Professional Documents
Culture Documents
*Central Institute of Brackishwater Aquaculture, 141, Marshalls Road, Egmore, Madras 600 008, India
Division of Agricultural Economics, Indian Agricultural Research Institute, New Delhi 110 012, India
ABSTRACT
Multi-objective analysis, via the constraint approach, of paddy-fishery enterprise system in the
Kuttanad region of Kerala State. lndia, is attempted to develop a trade-off analysis between paddy
and fishery systems and to suggest optimal operating policies for the Thaneernukhom salt-water
barrage for maximizing the returns from the region.
Both primary an secondary data were collected and used in formulating the linear
programming matrix, which formed the basis of the multi-objective analysis and the trade-off
analysis by way of transformation curves.
The trade-off analysis revealed a greater pay-off in gross area at the ‘ideal’ point of paddy
area ‘anti-ideal’ point of fish area. The shift in area from fish to paddy was greater in all the three
cases than vice versa. The net benefit-loss figures arrived from actual and normative values of area
and income generated showed an estimated maximum loss of 39.69% (1981-82) to a minimum loss
20.89% (1984-85) during the 1980s for the region.
Income was greater per unit area allocated from fish culture than from paddy cultivation.
The study offered three alternatives for planners. The highest income. Rs. 340.30 million, could be
generated from the region by keeping the barrage open year-round. The highest paddy production
could be obtained by keeping the barrage closed for 6 months. A via media solution is to keep the
barrage open for 3 months (mid-December to March). This would provide an income of Rs. 267 8
million.
The nature of conflicting objectives forces the economists and planners to have a
broader and more comprehensive notion of the project and its evaluation. The
nature of the problem and the type of empirical data dictate the specific model to
be used, although non-availability of reliable and relevant empirical data limits the
scope of mathematical modelling in farming system studies (Maji 1991).
Multiple-criteria decision making became acceptable to researchers in 1972
(Romero and Rehman, 1989). Its superiority over the standard linear programming
approach has been often repeated. The choice of appropriate approach of multi-
objective analysis in the attempt to model the Kuttanad farming system in the
context of enabling a trade-off between paddy and fish farming systems and
suitable operating policies for the Thaneermukhom salt-water barrage for
maximizing the returns from the region rests on the capacity of the methodology to
(1) maximize the returns from paddy farming, (2) maximize the returns from fish
farming, and (3) facilitate a trade-off between the two enterprises.
Of the available approaches to multi-objective analysis (Thampapillal, 1978;
Willis and Perlack. 1980; Sandiford, 1986), multi-objective programming via the
constraint approach (Romero and Rehman, 1989) offered the most appropriate
and correct methodology.
Primary data were collected from 150 farmers, 50 each from ‘Karapadam,
Kari’, and Kayal’ lands. All data relating to agricultural operations pertaining to
paddy, fish, paddy-fish-paddy, and paddy-cum-fish farming were collected.
Secondary data on macrovariables such as gross cropped area, availability of
human labour, bullock labour, tractor hours, fertilizers, and teed for each type were
collected from various sources for the period under analysis.
The standard linear programming matrix was developed and formed the
basis of the constraint approach to multi-objective programming. The general
nature of the multi-objective programming problem with q objectives can be stated
as follows (Romero and Rehman, 1989):
Sub to :XεF
where Eff search for the efficient solutions and F = feasible set.
The initial and useful information of the constraints approach to multi-
objective programming was generated by optimizing one of the objectives
(maximizing area paddy1) while the other (maximizing area under fish1) was
specified ass restraint. This mathematical programming problem can be stated as
Maximize Zk(X).
Sub to: XεF
Zj(X)≥Lj J=1,2,……k-1, k+1,……..q
where Zk (X) is the objective to be optimized. Through parametric variation of the
right side L1, the efficient set was generated.
This methodology enabled us to generate the pay-off matrix containing the
deal and the ‘anti-ideal’ values for maximization of paddy area and fish area. The
values when converted to monetary terms by using the farm harvest prices and
yields gave the range of returns from paddy and fish. The ‘ideal’ and ‘anti-ideal’
values defined the upper and lower bounds for the range where the parameter L1
can vary. Parameterizing L1 for values belonging to the interval, an approximation
of the efficient set was obtained. Efficient solutions were generated when the
parametric constraints were binding at the optimal solutions.
Trade-offs
The trade-off between two criteria meant the amount of achievement of one
criterion that was sacrificed to gain a unitary increase in the other one. Given two
efficient solutions X1 and X2 the trade-off between the j th or kth criteria is given
fj ( X 1 ) − fj ( X 2 )
Tjk=
fk ( X 1 ) − fk ( X 2 )
1
Statistics on paddy area and fish area have been used in lieu of production statistics owing to the
non-availability of the latter on a time series basis. Trial runs made with the objective of
maximization of paddy/fish production did not indicate any substantial difference in the levels of
production of either, between values of the efficient set, thus justifying the use of area figures as a
proxy for production in generating the efficient set.
The transformation curves
The area transformation curves were derived from the efficient set(s)
generated turns through the constraints approach of the multi-objective
programming. The iso-revenue curve for each year considered was derived as a
ratio of the farm harvest prices of the commodities under study. Graphically the
normative equilibrium of the region for each year with respect to area of paddy/fish
is defined by the point of tangency of the respective transformation curve and the
highest iso-relevant curve.
The region’s net loss/benefit was arrived at by subtracting the actual area under
paddy and fish farming from the transformation curve in real and monetary terms
on an annual basis.
The standard linear programming matrix was used to arrive at normative figures of
area under paddy and fish farming in each land type by suitably simulating the
objective function row for three different scenarios of barrage operations. Three
sets of these optimal plans were worked out for the three land types and put
together in a single pay-off matrix representing the region.
Table 1 represents the pay-off matrices of area transformation in the ‘ideal’ and the
‘anti-ideal’ situations in Karapadam, Kari, and Kayal lands. The best values are
given in bold along the main diagonal.
Table 2 accounts for the major factors that can be deduced from the pay-off
matrices of the different land types of Kuttanad. The gross cropped area was
51,489 ha at the ‘ideal’ point of paddy area and ‘anti-ideal’ point of fish area in
Karapadam lands. For the same parameter it s 20,024 and 13,757 ha in Kari and
Kayal lands respectively. This indicated a cropping intensity of 154.10% in
Karapadam lands, 166.05% in Kari lands, and 145.36% in Kayal lands. The gross
cropped area was 41,488 ha at the ‘ideal’ point of fish area and ‘anti-ideal’ point of
paddy area in Karapadam lands, indicating a cropping intensity of 149.59% in the
former and 142.86% in the latter land type. ‘ldeal’ returns were maximum in
Karapadam lands at Rs 330.61 million. The total ideal returns for Kuttanad would
be Rs 515.56 million. The pay-off matrices also provided an index of the
opportunity cost of paddy and fish farming in terms of shifts in area from one
enterprises to the other. The ratio of shift in area from fish to paddy was greater in
all land types than vice versa. This shift was maximum in the Kayal lands in the
ratio of 1:5.85 ha, and lowest in Kari lands at 1:3.92 ha. The shift from paddy to
fish was maximum in Karapadam lands in the ratio of 1:3.92 ha
Table 2. Factors emerging from pay-off matrices of different land types of Kuttanad
Particulars Karapadam Kari Kayal
GCA 1(ha) 51489 20024 13757
GCA 2 (ha) 41488 18039 13520
Ideal returns (Rs Million) 330.61 103.07 81.88
Shift in area 1 (ha) 1:5.38 1:3.92 1:5.85
Shift in area 2 (ha) 1:3.14 1:1.39 1:1.69
GCA 1 = gross cropped at the ‘ideal’ point of paddy area and ‘anti-ideal’ point of
fish area.
GCA 2 = gross cropped at the ‘Ideal’ point of paddy area and ‘anti-ideal’ point of
paddy area.
Ideal returns = maximum returns possible as a combination of ‘ideal’ points of
paddy and fish areas.
Shift in area 1= shift in area from fish to paddy
Shift in area 2 = shift in area from paddy to fish.
Table 4. Income area and production of paddy and fish in Kuttanad region under
different scenarios of barrage operations (area in ha, production in
thousand t , profits in Rs. Million)
actual figures was 39.69%. It was 32.54% in 1981-82 20.89% in 1984-85, and
27.62% in 1988-89. This analysis helped in estimation of the possible losses that
had been sustained by Kuttanad over the time considered.
In the context of both ex-ante and ex-post evaluation of prolects simulation
offers a comprehensive methodology for assessment of large scale and complex
systems (Budnick et al. 1988). In a broad sense simulation is a methodology for
conducting experiments using a model of the real system (Lal, 1990).
Using simulation income, area and production of paddy and fish for each
land type of Kuttanad for three different scenarios of barrage operations have been
optimized independently and presented in a single matrix (Table 4).
Optimizing for each of the scenarios independently ,for income, area, and
production of paddy and fish in each of the land types, it may be seen that income
for the region was maximum at As 340.13 million under scenario Ill with no closure
of the barrage. Income was maximum at Rs.267.84 million under scenario II
indicating a lower revenue of As 72.47 million than the income obtained under
scenario III. Under scenario I the income was Rs 127.38 million less than under
scenario Ill. Paddy area and production was the highest under scenario I with
barrage being closed from December to June. It was the lowest under scenario III.
Contrarily, area and production under fish farming was maximum under scenario Ill
and lowest under scenario I.
It is obvious that contribution to income was greater per unit area/production
from fish than from paddy farming. The analysis thus offered alternatives to policy
makers to (1) maximize income from the region by opting for no closure of the
barrage (2) enhance production of both paddy and fish while maximizing income
reasonably well by closing the barrage from December to March or (3) maximize
production of paddy from the region when the barrage remains closed for 6 months
from mid-December to June.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the assistance given to the project by the
Regional Research Station Kerala Agricultural University, Kumarakom and
Mancompu, the Fish Farmers Development Agency, Alleppey, Statistical Office,
Kuttanad Taluk and District Statistical Office, Alleppey The first author expresses
his thanks to the Post Graduate School, Indian Agricultural Research institute, New
Delhi for the senior fellowship awarded for conducting this project and also Dr.
E.G. Silas, former Vice Chancellor Kerala Agricultural University for his
constructive suggestions on an earlier draft of the paper.
REFERENCES
B.K.H. Consulting Engineers. (1989). Kuttanad Water Balance Study, Kingdom of
Netherlands, Ministry of Foreign Affairs, Amsterdam, 483 pp.
Budnick, Frank S., McLeavey Dennis and Mojena, Richard. (1988). Principles of
operations Research for Management. Richard D. Irwin Inc., Homewood,
Illinois, 988 pp.
Cohon, J.L., Church R.L., and Sheer, D.P. (1979). Generating multiobjective trade-
offs: An algorithm for bicriterion problem. Water Resources Res. 15:1001-
1010.
Gilbert, Sylvia. (1993). Market information and fisheries management. A
multiobjective analysis. Aquaculture and seafood markets: public policy,
consumer behaviour and industry relationships. University of Rhode
Island/Oregon State University Research Paper Series. OSU-93-103.
Krishnan, M. (1994). Multiobjective analysis of paddy-fishery enterprise systems in
the Kuttanad region of Kerala State. Ph. D. thesis, Division of Agricultural
Economics, Indian Agricultural Research Institute, New Delhi, 197 pp.
Kurup, Madhusoodhana B., and Samuel, C.T. (1985). Fish and fishery resources
of Vembanad Lake. In Harvest and Post Harvest Technoloqy of Fish.
Society of Fisheries Technologists Cochin, pp. 77-82
Lal Babu. (1990) Impact of Farakka barrage on the hydrological changes and
productivitv potential of Hooghly Estuary J. Inland Fish Soc. India, 22 (1 &
2):38-42.
Long, Roger B. (1990). A trade off analysis of alternative water uses. Water
Resources Bull., 26(1).
Maji, C.C. (1990). Farming systems approach to research Indian J. Agro. Econ.,
XLVI(3): 403-411.
Romero, C., and Rehman, T. (1989) Multiple criteria Analysis for Agricultural
Decisions. Elsevier Science Publishes B.V. Amsterdam, 255 pp.
Sandiford, Francis. (1986) An analysis multiobjective decision making for the
Scottish inshore fishery. J. Agric. Eon., 37(2):207-219.
Shetty, H.P.C. (1965). Observations on the fish and fisheries of the Vembanad
backwaters, Kerala Froc. Nat. Acad. Sci. India, Secion B 35(1):115-130.
Thampapillai, D.J. (1978) Methods of multiple objective planning. Word Agric.
Econ. Sociol. Abst., 20(12):803-813.
Vedulla, S., and Rogers, P.P. (1981). Multiobiective analysis or irrigation planning
in river basin development. Water Resources Res., 17(5):1304-1310.
Venugopal, P. (1992). A barrier that is fishermen’s bane. The Hindu, 18 March,
Coimbatore ed., p. 4.
Willis, C.E and Perlack, R..D. (1980). A comparison for generating techniques and
goal programming for public investment, multiple objective decision
making. Amer. J. Agric. Econ. 62:66-74