesirable outputs.
This indicator can also capture the time dimensions of a production
process. It credits production units for reduction of undesirable
outputs, providing a measure of productivity and indicates whether
the environmentally adjusted productivity has improved over time
9/7/2014
IIFET 2014, Brisbane, Australia
12
Directional distance function
DDF measures the maximum expansion of desirable
outputs in the gy direction with the largest feasible
proportional contraction in inputs and undesirable outputs
in the –gx and –gb directions, respectively, or:


Do  x, y, b; g y  gb ,  g x   sup  :  y   g y , b   g x   P( x   g x )
Output directional distance function
Do ( x, y, b; g y ,  gb )  sup  : ( y   g y , b   gb )  P( x)
9/7/2014
IIFET 2014, Brisbane, Australia
13
Directional distance function
Do ( x, y, b; g y ,  gb )  sup  : ( y   g y , b   gb )  P( x)



If the directional distance function value β is zero, the
farm operate on the frontier and state; otherwise
inefficiency
Based on directional distance function, Chambers et al.
(1996) introduced Luenberger Productivity Indicator
to measure the productivity growth in APEC countries
Luenberger productivity indicator is constructed in terms
of difference between directional output distance
functions and compares farm performances in adjacent
periods, t, and t+1.
9/7/2014
IIFET 2014, Brisbane, Australia
14
Luenberger Productivity Indicator

the Luenberger productivity indicator including
undesirable outputs can be expressed by the following
equation:




1 t 1 t t t
 Do ( x , y , b ; g )  Dot 1 ( xt 1, y t 1, bt 1; g )    Dot ( xt , yt , bt ; g )  Dot ( xt 1, y t 1, bt 1; g ) 
2
If the value of L(.) is positive it indicates productivity
growth while zero represents no change in productivity
Luenberger productivity indicator can be decomposed
additively into an efficiency change component (LECH),
and a technical change component (LTCH).
Ltt1 
9/7/2014
AARES 2013, Sydney, Australia
15
Luenberger Productivity Indicator

Luenberger Productivity Indicator can be decomposed
into efficiency change and technical change components
LECHtt 1  Dot ( xt , yt , bt ; g y , gb )  Dot 1 ( xt 1 , y t 1 , bt 1; g y , gb )
LETCtt 1 

1
 Dot 1 ( xt 1 , y t 1 , bt 1 ; g y , gb )  Dot ( xt 1 , y t 1 , bt 1; g y , gb ) 
2

  Dot 1 ( xt , y t , bt ; g y , gb )  Dot ( xt , y t , bt ; g y , gb )  

LECH captures the efficiency change between periods t and t+1, i.e.
catching up or falling behind relative to the best practice frontier. A
positive (negative) value of indicates increase (decrease) in
efficiency. LTCH measures shifts in the production possibilities
frontier itself. A positive (negative) value of indicates technological
progress (technological regress) between periods.
9/7/2014
IIFET 2014, Brisbane, Australia
16
The Luenberger productivity indicator
4/17/2015
IIFET 2014, Brisbane, Australia
17
Four Models

Model 1: credits farm for expanding desirable outputs and
simultaneously contracting undesirable outputs with directional
vector g= (1,-1)

Model 2: credits farm only for expanding desirable outputs while
undesirable outputs are present, with directional vector g= (1,0)

Model 3: credits farm for contracting undesirable outputs while
desirable outputs are held constant with directional vector g=(0,-1)

Model 4: excludes undesirable outputs from the production process
and only considers the expansion of desirable outputs with
directional vector g=(1). It is not environmentally adjusted.
9/7/2014
IIFET 2014, Brisbane, Australia
18
Data


Eight coastal districts and 31 sub-districts were chosen where each
sub-district acts as a representative farm. Each sub-district is
considered as an observation unit resulting with 62 observational
units for the two time periods analysed, i.e. 2000 and 2010.
The variables are per hectare cost and return, shrimp area in
hectare, % of salinity affected area, water salinity of river (EC) in
ds/m, fish catch (kg/household), mangrove area in hectare
Dot 1 ( x k 't , y k 't , bk 't ;1, 1)  max 
y1= gross return from shrimp enterprises
y2= average fish catch per household
x= total cost for shrimp enterprises
b1= soil salinity in shrimp farming sub-districts
b2=water salinity
b3= loss of mangrove forest in shrimp farming area
9/7/2014
IIFET 2014, Brisbane, Australia
19
Solving DDF

The maximization problem for the mixed-period distance
function,Dot 1 (t ) is, for ( xk , y k , bk ) with k  1,..., K observations,is
obtained by solving the following linear programming
technique:
Dot 1 ( x k 't , y k 't , bk 't ;1, 1)  max 
s.t.
K
t 1 t 1
t
z
y

(1


)
y
 k km
k ' m , m  1,..., M
k 1
K
t 1 t 1
t
z
b

(1


)
b
 k ki
k ' i , i  1,..., I
k 1
K
z
k 1
t 1 t 1
k
kn
x  xkt ' n , n  1,..., N
zkt 1  0, k  1,...,K
9/7/2014
IIFET 2014, Brisbane, Australia
20
Results
District Model 1(g=1, -1)
PC
EC
TC
Model 2 (g=1, 0)
Model 3 (g=0, -1)
Model 4 (g=1)
PC
PC
PC
EC
TC
EC
TC
EC
TC
Khulna
-0.073
0.021
-0.094
-0.126
0.039
-0.145
-0.271
-0.258
-0.012
0.029
0.240
-0.211
Bagerhat
-0.041
0.039
-0.080
-0.066
0.048
-0.111
-0.141
0.097
-0.239
-0.003
0.169
-0.172
Satkhira
0.043
0.135
-0.092
0.075
0.159
-0.161
-0.098
-0.027
-0.071
0.082
0.277
-0.195
Cox’s
Bazar
0.039
0.068
-0.029
0.042
0.070
-0.028
0.050
0.024
0.026
0.045
0.241
-0.196
Chittagong
0.138
0.162
-0.024
0.307
0.392
-0.084
0.240
0.277
-0.037
0.238
0.478
-0.241
Jessore
-0.067
0.000
-0.067
-0.136
0.000
-0.272
-0.293
0.000
-0.293
-0.307
0.017
-0.325
Patuakhali
-0.262
-0.227
-0.036
-0.564
-0.253
-0.092
-0.527
-0.583
0.056
-0.347
-0.175
-0.172
Pirojpur
-0.024
0.000
-0.024
0.001
0.000
0.002
0.161
0.425
-0.264
-0.179
0.107
-0.286
Average
-0.031
0.025
-0.056
-0.023
0.057
-0.080
-0.110
-0.006
-0.104
-0.178
-0.100
-0.065
9/7/2014
IIFET 2014, Brisbane, Australia
21
Results

The average productivity indicator for all sub-distrcits has a value of
-.031, indicating that the environmentally adjusted productivity
growth of the shrimp farming industry has decreased by -3.1% over
a decade between 2000 to 2010.

A comparison across districts indicates that sub-districts of Khulna and
Bagerhat districts in Southwest region experienced a negative
environmentally adjusted productivity growth of -7.3% and -4.1%
respectively during the sample time period.
On the other hand, sub-districts in Cox’s Bazar and Chittagong districts of
the Southeast region have experienced a positive environmentally adjusted
productivity growth, 3.9% and 13.8% respectively during the sample time
period.
The Southern region (Pirojpur and Patuakhali districts) has experienced a
negative productivity growth, without any improvement, neither in
technological features, nor in technical efficiency attributes.


9/7/2014
IIFET 2014, Brisbane, Australia
22
Results




Although the analyses generate a mixed response of productivity
change across farms, overall the farms have experienced negative
productivity growth in the Southwest and Southern regions and
positive productivity growth in Southeast region.
One of the significant reason for productivity decline between two
time points, 2000 and 2010 can be explained by the two major
cyclones called ‘Sidr’ and ‘Aila’ which hit these area in 2007 and
2009 respectively.
Since Model 4 ignores the environmental attributes by excluding
undesirable outputs, the productivity measurement underestimates
productivity growth from a welfare point of view.
The results obtained from the Luenberger productivity indicators
imply that the traditional technologies that are currently being used
in the shrimp farming industry in Bangladesh need to be improved
for the better performance of the industry;
9/7/2014
IIFET 2014, Brisbane, Australia
23
Results
0.3
Model 1
Productivity Growth
0.2
0.1
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
PC
EC
-0.1
TC
-0.2
-0.3
9/7/2014
Observation
IIFET 2014, Brisbane, Australia
24
Results
0.8
Model 2
0.6
Productivity Growth
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
PC
EC
-0.2
TC
-0.4
-0.6
-0.8
9/7/2014
Observation
IIFET 2014, Brisbane, Australia
25
Results
1
Model 3
0.8
Productivity Growth
0.6
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
PC
EC
-0.2
TC
-0.4
-0.6
-0.8
-1
9/7/2014
Observation
IIFET 2014, Brisbane, Australia
26
Results
0.6
Model 4
Productivity Growth
0.4
0.2
PC
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
EC
TC
-0.2
-0.4
-0.6
9/7/2014
Observation
IIFET 2014, Brisbane, Australia
27
Policy Implications




The productivity measures are essential for the farm operators as
well as for the policy makers for further initiatives to achieve the
sustainable development goal. When this measures incorporate the
environmental attributes it can provides the right indication for future
actions.
A well-defined shrimp policy which will focus on new environmental
friendly technologies can ensure the sustainability of this vital
industry.
The results could provide the basis for implementing more regionspecific policies for shrimp industry which will be useful to improve
environmental performance of shrimp farms in conjunction with their
economic performance.
The present study indicates the necessity of measuring
environmentally adjusted productivity growth for the sustainable
development of the shrimp industry and the government, business
and academic communities can draw some lessons from it to take
further necessary actions.
9/7/2014
IIFET 2014, Brisbane, Australia
28
Thank you
9/7/2014
IIFET 2014, Brisbane, Australia
29