j ourna l homepage: www.e lsev ie r .com/ locate /eco lecon
Analysis
Risk versus economic performance in a mixed fishery
S. Gourguet a,b,⁎, O. Thébaud a, C. Dichmont a, S. Jennings d, L.R. Little c, S. Pascoe a, R.A. Deng a, L. Doyen b
a CSIRO, Marine and Atmospheric Research, 41 Boggo Road, Dutton Park, QLD 4102, Australiab CNRS-MNHN, CERSP UMR 7204, CP 51, 55 rue Buffon, 75005 Paris, Francec CSIRO Marine and Atmospheric Research, PO Box 1538, Hobart, TAS 7001, Australiad School of Economics and Finance, University of Tasmania, Hobart, TAS 7001, Australia
Balancing bio-economic risks and high profit expectations is often amajor concern in fisheriesmanagement.Weexamine this trade-off in the context of the Australian Northern Prawn Fishery (NPF). The fishery derives itsrevenue from different prawn species with different dynamics and recruitment processes. A multi-species bio-economic and stochastic model is used to examine the trade-offs between mean profitability of the fishery andits variance, under a range of economic scenarios, fishing capacities and distributions of fishing effort acrossthe various sub-fisheries that comprise the NPF. Simulation results show that the current fishing strategy diver-sifying catch across sub-components of the fishery entails a compromise between expected performance andrisk. Furthermore, given the current economic conditions, increases infleet sizewould improve the expected eco-nomic performance of the fishery, but at the cost of increased variability of this performance.
Globally, many capture fisheries do not achieve their full econom-ic potential and are subject to excess capacity (Munro, 2010). Forsome fisheries, this may be due to failure in regulating the race tofish. Other fisheries may be managed to achieve Maximum Sustain-able Yield (MSY), rather than Maximum Economic Yield (MEY). Insome cases, social considerations may dominate the managementdecision process leading to the approval of even higher levels of ca-pacity. In other cases, differences between observed harvestinglevels of individual species and the levels which would ensure MEYmay be related to the fact that commercial fishers operate across arange of species, with varying ability to target these species sepa-rately, leading to difficulties in identifying optimal fishery-widelevels of fishing capacity and allocation of fishing effort. Moreover,revenues from fisheries may vary greatly from year to year owingto natural variation in fish stocks (Kasperski and Holland, 2013)that cannot be predicted with any reliability, leading to varyinglevels of economic risks for fishing operators (Sethi, 2010). Whilemaximising economic yield is usually seen as a desirable objectivefor fisheries management, industry stakeholders usually also valuestability over time. This may be due to risk aversion, but also to theneed to maintain markets, avoid market saturation and guide invest-ment decisions relating to non-malleable capital (Holland andHerrera, 2009). Successful fisheries management should therefore
P 51, 55 rue Buffon, 75005 Paris,
uet).
ghts reserved.
identify and cope with risk to minimize the effects of unpredictablevariability (Sethi, 2010). Indeed, as expressed by Hilborn et al.(2001): “if we are to succeed at management – if we are to maintainstable fishing communities –wehave to begin tomanage risk”. The pro-cess of dealing with risk in fisheries management involves the formula-tion of advice for fisheries managers in a way that conveys the possibleconsequences of uncertainty, but also handles the ways in which man-agers take uncertainty into account in making decisions (Francis andShotton, 1997).
Inmulti-speciesfisheries, thedifferentfish stocks contributing to theoverall catchmay present different levels of natural variability, such thatthe choice of fishing strategies can be associated with trade-offsbetween mean and variance of the fishery's economic yield. Portfoliotheory focuses on the selection of assets (such as species) to create abundle that provides the greatest expected economic performance(such as catch or annual income) at the least variation about the expect-ed performance (Markowitz, 1952; Roy, 1952). Mean-variance analysis,which is consistent with portfolio theory, is particularly important infinance (Epstein, 1985). While portfolio effects have been studied forfisheries management (Sanchirico et al., 2008; Sethi, 2010), mean-variance analyses have not been explicitly applied in the context ofallocating effort in a multi-species fishery.
This article focuses on amean-variance analysis of inter-temporalprofits of a fishery in which the set of target species have differentlevels of environmentally driven variability in recruitment. Thestudy is based on a dynamic bio-economic modelling approach, inline with capital theory (Clark and Munro, 1975), where fish stocksare taken as natural capital assets and where net present value ofprofits plays a major role. The analysis is applied to the case of the
111S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
Northern Prawn Fishery (NPF) in Australia. The bio-economic modelis used to explore the trade-offs betweenmean performance and riskassociated with alternative managements, taking into account thedistribution of fishing effort across its sub-components and differentassumptions regarding changes in fuel and prawn prices. Simulationresults show that the economic performance of the fishery could beimproved by increasing the capacity of the fleet, but at the cost of in-creased inter-annual variability of this performance. The current al-location of fishing effort between the sub-components of thefishery achieves a compromise between performance and risk.With likely changes in economic conditions, maintaining currentfishing capacity and effort allocation would achieve the highest eco-nomic yield, but for a similar level of performance this would comewith higher risk than were a reduced fleet to focus more on the rela-tively stable tiger-prawn component of the fishery.
2. Material and Methods
2.1. Case Study: The Northern Prawn Fishery
The Northern Prawn Fishery (NPF), which is located off Australia'snorthern coast (Fig. 1), is a multi-species trawl fishery based on severaltropical prawn species. It is one of Australia's most valuable federallymanaged commercial fisheries, and has regularly returned a positiveprofit (Rose and Kompas, 2004) since its establishment in the late1960s. However, in recent years the fishery has experienced a declinein value as a result of the increased supply of aquaculture-farmedprawns to both domestic and international markets, strong Australiancurrency and increasing fuel prices (Punt et al., 2011).
The NPF is currently managed using input controls in the form oflimited entry, gear restrictions, as well as time and spatial closures.Management of the fishery has been supported by the developmentand application of a full Management Strategy Evaluation (MSE) ap-proach (Dichmont et al., 2006, 2008; Venables et al., 2009). Followingseveral industry and government funded buy-back schemes, the NPFnow comprises 52 vessels, which is believed to be the number requiredto achieve Maximum Economic Yield (MEY) in the fishery (Barwick,2011). By comparison, more than 120 vessels operated in the fishery adecade ago, and over 300 vessels in the 1970s and 1980s.
The NPF operates over two ‘seasons’ spanning the period April toNovember with a mid-season closure of variable length from June toAugust. Seasonal closures are in place to protect small prawns (clo-sure from December to March), as well as spawning individuals(mid-season closure) (AFMA and CSIRO, 2012). The fishery consistsof two main sub-fisheries that are (to a large degree) spatially and
Fig. 1. Map of northern Australia showing the extent of the Northern Prawn Fishery(Milton, 2001).
temporally separate1. The banana prawn sub-fishery is a single-species fishery targeting the white banana prawn (Penaeusmerguiensis), while the tiger prawn sub-fishery is a mixed speciesfishery targeting grooved and brown tiger prawns (Penaeussemisulcatus and Penaeus esculentus, respectively), as well as blue en-deavour prawns (Metapenaeus endeavouri) which are caught as by-product (Woodhams et al., 2011). Two different fishing strategiescan be identified within the tiger prawn sub-fishery, one associatedwith catching grooved tiger prawns (called the grooved tigerprawn fishing strategy) and the other associated with catchingbrown tiger prawns (called the brown tiger prawn fishing strategy).
White banana prawn stocks are strongly influenced by weatherpatterns, seasons of higher seasonal catches generally following higherthan average rainfall during the preceding summer (Vance et al.,1985). The variability of white banana prawn stocks makes it difficultto set catch or effort limits in a way that protects spawning stocks butalso allows operators to profit fromyears inwhich prawns are abundant(Buckworth et al., 2013). Tiger prawn stocks are more stable andpredictable and these species are generally more dispersed relative towhite banana prawns. Consequently, even though the same vesselsare used in both sub-fisheries, the fishing gears and techniques differ.The banana prawn sub-fishery operates mostly during the first season.However, if banana prawns are still available in large enough numbers,somevesselswill continue to target themduring the second season. Thefleet then switches during the second season to the tiger prawn sub-fishery, for which catches per unit effort are lower than for whitebanana prawns, but less variable.
2.2. Bio-Economic Model
To date, bio-economic analysis of the fishery has been largelyfocused on the more predictable component of the fishery, namelythe tiger prawn sub-fishery (Dichmont et al., 2008, 2010; Puntet al., 2011). The bio-economic model developed here synthesizes,in a single model, previous modelling works by Dichmont et al.(2003, 2008) and Punt et al. (2010, 2011) on the NPF, and extendsit by integrating the more variable banana prawn resource. Themodel is based on recent developments in mixed fisheries bio-economic modelling (Gourguet et al., 2013). Our model capturesthe major components and interactions that characterise the NPF,as described in Section 1 and detailed in Fig. 2.
Population dynamics of tiger and blue endeavour prawns are basedon a multi-species weekly time-step, sex-structured population modelwith Ricker stock-recruitment relationship and environmental uncer-tainties. The population dynamics model allows for week-specificity inrecruitment, spawning, availability and fishing mortality. However,white banana prawns are represented without explicit density-dependence mechanisms, due to highly variable recruitment and ab-sence of a defined stock–recruitment relationship.
The bio-economic analysis is based on the sake of satisfying trade-offs between expectation and variability of profitability of the entireNPF. By profitability is meant net present value (NPV) of profits,in line with capital theory and optimal control approach (Clark andMunro, 1975). A mean-variance analysis is used to examine the trade-offs.
2.2.1. Tiger and Endeavour Prawns: Multi-Species, Stochastic andDynamic Models
The population dynamics of grooved and brown tiger prawns(species s = 1 and 2, respectively) and blue endeavour prawns(s = 3) are based on a sex- and size-structured model relying on a
1 A third sub-fishery exists in the Joseph Bonaparte Gulf in the far western part of thefishery based on red-leg banana prawns (Fenneropenaeus indicus). This sub-fishery isexploited by a relatively small number of vessels as it occurs at the same time as the (morevaluable) tiger prawn sub-fishery, and is not included in the subsequent analysis.
2 Year y(t) is a function of week t, where weeks are numbered 1,…, 52, 53,…, 102, 103,…
Fig. 2. Stylized representation of the Northern Prawn Fishery used as a basis to develop the bio-economicmodel. Thewidth of the arrows between the three fishing strategies and variousprawns are proportional to the proportion of the catch by species and by fishing strategy compared to the total catch of the fishery in 2010. The three fishing strategies correspond to thetwo tiger prawnfishing strategies (i.e. the groovedand brown tiger prawnfishing strategies) and the bananaprawn sub-fishery. Both tiger prawnfishing strategies result in by-catch of theother tiger prawn species as well as endeavour prawn species.
112 S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
weekly time-step as presented in Punt et al. (2010) and summarized byEq. (1):
N!s t þ 1ð Þ ¼ gs t; N!s tð Þ; F
!s tð ÞÞ; s ¼ 1;2;3:ð ð1Þ
where t corresponds to one time step, i.e. oneweek. N!s tð Þ is thematrix ofabundance Ns,x,l(t) of prawns of species s of sex x (with x = ♂ for maleand ♀ for female) in size-class l alive at the start of time t. Grooved andbrown tiger prawns are represented by 1-mm size-classes betweenlengths of 15 and 55 mm, while blue endeavour prawns are modelledas a single aggregated length class. F
!s tð Þ is the matrix of fishing mortal-
ity Fs,l(t) of animals of species s and size-class l at time t. Details on fish-ing mortality are given in Appendix A.1. The dynamic function gsaccounts for species recruitment and mortality mechanisms of speciess as detailed in Punt et al. (2010). Recruits in the fishery for speciess = 1,2,3 during a ‘biological’ year are assumed to be related to thespawning stock size index of species s for the previous year, accordingto a Ricker stock–recruitment relationship fitted assuming tempo-rally correlated environmental variability and down-weighting re-cruitments, as described in Punt et al. (2010) and Punt et al.(2011).
2.2.2. White Banana Prawn: An Uncertain ResourceAbundance of white banana prawns (species s = 4) appears to be
more heavily influenced by the environment than by fishing pres-sure (Die and Ellis, 1999; Venables et al., 2011) and its year to yearavailability is highly variable. More specifically, stocks are stronglyinfluenced by weather patterns, generally peaking in years inwhich there has been high rainfall. It is assumed that spawningstock biomasses of white banana prawns do not influence signifi-cantly the stock abundances the following years and that annual en-vironmental influences are independent. Therefore, in the presentstudy, white banana prawn annual biomass is modelled as a uniformi.i.d. random variable:
Bs y tð Þð Þ→U B−s ;Bþ
s
� �; s ¼ 4: ð2Þ
with Bs(y(t)) the stochastic biomass of white banana prawn for theyear y(t),2 and Bs
− and Bs+ the uniform law bounds, for s = 4.
2.2.3. Fishing Mortality and CatchFishing mortalities Fs,l,f(t) due to fishing effort of fishing strategy f
(with f = 1 and 2 for the grooved and brown tiger prawn fishing strat-egies, respectively; and f = 3 for the banana prawn sub-fishery) on an-imals of species s in size-class l during time t are given by:
Fs;l; f tð Þ ¼ us tð ÞE f tð Þ; s ¼ 1;2;3 and f ¼ 1;2: ð3Þ
where Ef(t) corresponds to the effort of fishing strategy f during time t.Fishing mortality functions us are detailed in Appendix A.1.
Weekly catches Fs,l,f(t) of species s = 1,2,3 in length-class l bytiger prawn fishing strategy (f = 1,2); and annual catches Ys = 4,
f = 3(y(t)) of white banana prawns (s = 4) by the banana prawnsub-fishery (f = 3) for the year y(t) are defined by the system ofEq. (4):
Ys;l; f tð Þ ¼Xx
υs;x;lNs;x;l tð ÞFs;l; f tð Þ1− exp −Ms−
Xf¼1;2
Fs;l; f tð Þ� �
Ms þX
f¼1;2Fs;l; f tð Þ scs ¼ 1;2;3 and f¼1; 2
Ys; f y tð Þð Þ ¼ qs; fBs y tð Þð ÞEyf y tð Þð Þ s ¼ 4 and f ¼ 3:
8>><>>:
ð4Þ
with υs,x,l the mass of an animal of species s = 1,2,3 and sex x insize-class l, Ms the natural mortality of an animal of species s, andEfy(y(t)) the annual effort of fleet f during year y(t).Sub-indices used in this study are summarized in Table 1 where
their symbols, values and descriptions are displayed.
Table 1Symbols, values and descriptions of the sub-indices used in the study.
Symbols Values Description
s 1 Grooved tiger prawn species2 Brown tiger prawn species3 Blue endeavour prawn species4 White banana prawn species
l 1 to 41 1-mm length-class between 15 and 55 mmf 1 + 2 Tiger prawn sub-fishery which comprises two fishing strategies
1 Tiger prawn fishing strategy targeting the grooved tiger prawns2 Tiger prawn fishing strategy targeting the brown tiger prawns3 Banana prawn sub-fishery which targets white banana prawns
Table 2Effort combinations (in each row) considered in this study. They differ in the annual effortEf = 1 + 2y (y(t)) allocated to tiger prawn sub-fishery.
Tadapt See Eq. (10). ‘Adaptive’ tiger prawn effortT90 ∝ tig = 90 %. Ef = 1 + 2
y (y(t)) = 0.9Ey(y(t))T100 ∝ tig = 100 %. Ef = 1 + 2
y (y(t)) = Ey(y(t))
113S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
2.2.4. Fishing Income and CostsThe annual gross incomes of the tiger prawn fishing strategies
(f = 1 and 2) and banana prawn sub-fishery (f = 3) are calculated asdescribed by the set of Eq. (5):
Incf y tð Þð Þ ¼X
t¼52 y tð Þ−1ð Þþ1
52y tð Þ X3s¼1
Xl
ps;l y tð Þð ÞYs;l; f tð Þ !
; scs ¼ 1;2;3 and f ¼ 1;2
Incf y tð Þð Þ ¼ ps y tð Þð ÞYs; f y tð Þð Þ; s ¼ 4 and f¼3:
8><>:
ð5Þ
where ps,l(y(t)) is the average market price per kilogramme for animalsof species s = 1,2,3 in size-class l (related to five market categories forthe tiger prawns and corresponding to an average price for the blue en-deavour prawns, as they are represented through an aggregated length-class) during the year y(t). Grooved and brown tiger prawns aremarketed together as ‘tiger prawns’ under a common size- and time-dependent price, therefore ps,l(y(t)) are identical for s = 1 and s = 2.The average prices per kilogrammeof blue endeavour andwhite bananaprawns are denoted ps = 3,4(y(t)) and are also time-, but not size-dependent.
Variable costs Cfvar(t) for the fishing strategy f during time t, andannual fixed costs by vessel Cvfix are detailed in Eq. (6):
Cvarf tð Þ ¼ cLInc f tð Þ þ cM
X4s¼1
Ys; f tð Þ þ cKf þ cFf y tð Þð Þ� �
E f tð Þ;
Cfixv ¼ Wv þ r þ dð Þψv
8><>: ð6Þ
where CL is the share cost of labour (crew are paid a share of theincome) and CM is the cost of packaging and gear maintenance(assumed to be proportional to the fishery catch in weight). Unitcosts cfK and cf
F(y(t)) are respectively the cost of repairs and mainte-nance and the cost of fuel and oil per unit of effort of fishing strate-gy f during the year y(t). The values of these costs are assumedconstant across grooved and brown tiger prawn fishing strategiesbut differ between tiger and banana prawn sub-fisheries. Wv arethe annual vessel costs (i.e. those costs that are not related to thelevel of fishing effort), r is the opportunity cost of capital and isassumed equal to the discount rate, set at 5% following Punt et al.(2011), d is the economic depreciation rate and ψv is the averagevalue of capital by vessel.3
3 The values of cost parameters (cL, cM, cfK, cfF,Wv and ψv) are derived from an economicsurvey of the fishery during 2007–2008 (Perks and Vieira, 2010) and were adjusted forknown changes in input prices to provide estimates of the costs in 2009–2010 values.All the underlying cost and price assumptions were discussed with, and validated by in-dustry representatives who were members of the NPF Resource Assessment Group(RAG) which has responsibility for assessing the dynamics and status of NPF species.The group comprises fishery scientists, industry members, fishery economists, and theAFMA manager responsible for the fishery.
2.2.5. Annual Profit and Net Present ValueThe total annual profit π(y(t)) for the entire NPF for year y(t) is given
by:
π y tð Þð Þ ¼X3f¼1
Inc f y tð Þð Þ−X52y tð Þ
t¼52 y tð Þ−1ð Þþ1
Cvarf tð Þ
0@
1A−Cfix
v K y tð Þð Þ: ð7Þ
where K(y(t)) is the number of vessels involved in the NPF during theyear y(t).
The net present value (NPV) of the flow of profits over simulationtime is calculated as the aggregated value of discounted annual profitsand is given by:
NPV ¼XTy tð Þ¼0
π y tð Þð Þ1þ rð Þy tð Þ : ð8Þ
where r is the discount rate, and T is the terminal year of the simulation.Details on the estimations of the bio-economic model parameters
are given in B, and further details on population dynamics are given inGourguet (2013).
2.3. Effort Combinations
In this paper, the economic performance of the NPF is comparedunder six effort allocations consisting of different effort combinationsbetween tiger and banana prawn sub-fisheries. The effort combinationsare described in terms of proportion of total annual effort allocated tothe tiger prawn sub-fishery (f = 1 + 2) and are summarized inTable 2. The annual proportion ∝ tig(y(t)) of effort directed towardsthe tiger prawn sub-fishery is expressed as in Eq. (9):
Ey y tð Þð Þ ¼ Eyf¼1þ2 y tð Þð Þ þ Eyf¼3 y tð Þð Þ;
∝tig y tð Þð Þ ¼Eyf¼1þ2 y tð Þð ÞEy y tð Þð Þ
8><>: ð9Þ
where Ey(y(t)) is the total annual fishing effort for the entire NPF,Ef = 1 + 2y (y(t)) corresponds to the annual effort of tiger prawn sub-
fishery, and Ef = 3y (y(t)) of banana prawn sub-fishery, during the year y(t).
In five of the effort combinations, the annual proportion of total ef-fort allocated to tiger prawns∝ tig(y(t)) is pre-defined. Two ‘banana ef-fort combinations’ (T0 and T10) consist of setting the annual proportionof tiger prawn effort∝ tig to 0% and 10% of total annual effort. Two ‘tigereffort combinations’ (T90 and T100) involve allocating 90% and 100% ofthe annual effort to the tiger prawn sub-fishery. A ‘balanced’ effortcombination (T50) is also analysed, in which total annual effort is splitequally between the two sub-fisheries. Finally, an ‘adaptive’ effortcombination (Tadapt), which reflects the current fishing behaviour inthe NPF, is studied. Under this combination, the allocation of the totalannual fishing effort between tiger and banana prawn fishing dependsdirectly on white banana prawn catch per unit effort CPUEs = 4 asexpressed in Eq. (10):
∝tig y tð Þð Þ ¼ aCPUEs y tð Þð Þ þ b; s ¼ 4: ð10Þ
114 S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
Details on CPUE are given in Appendix A.2. Parameters a and b areestimated from a linear regression model using historical data from1994 to 20104which is displayed in Fig. C.2 in Appendix C. The resultingproportion of total annual effort directed to the tiger prawns ranges be-tween 60 and 76%.
For each of the six effort combinations, the annual tiger prawn effortis then allocated byweek and between grooved and brownfishing strat-egies as described in Appendix A.3.
2.4. Fleet Sizes
The total annual effort Ey(y(t)) for the entire NPF can be expressed asin Eq. (11):
Ey y tð Þð Þ ¼ eK y tð Þð Þ: ð11Þ
where e is the annual average effort per vessel (set to the value estimatedfor 2010: 162 days at sea) andK(y(t)) the number of vessels for year y(t).
We assess the effects of changes in fishing capacity, in terms ofthe number of vessels K(y(t))A status quo fleet size SQ correspondingto an annual number of vessels equal to the one observed in 2010,i.e. K(y(t)) = 52 vessels, but also increased and decreased fleet sizesare studied, as summarized in Table 3.
2.5. Economic Scenarios
The key economic outputs from the bio-economicmodel are the an-nual profits for the entire NPF and the associated net present value, all ofwhich are sensitive to assumptions about the values of biological andeconomic parameters. Sensitivity to economic parameters is exploredthrough the analysis of scenarios incorporating different assumptionsabout changes in fuel and prawn prices. All other economic parametersare assumed to remain constant over the simulation period. We reportresults for only two economic scenarios,5 these being a ‘base case’ sce-nario (BC) and a ‘likely’ scenario (L) detailed in Table 4.
The BC scenario assumes that prawn and fuel prices remain constantat their estimated 2010 levels over the simulation period. Variable andfixed costs are set to the average values estimated for the 2010–2012period. The likely scenario represents a possible evolution over the sim-ulation period of key economic parameters for this fishery (based onhistorical trends). Except for banana prawn, the main market for NPFprawns is Asia6 (especially Japan), and the price received is largely de-
4 Only historical data after 1993 are taken into account due tomajor changes in thefish-ery structure that occurred in that year (Shotton, 2001).
5 We tested different combination of scenarios including increase and decrease ofprawn and fuel prices. For the sake of clarity and synthesis, only the results relating to astatus quo and a likely scenario, which involve two extremes of the scenario combinationstested, are considered.
6 Until recently themainmarket for banana prawns was also Asia, however most of ba-nana prawns are now sold in domestic market.
Table 4Economic scenarios (in each row) considered in this study.
Scenarios Description
BC Base case scenario: constant prawn and fuel pricesL Likely scenario: prawn prices decrease by 3% per year
and fuel price increases by 5% per year
pendent on the Yen–AU$ exchange rate and the total supplies tothis market (Punt et al., 2010). Therefore prawn prices are assumedto be independent of the landings in our model. The likely economicscenario assumes a progressive decrease in prawn prices, by 3%annually, based on historical trends, and fuel price is assumed to followa progressive increase of 5% per year. The later assumption is supportedby a linear model adjusted to the historical data given in Fig. C.1 inAppendix C.
The economic performance of the fishery for the six effort combina-tions (c.f. Section 2.3), the three fleet sizes (c.f. Section 2.4) and underthe two economic scenarios (c.f. Section 2.5) is analysed accountingfor the stochastic nature of the model (i.e. environmental variabilitiesapplied to annual recruitments of tiger and blue endeavour prawnsand to white banana prawn annual biomasses). For every combinationof effort allocation, fleet sizes and economic scenarios, 1000 trajectoriesare simulated over a 10 year period from 2010. Each trajectoryrepresents a possible state of nature for each year of the simulation,ω(.) =(ω1(.), ω2(.), ω3(.), ω4(.)); which stands for the set of annualrecruitments of tiger and blue endeavour prawns as detailed inPunt et al. (2011) and annual biomasses of white banana prawns asin Eq. (2). The different ωi(∙) are assumed to be independent byspecies. Each combination of strategies and scenarios is simulatedwith the same set of ωi(∙). The numerical implementations and compu-tations of the model have been carried out with the scientific softwareSCILAB.7
3. Results
3.1. Mean Performance Versus Risk Trade-off in the Status quo
Fig. 3 presents the average annual profit of the fishery versus itsstandard deviation under a status quo fleet size SQ and with currenteconomic conditions in the fishery (i.e. BC economic scenario), for thesix effort combinations. For each combination, the annual variability ofincome associated with each of the different species is also displayedusing bar charts.
Overall, variability in profits increases with the increase inthe proportion of effort that is allocated to the more variablebanana prawn component of the fishery. Average annual economic
7 SCILAB is a free software http://www.scilab.org/ dedicated to engineering and scientificcalculus. It is especially well-suited to deal with dynamic systems and control theory.
Fig. 3. Economicmean-variance analysis for the six effort combinations under a status quofleet size and a BC economic scenario. For each effort combination, the bar chart representsthe standard deviations of the annual incomes estimated from the catches of each of thefour species (s = 1, 2, 3, 4).
115S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
performance increases as the proportion of banana prawn fishingincreases, up to a maximum which is reached with mixed effortcombinations, while strategies focusing on banana prawn fishingachieve lower levels of average economic performance. Of the ef-fort combinations explored, the Tadapt one, which mimics the effortcombination currently observed in the fishery, provides what ap-pears to be the best compromise between average performanceand risk. Strategies specialized on tiger prawn fishing would leadto reduced risk (originating from the two tiger and blue endeavourprawn species), at the cost of reduced average performance. Strat-egies specialized further on banana prawn fishing would only in-crease the levels of risk, with no offsetting benefit in terms ofaverage performance.
3.2. Performance of the NPF with Alternative Fleet Sizes
The mean-variance performance of the fishery with the three fleetsizes described in Section 2.4 and with alternative effort combinationsis illustrated in Fig. 4.
Comparisons across alternative fleet sizes show that a larger fleetwould achieve higher average economic performance withmixed ef-fort combinations including ‘adaptive’ (Tadapt), ‘balanced’ (T50) and‘banana specialisation’ (T10). However, such average performancewould entail much larger levels of variability in economic perfor-mance. Inversely, reducing the size of the fleet leads to a reductionof this variability, but is associated with lower mean economic per-formance levels.
Fig. 4. Average annual total profit π(y(t)) (over the years and the 1000 trajectories simulated) vereffort combination (Tadapt, T0, T10, T50, T90 and T100)with differentfleet sizes. The blue circles corresand the green crosses to a decrease in the number of vessels K−. (For interpretation of the refere
Table 5Predicted mean net present value of fleet profits (NPV) achieved by the fishery under a BC ecoalternative fleet sizes (columns). Their standard deviations are displayed in parenthesis. Mean
Table 5 presents themean (among the1000 trajectories) net presentvalue of fleet profits (NPV) associated with the six effort combinationsfor each of the three fleet sizes studied. The standard deviation ofthese NPV is displayed in parentheses.
As expected from Fig. 4, the strategy that increases fleet size al-lows higher levels of economic yield to be achieved for effort com-binations that include banana prawn fishing as a significantcomponent of the overall activity of the fleet (the T50 effort combi-nation achieving the highest yield). This is because the greater fish-ing capacity allows fishers to make the most of the peak abundanceyears in banana prawns. However, this is associated with a highlevel of inter-annual variability in economic yield. As the capacityof the fleet decreases, the ability to capture the full benefits ofhigh abundance years for banana prawns is reduced, and strategiesthat achieve greater yield (among a reduced fleet size) involve al-locating more effort towards the tiger prawns, with reduced levelsof economic variability.
3.3. Mean-Variance Analyses Under a Likely Economic Scenario
To explore the sensitivity of our results to potential economicchanges to the fishery, Fig. 5 displays the mean-variance analyses ofthefisherywith three alternative fleet sizes, and six effort combinations,under a likely (L) economic scenario.
Fig. 5 shows that under a likely economic scenario, it will notbe possible to maintain the current levels of economic yield. Givencurrent fleet size, maintaining current effort allocation would still
sus the standard deviation associated under a BC economic scenario. Results are featured bypond to a status quofleet size SQ, the red triangles to an increase in the number of vessels K+
nces to color in this figure legend, the reader is referred to the web version of this article.)
nomic scenario, associated with the different effort combinations (rows) for each of threes and standard deviations are expressed in AU$ million.
Fig. 5.Average annual total profit π(y(t)) (over the years and the 1000 trajectories simulated) versus the standarddeviation associated under a L economic scenario. Results are featured byeffort combination (Tadapt, T0, T10, T50, T90 and T100) with different fleet sizes. The blue circles correspond to a status quo fleet size SQ, the red triangles to an increase in the number ofvessels K+ and the green crosses to a decrease in the number of vessels K−. (For interpretation of the references to color in this figure legend, the reader is referred to the web versionof this article.)
116 S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
result in the highest level of average economic yield. However, effortcombinations which have similar economic yield, but lower risk areachievable for reduced fleet size.
As illustrated in Table 6, the fleet-wide economic yield measured interms of total net present value over the simulation horizon is highestwith the currentfleet size. Reducing the number of vessels, and focusingthe effort on the tiger-prawn sub-component of thefisherywould entaila modest reduction in the overall yield, but a significant reduction ininter-annual yield variability. A reduced fleet size limits the fleet'sability to gain from the years of high banana prawn abundance, hencethe fleet's effort is best targeted at the less variable tiger prawn compo-nent of the fishery.
4. Discussion
The bio-economicmodel of the NPF presented in this article is basedon the synthesis of a complex set of models developed in support ofthe Management Strategy Evaluation (MSE) approach to managingthis fishery (Dichmont et al., 2006, 2008; Venables et al., 2009). Themodel, presented here, allows for the explicit representation of both
Table 6Predicted mean net present value of fleet profits (NPV) achieved by the fishery under a Leconomic scenario, associated with the different effort combinations (rows) for each ofthree alternative fleet sizes (columns). Their standard deviations are displayed in paren-thesis. Means and standard deviations are expressed in AU$ million.
the tiger and the banana prawn sub-fisheries (which are not usuallymodelled together) and it includes a representation of seasonal alloca-tion of fishing effort, as advocated by Anderson and Seijo (2010).While many features of the original models have been simplified, keyaspects of model structure have been maintained where these wereconsidered crucial to the understanding of the bio-economic systemunder study.
4.1. Implication of Economic Risk When Managing a Mixed Fishery
The analysis illustrates an important aspect of managing mixedfisheries under the objective of maximising the net present value ofprofits, as usually proposed in optimal control or capital theoryapproach for fisheries (Clark and Munro, 1975), where fish stocksare taken as natural capital assets. Given the biological variability ofsome of the target resources in a mixed fishery, increased averageprofits may be associated with increased inter-annual variabilityin these profits, which may be perceived as a negative outcome bythe industry. This is in line with results encountered in the finance lit-erature (Fishburn, 1977; Jiang and Lee, in press; Ludvigson and Ng,2007; Rossi and Timmermann, 2010), where the higher (lower) thereturns aimed at, the higher (lower) the associated risk. More specif-ically, this study illustrates the trade-off between risk and averagereturns associated with the allocation of fishing effort across specieswith different levels of biological variability. It also illustrates thetrade-off in terms of managing the capacity of the fleet, betweenbeing able to take higher catches in years of high banana prawnabundance, but having excess capacity in the low-abundance years;or having a fishing capacity tailored towards the low banana prawnabundance years, with a risk of missing out on some of the potentialcatch in the high abundance years.
Risk aversion of the industry may lead to management optionswith lower levels of performance, but reduced economic risk, beingpreferred. Risk aversion of key stakeholders (fishers, industry, andmore broadly, society) should therefore be included in the evalua-tion of management strategies. As Mistiaen and Strand (2000) point-ed out, it is widely agreed that fisher's risk preference is a major
117S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
determinant of their responses to various changes in fishing stock,market, and weather conditions. Therefore, it is important to inte-grate fisher's risk preference in modelling and to analyse theirdecision-making behaviour (Nguyen and Leung, 2009).
The NPF operates under a strong co-management structure, withmost of the industry incorporated into a single company which isrepresented in the management decision process. Moreover industryis involved in setting the number of vessels to achieve MEY. Manage-ment and industry objectives can therefore be considered to be relative-ly aligned (Pascoe et al., 2009). Simulation results show that the totaleconomic yield could actually be increased with a larger number of ves-sels in the fleet (than the actual 52 vessels), provided the fleet allocatedhalf of its effort to the banana prawn sub-fishery (the T50 combination).However, a strategy of increased fleet size would come at the cost ofhigher economic risk. Results of this study highlight a risk trade-off as-sociated with the Tadapt effort combination. With a status quo fleetsize, this effort combination leads to maximum average annual yieldand intermediate levels of inter-annual variability in yield. This scenario,which is based on the historical pattern of effort allocation in the fishery,may thus reflect the degree to which the decision makers (industrytogether with government managers) are willing to trade of averageprofit against variability.
While the results obtained are specific to the case study consid-ered in this analysis, the methods proposed would apply to anymixed fishery where information allows calibration of a dynamicbio-economic model of fishing across a range of species presentingdifferent levels of natural variability. It is likely that most of themixed fisheries in the world would be subject to similar trade-offsbetween average economic performance and variability of this per-formance from year to year. If this is the case, we argue that the ques-tion of variability in returns of a fishery should also be consideredwhen discussing the identification of management strategies aimedat MEY. This would raise the question of the degree of risk aversionof key stakeholders, including industry, the fishers, and more broad-ly, society.
4.2. Likely Economic Scenario and Possible Adaptation Options
Analyses of economic performance of the fishery under differenteconomic scenarios illustrate the importance of sensitivity analysesto key economic parameters in bio-economic assessments. Whereasit is difficult to predict the future evolution of prices and costs as theyare influenced principally by external drivers, scenarios and projec-tions – based on the best available knowledge of these drivers –
show that the fishery is likely to encounter economic challenges,with higher fuel and lower product prices resulting in average annu-al profit levels expected to be low compared to the current situation(and even negative in some periods). Increased fleet size, given thecosts per vessel, would in this case fail to improve fishery perfor-mance, in terms of mean annual profits, net present values and eco-nomic risk. However, reducing fleet size below its historical levelwould induce a reduction in economic risk, while involving only alimited loss in economic yield. Furthermore, increase in fuel priceswould lead to a relatively more important increase in variable costsfor the banana prawn sub-fishery, compared to the tiger prawnsub-fishery, as the banana prawn sub-fishery uses a greater amountof fuel per effort unit. This, and the limited capacity of the fleet totake advantage of the good years of banana prawn abundance,would lead to favour tiger prawn effort specialisations.
4.3. Perspectives
Fisheries management increasingly acknowledges that fish popu-lation dynamics are complex and influenced by factors that are often
poorly understood. This is the case with the white banana prawn dy-namics. It may be that the conclusions of our analysis would changeif the patterns of variability in abundance of banana prawns changed,due for example to changes in the environmental drivers which de-termine its year-to-year fluctuations in abundance. In particular,rainfall and sea level rise have been identified by Hobday et al.(2008) as key impacts of climate change in the NPF region, whichmay have an impact on the dynamics of the different species targetedby the NPF, notably on white banana prawn abundance. Climatechange projections for rainfall are highly uncertain; rainfall isprojected to decrease across parts of northern Australia, with someareas showing a slight increase which may have a positive impacton white banana prawn catches (Hobday et al., 2008). Climatechange may also have an impact on seagrass beds and mangroveforests, which are important nursery grounds for tiger prawns andbanana prawns, respectively (Sands, 2011). Coupling the bio-economic model presented in this paper with projections derivedfrom models relating climate change to the environmental driversof prawn abundance could therefore allow a more informed evalua-tion of the future trade-offs between mean performance and eco-nomic risk in this fishery.
Finally, while our analysis has focused exclusively on the bio-economic trade-offs associated with species with commercialvalue, another key dimension of mixed fisheries which may alsoneed to be considered is the impact of effort combinations on by-catch species of low commercial value, as well as on threatened, en-dangered and protected species and on habitats (Woodhams et al.,2011). Different levels of fishing capacity and alternative effortcombinations, impacting differently the surrounding ecosystem,will potentially lead to different outcomes in terms of the ecologicalimpacts of fishing. This will be the focus of further research usingthe bio-economic model presented in this article.
Acknowledgement
This work was carried out as part of a co-tutelle PhD project jointlyfunded by Ifremer and the joint CSIRO/UTAS PhDProgram inQuantitativeMarine Sciences. Additional supportwas providedby the FrenchResearchAgency ANR as part of the Adhoc project, as well as by the AustralianFisheries Research and Development Corporation (FRDC). We areextremely grateful to the CSIRO researchers for providing access tothe models and data relating to the NPF, and for their guidance in theanalyses presented in the paper. We also thank three anonymousreviewers for their helpful comments on an earlier version of themanuscript.
Appendix A. Dynamics Details of the Bio-Economic Model
Appendix A.1. Fishing Mortality
Fishing mortalities of species s = 1,2,3 are given by:
Fs;l; f tð Þ ¼ As tð ÞSels;lqs; f E f tð Þ; s ¼ 1;2 and f¼1;2F3; f tð Þ ¼ As tð Þqs; f E f tð Þ; s ¼ 3 and f ¼ 1;2:
�ðA:1Þ
where As(t) is the relative availability of animals of species s duringtime t and Ef(t) is fishing effort (days at sea) associated withgrooved or brown tiger prawn sub-fishery f = 1,2 at time t.Catchability qs,f corresponds to the fishing mortality of species s as-sociated with one unit of fishing effort of fishing strategy f (as in2010) and is assumed constant over the simulation period. Sels,l isthe selectivity of the fishing gear on animals of species s in size-class l as described in Punt et al. (2010).
Fig. C.1. Linear regression of fuel price index relying on historical data from 1971 to 2010.R2 = 0.9033 and P value = 2.2 ∗ 10−16 (b 0.05, significant). The slope of the regressionline is 4.9831 (meaning an increase of 5% per year).Data source: ABARES (2010).
Fig. A.1. Flowchart of the algorithm used to determine the weekly effort (days at sea)during year y(t) of grooved and brown tiger prawn fishing strategies (f = 1, 2), accordingto the proportion of effort directed to tiger prawn sub-fishery. The variables next to thearrows represent the output from one box and input into another box.
8 The values of ϒgrooved(t) correspond to the predicted proportion of tiger prawn effortdirected towards the grooved prawns during week (t modulo 52) in 2010 derived fromthe CSIRO operating model.
118 S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
Appendix A.2. CPUE
Annual average banana catch per unit effort (CPUE) iscomputed from white banana prawn annual biomass Bs = 4(y(t)),as:
CPUEs y tð Þð Þ ¼ qs; f Bs y tð Þð Þ; scs ¼ 4andscf ¼ 3: ðA:2Þ
where qs,f is the catchability of the white banana prawn (s = 4) bythe banana prawn sub-fishery (f = 3). Estimated values of qs = 4,
f = 3 are given in Appendix B Table D.1.
Appendix A.3. Weekly Effort Allocation Model
To capture what happens currently in the NPF, the total annualfishing effort is allocated weekly between tiger and banana prawnsub-fisheries, and then between the two tiger prawn species througha simplified, two-steps, effort allocation model.
Step 1 Weekly tiger prawn sub-fishery effort allocation.An empirical approach is taken to predict the weekly allocation ofthe tiger prawn sub-fishery effort for year y(t). Because of thegreat variability of the variousweekly effort patterns of the histor-ical years, using a fixed weekly pattern is not relevant. Due to theinfluence of the season start dates and annual effort of bananaprawn sub-fishery on the tiger prawn weekly effort patterns,the solution to randomly select a year among the historicalyears 1994 to 2010 was estimated as not optimal for this study.Sensitivity analyses of weekly patterns would indeed be neces-sary, and would increase the number of simulations to run, in-creasing significantly the level of complexity of the model. Forsimplicity, weekly tiger prawn effort are estimated as describedin Eq. (A.3).
minEHistf¼1þ2
Xt
jjE f¼1þ2 tð Þ−EHistf¼1þ2 tð Þjji"
ðA:3Þ
E f¼1þ2 t�i� � ¼ 0X
t
E f¼1þ2 tð Þ ¼ Eyf¼1þ2 y tð Þð Þ
8<:with ti
∗ corresponding to the weeks which are closed to thefishery for management constraints (the pattern of
open and closed weeks is set to that which occurred in 2010),and Ef = 1 + 2
Hist standing for the weekly historical effort registeredfor the tiger prawn sub-fishery.
Step 2 Grooved and brown tiger prawn fishing strategies effortallocation.The second step of the effort distribution model (arrow (3)in Fig. A.1) allocates the weekly tiger effort to the groovedand brown tiger prawns. This is achieved using a fixed pat-tern8 ϒgrooved (t modulo 52) of proportion of weekly tigerprawn effort directed towards grooved tiger prawns(f = 1) at time (t modulo 52). The effort by week directedtowards grooved ( f = 1) and brown ( f = 2) tiger prawnsis described by Eq. (A.4):
E f¼1þ2 tð Þ ¼ E f¼1 tð Þ þ E f¼2 tð Þ;E f¼1 tð Þ ¼ ϒgrooved tmodulo52ð ÞE f¼1þ2 tð Þ:
(ðA:4Þ
Appendix B. Parameter Estimation
Dichmont et al. (2003) and Punt et al. (2010) describe theapproaches used to estimate parameter values for the dynamic pop-ulation models. The impact of parameter uncertainty was exploredin Punt et al. (2010). A non-linear least-squares method was usedfor the estimation of the parameters (Bs = 4
− , Bs = 4+ and qs = 4,f = 3)
related to white banana prawn by fitting observed data of white ba-nana prawn catches (in weight) and annual banana fishing effortover 17 years, from 1994 to 2010 (c.f. Fig. C.2 in Appendix C). Basecase values of all biological and economic parameters are given inAppendix D.
Appendix C. Statistical Analyses
This appendix displays the outputs of statistical analyses used tocalibrate the bio-economic model and scenario projections. Fig. C.1represents the linear regression used for the projection of the fuel pricesand Fig. C.2 the one used for the Tadapt effort combination described inSection 2.3.
Fig. C.2. Linear regression (∝ tig(y) = aCPUEs = 4(y) + b) between the average annual banana catch per unit effort, CPUEs = 4 and the annual proportion of tiger prawn sub-fishery effort,∝ tig. Model relies on historical catches and effort data from 1994 to 2010. R2 = 0.7016 and P value = 1.657 ∗ 10−5 (b 0.05, significant). a = −1.172 ∗ 10−4 and b = 0.813.
Table D.4Economic parameters.
(a) Variable costs
Parameters Fishing strategies
tiger (f = 1,2) banana (f = 1,3)
Unit cost of repairs and maintenance, cfK 332 (AU$/day) 529 (AU$/day)Base unit cost of fuel and grease, cfF(2010) 1815 (AU$/day) 2236 (AU$/day)Share cost of labour, cL 0.24 0.24Cost of packaging and gear maintenance, cM 0.92 (AU$/kg) 0.92 (AU$/kg)
(b) Fixed costs and rates
Parameters Value
Annual vessel costs, Wv 296,847 (AU$/vessel)
119S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
Appendix D. Bio-Economic Parameter Values
This appendix displays the values of the biological and economicparameters used to calibrate the bio-economic model presented inSection 2. Table D.1 displays the parameters related to the white ba-nana prawn, while Table D.2 summarizes the catchability parametervalues for the grooved and brown tiger and blue endeavour prawns.Tables D.3 and D.4 summarize the values of parameters involved inthe profit equation. Table D.5 exhibits the weekly proportion oftiger prawn sub-fishery effort directed towards grooved andbrown tiger prawn fishing strategies used to split the tiger prawnsub-fishery effort into grooved and brown tiger prawn fishing strat-egies as described in Appendix A.3.
Table D.1Estimated parameters related to white banana prawn (s = 4 and f = 3).
Bs−
(in thousand tonnes)Bs+
(in thousand tonnes)Catchability,qs,f
White bananaprawn
28.72 125.8 0.0000142
Table D.2Estimated values of catchabilities qs,f by species s and by tiger prawn fishing strategiesf = 1,2 for a fishing power of the fishery as in 2010.
120 S. Gourguet et al. / Ecological Economics 99 (2014) 110–120
References
ABARES, 2010. Australian Commodity Statistics 2010. Technical Report. Canberra.AFMA, CSIRO, 2012. Harvest strategy for the Northern Prawn Fishery under input con-
trols. Technical Report. Canberra.Anderson, L., Seijo, J., 2010. Bioeconomics of Fisheries Management. Wiley-Blackwell.Barwick, M., 2011. Northern Prawn Fishery Data Summary 2010. Technical Report. NPF
Industry Pty Ltd, Australia.Buckworth, R., Ellis, N., Zhou, S., Pascoe, S., Deng, R., Hill, F., O'Brien, M., 2013. Comparison
of TAC and current management for theWhite Banana Prawn fishery of the NorthernPrawn Fishery. Final Report to the Australian Fisheries Management Authority. Tech-nical Report. CSIRO Marine and Atmospheric Research, Brisbane.
Clark, C., Munro, G., 1975. The economics of fishing and modern capital theory: a simpli-fied approach. J. Environ. Econ. Manag. 2, 92–106.
Dichmont, C., Punt, A., Deng, A., Dell, Q., Venables, W., 2003. Application of a weeklydelay-difference model to commercial catch and effort data for tiger prawns inAustralia's Northern Prawn Fishery. Fish. Res. 65, 335–350.
Dichmont, C., Deng, A., Punt, A., Venables, W., Haddon, M., 2006. Management strategiesfor short-lived species: the case of Australia's Northern Prawn Fishery: 1. Accountingfor multiple species, spatial structure and implementation uncertainty when evaluat-ing risk. Fish. Res. 82, 204–220.
Dichmont, C., Deng, A., Punt, A., Ellis, N., Venables, W., Kompas, T., Ye, Y., Zhou, S., Bishop, J.,2008. Beyond biological performance measures in Management Strategy Evaluation:bringing in economics and the effects of trawling on the benthos. Fish. Res. 94, 238–250.
Dichmont, C., Pascoe, S., Kompas, T., Punt, A., Deng, A., 2010. On implementing maximumeconomic yield in commercial fisheries. Proc. Natl. Acad. Sci. 107, 16–21.
Die, D., Ellis, N., 1999. Aggregation dynamics in penaeid fisheries: banana prawns (Penaeusmerguiensis) in the Australian Northern Prawn Fishery. Mar. Freshw. Res. 50, 667–675.
Epstein, L., 1985. Decreasing risk aversion and mean-variance analysis. Econometrica945–961.
Fishburn, P., 1977. Mean-risk analysis with risk associated with below-target returns. Am.Econ. Rev. 67, 116–126.
Francis, R., Shotton, R., 1997. ‘Risk’ in fisheries management: a review. Can. J. Fish. Aquat.Sci. 54, 1699–1715.
Gourguet, S., 2013. Ecological and economic viability for the sustainable management of mixedfisheries. Ph.D. thesis. Université de Bretagne Occidentale and University of Tasmania.
Hilborn, R., Maguire, J., Parma, A., Rosenberg, A., 2001. The precautionary approach andrisk management: can they increase the probability of successes in fishery manage-ment? Can. J. Fish. Aquat. Sci. 58, 99–107.
Hobday, A., Poloczanska, E., Matear, R., et al., 2008. Implications of climate change forAustralian fisheries and aquaculture: a preliminary assessment. Technical Report.Report to the Department of Climate Change. Canberra, Australia.
Holland, D., Herrera, G., 2009. Uncertainty in the management of fisheries: contradictoryimplications and a new approach. Mar. Resour. Econ. 24, 289.
Jiang, X., Lee, B.S., 2013. The intertemporal risk–return relation: a bivariate model ap-proach. J. Financ. Mark. Available online 27 February 2013, ISSN 1386-4181, http://dx.doi.org/10.1016/j.finmar.2013.02.002. (in press).
Kasperski, S., Holland, D., 2013. Income diversification and risk for fishermen. Proc. Natl.Acad. Sci. 110, 2076–2081.
Ludvigson, S., Ng, S., 2007. The empirical risk–return relation: a factor analysis approach.J. Financ. Econ. 83, 171–222.
Markowitz, H., 1952. Portfolio selection. J. Financ. 7, 77–91.Milton, D., 2001. Assessing the susceptibility to fishing of populations of rare trawl by-
Mistiaen, J., Strand, I., 2000. Location choice of commercial fishermenwith heterogeneousrisk preferences. Am. J. Agric. Econ. 82, 1184–1190.
Munro, G., 2010. Theway forward: getting the economic theory right—the first steps. Pro-ceedings of the Fifteenth Biennial Conference of the International Institute of Fisher-ies Economics and Trade.
Nguyen, Q., Leung, P., 2009. Do fishermen have different attitudes toward risk? An applica-tion of prospect theory to the study of Vietnamese fishermen. J. Agric. Resour. Econ. 34,518.
Pascoe, S., Proctor, W., Wilcox, C., Innes, J., Rochester, W., Dowling, N., 2009. Stakeholderobjective preferences in Australian Commonwealth managed fisheries. Mar. Policy33, 750–758.
Perks, C., Vieira, S., 2010. Australian fisheries surveys report 2010, results for selected fish-eries, 2007–08 and 2008–09, preliminary estimates for 2009–10. Technical Report.ABARES report prepared for the Fisheries Resources Research Fund, Canberra,December.
Punt, A., Deng, R., Dichmont, C., Kompas, T., Venables, W., Zhou, S., Pascoe, S., Hutton, T.,Kenyon, R., van der Velde, T., et al., 2010. Integrating size-structured assessmentand bioeconomic management advice in Australia's Northern Prawn Fishery. ICESJ. Mar. Sci. 67, 1785–1801.
Punt, A., Deng, R., Pascoe, S., Dichmont, C., Zhou, S., Plagányi, É., Hutton, T., Venables, W.,Kenyon, R., van der Velde, T., 2011. Calculating optimal effort and catch trajectoriesfor multiple species modelled using a mix of size-structured, delay-difference andbiomass dynamics models. Fish. Res. 109, 201–211.
Rose, R., Kompas, T., 2004. Management Options for the Australian Northern Prawn Fish-ery. Australian Bureau of Agricultural and Resource Economics, Commonwealth ofAustralia, Canberra 47.
Rossi, A., Timmermann, A., 2010. What is the shape of the risk–return relation? AFA 2010Atlanta Meetings Paper.
Roy, A., 1952. Safety first and the holding of assets. Econometrica 431–449.Sanchirico, J., Smith, M., Lipton, D., 2008. An empirical approach to ecosystem-based fish-
ery management. Ecol. Econ. 64, 586–596.Sands, A., 2011. Fisheries and climate change: potential challenges and opportunities for
Sethi, S.A., 2010. Risk management for fisheries. Fish Fish. 11, 341–365.Shotton, R., 2001. Changes in fleet capacity and ownership of harvesting rights in
Australia's Northern Prawn Fishery. FAO Fisheries Technical Paper 412.Vance, D., Staples, D., Kerr, J., 1985. Factors affecting year-to-year variation in the catch of
banana prawns (Penaeus merguiensis) in the Gulf of Carpentaria, Australia. ICESJ. Mar. Sci. 42, 83–97.
Venables,W., Ellis, N., Punt, A., Dichmont, C., Deng, R., 2009. A simulation strategy for fleetdynamics in Australia's Northern Prawn Fishery: effort allocation at two scales. ICESJ. Mar. Sci. 66, 631–645.
Venables, B., Hutton, T., Lawrence, E., Rothlisberg, P., Buckworth, R., Hartcher, M., Kenyon,R., 2011. Prediction of common banana prawn potential catch in Australia's NorthernPrawn Fishery. Prepared for Australian Fisheries Management Authority. TechnicalReport. CSIRO.
B.G.Fishery status reports 2010: status of fish stocks and fisheries managed by theAustralian Government. In: Woodhams, J., Stobutzki, I., Vieira, S., Curtotti, R.,Woodhams, J., Stobutzki, I., Vieira, S., Curtotti, R. (Eds.), Technical Report. AustralianBureau of Agricultural and Resource Economics and Sciences.
