Trends in capacity utilisation in the English Channel1
Diana Tingley, Sean Pascoe and Simon Mardle
Centre for the Economics and Management of Aquatic Resources (CEMARE),
University of Portsmouth, UK.
A key component of the Structural Policy of the CFP involves the reduction of fishing
capacity in order to bring this in line with the reproductive capacity of the stocks. Capacity
reduction targets, defined in terms of physical input use under the multi-annual guidance
programme (MAGP), are largely based on the believed level of biological overexploitation of
the stocks. An alternative indicator of the extent of excess capacity in a fishery is the level of
capacity utilisation. This is the ratio of actual to potential catch, and is an output, rather than
input, approach to capacity measurement. In this study, trends in capacity utilisation are
examined for some key fleet segments operating in the English Channel. The level of capacity
utilisation is estimated using Data Envelopment Analysis.
Paper presented at the XII Conference of the European Association of Fisheries Economists,
Salerno, Italy, 18-20 April 2001.
This study is part of the EU funded project ‘Measuring capacity in fisheries industries using
the data envelopment analysis (DEA) approach’ (DGFISH-99/005).
1
Introduction
The measurement and management of fishing capacity has become a major international
theme in fisheries management over the last few years. This is reflected in the number of
international conferences and workshops dedicated to capacity measurement and management
(e.g. FAO 1998, 2000) and the development of an "International Plan of Action on the
Measurement of Fishing Capacity" (FAO 1999).
In the EU, capacity management has been an important feature of the Structural Policy of the
Common Fisheries Policy (CFP). In most EU countries, fleet reduction has been required
through a decommissioning scheme known as the Multi-Annual Guidance Programme
(MAGP). Several separate (but consecutive) programmes have been run since 1983. Prior to
1992, the aim of the programme was largely to contain fleet capacity and prevent effort from
expanding. Since 1992, the aim of the programme has been to reduce the fleet capacity in
each member state, measured in terms of total engine power and gross tonnage, to target
levels that are assumed commensurate with the restriction on harvesting imposed through the
total allowable catches. Generally, older vessels have been targeted for removal. Countries
that exceed their capacity reduction targets may allow new vessels into the fishery, with fleet
modernisation being another feature of the Structural Policy.
Capacity management requires a measure of the existing level of capacity, as well as some
target level of capacity. In the EU, capacity is measured in terms of physical inputs,
principally engine power, gross tonnage and days fished. Capacity targets are set for each
country for different fleet segments in terms of these three input levels under the MAGP, with
particular emphasis on the first two inputs.
Implicit in these measures is a relationship between the level of inputs and outputs of the
fishery. Capacity reduction targets are primarily set on the basis of the level of
overexploitation of stocks harvested by different fleet segments. It is assumed that a
percentage decrease in physical inputs will result in a proportional decrease in outputs (i.e.
effectively assuming constant returns to scale).
The effects of a reduction in the physical inputs employed in the fishery on the level of output
will, to a large extent, depend on the level of utilisation of these inputs. If the inputs are not
fully utilised, then fleet reduction may have little or no effect on the output of the fishery, as
the remaining boats may increase their individual outputs through increased capacity
utilisation. As boat numbers decrease in a fishery, crowding effects also decrease, resulting in
an increased output per unit of (nominal) effort and hence encourage an increase in individual
Page 1
fishing effort2. As a result, fleet reduction programmes are only successful in reducing catch if
average capacity utilisation is high in the fishery (such that the remaining boats are unable to
increase their effort).
Capacity utilisation refers to the ratio of actual to potential output. A measure of capacity
utilisation less than one implies that the same fleet, if fully utilised, could produce more than
it is currently doing. Conversely, the same level of catch could have been taken by a smaller
fleet if fully utilised. As a result, capacity underutilisation is also an indicator of existence of
excess capacity in a fishery, and the measure can be used to provide an indication of the
extent of excess capacity.
From the above, the measurement of capacity utilisation can provide valuable information
relevant to capacity management. A range of methods have been developed to estimate
capacity utilisation, although the most common is Data Envelopment Analysis (DEA). The
DEA technique has been suggested as the preferred approach to capacity measurement in
fisheries largely as a consequence of being able to measure capacity at the individual species
level in a multispecies fishery (FAO, 2000). In fisheries, the technique has been applied to the
Malaysian purse seine fishery (Kirkley, Squires et al., 1999), US Northwest Atlantic sea
scallop fishery (Kirkley, Färe et al., 1999), Atlantic inshore groundfish fishery (Hsu, 1999),
pacific salmon fishery (Hsu, 1999), the Danish gillnet fleet (Vestergaard et al., 1999), and the
total world capture fisheries (Hsu, 1999).
In this study, capacity utilisation is estimated using DEA for a number of different UK fleet
segments operating in the English Channel. A range of different measures of capacity
utilisation are made based on different output measures (both composite single outputs as well
as multiple outputs). Trends in capacity utilisation over the period for different size classes of
boats are also examined.
The Fisheries of the English Channel
The English Channel contains of a wide variety of fishing activities that are aimed at targeting
a variety of species. Approximately 4000 boats operate within the English Channel, over half
of which are UK boats (Tétard et al., 1995). These broadly fall into 7 gear types: beam trawl,
otter trawl, pelagic/mid-water trawl, dredge, line, nets and pots. In total, 92 species are landed
by boats operating in the English Channel. However, the majority of the landed weight and
value are made up of less than 30 species.
The optimal (profit maximising) level of effort employed in the fishery by an individual is
the level of effort at which marginal revenue per unit of effort equals its marginal cost.
Decreased crowding increases the marginal revenue per day fished, thereby increasing the
optimal number of days fished (assuming marginal cost per day does not change).
2
Page 2
Capacity management in the Channel is based primarily on a unitisation scheme. Each vessel
is required to hold a number of vessels capacity units (VCUs) based on the size and engine
power of their boats, such that
VCU  l  b  0.45kW
(1)
where l is length of the boat (in metres), b is the breadth (in metres), and kW is the engine
power (in kilowatts). In order for a new boat to enter the fishery, sufficient VCUs need to be
purchased from other fishers to meet the requirements of the new boat. Also, an additional
number of VCUs need to be purchased and surrendered under the unitisation policy, the
number of which varies depending on the size of the new boat. The objective of this is to
ensure that total fleet capacity (in terms of VCUs) does not increase as a result of boat
replacement, and is, in fact, reduced to allow for the (presumed) greater efficiency of the
newer boat. The VCUs have also been the basis of the fleet reduction programme in the UK.
A decommissioning programme was established in the UK to meet the capacity reduction
targets set under the MAGP. Under this programme, VCUs were bought back by the
Government with the intention of reducing the overall fleet capacity.
Pascoe, Coglan and Mardle (2001) examined the relationship between VCUs and the
harvesting capacity of two fleet segments in the Channel (gillnetters and otter trawlers in the
western Channel) and found that the capacity output per VCU varies considerably (in terms of
catch composition) between the segments. As a result, transfer of units from one segment to
another may result in a substantial change in the overall catch composition in the Channel.
Consequently, the impact of the decommissioning scheme on the output of key species in the
Channel will depend on from which fleet segments the VCUs were removed. Further, the
relationship between capacity output and the number of VCUs also varied considerably
between the two segments. While VCUs were related to capacity output for the trawlers (at
least for some species), there was little correlation between output and VCUs for the
gillnetters.
As noted above, the objective of this study is to examine the level of capacity utilisation for
different fleet segments in the Channel. This will provide further information on the potential
effectiveness of fleet reduction scheme as a means of reducing the overall output from the
fishery. Low levels of capacity utilisation will reduce the effectiveness of the scheme, as the
remaining vessels can increase their utilisation rate, thereby increasing their output.
Capacity and capacity utilisation measurement using DEA
The measurement of capacity of a firm (e.g. boat) can be described as its potential output
given its fixed factors of production. Therefore, to measure this level of overall capacity, in
practice the potential output of a firm is determined by a comparative analysis of the output
Page 3
levels achieved by other firms of similar size with similar activities. Differences in output
between similar firms can be due to either differences in capacity utilisation or differences in
technical efficiency, both of which are relative measures. Capacity utilisation is the level at
which the firm operates given its level of variable input usage, which may be less than
possible under normal working conditions. Technical efficiency on the other hand is the
degree to which the potential output is achieved given the amount of both variable and fixed
inputs employed. For example, in the case of a fishery, differences in the catch of two boats of
the same size may be due to a difference in the number of days fished (capacity utilisation), or
a difference in the ability of the skipper in harvesting the resource (technical efficiency).
Therefore, in order to determine the potential output of a boat under normal operating
conditions, these effects need to be separated out.
DEA is a non-parametric approach to the estimation of capacity and technical efficiency. An
advantage of DEA is that it is able to incorporate multiple outputs directly in the analysis.
Further, the technique does not require any pre-described structural relationship between the
inputs and resultant output, which allows greater flexibility in the frontier estimation. A
disadvantage of the technique, however, is that it does not account for random variation in the
output(s), and so attributes any apparent shortfall in output to either capacity under-utilisation
or technical inefficiency.
The following example takes a two output example to demonstrate DEA for the estimation of
capacity and capacity utilisation. The illustrated example describes five boats (j =
{A,B,C,D,E}) targeting. In terms of fixed input use, the fleet is homogeneous. Therefore, the
level of catch is determined by the extent to which the fixed inputs are fully utilised. Figure 1
shows the catch (uj,m) achieved by the boats for both species (m = {1,2}). The production
possibility frontier is defined by boats A, B C and D, which as they lie on the frontier are
assumed to be operating at full capacity. However, boat E is producing less of both species
relative to the frontier and is therefore assumed to be operating at less than full capacity. The
production potential of boat E can be found by expanding the output of both species radially
from the origin until it reaches the frontier (point E*). OE*/OE is the expansion factor () by
which output of boat E could be increased. Ccapacity utilisation of boat E is given by
OE/OE* (i.e. 1/).
The shape of the frontier will differ depending on the scale assumptions that underlie the
model. Two scale assumptions are generally employed: constant returns to scale (CRS) and
variable returns to scale (VRS). The latter encompasses both increasing and decreasing
returns to scale. However, there are generally a priori reasons to assume that fishing would be
subject to variable returns, and in particular decreasing returns to scale. Figure 2 shows the
differences between these alternative measures for the five boats in the example above. In the
analysis in this paper, the frontier is assumed to follow the form of a VRS model where zero
Page 4
inputs equates to zero outputs. Hence, the frontier would go through the points OBCD and
would not be defined by the standard VRS envelope ABCD as shown.
Figure 1. Two output production
possibility frontier
uj,1
Figure 2. CRS and VRS efficient frontiers
Output
A
CRS frontier
B
C
E*
D
VRS frontier
O3
O2
B
E
C
E
O1
A
D
O
uj,2
Fixed input
The VRS DEA model is formulated as a linear programming (LP) model, where the value of
 for each vessel can be estimated from the set of available data. Following Färe et al. (1989,
1994) this DEA model of capacity output given current use of inputs is given as:
Max 1
subject to
 1u 0 , m   z j u j , m
m
j
z
n 
j
x j ,n  x0,n
z
j
x j ,n   0,n x 0,n
z
j
j
n  ˆ
(2)
j
1
j
z j  0,
 j ,n  0 n  ˆ
where  1 is a scalar showing by how much the output of each boat can be increased, uj,m is the
output m produced by boat j, xj,n is the amount of input n used by boat j and zj are weighting
factors measuring the distance boat j is from the frontier. The value of  1 is estimated for each
vessel separately, with the target vessel’s outputs and inputs being denoted by u0,m and x0,n
respectively. Inputs are divided into fixed factors (i.e. set  ) and variable factors (i.e. set ̂ ).
The measure of capacity output is calculated by relaxing the bounds on the sub-vector of
variable inputs, x̂ . This is achieved by allowing these inputs to be unconstrained through
introducing an input utilisation rate (  j , n ). This is estimated in the model for each boat j and
Page 5
variable input n (Färe et al., 1994). The restriction  z j  1 allows for variable returns to
j
3
scale . Hence, capacity utilisation (CU) is defined as:
CU  1/ 1
(3)
The measure of CU ranges from zero to 1, with 1 being full capacity utilisation (i.e. 100 per
cent of capacity).
Due to random variations in the catch being measured as under-utilisation rather than
stochastic error, the estimated capacity utilisation may be biased downward (and capacity
output biased upwards). Further, the observed outputs may not be produced efficiently (Färe
et al., 1994), and hence some of the apparent capacity under-utilisation may be due to
inefficiency (i.e. not producing the full potential given the level of fixed and variable inputs).
If all inputs (both fixed and variable) are not being used efficiently, then it would be expected
that output could increase without an increase in the level of variable inputs through the more
efficient use of these inputs. By comparing the capacity output to the technically efficiency
level of output, the effects of inefficiency can be separated from capacity under-utilisation. As
both the technically efficient level of output and capacity output can be upwardly biased due
to random variability in the data, the ratio of these measures is a less biased (both statistically
and theoretically) measure of capacity utilisation.
The technically efficient level of output requires an estimate of technical efficiency of each
boat, and requires both variable and fixed inputs to be considered. The VRS DEA model for
this technically efficient measure of output is given as:
Max  2
subject to
 2 u 0,m   z j u j ,m  m
j
z
j
x j ,n  x0,n
z
n
(4)
j
j
1
j
zj  0
where  2 is a scalar outcome showing how much the production of each firm can increase by
using inputs (both fixed and variable) in a technically efficient configuration. In this case,
In contrast, excluding this constraint implicitly imposes constant returns to scale while zj1 imposes nonincreasing returns to scale (Färe et al., 1989).
3
Page 6
both variable and fixed inputs are constrained to their current level. In this case,  2 represents
the extent to which output can increase through using all inputs efficiently. The technically
*
efficient level of output ( uTE
) is defined as  2 multiplied by observed output (u). As the level
of variable inputs is also constrained,  2  1 and the technical efficient level of output is less
*
than or equal to the capacity level of output (i.e. uTE
 u * ). The level of technical efficiency is
estimated as:
TE  1 /  2
(5)
Consequently, the unbiased estimate of capacity utilisation (CU*) is estimated by:
CU * 
CU 1

TE  1
1
2

2
1
(6)
As 1   2 , the unbiased estimate CU*  CU.
Data
An extensive database of trip level log-book data covering the period 1993-1998 was
disaggregated into 8 different fleet segments based on recorded fishing activity (beam trawl,
otter trawl, scallop dredging, lining, netting, crab potting, whelk potting and ‘other’
activities). The trip level data were aggregated to provide monthly levels of output and effort
by vessel over the period examined. In total, the combined sub-data sets contain over 150,000
observations (Table 1).
Table 1. Summary of available data
Gear
Total data set
Boats Observations
Average per year
Boats Observations
Catcha
Valueb
(tonnes)
(£'000)
Beam trawl
247
18,702
139
3117
1595
14,953
Otter trawl
529
99,594
243
16,599
1913
10,451
Scallop Dredge
215
12,504
84
2084
5277
8728
Pots
178
5021
59
837
1784
3084
Gillnets
337
13,451
140
2242
707
1959
Longline
209
5142
63
857
628
366
a) The catch has been weighted by revenue shares. b) Values have been inflated to 1998 values using a Fisher
price index
The key inputs used in the analysis were days fished, 'deck' size (estimated as
length4*breadth, comparable to the first part of the VCU definition in equation 1) and engine
4
Overall vessel length was used as opposed to regulation length
Page 7
power (in kW). A range of alternative output measures were used in the analysis. CU was
estimated using both single composite outputs and multiple outputs. The two composite
outputs were catch weight (a composite measure of individual catch of each species weighted
by its average revenue share for all boats operating with that gear in that year), and revenue.
The latter was inflated to 1998 values using a Fisher price index. A different Fisher price
index was estimated for each fleet segment, representing the different combination of species
in the catch.
For the multiple output measures, the top five species in terms of value for each type were
used individually, with the other species aggregated into a composite 'other' category. Again,
the 'other' catch (in weight terms) category was derived using revenue shares, and all revenues
were inflated to 1998 values.
The number of observations varied from year to year. Many boats are multi-purpose,
particularly the smaller boats, so the number of boats using a particular gear type varied from
year to year and over the year. As some boats will operate for only a relatively short time
period using a particular gear type trawl gear, the final data set used in the analysis was a
subset of the data set presented in Table 1. Only boats that used the gear for at least 4 months
a year and in at least 3 of the six years were used in the final analysis. This resulted in many
of the multi-purpose boats being excluded from the analysis, such that the resultant data set
consisted of boats that primarily used one gear throughout the year. As a result, the data set
used in the analysis was substantially smaller that the set of available data (Table 2).
Table 2. Summary of consolidated data sets used in the analyses
Total in data set
Gear
Boats
Number of
Catchb
Obs.
(Kg)
Beam trawl
101
6840
4801
Otter trawl
171
8215
1141
Scallop Dredge
37
1553
14,806
Pots
28
916
6030
Gillnets
51
2276
2618
Longline
15
452
6270
a) The catch has been weighted by revenue shares. b) Values
price index
Average per boat per month
Valueb
Days
Deck area
Engine
(£)
fished
power (Kw)
10,352
6
160
422
6454
12
56
155
21,860
11
143
375
8121
9
43
95
3444
5
50
127
2910
8
40
186
have been inflated to 1998 values using a Fisher
Results
The results of DEA calculations of CU, TE and unbiased CU for five major gear types
averaged over the period 1993-98 are shown in figure 3. The number of observations in the
data sample for longliners was felt to be too small and so results are not presented for this
gear type.
Page 8
The results were produced using a linear programming model developed in GAMS. As data
on stock abundance are not available the model is run separately for each time period, i.e. one
month, with all vessels fishing in the same area5, in the same month, being compared to each
other to determine which vessels lie on the full efficiency or full capacity utilisation frontier
and for those that lie within it, how far inside it they are found. It is assumed that stock levels
will not vary considerably during one month hence lack of stock abundance data is not
perceived to be a significant problem.
It is apparent from Figure 3 below that the results for CU and TE calculated using multipleoutput measures based on revenues and weights (‘multiple, revenue’ and ‘multiple, weight’)
are higher than those calculated using composite single-output measure indexes (‘single,
revenue’ and ‘single, weight’) across all gear types. This difference is particularly notable
between CU and TE scores generated for each gear type.
When unbiased CU scores are calculated the difference between single-output and multioutput measure results is less dramatic. However, the latter scores were still higher than the
former across each gear type: between 5% and 13% higher. The smallest difference of 5%
was found for potting, whilst the largest difference was for gillnetting. The difference for
scallop dredging was 8%, otter trawling 9%, and beam trawling 11%. No pattern to these
results between gear types is discernible.
5
While it is possible for boats to operate in two areas in the same month, resulting in a lower CU in each area
relative to the boats that only fished in one area, the incidence of this in the consolidated data sets used was
relatively small (i.e. less than 10 percent of the observations).
Page 9
Figure 3 Comparison of DEA results, by type of output measure and gear type (1993-98)
Otter trawl
Beam trawl
1.0
1.0
0.9
0.9
0.8
0.8
0.7
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
single, revenue
single, weight
multiple, revenue
multiple, weight
single, revenue
single, weight
multiple, revenue
multiple, weight
Scallop dredge
Key:
1.0
0.9
CU
0.8
TE
unbiasedCU
0.7
0.6
0.5
Single, revenue – single composite output based on revenues
0.4
Single, weight – single composite output based on weights
0.3
Multiple, revenues – multiple outputs based on revenues
0.2
Multiple, weights – multiple outputs based on weights
0.1
0.0
single, revenue
single, weight
multiple, revenue
multiple, weight
Pots
Gillnets
1.0
1.0
0.9
0.9
0.8
0.8
0.7
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
single, revenue
single, weight
multiple, revenue
multiple, weight
single, revenue
single, weight
multiple, revenue
multiple, weight
The unbiased CU scores for each gear type in turn are shown in figure 4. These have been disaggregated by vessel length categories to focus on small inshore vessels (less than 10m in
overall length), medium-sized vessels (10 to 15.9m) and large vessels (greater than 16m).
Note should be taken of the sample sizes used to provide average results for each vessel
length category; for example, the average unbiased CU results for medium-sized otter
trawlers was calculated from the results of between 102 to 151 vessels on average for each
year, whilst results for the larger vessels were produced using data from between 13 to 21
vessels each year over the period 1993-98.
Page 10
Beam trawl
Otter trawl
The results1.0 for the mobile gear types (otter trawl, beam trawl and scallop dredge) show that
Scallop dredge
the1.0 medium
sized
vessels were operating more closely to optimum unbiased CU levels than
0.9
5-14 vessels
8-17 vessels The difference was very significant for the beam trawl and otter trawl
the0.9 larger
vessels.
102-151 vessels
0.8
vessels
across
the whole period. The difference was also clear for scallop dredges.
0.7
0.8
1.0
0.9
0.8
0.7
0.6
0.7
13-21 vessels
13-20 vessels
0.6
70-74 vessels
0.5
Pots results, single output revenue index, by gear type and vessel length
Figure
4 Unbiased
CU
0.6
1.0
0.4
3-6 vessels
category
(1993-98)
0.5
1993
1994
1995
1996
1997
1998
0.5
0.4
1993
1994
1995
1996
1997
1998
0.9
0.8
0.7
9-14 vessels
0.4
1993
1994
1995
1997
1998
<10m
10-16m
>16m
all
1-4 vessels
0.6
0.5
0.4
1993
1996
1994
1995
Key to vessel lengths:
1996
1997
1998
Gillnets
1.0
0.9
20-33 vessels
3-9 vessels
0.8
0.7
0.6
1-6 vessels
0.5
0.4
1993
1994
1995
1996
1997
1998
Page 11
Results are less clearly defined between vessel length categories for the fixed gear types. It
seems that unbiased CU was highest for medium-sized potting vessels and generally lowest
for smaller potters. As many of the smallest potters operate on a part-time basis, this result is
not unexpected. However the smaller gillnetting vessels have higher unbiased CU scores as
compared to the largest gillnetters which have the lowest scores. Attention should be paid to
the numbers of vessels proving data for analysis in each length category.
While capacity utilisation fluctuated from year to year, there appeared to be a general, gentle
upwards trend in average annual unbiased CU for all major gear types between 1993 and
1998 (Figure 5). The rise between 1993 and 1998 was only 1% for scallop dredgers, but 7%
for otter trawlers, 9% for beam trawlers, 10% for potters and 12% for netters.
Overall potters achieve the highest unbiased CU scores, consistently throughout the period,
with the exception of 1996. In 1998 potters appear to achieve, on average, unbiased capacity
utilisation levels of over 90%. Scallop dredgers appear to achieve the next highest consistent
score although are just overtaken in 1998 by otter trawls being the third nearest type to
achieve full capacity utilisation levels. Gillnetters achieve unbiased capacity utilisation levels
of just over 80% in 1998 whilst beam trawlers appear to have the lowest unbiased CU
averaging a score of 68% over the period and 75% in 1998.
Figure 5 Average annual unbiased CU by major gear type (1993-98)
1.0
unbiased CU score
0.9
0.8
scallop dredgers
beam trawlers
netters
potters
otter trawl
0.7
0.6
0.5
1993
1994
1995
1996
1997
1998
It is interesting to note that numbers of vessels included in the analysis generally decreased
between 1993 and 1998 numbers. As only records of vessels fishing a particular gear type for
Page 12
4 or more months per year and for at least 3 years in the 1993-98 period were included in the
analysis, their numbers used in the analysis provide a very crude measure of fishing effort.
With the exception of scallop dredging and beam trawling, vessel numbers used in the
analysis decreased between 1993 and 1998; by 28% of otter trawlers, 22% of potters and 32%
for netters. Numbers of beam trawlers in the sample increased by only 3% over the period
whilst numbers of scallop dredgers increased by 44%.
Discussion and conclusions
The results of the above analysis show some interesting features that are relevant to such an
analysis in other fisheries. Foremost of these is the similarity between the unbiased capacity
utilisation scores for both the revenue and catch based measures. In all cases, the difference
between the unbiased CU for the two output measures was small. A greater difference was
observed between single and multi-output measures, with the later demonstrating higher
average unbiased CU.
This last result is likely due to the fact that that the multiple output measures provide six
pieces of output information for the main species landed (five main species and a sixth
composite measure of all other landings) against which each vessels’ activity can be ranked to
determine where the efficient ‘frontier’ lies and how each vessel compares to it, given its
input level. As compared to the single output measures, much more information is available in
multi-output analysis thus allowing the estimation of efficiency and capacity utilisation to be
determined more accurately.
Further to this, single output measures are more vulnerable to random fluctuations in the catch
of one particular species, whereas multi-output data incorporates information across the range
of key species into the analysis. This has the effect of reducing the influence of random
fluctuations on the comparative process, fundamental to DEA, so providing more accurate
results. For example, under analysis using a single measure revenue output, if one vessel
caught very large amounts of a high value species whilst other vessels fishing the same gear
in the same month caught ‘normal’ amounts of this species (and all vessels caught normal
amounts of other species), the other vessels would be ranked much less efficient if a single
measure output is used as compared to a multi-output measure. In this situation a multi-output
measure would determine that the other vessels caught normal amounts of the other species,
but just were not as lucky to catch so much of the high value species, thus their DEA scores
would be higher and more accurate under the multi-output measure analysis, as compared to
with the single-output measure.
The results of the analysis suggest that the existing fleet could increase its output by up to 30
percent by increasing its capacity utilisation. Conversely, the same output could have been
Page 13
taken by a smaller, fully utilised fleet. The general upward trend in unbiased CU is perhaps
due to decommissioning schemes in the English Channel which operated up to 1997. The
reduced crowding as a result of the scheme would have resulted in economic incentives for
the individuals remaining in the fishery to increase their level of effort.
Future analysis
Of key interest to managers are the factors that cause CU to change. Results from the
estimation of unbiased CU scores can be regressed6 against a range of factors to determine if
any factors, other than the inputs used (days at sea, ‘deck’ area and engine power) drive the
results. Factors which may influence the results include: total effort in the fishery which
provides a representation of the extent to which it is crowded; changes in key prices, i.e. of
fuel and fish; and, port data detailing localities to fishing grounds.
The possible impact of management measures implemented in English Channel fisheries,
either by the local Sea Fisheries Committees or at the national or international (EU) level,
needs to be examined. Particular features to be considered include the introduction of square
mesh size limits, decommissioning programmes implemented as part of the Multi-annual
Guidance Programmes and restrictions on ‘days at sea’.
6
As the value of CU is limited to be less than 1, appropriate limited dependent variable regression techniques
need to be applied.
Page 14
References
Brooke, A.D., Kendrick, D. and Meerhaus, A. 1992. GAMS: A User’s Guide. Scientific Press,
California.
Campbell, H.F. and Lindner, R.K. 1990. The production of fishing effort and the economic
performance of license limitation programmes. Land Economics, 66:55-66.
Coglan, L., Pascoe, S. and Harris, R.I.D. 1999. Measuring efficiency in demersal trawlers
using a frontier production function approach. In Proceedings of the Xth Annual
Conference of the European Association of Fisheries Economists, The Hague,
Netherlands, 1-4 April 1998, pp. 236-257 Ed by P.Salz, Agricultural Economics
Research Institute (LEI), The Netherlands.
Cooper, W.W., Seiford, L.M and Tone, K. 2000. Data Envelopement Analysis: a
comprehensive text with models, applications, references and DEA-solver software.
Kluwer Academic Publishers, USA.
FAO, 1998. Report of the Technical Working Group on the Management of Fishing Capacity,
La Jolla, Callifornia, United States, 15-18 April 1998, FAO Fisheries Report No. 586,
FAO, Rome, 1998.
FAO, 1999. International Plan of Action on the Measurement of Fishing Capacity, FAO
Rome, 1999.
FAO. 2000. Report of the Technical Consultation on the Measurement of Fishing Capacity.
FAO Fisheries Report No. 615 (FIPP/R615(En)). FAO, Rome.
Färe, R., Grosskopf, S. and Lovell, C.A.K. 1994. Production Frontiers. Cambridge University
Press, UK.
Färe, R., Grosskopf, S.and Kokkelenberg, E.C. 1989. Measuring plant capacity, utilization
and technical change: a non-parametric approach. International Economic Review,
30(3): 655-666.
Hannesson, R. 1983. Bioeconomic production function in fisheries: Theoretical and empirical
analysis. Canadian Journal of Fisheries and Aquatic Science, 13(3): 367-375.
Holland, D.S. and Lee, S.T. 2001, Impacts of random noise and specification on estimates of
capacity derived from Data Envelopment Analysis. Journal of the Operational
Research Society, [in press].
Hsu, T., 1999. Simple capacity indicators for peak to peak and data envelopment analyses of
fishing capacity, FAO Technical Consultation on the Management of Fishing
Capacity, Mexico City, Mexico 29 November – 3rd December, 1999.
Johansen, L. 1968. Production functions and the concept of capacity. Recherches recentes sur
la fonction de production, Centre d’Etudes et de le Recherche Universitaire de Namer.
Kirkley, J.E. and Squires, D.E. 1999. Measuring capacity and capacity utilization in fisheries.
In: Gréboval, D. (ed) Managing Fishing Capacity: Selected Papers on Underlying
Concepts and Issues. FAO Fisheries Technical Paper 386. FAO, Rome.
Page 15
Kirkley, J.E., Färe, R., Grosskopf, G., McConnell, K., Squires, D.E., and Strand, I. 1999.
Assessing capacity and capacity utilization in fisheries when data are limited, FAO
Technical Consultation on the Management of Fishing Capacity, Mexico City, Mexico
29 November – 3rd December, 1999.
Kirkley, J.E., Squires, D. and Strand, I.E. 1995. Assessing technical efficiency in commercial
fisheries: The Mid-Atlantic Sea Scallop Fishery. American Journal of Agricultural
Economics 77: 686-97.
Kirkley, J.E., Squires, D. and Strand, I.E. 1998. Characterizing managerial skill and technical
efficiency in a fishery. Journal of Productivity Analysis 9: 145-160.
Kirkley, J.E., Squires, D.E., Alam, M.F., and Omar, I.H. 1999. Capacity and offshore
fisheries development: the Malaysian purse seine fishery, FAO Technical Consultation
on the Management of Fishing Capacity, Mexico City, Mexico 29 November – 3rd
December, 1999.
Pascoe, S. and Coglan, L. 2000. Implications of differences in technical efficiency of fishing
boats for capacity measurement and reduction. Marine Policy, 24(4): 301-307.
Pascoe, S., Coglan, L. and Mardle, S. 2000. Physical versus harvest based measures of
capacity: the case of the UK vessel capacity unit system. ICES Journal of Marine
Science [in press]
Pascoe, S. and Robinson, C. 1998. Input controls, input substitution and profit maximisation
in the English Channel beam trawl fishery. Journal of Agricultural Economics, 49(1):
16-33.
Pascoe, S., Andersen, J.L. and de Wilde, J.W. 2001. The impact of management regulation on
the technical efficiency of vessels in the Dutch beam trawl fishery. European Review
of Agricultural Economics, 28(2) [in press]
Sharma, K.R and Leung, P. 1999. Technical Efficiency of the Longline Fishery in Hawaii: An
application of a Stochastic Production Frontier. Marine Resource Economics 13: 259274.
Squires, D. 1987. Fishing effort: Its testing, specification, and internal structure in fisheries
economics and management. Journal of Environmental Economics and Management.
14(September): 268-282.
Tétard, A., M. Boon, et al. 1995. Catalogue International des Activities des Flottilles de la
Manche, Approche des Interactions Techniques, Brest, France, IFREMER.
Vestergaard, N., Squires, D. and Kirkley, J.E. 1999. Measuring capacity and capacity
utilization in fisheries: the case of the Danish gillnet fleet, FAO Technical
Consultation on the Management of Fishing Capacity, Mexico City, Mexico 29
November – 3rd December, 1999.
Page 16