Anticipating fisher response to management: can economics help?

fancyfantasicΤεχνίτη Νοημοσύνη και Ρομποτική

7 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

123 εμφανίσεις


1

CM 2005/


Anticipating fisher response to management: can economics
help?


Sean Pascoe and Simon Mardle

CEMARE, University of Portsmouth, Boathouse 6, College Road, HM Naval Base,
Portsmouth PO1 3LJ. Tel: +44 23 92844242; Fax: +44 23 92844614; Email:
sean.
pascoe@port.ac.uk; Web: www.port.ac.uk/cemare



Abstract

Although a key objective of fisheries management is to conserve the fisheries resource, management cannot
directly influence the state of the resource. Instead, management can only influence the acti
ons of those
exploiting the resource. This may be through limiting catches, effort, area fished or gear used. In each case, the
response from the fisher may not be exactly as anticipated by management, as the fisher responds to the new set
of economic ince
ntives created by the change in management. In this paper, recent developments in modelling
fisher response to changes in management are reviewed. The key areas examined in the paper are effort
allocation, entry and exit behaviour, and compliance with fish
eries regulations.


Introduction


Fishing in anything other than a subsistence
-
based economy is an economic activity. Fishing
effort is targeted towards species that have a value to consumers (represented by the price
consumers are willing to pay), with s
pecialist fishing gears (e.g. prawn trawls, scallop
dredges, lobster pots) being developed to target the most valuable species. Similarly, the
development and adoption of new technologies and increase in average boat size and power
are direct consequences
of fishers’ desire to increase the profitability of their activity.


A consequence of this economic activity has been, perversely, a reduction in the potential
yields that can be achieved from the stock and the dissipation of potential economic benefits
th
at the fishery could produce. FAO (2004a) estimated that in 2003, 25 per cent of world fish
stocks were overexploited or depleted and in need of rebuilding. Other studies suggest that
this is an underestimate of the level of overexploitation. Myers and Wor
m (2003) estimated
that the total biomass in most commercial fisheries is less than 20 per cent of its unexploited
level, while stocks of larger predators are less than 10 per cent of their pre
-
industrial levels.


Fisheries management has been introduced i
n some form in most countries in an attempt to
either limit further resource degradation or aid in the recovery of depleted stocks. A recent
survey of FAO Member States (FAO, 2004b) found that 90% of respondents had introduced
at least some form of limited

access to their fisheries resources. A key objective in most
fisheries management plans is the conservation of the resource, although social and economic
factors are also often considered when developing management targets. Economic
considerations, howeve
r, have largely been limited to short term regional economic impacts
of effort reduction (employment in particular), and these have often been used as a reason to
moderate effort or catch reductions proposed by fisheries scientists.


Where management rest
rictions have been imposed, these have largely focused on limiting
the level of activity, either by restricting fishing effort, level of landings, area fished or the
type of gear used. From the FAO (2004b) survey, of those Member States that had introduced


2

some form of management, between 80% and 90% had implemented area/seasonal closures
and/or gear restrictions. Over 50% had also introduced some form of catch restriction.


These types of restrictions have generally been termed “incentive blocking” (Grebov
al and
Munro, 1999) as they aim to prevent fishers from behaving in a way that the economic
incentives that they face would otherwise encourage. However, a more appropriate
terminology may be “incentive altering”, as these measures change the incentives fa
cing
fishers, resulting in behaviour unintended by the managers. For example, a restriction on time
fished may result in fishers adopting more powerful engines to overcome the restriction (the
Pacific halibut fishery is an obvious example of such a phenome
non). How fishers react in
response to such regulations can greatly affect the success of management in achieving its
objectives.


The response of fishers to management change is a major source of uncertainty in
implementing fishery management decisions (L
ittle et al., 2004). The way in which
management alters the incentives facing fishers, and their consequent behaviour, has been an
area of considerable interest in fisheries economics. Considerable attention has been focused
on developing theoretical model
s of behaviour such as entry/exit, input substitution, capital
stuffing and effort allocation. In many cases, these theoretical models have been used to
explain actual observed behaviour in fisheries. However, these models have only been useful
in predicti
ng response to proposed management changes in a qualitative manner. That is, the
direction of change can be estimated, but the magnitude and impact of the change on the
fishery system has remained highly uncertain.


The development of models of fisher beha
viour is essential for predicting, understanding and
designing efficient regulatory programs (Wilen, 1979). Ideally, these models should be able to
be integrated into management evaluation frameworks in order to provide improved estimates
of the biological

and economic impact of management. In this paper, recent experiences in
developing empirical models of fisher behaviour that could potentially be incorporated into
bioeconomic models for management evaluation are reviewed.


The basic economic premise of
fisher behaviour


Economic models of fisher behaviour


both theoretical and empirical


are based on the
general premise that the key objective of the individual fisher is to maximise his/her
individual profits. Profit maximising behaviour does not necess
arily mean that fishers actually
obtain the highest level of profits possible. Instead, they respond in a way that would
potentially increase their profitability. For example, fishers will switch gear if the benefits
from the use of an alternative gear exc
eed the benefits of the current gear and the costs of
switching gear. Similarly, fishers will not go to sea if the expected revenue from the catch
does not cover the fuel and other running costs associated with the trip (as doing so would
reduce their prof
it).
1



There is considerable evidence to support the profit maximising behaviour. Robinson and
Pascoe (1998) interviewed fishers about their motivation and responses to certain situations,
and found these responses were consistent with profit maximising b
ehaviour even though the
fishers did not consider themselves to be profit maximisers. Pascoe and Robinson (1998)



1

In economic terms, short run profit maximizing behaviour involves continuing to

operate as long as the
marginal revenue exceeds the marginal cost.


3

demonstrated that input substitution observed in the English Channel beam trawl fleet
following restrictions on engine power was also consisten
t with profit maximising behaviour.
Similarly, Pascoe and Tingley (2005), looking at levels of capacity utilisation in the Scottish
fleet, found that most fishers operated at the point where marginal revenue was equal to
marginal cost


consistent with pro
fit maximisation.


The assumption of profit maximising behaviour does not preclude the existence of other
social constraints. For example, Roy (1998) considered the marginal cost of an additional
week fished to include not only the cost of the fuel and ot
her variable inputs, but also the
forgone leisure time (the value of which increased as the quantity available decreased).
Similarly, other social restrictions (e.g. birthdays, holidays etc) can affect the level of fishing
activity. However, within these c
onstraints, behaviour has generally been considered profit
maximising.


The assumption of profit maximising behaviour does not mean that all fishers will be
profitable nor that all fishers with similar vessels will operate at the same level or in the same
way. Numerous studies have been undertaken looking at variations in performance of fishing
vessels in terms of the technical efficiency and capacity utilisation. These studies have
concluded that a major cause of variation in output and activity level of i
ndividual vessels is
the level of technical efficiency (see for example Vestergaard
et al,

2003; Tingley et al., 2003;
Kirkley et al.,

2003), while studies of technical efficiency have concluded that skipper skill is
a major factor (see for example Kirkley

et al, 1998; Pascoe and Coglan, 2002). Variations in
performance could be attributed to the lack of perfect information and the existence in many
cases of deliberately imperfect information (for example, by fishers lying to each other about
where they cau
ght their catch) (Allen and McGlade, 1987). Hence, skippers may be assumed
to attempt to maximise profits, although some will be better at this than others.


A number of alternative hypotheses have been proposed to explain fisher behaviour. In
particular,
habit has been thought to be characteristic of fisher behaviour (e.g. Holland and
Sutinen, 2000). That is, fishers are assumed to prefer to fish in the same areas with the same
gear year after year. Similarly, Shepherd and Garrod (1981) and Placenti et al
(1992) assumed
“inertia” existed in the fishery, with major improvements being necessary to encourage
fishers to change their behaviour. In some studies, this “habit” or “inertia” has been linked to
risk aversion (e.g.
Strand, 2004). That is, fishers are a
ssumed to prefer to go where they know
the likely outcome rather than try somewhere new where the outcome is unknown.

However,
while habit, inertia and risk aversion may influence fisher behaviour in the absence of any
changes in their regulatory, economic

or natural environment, any disruption to this
environment is likely to result in a response based on the economic incentives that exist.
Wilen et al (2002) demonstrated that assuming effort allocation or level is relatively fixed (i.e.
due to habit) coul
d result in erroneous conclusions about the effects of fisheries management
changes.


Anderson (2004) examined the implications for fisheries management if fishers operated
under satisficing rather than profit maximising behaviour. That is, they aim to ach
ieve a
certain level of profit rather than maximise profit. Under such an assumption, Anderson
(2004) found that higher levels of fishing effort at low stock levels are likely to occur than if
profit maximising behaviour was assumed.
2

The assumption of sat
isficing behaviour has also



2

Similar outcomes were found in studies where multi
-
objective models were used to determine “optimal” fleet
configurations subject to a variety of management objectives (e.g. see Pascoe a
nd Mardle, 2001; Mardle and
Pascoe, 2002).


4

been applied by Tingley (2005). Given “acceptable” levels of vessel profit and skipper
income


agreed with the fishing industry


Tingley (2005) estimated the maximum fleet size
that could be sustained in the fishery.


While t
he satisficing assumption may be suitable for determining “optimal” vessel numbers, it
is not necessarily an appropriate assumption for considering fisher reaction to management
change. In such cases, profit maximisation remains the most appropriate assump
tion for
estimating the short run response to management change (see Anderson, 1999).


Modelling fisher behaviour


In terms of assessing the impacts of fisheries management, three main areas of research have
attracted the most attention over recent years.
These are fisher location choice (effort
allocation), entry/exit behaviour and compliance. Empirical models of these three types of
fisher behaviour have been based (in some cases implicitly) on the assumption that fishers
will act in a way that is expecte
d to improve their profitability. However, information on
profitability of different activities is limited, so alternative indicators are often used as proxies
for profitability. At a very basic level, models may be based solely on revenue on the
assumptio
n that costs are constant (and hence changes in revenues are directly related to
changes in profitability). In other cases, distance to a fishing ground is used as a proxy for
costs of fishing where cost information is not available.


In this section, the

methodologies that have been applied to analyse these types of behaviour
and examples of their applications are reviewed.


Effort allocation (location choice)


Considerable attention has been focused in recent years on how fishers might respond to an
are
a closure. This has largely been driven by an increasing interest in the use of marine
protected areas or fishing exclusion zones as fisheries management tools, but has also been
examined in the context of fisher reactions to pollution events such as an oi
lspill (e.g. Collins
et al 2003). While fishers may tend to fish in the same areas due to habit, closing these areas
will force them to either move elsewhere or cease fishing. However, assuming that the fishing
effort previously expended in an area will ev
aporate following the area closure is, more than
likely, a naive assumption. Instead, the effort will move to the next best available fishing
ground. For example, in Figure 1, closure of the centre area will force fishers to move to some
new area outside t
he closure. Where the fishers move to will have an effect on their own
economic performance as well as the economic performance of the boats already in those
areas. This can also reduce the effectiveness of the management plan. For example, where the
area
closure is implemented for resource conservation benefits, the effectiveness of the
closure will be compromised if total catches are not reduced due to increased activity outside
of the area.



5

Figure 1. Potential effort displacement from an area closure





A difficulty when examining location choice of fishers is that they are not homogeneous


they come from different locations (as well as fish in different locations), and their
characteristics affects their cost structure. Hence, complications exist


e
xpected economic
returns are not only determined by revenue of catch (i.e. highest catch rates), but also by the
costs associated with the fishing trip. Costs increase as distance travelled and steaming time
increases. As a result, fishers are (within reas
on) able to select which port they fish from and
where they land their catch to maximize the returns for species captured. As noted above,
information on costs of individual vessels and variations in profitability due to spatial activity
is often difficult

to obtain. In the modelling of spatial dynamics, several assumptions have
been proposed. For example, the distribution of fishing effort could be assumed to move
towards areas of highest catches (i.e. reflecting differences in revenues assuming constant
c
osts, e.g. Maury and Gascuel, 1999), highest catch rates modified for distance to port (i.e.
taking into consideration revenues and costs implicitly, e.g. Sampson, 1991) or greatest profit
(Bockstael and Opaluch, 1983; Chakravorty and Nemato, 2001).


A me
thod that allows for heterogeneity in both fishing activity and fisher characteristics is
discrete choice modelling, or the random utility model (RUM). The key features of RUM is
that it models discrete decisions with no assumption of homogeneity amongst i
ndividuals
being required. It is based on the concept of utility, which, in economics, is a measure of the
happiness or satisfaction gained from a good, service or activity. Rational decisions makers
are assumed to make decisions that maximise their level
of utility subject to any constraints.
In the case of effort allocation in fisheries, utility is assumed to relate to profitability (subject
to any constraints the fisher may face). Details of the discrete choice modelling methodology
are presented in the
Annex. The methodological difficulties of applying RUM to the fishing
location choice problem are discussed in detail by Smith (2000).


The method is probabilistic in nature, in that the model estimates the probability of a fisher
operating in a given area

based on the characteristics of the area (e.g. average revenue per unit
effort, distance from port etc) and the characteristics of the fisher. This probability is,
therefore, specific to an individual fisher. The allocation of effort of the individual fis
her to
each area is estimated as the product of the total effort expended by the fisher and the
probability that effort will be applied to each area. The total spatial effort allocation is derived
by summing the effort in each area of the individual fisher
s.



6

Several studies in the field of fisheries have been undertaken that implement a random utility
methodology for the prediction of individual fishery and location choice.
(e.g. Bockstael and
Opaluch, 1983; Eales and Wilen, 1986;
Morey et al, 1993;
Hollan
d and Sutinen, 1999; Wilen
et al, 2002; Hutton et al, 2004).

Bockstael and Opaluch (1983,1984), generally accepted to be
the first to employ a discrete choice framework for the analysis of fisheries issues, investigate
the transfer of vessels from one fish
ery to another, reasoning about the choice of a particular
fishery.

Morey
et al.

(1993) Holland and Sutinen (1999) considered individual vessels using
RUM for location choice of 400 heterogeneous trawlers (trip data) including lagged average
revenue rates
for alternatives as well as past behaviour for the prediction of location choice
and species mix. Overall, aggregate effort levels were predicted for different fisheries and
areas over time. In another application of RUM, Wilen et al. (2002) modelled the C
alifornia
red sea urchin dive fishery, where a diver chooses a home port at beginning of season,
chooses to participate based on weather, prices, expected abundance, diver traits and
processor commitments, then chooses dive locations as well as diving hour
s. Hutton et al.
(2004) examined the UK beam trawl fleet operating in the North Sea, and found that the
number of trips, the average trip length, and the average effort in each ICES rectangle were
significant variables affecting location choice, in additio
n to catch rates (weighted by value)
for the previous year.


Little et al (2004) developed a probabilistic based model of effort allocation, with the
probability based on the value per unit effort of fishing in a given area less the cost of getting
to that

area. This probability, however, was not estimated using the RUM approach. The
potential net profit was estimated for each area, and the results scaled such that the resultant
“probability” was equal to one. Hence, the probabilities did not reflect actual

observed
choices. These probabilities were further updated using a model of information flow between
skippers, dependent upon an estimated “link parameter”.


Although the RUM approach has proved successful in estimating the fisher response to area
closur
es, few attempts have been made to integrate these models into bioeconomic models for
a more broad based assessment of fisheries management changes. Smith and Wilen (2003,
2004) incorporated the probabilities of fishing in each area into a bioeconomic mode
l. Most
of the other analyses cited above were extremely short run in nature, and considered the
immediate impact on profitability and catch of effort relocation following a management
change.


Most attempts at incorporating spatial behaviour into bioecono
mic models have adopted
approaches based on the set of economic incentives facing fishers. An early approach to
modelling spatial movement of fishing effort was proposed by Hilborn and Walters (1987). In
their model, effort was originally allocated spatial
ly, and movement of effort was based on an
iterative approach where the expected profitability (measured in terms of value per unit of
effort less an area
-
specific cost per unit of effort) in alternative areas was considered. The
catch rate estimates (and
hence value and profitability per unit of effort) were endogenous in
the model based on total effort allocated to an area.


Several studies (e.g. Aswani, 1998; Dorn, 2001) have applied the concept of optimal foraging
theory to model location choice. The co
ncept of an optimal foraging model is, in essence, a
model of rational economic behaviour. In the fisheries context, “energy” is replaced by
economic variables in the objective function. That is, the fishers are assumed to try to
maximise economic returns
(energy in) while minimising the costs of obtaining the returns
(energy out). A fisher is assumed to leave the patch (fishing area) when the marginal rate of

7

return of the patch is equal to the mean return (including the travel cost) for the entire set of
patches (Aswani, 1998).


The Hilborn and Walters (1987) model and the optimal foraging studies considered above
assumed that effort moves from one area to the next based on economic incentives, but do not
provide information on the initial allocation of e
ffort. An alternative approach has been
employed by Sanchirico and Wilen (1999), and Sanchirico (2004) that allows both initial
allocation and effort movement to be determined endogenously. These models used
optimisation procedures to determine the spatial

allocation of effort across the fishery based
on the relative profitability in each patch, with movement between the patches based on
changes in the stock status. This model was extended by Collins et al (2003), who developed
a dynamic form of the model t
o considered changes in effort allocation, stocks and
profitability over time following a pollution event in one of the areas.


A major difficulty facing the development of spatial bioeconomic models is limited
information on economic performance measures
and their relationship with area fished. A
more generic difficulty facing the estimation of location choice models of fishers is lack of
detailed spatially disaggregated data in general. The (annual) process of stock assessments
does not generally include
the spatial detail required for inclusion in fisher choice models.
Also, logbooks that fishers complete may also suffer from similar problems of spatial
definition. In Europe, where location is recorded, it is generally done at the ICES rectangle
level. He
nce, more often than not in bioeconomic models of fisheries, spatiality is either not
included or included for large areas defining a fishery. The increasing interest in spatial
management of fisheries, however, will lead to greater activity in collecting
spatial
information. It is likely that spatial location choice will continue to be a major area of
investigation over the next decade.


Entry and exit


Early models of entry and exit essentially assumed that entry and exit was unrestricted and
linked to th
e profitability of the fishery. The earliest fisheries economics models (Gordon,
1954; Scott, 1955) suggested that fishing effort would increase in an unregulated fishery
through the entry of new vessels as long as the profitability in the fishery exceeded

the rate of
return the fisher might expect on their investment elsewhere in the economy. Conversely,
vessels were assumed to exit the fishery if they could achieve greater returns on their capital
investment elsewhere.


Based on this theory, simple dynam
ic models of entry and exit have been developed that link
the entry and exit to the level of profitability. Bjorndal and Conrad (1987) developed a model
of the North Sea Herring fishery, with changes in effort (i.e. entry and exit) modelled as a
function o
f total profitability. Mackinson et al (1997) also linked changes in effort levels to
profitability. Pascoe and Revill (2004) used a similar principle when modelling effort flows
into and out of the North Sea Crangon fishery, where an optimisation model wa
s used to
estimate the impacts of introducing bycatch reduction devices. Effort was assumed to flow
into the fishery if the marginal profitability of effort exceeded its estimated shadow price,
assumed to represent the opportunity cost of participating in
other fishing activities.
Conversely, effort flowed out of the fishery if the marginal profitability was less that the
opportunity cost.



8

The above models assume that fishing effort (and vessels) are free to move in and out of the
fishery as a result of f
ree and open access to the fishery or stock (the latter being the case for
many non
-
quota stocks in many European fisheries until relatively recently). With the advent
of fisheries management, the ability of fishers to move in and out of fisheries has beco
me
more constrained. Entry is restricted by the availability of licences or quota, while exit is
made more difficult as there is no alternative use for the boat.


The issue of “non
-
malleability” of capital was first raised by Clark, Clarke and Munro (1979
),
and models of irreversible investment further developed by
McKelvey (1985) and
Boyce
(1995). These models demonstrated how unregulated fisheries become overcapitalised and
overexploited, contrary to the outcomes under the traditional models of Gordon (1
954). These
models, however, where largely theoretic, and could not be applied in practice. Attempts at
developing empirical models for particular fisheries have been limited.
Ward and Sutinen
(1994) developed a dynamic bioeconomic model with exit/entry of

vessels based on
profitability, with non
-
malleability incorporated as a constraint on the rate of exit.
Tomberlin
(2001) developed a dynamic programming model of the Californian commercial salmon
fishery that included an exit decision based on the scrap v
alue of the vessel and the expected
future flow of returns from staying in the fishery.


Few studies exist that model the decision making process to either exit or enter fisheries. For
those studies that have been conducted, the majority use a discrete ch
oice framework (see
Annex). As noted previously, Bockstael and Opaluch (1983,1984) employed a discrete choice
framework to model the transfer of vessels from one fishery to another. Although this is not
directly an exit/entry study, it established the fram
ework used in later studies. Ward and
Sutinen (1994) also developed a discrete choice model that they incorporated into the
bioeconomic model mentioned previously. The model included prices, costs and measures of
stock abundance as explanatory variables. P
radhan and Leung (2004) used revenue by gross
tonnage in a binomial logit framework to consider exit/entry strategies versus staying in the
fisheries for Hawaiian longliners.
Mardle et al (2005) modelled the decision to enter, stay in
or exit the North Sea

beam trawl fishery based on vessel characteristics, revenue and stock
size using a multinomial logit model.


Compliance


Compliance or non
-
compliance with fishery regulation is an area of fisher behaviour that has
attracted generally less interest by eco
nomists than the previous two areas. Many studies of
fisher compliance have focused on what might be considered sociological rather than
economic factors. However, economic factors have been demonstrated to be important in the
decision to comply or not com
ply with regulations.


Early empirical studies of individuals’ compliance with fishery regulations examined standard
deterrence model, which considered the potential benefits of not complying, the risks of
detection and the fine if c
aught. These studies recognised factors other than those directly
related to the monetary costs and benefits of violation affected compliance behaviour. Sutinen
and Gauvin (1989) included personal characteristics such as age, years in the fishery and the
e
xtent of fishery income dependence as well as the economic factors. These “characteristic”
variables were all found to be significant in their estimated model. Furlong (1991) included in
his theoretical compliance model a vector of variables to capture “pe
rsonal and household”
characteristics. In his estimation of the model using data from a survey of Canadian fishers he
included variables for age, the proportion of the family currently unemployed and the

9

proportion of family income derived from fishing. Th
ese were designed to serve as proxies
for individual differences in “attitudes and proclivities” towards violation. These variables
were found to have the (intuitively) expected signs, but only the age of the fisher was
statistically significant.


More re
cent studies have considered the role of attitudes to compliance directly, and many
have found these to be major influencing factors. Sutinen and Kuperan (1999) developed an
extended compliance model that included, alongside monetary incentives and deterre
nce
variables, variables relating to social influence, moral norms and the perceived legitimacy of
the regulator and the regulations. Kuperan and Sutinen (1995, 1998) estimated a similar
supply of violations model using data from a survey of fishers in Mal
aysia and their
compliance with fishery zoning regulations. They found certain social, moral and legitimacy
variables to be significant in determining levels of compliance. Gezelius (2002) found that
compliance by Norwegian fishers was associated with an i
nformally enforced (i.e. by other
fishers) set of norms. These norms themselves permitted violation of some regulations that
were not considered legitimate. Social factors were found to be major influences of
compliance in Italian fisheries. Gambino
et al
.

(2003) found that violation behaviour of the
Italian fishers interviewed was mainly affected by (in order of importance) social pressure,
their judgement about legitimacy, moral influence and their judgement about the effectiveness
of the enforcement syst
em. However, in the Italian case, enforcement was believed to be
ineffective, so the probability of being caught and prosecuted was believed negligible. Hence,
only “moral” restraints were binding.


In other studies, economic factors have been found to be
the dominant motivators to comply.
Hatcher
et al

(2000) estimated an empirical model of compliance with quota restrictions
among fishers in the UK. Significant explanatory variables for compliance in their model
were the perceived risk of detection and the

size of the expected fine if detected, but also the
feeling of involvement in the design and implementation of regulations, an indicator of a
norm of compliance and the perceived attitude of others. Nielsen and Mathiesen (2003) found
economic benefits fro
m non
-
compliance to be the most important single factor influencing
compliance of the Danish fishers. However, norms, inclusion in the decision
-
making process
(affecting legitimacy), beliefs about the behaviour of others and belief (or disbelief) that the
system would provide conservation benefits were all found to be contributing factors. A more
recent UK study by Hatcher and Gordon (2005) found that “conventional” economic
incentives predominated, although still found some evidence of normative influences

on quota
violation levels. Mason and Gullet (2005) found that licence cancellation as a penalty for non
-
compliance provided a major incentive to comply, more so than financial penalties as there
was a perception that these were less likely to be imposed b
y the courts.


The studies cited above essentially employed discrete choice based modelling, with a binary
choice to either comply or not comply. The estimation procedures in some cases, however,
differed from those used to estimate the location choice and

entry/exit models. Further details
on the estimation process are provided in the Annex.


Attempts at incorporating compliance behaviour into fisheries bioeconomic models have been
very limited. In most bioeconomic models, compliance has been implicitly as
sumed. For
example, in a model of the Dutch beam trawling fleet (Ulrich
et al
., 2002), it was assumed
that fishing would cease as soon as quota of one of the several species was achieved, even if
quota on the other species was available. This resulted in a
n underestimate of the overall catch
and economic benefits from the fishery, as fishers would have continued to fish and discard

10

overquota catch (again, assuming compliance behaviour) provided the value of the catch that
could be landed exceeded the additi
onal cost of catching it.


Where non
-
compliance has been explicitly considered in fisheries models, it has generally
involved a comparison of compliance with restrictions
versus

the case with no restrictions
(i.e., effectively zero compliance). For example
, Hutton
et al
. (2001) modelled the benefits of
compliance versus non
-
compliance
3

with various minimum size and effort restrictions for a
South African fishery using a bioeconomic simulation model. Non
-
compliance was modelled
through removing the restricti
ons imposed in the compliance simulation.
4



Such an approach effectively represents the two extremes of compliance: perfect compliance
and perfect non
-
compliance. However, the studies of compliance in fisheries have found that
compliance generally falls s
omewhere between these extremes, with part of the fishery
complying and other parts not complying, or complying to varying degrees. As a result, the
“true”
5

outcome from a management strategy will fall somewhere between the two extremes
that can be readily

modelled. Kritzer (2004) attempted to overcome this through assuming, in
the case of marine protected areas (MPAs), that some of the fishing effort that should be
displaced through the designation of the MPA would remain within the protected area (i.e.,
i
mperfect compliance). The effects of varying amounts of illegal fishing activity were
examined.


Modelling studies in other fields have attempted to incorporate the effects of non
-
compliance
through assuming compliance with higher levels than the regulatio
n permits. For example,
Meyburg
et al
. (1998) modelled the economic impacts of weight limits for heavy vehicles by
assuming perfect compliance with a limit above that actually imposed (i.e. non
-
compliance
with the actual regulation). Similarly, Sadiq
et al
. (2004) modelled wastewater treatment
effectiveness by assuming an amount of agricultural pollution above that that the regulations
prescribed. In both examples, the determination of the ‘excess’ above the regulated restriction
(i.e. the effective complia
nce limit) was relatively subjective, based largely on anecdotal
evidence.


In theory, it should be possible to determine the effective compliance limits based on the
likelihood of compliance. The previous studies of compliance in fisheries effectively
det
ermined the probability of compliance based on the characteristics of the fishery, the
fishers and the management regime. Given this, the effective compliance limit could be
estimated as the expected outcome given the probability of compliance. That is, av
erage of
the limit multiplied by the proportion expected to comply and the unconstrained outcome
multiplied by the proportion expected not to comply (for example, the capacity output of the
fleet in the case of quotas or the potential effort in the case of

effort controls). To date, such
an approach has not been applied to any fisheries bioeconomic model, nor to any model of
any other sector.





3

The model was actually based on cooperation and non
-
cooperation rather than compliance per se, but
cooperation was defined in terms of compliance with the regulations and non
-
cooperation was defined in terms
of
non
-
compliance with the regulations.

4

Similar approaches are used in other fields also. For example, Pacini et al. (2004) estimated the costs of
compliance with EU agricultural pollution regulations by comparing a perfect compliance scenario with a perfec
t
non
-
compliance (i.e. modelled as no regulation) scenario.

5

“True” in this context is only relative, as models contain many other simplifications and assumptions that may
result in a divergence between the modelled and real
-
world impact.


11

Discussion and Conclusions


A key feature of all of the above studies was the reliance on the profit maximising beh
aviour
assumption to model fisher behaviour. In many cases, this was implicit, such as in the optimal
foraging models of location choice. In other cases, information on profitability
per se

was not
readily available, so proxy measures were used that produc
ed equivalent results. For example,
revenue or catch rates (assumed to be correlated with revenue), and costs of travel or distance
to the area (correlated with the travel cost). Variations in economic performance were taken
into consideration through inco
rporating vessel or skipper characteristics. Hence, while no
“economic” information as such may appear in some models, the data that is used either
explicitly or implicitly reflects profitability of different activities.


Much of the most recent modelling
activity has focused on fisher location choice. This has
largely been a consequence of the increased use of area closures for fisheries management
purposes, or the introduction of marine protected areas for a variety of purposes. Economic
based models of l
ocation choice have been particularly successful in assessing the impact of
these management measures largely because the impact of the measure on the fisher is fairly
obvious. If an area is closed, then the fisher must go somewhere else. Other management
measures have less obvious impacts on the fisher, making the development and use of
behavioural models difficult. Entry and exit decisions depend on a range of factor such as the
availability of alternative fisheries or use of the boat, and management meas
ures themselves
to assist in overcoming the non
-
malleability problem (e.g. decommission schemes). The
increased interest in property rights based measures such as ITQs has attracted some attention
(e.g. Dupont, 2000), although development of such models is

less straightforward.
Nevertheless, economic analysis of a more qualitative nature has been successful in
“forecasting” the outcomes of such management systems, again based on the profit
maximising behaviour assumption.


Although considerable effort has b
een made in recent years to model fisher behaviour, there
are still a number of areas where models have not been developed. Foremost of these is in the
area of input substitution, with relatively few attempts at modelling changes in input use
following reg
ulations.
Dupont (1991) and Pascoe and Robinson (1998) considered the
potential for input substitution between key vessel characteristics (e.g. labour, fuel, gear, boat
size and engine power). Eggert and Tveterås (2004) developed a choice based model of fi
sher
gear choice in response to economic factors, although this was not related to fisheries
management change. Given the dominance of input controls as a management measure
internationally, and the general acceptance that input substitution following a ch
ange in input
controls will occur, this lack of empirical modelling research is surprising.


A possible explanation for this is that the direction of substitution can be highly varied, with
perhaps too many choices to consider (some of which may not have b
een observed prior to
the restriction). For example, in the study by Pascoe and Robinson (1998), fishers responded
to area restrictions linked to engine power by reducing engine power to remain in the area
rather than move out of the restricted area. In ot
her fisheries, a restriction on gear size may
result in increased tow time, larger engines (to increase area swept), or adoption of alternative
fishing gears that are not restricted, with different fishers responding in different ways. As a
consequence, de
veloping models to estimate how fishers may respond to an input control
change may not be feasible, even though a response would be expected.



12

Another area where further work needs to be undertaken is linking the models of fisher
behaviour to bioeconomic m
odels. As noted above, several attempts have been made in this
area, although most studies of fisher behaviour are effectively “stand alone”. Given the
probabilistic nature of most fisher behaviour models, incorporating these models into
bioeconomic models

is not straightforward, particularly models of compliance. Location
choice models are less problematic, as the probability distribution can be readily converted
into an expected effort allocation in spatial models. There is considerable potential for
util
ising Bayesian networks, however, with some attempts already having been made in this
area (e.g. Little et al, 2004). Model outcomes could be derived using different assumptions
about compliance behaviour, with these outcomes feeding into a Bayesian net. S
uch an
approach is being examined in the European Funded COMMIT project.


It is generally recognised that the evaluation of management options for fisheries systems
should include the ecological, biological, social and economic status of the fishery. Howev
er,
most evaluation frameworks assume that the management options achieve their stated
objectives. From both practical experience and the empirical work undertaken on fisher
behaviour, management may not always result in the expected outcome, either as a r
esult of
non
-
compliance, or as a result of fishers modifying their behaviour in response to the
management change. These responses can largely be explained in terms of fishers operating in
a manner consistent with profit maximisation.
Even if fishery manag
ers are not interested in
maximising fishers’ profit per se, consideration needs to be given to the implications of
changing profitability in the fishery on fisher behaviour (Wilen et al, 2002).


Acknowledgements

This review has been carried out with the f
inancial support of the Commission of the
European Communities, as part of Fifth Framework project Q5RS
-
2002
-
01291, “Technical
development and tactical adaptations of important EU fleets (TECTAC)” and the Sixth
Framework project SSP8
-
CT
-
2003
-
502289, “Creat
ion of Multiannual Management Plans for
Commitment”. It does not necessarily reflect the Commission’s views and in no way
anticipates future policy in this area.


Annex: Discrete choice modelling


As in most economic
-
based choice models, utility is assumed

to derive from an individual’s
choice, while the choice itself is assumed to be made on the basis of the characteristics of the
option chosen. The individual choice (and the derived utility) is assumed to have both a
deterministic component and a stochast
ic error component (thereby giving the term “random
utility model”). Utility is typically defined as a (linear) combination of a set of explanatory
variables that together are surmised to form (for the most part) the non
-
random components of
the utility, a
nd a stochastic error component:



U
ij

z
ij

ij

(1)


where for a given person time
-
event,
i
, (such as a fishing trip) choice j is made. The
explanatory variables
z
ij

can be comprised of attributes of the choice,
x
ij
, and characteristics of

13

the individual,
w
i
. Through a choice of distribution o
n disturbances of the error term, typically
logit, then the model can be estimated.
6



Conditional logit (McFadden, 1974; 1981) has long been used for the estimation of these
parameters,

. Recently, mixed and multinomial logit have also been applied for d
iscrete
choice analyses. These three approaches are outlined below. A brief description of the
compliance models is also presented.


Conditional Logit (McFadden’s choice model)

In a conditional logit formulation, a set of unordered choices (e.g.
j = 1, 2,
…, n
) is assumed.
So, the probability that individual
i

makes choice
j

is given by,







J
j
ij
ij
i
z
z
j
Y
1
)
'
exp(
)
'
exp(
)
Pr(



(2)


where
y
i

are 0
-
1 variables indicating 1 where the choice was made and 0 otherwise, and the
independent variables
z
ij
=[
x
ij

w
i
]

where
x

are attrib
utes of choice
j

of individual
i

and
w

are
attributes of individual
i
. The residuals,

ij
, are independently and identically distributed (iid)
with the type I extreme
-
value distribution. The presumption of independence of

ij

leads to the
assumption of the

property of independence of irrelevant alternatives (IIA). That is, the ratio
of any two alternatives will remain the same irrelevant of other alternatives. In the case where
the independent variables consist only of individual attributes, then the multin
omial logit
results.
7



However, for spatial analyses as experienced in fisheries, Wilen et al. (2002) state that
conditional logit is inappropriate as the independence of irrelevant alternatives (IIA)
assumption that it imposes could be potentially invali
dated and therefore influence policy
analyses that consider spatial aspects. IIA implies that a change in the choice set would not
affect the relative choice probabilities, as the choices are assumed independent. To take this
into account, they use the nes
ted logit model (e.g. McFadden, 1981; Morey et al., 1993) that
although in basic structure is the same (but with the imposition of a hierarchical structure on
the decision process) does not impose IIA across nests. An alternative to the nested approach,
wh
ich can be difficult to estimate for more than 3 or 4 hierarchical levels, is the mixed logit
model which does not impose IIA but also includes choice attributes and individual
characteristics.


Mixed Logit

As noted above, a conditional logit model is used

to estimate the probability that an
individual,
i
, chooses alternative,
j
, from some set of alternatives. However, mixed logit
generalizes this by allowing the estimated coefficients,

, to be random rather than fixed. So,
in this case, the probability (
P
) that individual,
i
, chooses alternative,
j

is estimated using,








d
f
L
P
i
i
)
(
)
(

(3)




6

See for exam
ple Greene (2003) for a more complete discussion surrounding the derivation of the model. It
should be noted that most good statistical/econometric packages include some/all variants of logit methodology
(e.g. SAS, STATA, LIMDEP, SPSS).

7

See McFadden (197
4) and Greene (2003) for more details.


14


where
L
i
(

)

is the logit function estimated for parameters,


and
f(

)

is a density functi
on
on the parameters,

.


In some studies, the mixed logit model is estimated without random
-
coefficients, but with
error components that are correlated amongst the utilities for the alternatives. In this case,
utility is given as



i
i
i
i
z
x
U







(
4)


where
x

and
z

are observed variables,


is a vector of fixed coefficients,


is a vector of
random terms with zero mean and


is distributed iid extreme value. Hence, if


is zero then a
conditional logit formulation results. In fact, McFadden and Trai
n (2000) showed that any
random utility method can be approximated using mixed logit.
8



The conditional logit depends on the assumption of IIA. So, this implies that a change in the
attributes will not change the ratio of the two alternatives, and moreove
r will result in the
same percentage change across alternatives. Mixed logit does not assume IIA, as the
percentage change across alternatives may vary given changes in attributes. Hence, mixed
logit overcomes many of the limitations of conditional logit.


Furthermore, extra information is provided by mixed logit over conditional logit. This is due
to the fact that mixed logit can estimate the extent to which individuals differ in their
preferences for attributes. In the case where coefficients on the stand
ard deviations of
observed variables differ significantly, then mixed logit represents the choices made better
than conditional logit. This is because it is assumed in conditional logit that coefficients are
the same for all individuals. Mixed logit can al
so take into account several choices made by
the same individual. Also, where applicable, willingness to pay for individuals towards
changes in attributes can be estimated from a mixed logit solution.


In fisheries, two studies that have considered a mixed

logit variant in the analysis of
heterogeneous risk preferences of fishers are Eggert and Tveterås (2004) and Mistaien and
Strand (2000).


Multinomial logit


The multinomial logit formulation differs from the conditional logit model only in the
determini
stic part of the utility,



ij
i
j
ij
w
U





(5)


So, the probability that individual
i

makes choice
j

is given by,







J
j
j
i
j
i
i
w
w
j
Y
1
)
'
exp(
)
'
exp(
)
Pr(



(6)




8

Particularly important for fisheries, nested logit can be approximated by specifying dummies in a mixed logit
formulation.


15


Mardle et al (2005) used the multinomial logit formulation to estimate the exit/entry model
with multiple ch
oices (i.e. enter, exit or stay).


Supply of Violation functions


The analysis of factors affecting compliance utilises another form of discrete choice
modelling, where the decision is a binary choice to either to comply or not comply. The
studies cited i
n the compliance section, however, used a wide variety of estimation techniques
(but all variations on a similar basic model). The model estimated by Hatcher et al (2000) is
used as an example of the modelling technique.


The individual fisher’s expected u
tility from violating a regulation is assumed to be a function
of the benefits of not complying, the probability of being caught, the expected fine, social
factors and individual fisher characteristics. This can be expressed as



EU(V
i

)= f(B
i
, p
i
, F
i
, S
i
, X
i
)

(7)


where
V
i

is the self reported violation rate;
B
i

is the expected benefit of not complying;
p
i

is
the subjective probability of detection and prosecution; F
i

is the expected size of the penalty if
prosecuted;
S
i

is a vector of social
-
related va
riables include moral obligation to comply,
perceptions of regulatory legitimacy and other social influences; and
Xi

is a vector of
variables to describe the characteristics of the fishermen and his/her vessel.


The model can be estimated using either log
it or probit regression. In the case study presented
by Hatcher et al (2000), probit analysis was used.



References


Allen, P.M. and McGlade, J.M. 1987,

Modelling complex human systems: a fisheries
example.
European Journal of Operational Research

30, 14
7
-
167.

Anderson, L.G. 1999. The microeconomics of vessel behavior: a detailed short
-
run analysis of
the effects of regulation.
Marine Resource Economics

14(2), 129
-
150.

Anderson, L.G. 2004. Open
-
access fishery performance when vessels use goal achievement
behaviour.
Marine Resource Economics

19(4), 439
-
458.

Aswani, S. 1998. Patterns of marine harvest effort in southwestern New Georgia, Solomon
Islands: resource management or optimal foraging?
Ocean & Coastal Management

40(2
-
3), 207
-
235.

Bjorndal, T. and Con
rad, J.M. 1987. The Dynamics of an Open Access Fishery.
Canadian
Journal of Economics

20(1), 74
-
85

Bockstael, N.E. and J.J. Opaluch, 1983, Discrete modelling of supply response under
uncertainty: The case of the fishery,
Journal of Environmental Economics
and
Management
, 10: 125
-
137.

Bockstael, N.E. and J.J. Opaluch, 1984, Behavioral modeling and fisheries management,
Marine Resource Economics

1(1): 105
-
115.


16

Boyce J.R. 1995. Optimal Capital Accumulation in a Fishery: A Nonlinear Irreversible
Investment Mode
l.
Journal of Environmental Economics and Management
28(3), 324
-
339.

Chakravorty, U., and Nemato, K. 2001. Modelling the effects of area closure and tax policies:
a spatial
-
temporal model for the Hawaii longline fishery.
Marine Resource Economics
, 15:
179

204.

Clark, C.W., Clarke, F.H. and Munro, G.R. 1979. The optimal explotation of renewable
resource stocks: problems of irreversible investment.
Econometrica

47(1), 25
-
47.

Collins, A., Pascoe, S. and Whitmarsh, D. 2003. Pollution externalities and fisheries
: insights
from a spatially
-
explicit bioeconomic model.
Marine Resource Economics,
18(4) 313
-
328.

Dorn, M.W. 2001. Fishing behavior of factory trawlers: a hierarchical model of information
processing and decision
-
making.
ICES Journal of Marine Science
58(1
), 238
-
252.

Dupont, D.P. 1991. Testing for Input Substitution in a Regulated Fishery.
American Journal
of Agricultural Economics

73(1), 155
-
64

Dupont, D.P. 2000. Individual transferable quota and efficient restructuring of the primary
harvesting sector.
An
nals of Operations Research

94: 275
-
294

Eales, J. and Wilen, J.E. 1986. An examination of fishing location choice in the pink shrimp
fishery.
Marine Resource Economics
, 2(4):331
-
351.

Eggert, H. and Tveterås, R. 2004. Stochastic production and heterogeneous

risk preferences:
Commercial fishers’ gear choices.
American Journal of Agricultural Economics

86(1):
199
-
212.

Food and Agriculture Organization of the United Nations (FAO)
. 2004a.

The State of World
Fisheries and Aquaculture, 2004
. FAO, Rome.

Food and Ag
riculture Organization of the United Nations (FAO). 2004b. An evaluation of
responses and action taken by FAO members to implement the international plan of action
for the management of fishing capacity (IPOA
-
CAPACITY). Paper presented at the
Technical Con
sultation to Review Progress and Promote the Full Implementation of the
IPOA to Prevent, Deter and Eliminate IUU Fishing and the IPOA for the Management of
Fishing Capacity, Rome, Italy, 24
-
29 June 2004.

Furlong, W.J. 1991. The deterrent effect of regulato
ry enforcement in the fishery.
Land
Economics
, 67(1): 116
-
129.

Gambino, M., Malvarosa, L. and Placenti, V. 2003. Fishery regulation, perceptions and
compliance: the fishers’ responses. Paper presented at the XVth Annual conference of the
European Associati
on of Fisheries Economists, IFREMER, Brest, 14
-
16 May 2003.

Gezelius, S.S. 2002. Do norms count? State regulation and compliance in a Norwegian
fishing community.

Acta Sociologica
, 45(4): 305
-
314.

Gordon, H.S. 1954. The economic theory of a common property

resource: the fishery.
Journal
of Political Economy

62, 124
-
42.

Gréboval, D. and Munro, G. 1999. Overcapitalisation and excess capacity in world fisheries:
underlying economics and methods of control.
In: Gréboval, D (Ed).
Managing Fishing
Capacity: Selec
ted Papers on Underlying Concepts and Issues
. FAO Fisheries Technical
Paper No 386, 1
-
48. Rome: FAO.

Greene, W. H. 2003.
Econometric Analysis
, 5th ed. Prentice
-
Hall, USA.


17

Hatcher, A. and Gordon, D. 2005. Further investigations into the factors affecting co
mpliance
with UK fishing quotas.
Land Economics

81(1): 71
-
86.

Hatcher, A., Jaffry, S., Thébaud, O. and Bennett, E. 2000. Normative and social influences
affecting compliance with fishery regulations.
Land Economics

76(3): 448
-
461.

Hilborn, R. and Walters,
C.J. 1987. A general model for simulation of stock and fleet
dynamics in spatially heterogeneous fisheries.
Canadian Journal of Fisheries and Aquatic
Science

44, 1366
-
1369.

Holland, D.S. and J.G. Sutinen, 1999, An empirical model of fleet dynamics in New E
ngland
trawl fisheries,
Canadian Journal of Fisheries and Aquatic Science

56, 253
-
264.

Holland, D.S. and J.G. Sutinen, 2000, Location choice in New England trawl fisheries: old
habits die hard.
Land Economics

76(1), 133
-
49.

Hutton, T., Griffiths, M.H., Sum
aila, U.R. and Pitcher, T.J. 2001. Cooperative versus non
-
cooperative management of shared linefish stocks in South Africa: an assessment of
alternative management strategies for geelbek (Atractoscion aequidens),
Fisheries
Research

51(1): 53
-
68.

Hutton, T.
, Mardle, S., Pascoe, S. and Clark, R.A. 2004 Modelling fishing location choice
within mixed fisheries: English North Sea beam trawlers in 2000 and 2001,
ICES Journal
of Marine Science

61(8), 1443
-
1452.

Kirkley, J., Squires, D. and Strand, I.E. 1998. Chara
cterizing Managerial Skill and Technical
Efficiency in a Fishery.
Journal of Productivity Analysis

9(2), 145
-
60.

Kirkley, J.E., Squires, D., Alam, M.F., and Ishak, H.O., 2003, Excess capacity and
asymmetric information in developing country fisheries: the
Malaysian purse seine fishery.
American Journal of Agricultural Economics

85(3), 647
-
662.

Kritzer, J.P. 2004. Effects of noncompliance on the success of alternative designs of marine
protected area networks for conservation and fisheries management.
Conser
vation Biology

18(4): 1021
-
1031.

Kuperan, K. and Sutinen, J.G. 1995. Compliance with zoning regulations in Malaysian
fisheries. In: Liao, D.S., ed.,
International Cooperation for Fisheries and Aquaculture
Development: Proceedings of the VIIth Conference of

the International Institute of
Fisheries Economics and Trade
, Vol. I. Keelung, Taiwan: National Taiwan Ocean
University.

Kuperan, K. and Sutinen, J.G. 1998. Blue water crime: deterrence, legitimacy and compliance
in fisheries.
Law and Society Review
, 32(2
): 309
-
337.

Little, L.R., Kuikka, S., Punt, A.E., Pantus, F., Davies C.R. and Mapstone, B.D. 2004.
Information flow among fishing vessels modelled using a Bayesian network,
Environmental Modelling & Software

19(1), 27
-
34.

Mackinson, S., Sumaila U.R. and Pi
tcher, T.J. 1997. Bioeconomics and catchability: fish and
fishers behaviour during stock collapse.
Fisheries Research

31(1
-
2), 11
-
17.

Mardle, S. and Pascoe, S. 2002. Modelling the effects of trade
-
offs between long and short
-
term objectives in fisheries ma
nagement.
Journal of Environmental Management

65(1),
49
-
62.

Mardle, S., Hutton, T., Wattage, P and Pascoe, S. 2005. Entering and exiting a fishery: A
strategic choice? Paper presented at the
third biennial NAAFE Forum, University of
British Columbia, Vanco
uver, BC, Canada, 25
-
27 May 2005.


18

Mason, R. and Gullett, W. 2005. Cancellation provisions in Australia’s Commonwealth
-
managed fisheries.
Marine Policy

[In Press, Corrected Proof, Available online 31 March
2005]

Maury, O., and Gascuel, D. 1999. SHADYS (‘si
mulateur halieutique de dynamiques
spatiales’), a GIS based numerical model of fisheries. Example application: the study of a
marine protected area.
Aquatic Living Resources
, 12: 77

88.

McFadden, D. 1974. Conditional logit analysis of qualitative choice b
ehavior. In:
Frontiers in
Econometrics
, P.Zarembka (ed.), New York: Academic Press.

McFadden, D. 1981. Econometric models of probabilistic choice. In: C.F.Manski and
D.McFadden (eds.),
Structural Analysis of Discrete Data with Econometrics Applications
,
19
8
-
272, Cambridge US: MIT Press.

McFadden, D. and Train, K.E. 2000. Mixed MNL models of discrete response.
Journal of
Applied Econometrics

15, 447
-
470.

McKelvey, R. 1985. Decentralized regulation of a common property renewable resource
industry with irrever
sible investment.
Journal of Environmental Economics and
Management

12(4), 287
-
307.

Meyburg, A.H., Saphores, J
-
D.M. and Schuler, R.E. 1998. The economic impacts of a
divisible
-
load permit system for heavy vehicles.
Transportation Research Part A: Policy
an
d Practice

32(2): 115
-
127.

Mistiaen, J.A. and I.E. Strand, 2000, Location choice of commercial fishermen with
heterogeneous risk preferences,
American Journal of Agricultural Economics
, 82(5): 1184
-
1190.

Morey, E.R., Rowe, R.D. and Watson, M. 1993. A repea
ted nested
-
logit model of Atlantic
salmon fishing.
American Journal of Agricultural Economics
, 75 9(3): 578
-
592.

Myers, R.A. and Worm, B. 2003. Rapid worldwide depletion of predatory fish communities.
Nature

423, 280
-
283.

Nielsen, J.R. and Mathiesen, C. 20
03. Important Factors Influencing Rule Compliance in
Fisheries: Lessons from Denmark.
Marine Policy,

27: 409
-
416.

Pacini, C., Woosink, A., Giesen, G. and Huirne, R. 2004. Ecological
-
economic modelling to
support multi
-
objective policy making: a farming sys
tems approach implemented for
Tuscany.
Agriculture, Ecosystems and Environment
, 102:349
-
364.

Pascoe, S. and Coglan, L. 2002. Contribution of unmeasurable factors to the efficiency of fishing
vessels: an analysis of technical efficiency of fishing vessels i
n the English Channel.
American Journal of Agricultural Economics
84,
45
-
57.

Pascoe, S. and Mardle, M. 2001. Optimal fleet size in the English Channel: a multi
-
objective
programming approach.
European Review of Agricultural Economics

28(2) 161
-
183.

Pascoe,

S. and
Revill, A. 2004. Costs and benefits of bycatch reduction devices in European
brown shrimp trawl fisheries.
Environmental and Resource Economics,
27(1): 43
-
64.

Pascoe, S. and Robinson, C. 1998. Input controls, input substitution and profit maximisa
tion
in the English Channel beam trawl fishery.
Journal of Agricultural Economics
, 49(1): 16
-
30.

Pascoe, S. and Tingley, D. 2005. Economic capacity estimation in fisheries: An application of
a nonparametric ray economic profit approach.
Resource and Energ
y Economics

[in press].


19

Placenti, V., Rizzo, G. and Spagnolo, M. 1992. A Bio
-
economic Model for the Optimization
of a Multi
-
species, Multi
-
gear Fishery: The Italian Case.
Marine Resource Economics

7(4),
275
-
95.

Pradhan, N.C. and Leung, P. 2004. Modeling en
try, stay, and exit decisions of the longline
fishers in Hawaii.
Marine Policy

28: 311

324.

Robinson, C. and Pascoe, S. 1998.
Fisher behaviour: Exploring the validity of the profit
maximising assumption, In: J.P. Boude and J. Boncoeur (Eds)
Proceedings of t
he IXth
conference of the European Association of Fisheries Economists
, UBO
-
CEDEM Quimper
and ENSA Rennes, France, 167
-
183.

Roy, N. 1998. Fishing behaviour and the length of the fishing season.
Marine Resource
Economics

13(3) 197
-
214.

Sadiq, R., Al
-
Zahrani
, M.A., Sheikh, A.K., Husain, T. and Farooq, S. 2004. Performance
evaluation of slow sand filters using fuzzy rule
-
based modelling.
Environmental Modelling
and Software
. 19(5): 507
-
515.

Sampson, D.B. 1991. Fishing tactics and fish abundance, and their infl
uence on catch rates.
ICES Journal of Marine Science
, 48(3): 291
-
301.

Sanchirico, J.N. 2004. Designing a cost effective marine reserve network: a bioeconomic
metapopultion analysis.

Marine Resource Economics

19(1), 41
-
66.

Sanchirico, J.N. and Wilen, J.E. 1
999. Bioeconomics of spatial exploitation in a patchy
environment.
Journal of Environmental Economics and Management

37, 129
-
150.

Scott, A. 1955. The fishery: the objectives of the sole owner.
Journal of Political Economy

63, 116
-
124.

Shepherd, J.G. and Ga
rrod, D.J. 1981. Modelling the response of a fishing fleet to changing
circumstances, using cautious non
-
linear optimisation.
Journal Conseil International pour
l'Exploration de la Mer

39(3) 231
-
238.

Smith, M.D. 2000. Spatial search and fishing location ch
oice: Methodological challenges od
empirical modeling.
American Journal of Agricultural Economics
, 82(5): 1198
-
1206.

Smith, M.D. and Wilen, J.R. 2003. Economic impact of marine reserves: the importance of
spatial behaviour.
Journal of Environmental Economi
cs and Management

46(2), 183
-
206.

Smith, M.D. and Wilen, J.R. 2004. Marine reserves with endogenous ports: empirical
bioeconomics of the California sea urchin fishery.
Marine Resource Economics

19(1), 85
-
112.

Strand, I.E., 2004, Spatial variation in risk p
references among Atlantic and Gulf of Mexico
pelagic longline fishermen,
Marine Resource Economics
, 19(1), 145
-
160.

Sutinen, J.G. and Gauvin, J.R. 1989. An econometric study of regulatory enforcement and
compliance in the commercial inshore lobster fishery

of Massachusetts. In: Neher, P.,
Arnason, R. and Mollet, N., eds.
Rights Based Fishing
, NATO ASI Series E: Applied
Sciences, 169. Dordrecht: Kluwer.

Sutinen, J.G. and Kuperan, K. 1999. A socio
-
economic theory of regulatory compliance.
International Journa
l of Social Economics
, 26(1
-
3): 174
-
193.

Tingley, D. 2005 Northern Ireland Fisheries Futures Study. Northern Ireland Department of
Agriculture and Rural Development, August 2005.


20

Tingley, D., Pascoe, S., and Mardle, S., 2003, Estimating capacity utilisatio
n in multi
-
purpose, multi
-
metier fisheries.
Fisheries Research

63(1), 121
-
134.

Tomberlin, D. 2001. Modeling California Salmon Fleet Dynamics, Proceedings of the First
North American Fisheries Economics Forum, April 1
-
4, 2001, New Orleans, LA.

Ulrich, C.,

Marchal, P., Sparre, P.J. and De Wilde, J
-
W. 2002. Influence of trends in fishing
power on bio
-
economics in the North Sea flatfish fishery regulated by catches
-

or by effort
quotas.
Canadian Journal of Fisheries and Aquatic Science,
59(5): 829
-
843.

Veste
rgaard, N., Squires, D., and Kirkley, J.E., 2003, Measuring capacity and capacity
utilization in fisheries: the case of the Danish gillnet fleet.
Fisheries Research

60(2
-
3):
357
-
368.

Ward, J.M. and J.G. Sutinen, 1994, Vessel entry
-
exit behaviour in the Gul
f of Mexico Shrimp
fishery,
American Journal of Agricultural Economics

76(4): 916
-
92.

Wilen, J.E. 1979. Fisherman behavior and the design of efficient fisheries regulation
programs.
Journal of the Fisheries Research Board of Canada
, 36: 855
-
858.

Wilen, J.
E., Smith, M.D., Lockwood, D. and Botsford, F.W. 2002. Avoiding surprises:
Incorporating fisherman behavior into management models,
Bulletin of Marine Science

70(2): 553
-
575.