INTRODUCTION
Efficiency in production is a way to ensure that products of firms are produced in the best and most profitable way. To prevent waste of resources, efficiency is of great importance for every sector in the economy. In addition, fishermen in Sistan region are facing with problems of unemployment, low income, drought, financial limitation, implying that the sensitivity of efficiency gap will become more sever.
Following Farrel (1957) one can describe technical and allocative efficiency of firms. From the output perspective, technical efficiency measures the potential increase in output, keeping the inputs constant. Allocative efficiency from the output perspective is simply the revenue maximizing problem. Technical efficiency from the input perspective measures the ability of the firms to produce a given output using the smallest set of inputs. Allocative efficiency in this case measures the firm’ ability to allocate the input bundle in the cost minimizing way. Combining measures of technical and allocative efficiency yields a measure of economic efficiency. The output and input perspective will coincide when measuring technical efficiency under constant returns to scale. The allocative and economic efficiency measures however are completely different in nature and are not likely to coincide for other reasons than by chance.
Various degrees of inefficiency in production seem to be the rule rather than
the exception. Bailey et al. (1989) estimated technical, allocative and
economic efficiency on a sample of Ecuadorian dairy farms. They found a positive
relationship between size and technical efficiency. In contrast to the New England
study, mediumsized Ecuadorian farms were found to be as allocatively efficient
as large farms. Bravo and Rieger (1991) examined technical, allocative and economic
efficiency of a sample of New England dairy farms, using the Stochastic Frontier
Approach (SFA) and a CobbDouglas production function. They found overall economic
inefficiencies of on average 30%. However it was little difference between mean
technical (83.0%) and mean allocative efficiency (84.6%). Heshmati and Kumbhakar
(1994) examined the technical efficiency of Swedish dairy farms, during period
of 19761988, excluding 1985, using the stochastic frontier approach and a translog
production function. They found that the mean technical efficiency indices were
between 0.81 and 0.83. This indicates technical inefficiencies of almost 20
% in the Swedish dairy farms. Jonasson (1996) used DEA approach and measured
various output efficiencies of a sample of Swedish farms during 19891991. He
found that the average technical and allocative output efficiencies where 0.95
and 0.92, respectively. Lansink et al. (2002) studied technical efficiency
of Finnish farms, using the Data Envelopment Analysis (DEA). They found that
the conventional livestock farms had technical efficiency scores of 69%. A possible
reason for the great difference between the two studies in Sweden is that Jonasson
didn’t aggregate output in DEA. Adding an extra output or input in DEA
will never cause a reduction of the efficiency scores and a greater number of
outputs and inputs compared to the total number of observations will always
cause greater efficiency scores. Thus, the difference is much likely to depend
on the differences in the methods. Although data envelopment indices should
not be used for comparison between different studies (Coelli et al.,
2002), since the scores only measure the relative efficiency within the sample,
there are evidence of technical, allocative and economic inefficiencies in dairy
farms (Coelli et al., 2002).
Although many researches conducted to investigate the efficiency of fishery in the world (Kirkley et al., 1995; Sharma and Leung, 1999; Pascoe et al., 2001; Fousekis and Kolonaris, 2003; Kompas et al., 2003; Tigley et al., 2005) only few studies could be found in Iran (Yazdani and Esmaeili, 1995; Esmaeili, 2006). Data Envelopment Analysis (DEA) is recently developed approach for measuring efficiency in fisheries (Felthoven; 2002; Pascoe and Herrero, 2004).
Hamoon Lake is located in southeastern of Iran. This Lake is the biggest freshwater Lake in Iran, which its water source is Hirmand River from Hendokosh Mountains in Afghanestan. The Hamoon Lake area is around 5000 km^{2} which share between Iran and Afghanestan. Fishery from Hamoon Lake is important for local economy.
The main purpose of this study is to examine the technical efficiency for the fishery industry in Iranian part of Hamoon Lake.
MATERIALS AND METHODS
The idea behind efficiency studies is to measure a firm’ position relative to an efficient frontier, resulting in an efficiency score of the firm. The efficiency scores will be bounded between zero and one, where a score of one indicates full efficiency. Measurement of efficiency requires knowledge of the efficient production function, which thus has to be estimated from the sample data.
As was pointed out in the previous section, DEA is technique of estimating
a firm’ relative position to the frontier. When using DEA, estimation via
the production, cost or profit function is possible. The cost and profit functions
are both dual to the production function and thus they can be derived from the
estimates. Cost and profit functions have the advantage of allowing for multiple
outputs (Coelli, 1996). Data envelopment analysis developed by Charnes et
al. (1978), is a nonparametric approach. The production frontier in DEA
is deterministic, so any deviations from the frontier are related to inefficiency.
DEA approach is very useful for multi product activities such as fishery in
Hamoon Lake.
The idea behind DEA is to use linear programming methods to construct a surface,
or frontier around the data. Efficiency is measured relative to this frontier,
where all deviations from the frontier are assumed to be inefficiency. Consider
n firms producing m different output using h different inputs. Thus, Y is an
mxn matrix of outputs and X is an hxn matrix of inputs. Both matrices contains
data for all n firms. The Technical Efficiency (TE) measure under the assumption
of Constant Returns to Scale (CRS), can be formulated as follows:
and solved
for each firm in the sample. θ_{i} is firm i’ index of technical efficiency relative
to the other firms in the sample. y_{i} and x_{i} represents
the output and input of firm i respectively. Yλ and Xλ are the efficient
projections on the frontier.A measure of θ_{i} = 1 indicates that
the firm is completely technically efficient. Thus, 1  θ_{i} measures
how much firm i’ inputs can be proportionally reduced without any loss
in output. However, the assumption of CRS is correct only as long as firms are
operating at an optimal scale (Coelli et al., 2002). Various constraints
on inputs like financing and the goals of the owner may cause the firm to operate
at a nonoptimal scale. Using the CRS DEA model when firms are not operating
at their optimal scale will cause the TEmeasures to be influenced by scale
efficiencies and thus the measure of technical efficiency will be incorrect.
By adding a convexity constraint to the model above VRS is instead assumed:
The new constraint is NI’λ = 1 where N1 is a nx1 vector of ones.
This constraint makes the comparison of firms of similar size possible, by forming
a convex hull of intersecting planes, so that the data is enveloped more tightly.
The technical efficiency measures under VRS will always be at least as great
as under the CRS assumption. In order to derive the economic efficiency of the
firm, the following model is solved:
Where:
x_{i}ι 
: 
The firm i’ vector of input prices 
x_{i}* 
: 
The cost minimizing input bundle faced by firm i. The economic efficiency
for firm 
i 
: 
Then solved by the following computation: 
that is, the observed cost is compared to the minimum cost the firm would face
if using the optimal input bundle. Furthermore, the Allocative Efficiency (AE)
of firm i can be calculated as follows:
AE measures firm i’ relative ability to allocate the inputbundle in the cost minimizing way, given the estimated technology.
Dataset applied in this study include cross sectional data from Iranian part of Hamoon Lake. The information was collected through facetoface interviews with skippers in 2004.
RESULTS
The data of this study was obtained by completing 74 questionnaires among a randomly selected sample of Hamoon fish harvesters in 2004. However, the selected sample accounts for 40% of fishermen population. A harvesting unit contains a head  harvester and some of the normal harvesters, competing with each other.
The income share of normal harvesters from fish sale is as a percentage of
total income, since harvesting units, instead of the number of harvesters, was
applied. So, their income depends on the total amount of harvesting fish. But,
the head harvester due to providing harvesting facilities, such as boat and
automobile for selling, gets considerable part of income. Different species
of harvested fishes, regarded as outputs, are Common carp (Cyprinus carpio;
Y_{1}), Grass carp (Ctenopharyngoden idella; Y_{2}),
Silver carp (Hypophthalmichthys molitrix; Y_{3}), Big head (Aristichthys
nobilis; Y_{4}) and Schizothorax (Schizothorax zarudnyi;
Y_{5}. The inputs are also as follow:
X_{1}: 
The number of harvesters of group 
X_{2}: 
The number of motorpowered boats 
X_{3}: 
The number of harvesters transferring automobiles. 
X_{4}: 
Net size. 
X_{5}: 
Annual cost of timber. 
X_{6}: 
Annual insurance cost of the group's members 
X_{7}: 
Annual cost of harvesting permission 
X_{8}: 
Annual costs of strip. 
X_{9}: 
Annual cost of fuel. 
The normal harvesters are managed by their head harvester. The harvested output is transferred by motorpowered boats from the lake to the coast and by truck from coast to sale to cooperatives and retailers. Net size is the most important input in harvesting. The wider the net size the more output will be obtained. Timber and annual permission are of other inputs that were used in terms of their costs. Insurance was also considered in terms of its cost. The rest of inputs are strip length and amount of fuel.
Based on the relationship presented in methodology for DEA approach, technical, allocative and economical efficiencies under two assumptions, constant and variable returns to scale, were estimated. The results revealed that the differences between technical efficiencies under two assumptions are not statistically significant. Under variable returns to scale, technical efficiency is divided into net technical and scale efficiencies.
Findings of Table 1, shows that harvesters have a good
performance from managerial viewpoint and their net technical efficiency is
near to one. Therefore, harvesters with respect to their activities and returns
to scale can increase technical efficiency by changing the amount of inputs.
This finding is similar to Esmaeili (2006), who analyzed fishery industry in
the northern Persian Gulf using frontier model. As the results showed 44 farms
have increasing returns to scale, 29 of them perform under constant returns
to scale, while only one farm has decreasing returns to scale (Table
2). That is, for half of harvests, increasing the amount of inputs will
result in decreased output, while those with constant returns to scale can expect
the same change rate for output as they change the amount of the input. Most
researchers found increasing return to scale for fishery industry. For instance,
Fousekis and Kolonaris (2003), Garcia Del Hoyo et al. (2004) and Esmaeili
(2006) calculated returns to scale of 1.26, 2.65 and 1.42, respectively.
Technical efficiency was estimated under input minimizing assumption. Average
technical, allocative and economical efficiencies for selected sample are 82.7,
75.5 and 62.7%, respectively (Table 3). Hence, economical
inefficiency comes from both of the technical and allocative inefficiencies;
although allocative efficiency is lower than technical one. Frequency distributions
of efficiencies are showed in Table 4. The result of the
DEA analysis does not only indicate the efficiency scores of the units, but
also the reference frequencies of units which are the last performing.
In the case of technical efficiency the highest frequency accounts for range
of 0.91, while for allocative efficiency range of 0.70.8 has the most frequency,
indicating that Sistanian harvesters fall short to allocate inputs in a costreducing
way (Table 4).
Due to the nonparametric structure of the data (unbalanced distribution),
the Mann Whitney Utest was used (instead of regular ttest) for statistical
analysis in order to compare the mean efficiency between groups of vessels.
Table 2: 
Returns to scale of the harvests 

Table 3: 
Mean, minimum and maximum of technical, allocative and economical
efficiencies 

Table 4: 
Frequency distribution of technical, allocative and economic
efficiency 

The result of statistical analysis indicates that, on the whole, vessels with
higher than 2 tons capacity technically and economically are more efficient
(p<0.01). This result is similar to Sharma and Leung (1999), Tigley et al. (2005)
and Esmaeili (2006) findings. In addition, the technical efficiency in vessels
which their skippers participated in extension classes were significantly more
that other ones (p<0.05). Finally, vessels with financial security are allocatively
and economically less efficient than other group (p<0.05). This difference
might be due to risk averse of skippers who had financial limitation.
DISCUSSION
Estimation of the efficiency in fisheries is becoming important as the distribution of efficiency has a considerable impact on the effectiveness of effort controls and profitability of fisheries. In this paper, using DEA approach and under two assumptions of constant and variable returns to scale, technical, allocative and economic efficiencies of Sistanian fish harvesters were estimated. In addition, two components of technical efficiency including net technical efficiency and scale efficiency were calculated. Relatively few examples of the case of DEA approach in fisheries exist for the purpose of estimating efficiency. The use of the DEA index is free of distributional and production related assumptions in its derivations and is therefore not subject to the same potential for bias (Pascoe and Herrero, 2004).
Based on the results, the mean of technical, allocative and economical efficiencies were obtained 82.7, 75.5 and 62.7%, respectively. However, differences in efficiency among vessels were small in this study; differences in efficiency are often attributed to differences in technology and the skill of the skipper (Sharma and Leung, 1999; Herrero and Pascoe, 2003). The technology used in the vessels is similar, so skipper skill is likely to be the major factor affecting vessels efficiency. Although, all information on skipper specifications is not available for all vessels in this study, but according to available data, vessels which their skippers participated in extension classes and those did not have financial limitation were more efficient. In addition, bigger capacity vessels technically and economically are more efficient. This result is similar to Sharma and Leung (199), Tigley et al. (2005) and Esmaeili (2006) findings.
Result from both technical and allocative inefficiencies, indicated that harvests could perform economically efficient by rising technical and allocative efficiencies.
Present findings also showed that allocative efficiency is less than technical one by a little, indicating the disability of harvesters in reducing costs. Lower amount of scale efficiency is of main cause of low technical efficiency, therefore the fisherman can enhance their technical efficiencies by using more inputs.