Journal of Environmental Management 131 (2013) 270e279

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Environmental cost-effectiveness analysis in intertemporal natural resource policy: Evaluation of selective fishing gearq Lone Grønbæk Kronbak*, Niels Vestergaard Department of Environmental and Business Economics, University of Southern Denmark, Niels Bohrs Vej 9, DK-6700 Esbjerg, Denmark

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 November 2012 Received in revised form 4 July 2013 Accepted 24 September 2013 Available online 31 October 2013

In most decision-making involving natural resources, the achievements of a given policy (e.g., improved ecosystem or biodiversity) are rather difficult to measure in monetary units. To address this problem, the current paper develops an environmental cost-effectiveness analysis (ECEA) to include intangible benefits in intertemporal natural resource problems. This approach can assist managers in prioritizing management actions as least cost solutions to achieve quantitative policy targets. The ECEA framework is applied to a selective gear policy case in Danish mixed trawl fisheries in Kattegat and Skagerrak. The empirical analysis demonstrates how a policy with large negative net benefits might be justified if the intangible benefits are included. Ó 2013 Elsevier Ltd. All rights reserved.

JEL classification: D61 D99 Q20 Q22 Keywords: Bio-economic model Cost-effectiveness analysis Evaluation Selective fisheries Fishing gear

1. Introduction Recent years have seen increased interest in endangered marine species.1 Nearly 1900 species are listed under the Endangered Species Act of 1973 (ESA),2 and several important fish stocks are being overfished and are below long-term sustainable levels (Commission of the European Communities, 2009). For many of these species, active recovery plans typically determined by noneconomic expert opinions, such as those of the ICES,3 are implemented with the aim of rebuilding depleted stocks. Several

q The authors thank the participants at the Danish Environmental Economic Council Conference 2010, the IIFET 2012 Conference, colleagues from our research group and two anonymous referees for useful comments and constructive suggestions. * Corresponding author. Tel.: þ45 6550 1000. E-mail addresses: [email protected] (L.G. Kronbak), [email protected] (N. Vestergaard). 1 E.g., mammals, turtles, fish, invertebrates and plants. 2 http://www.nmfs.noaa.gov/pr/pdfs/laws/esa.pdf (last accessed March 2013). 3 International Council for the Exploitation of the Sea, www.ices.dk (last accessed March 2013). 0301-4797/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jenvman.2013.09.035

recovery plans exist within the European Union for different endangered fish species, including sole in the Bay of Biscay, Southern hake and Norway lobster stocks in the Cantabrian Sea and Western Iberian Peninsula, and cod stocks in Kattegat and the North Sea (Council Regulations (EC) 2166/2005, 388/2006, 1342/2008). While many of the recovery and conservation plans4 are aimed at a single endangered species, they often have broader ecosystem consequences. If recovery plan regulations are interpreted as an increased political willingness to pay for conserving species, economic consequences for the fisheries of other species must also be taken into account. Otherwise, the total costs and benefits of the regulation are unknown, and the regulations are set by fumbling in the dark. Cost-Benefit Analysis (CBA) is, in principle, an appropriate approach for evaluating changes in fishery policies to achieve a

4 The simple scientific justification for drafting recovery plans is that by limiting the fishery in the short-run, the stock biomass will grow over time and will provide the basis for a better fishery in the long-run. This classical bio-economics reasoning was formulated 40e50 years ago (see Clark, 1976) and has been promoted regularly since then (see Grafton et al., 2007 for a recent example).

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

first-best solution.5 In most cases, this approach involves a dynamic bio-economic model, including the interaction between the ecosystem and the fishery. In reality, such model exercises are not performed, most likely because doing so is too complex and timeconsuming. Furthermore, the targets for the fishery policy may be predetermined and even non-monetary, such as changes in stock biomass. However, a recent trend in the literature applies bioeconomic modeling approaches in simulation exercises, where management strategies are evaluated in terms of biological, economic and management criteria (Kraak et al., 2008; Hoff and Frost, 2008). Several of these criteria are important. Because no systematic statement of benefits and costs is made, it is difficult to evaluate whether the management strategies are improving the economic welfare. Furthermore, concepts such as biodiversity and protection of habitat are often used to justify policies and are implicitly added to the benefit side without quantification (Armstrong, 2007). These benefits, called intangible benefits, are part of the total economic value and are important to include, but difficult to assess. Thus, if there is no good measure of all the benefits, it is impossible to make an efficient policy choice. In some cases, however, measures of the intangible benefits are not needed. Such is the case where the tangible net benefits (NB) are positive and intangible net benefits are expected to be positive. For example, it has been argued by Grafton et al. (2007) that reducing the fishery effort in a single species fishery below the Maximum Sustainable Yield level will pay off. In some cases with positive tangible net benefits, one could imagine that the intangible cost increases and that the assessment of these costs will be necessary. However, our focus is on cases where the net benefits of the tangible benefits and costs are negative. If, for instance, the reduction of effort is made in a mixed fishery,6 it might result in a reduction in catches of several economically important species for which the future tangible benefits are lower than the tangible costs due to a reduction in catches. In these cases, assessment of the intangible benefits and costs becomes important.7 The objective of recovery plans is to increase the stock biomass with a given rate (Council Regulation (EC) 1342/2008). The framework suggested in this paper is rooted in the costeffectiveness analysis (CEA) tradition, in which the best action among alternatives that achieves the objective with the least cost is identified. CEA is appropriate as a criterion for assessing management actions in at least two cases. In the first case, actions are determined primarily for conservation reasons (e.g., fishery managers want the fish stock to increase by a certain percentage rather than choosing the stock level that maximizes economic rents). In the second case, already-determined management actions have a highly uncertain or non-monetary benefit. Fishery conservation policies fall within both categories because they are often based on opinions with no economic expertise, and the benefits are often non-monetary. The framework studied here attempts to answer

5 For a comprehensive discussion of CBA’s role in the evaluation of natural resource policy, see Van Kooten and Bulte (2000). 6 It is well known that in a multispecies fishery, optimal equilibrium stock may be below the single species MSY stock level, and economically optimal management involves a trade-off in exploitation of different species. 7 These intangible benefits could be included in a CBA by using the non-market valuation. The aim of the proposed methodology is to provide recommendations on policy alternatives where non-monetary targets for the fishery policy may be predetermined. The overall societal value is not in the core but is instead the ranking of different alternatives. Furthermore, non-market valuation is not straightforward because the information to describe the change in the marine ecosystem in terms of services people care about is often unavailable. According to Innes and Pascoe (2010), the same is true for other non-market benefits such as reductions in the level of habitat change or the mortality of bottom fauna due to gear passage across the seabed.

271

questions such as the following: Which policy should be implemented if we want to have, for example, a stock biomass that is 20% higher than the current one 10 years from now? What are the costs per unit of the stock biomass? The outcome of applying CEA to conservation fishery policies is a cost-effectiveness ratio, which consists of a measure in monetary terms of costs per physical (non-monetary) unit change in the relevant fish stock (called effects). Thus, CEA has the advantage of measuring the effects of a policy alternative in quantitative, nonmonetary units and relating these effects to the costs of the policy. CEA has been applied in several different areas of environmental management, for example, in water quality (Hajkowicz et al., 2008), waste management (Van Beukering et al., 2009), nitrogen emission (Schou et al., 2000) and the mitigation of climate changes (Berndes and Hansson, 2007). Goulder et al. (1999) find the cost-effectiveness ratio of different environmental instruments in a second-best setting. To the knowledge of the authors, CEA has not yet been applied to the evaluation of fishery regulations. Such application would be an appropriate way to assess different management instruments under fishery conservation policy, where the objective of the policy is predefined and the objective of the evaluation is to find the least cost option. In summary, the aim of the paper is to formalize a framework for the economic evaluation of management actions in dynamic conservation policies where the objective of the policy has been identified. The approach is developed for cases where tangible net benefits from the policy are negative but with some intangible benefits from the policy that have to be allowed for. A second aim is to open a discussion about the measurement of effects in fisheries after implementing resource-saving technologies and then to relate these effects to the costs in a cost-effectiveness framework. The evaluation approach and the effects measured are empirically applied to the Kattegat and Skagerrak mixed trawl fishery, where analysis of the bio-economic consequences of the implementation of selective gear show negative tangible net benefits (Kronbak et al., 2009). The paper continues as follows. Section 2 introduces the methodology. Section 3 frames the methodology in an economic fishery context, including definitions of effect measures and decision rules. Section 4 introduces an example in which the framework developed in Sections 2 and 3 is applied. Section 5 presents the discussion, and Section 6 concludes the paper offers suggestions for further research. 2. Methodology A comprehensive economic evaluation analysis includes every change in costs and benefits of a policy implementation. A costbenefit analysis is an example of such an evaluation. By comparing the costs and benefits based on the net present value criteria, it is possible to give recommendations about the changes in the economic welfare of the policy. Formally, the net benefits are represented by:

DNB ¼

T X DBt  DCt t¼0

ð1 þ rÞt

(1)

where DBt measures the changes in benefits compared with the baseline scenario at time t, DCt measures the changes in costs compared with the scenario baseline at time t, r is the discount rate and T is the last period included in the evaluation, or the project’s lifetime. In the area of natural and environmental resources, however, some of the environmental effects or impacts are difficult to

272

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

measure in monetary units; they are, to a certain extent, intangible. Examples are better air quality, improved biodiversity and more robust stocks of renewable resources, for which it might be possible to measure the effects but difficult to determine the value of these effects. A slightly different approach to the CBA can be applied, where the non-environmental impacts are measured in the traditional CBA framework and the environmental consequences are measured as the net present value of the stream of environmental consequences for the considered period. Perman et al. (2003) refer to this type of analysis as an environmental cost-benefit analysis (ECBA). By comparing changes with the baseline scenario, ECBA is represented by:

DNB ¼

T X DBt  DCt t¼0

ð1 þ rÞ

t

þ

T X DEBt t ¼ 0 ð1

þ rÞt

¼ DB  DC þ DEB

(2)

where DEBt refers to the changes in environmental benefits. DB is the present value of changes in benefits, and DC is the present value of changes in costs. Thus, DB  DC refers to the change in net present value when ignoring environmental impacts, and DEB refers to the change in net present value of environmental benefits compared with the baseline. From the net present value criteria, the necessary condition for recommending a given policy is DNB  0, or DC DB  DEB. Environmental effects may be difficult to assess monetarily and, in many cases, impossible to quantify. However, it may still be possible to quantify the environmental effects of a policy alternative. The value from the change in environmental effects is more complex because it includes values from both active and passive (or non-) uses of the environment. One way to categorize the environmental effects in a total economic value framework is to add up the active and passive uses. Active use might be divided into rivalrous and non-rivalrous use of the environment. Passive use is often divided into subcategories (the existence value, the option value and the bequest value). If we divide the change in net present value of net costs (or negative net benefits) with the change in effect (and similarly on the right-hand side of the inequality (2)), a cost-effectiveness measure is formulated:

DC  DB DEB  DE DE

(3)

where DE is defined as the present value of changes in effects. The left-hand side of Eq. (3) is a cost-effectiveness-ratio, for example, costs per unit of effect, where costs are interpreted as the costs in excess of the benefits. The criterion basically states that the tangible net present value of costs minus benefits per unit of effect achieved from a given policy should not exceed the environmental benefit per unit of effect. In summary, our proposal is to combine CEA with the ECBA approach.8 This approach is called environmental costeffectiveness analysis (ECEA). The environmental cost-effectiveness ratio can be used in different ways in the decision-making process. When a decision has been made to achieve a given effect level, the policy alternative with the lowest ratio should be chosen. For the policy under consideration, the ratio will provide information about the costs of achieving the effects, which the decision-makers can compare with their “willingness to pay for one unit of effect” (wtp). The wtp can

8 To see this connection, assume that the marginal value of an additional unit of effects is constant and can be lumped into a single ‘price’, which we can define as the willingness to pay (wtp) per unit of effect. By this, the change in environmental benefits can be defined as the willingness to pay per unit of effect times the change PT t in effects, or: DEB ¼ t¼0 ðwtp$DEt Þðð1 þ rÞ Þ ¼ wtp$DE. By applying this definition, rearranging the formula for ECBA and assuming a positive change in effects (DE > 0), the criterion for a policy decision is determined as: ððDC  DBÞ=DEÞ  wtp.

be interpreted as a political wtp price, or cut-off price, per unit of additional effect achieved by the policy. If several policy options are available, only those with cost-effectiveness ratios below a given wtp-level (benchmark) should be implemented. Finally, the costeffectiveness ratio can be used to determine, given a fixed overall acceptable loss, which policies to implement. Because environmental cost-effectiveness analysis measures the tangible costs and benefits of a policy and relates it to the effects, measured in quantitative non-monetary units, the analysis cannot answer whether a policy is socially worthwhile (this is the aim of the full cost-benefit analysis), but it can indicate which of the alternatives offers the most cost-effective approach to achieving a certain objective. In a cost-effectiveness analysis, the effects are traditionally limited to a one-dimensional type of effect. To expand the dimension, a preference measure or weight can be multiplied to each effect, and then the weighted effects can be aggregated into a single common measure. In public health, a well-defined, though controversial, example is the Quality-Adjusted Life Years (QALY), which captures both reduced morbidity and reduced mortality in a single measure (Drummond et al., 2005). In other areas, such as fisheries, there are no predefined weights between, for example, species or age classes that merges these into a single common effect measure. Such weights will define the relative importance of species or age-classes with respect to each other and over the time horizon. From an economic analysis perspective, one way to describe the relative importance of species is to use the market price, which defines weights dependent on the relative economic importance of the specie/age.

3. Proposed ECEA framework applied to a fishery Measuring the environmental effects of different policies aimed at protecting species is not an easy task. One might ask, for example, what the aim is of a fishery policy introducing moreselective gear. In general, there may be several answers to this question, such as the following: (1) having a stock biomass that exceeds the current biomass level (and, if so, determining whether a longer time horizon should be considered, perhaps infinite or in some future equilibrium, e.g., ten years); (2) having an improved composition of age classes in a stock; and (3) having improved biomass levels of several stocks. Applying ECEA to fishery management problems does raise the issue of clearly defining the effects on the objective; thus, an appropriate measure of achievement9 is required for the policy objective. Implementing resource-saving technologies is one way for managers to achieve larger and more robust stock(s), which is a likely (non-economic) objective in a fishery conservation policy. There are no previous studies evaluating fishing policies by costeffectiveness analysis, and therefore, there is no reference list such as those used in public health, where broader objectives, such as maximizing health gains, are the most important (Drummond et al., 2005). There are applications of CEA studies in the environmental economics literature, but they primarily use static general equilibrium models (see, for example, Goulder et al., 1999) or static partial equilibrium models (e.g., Schou et al., 2000). Thus, the effect measure is limited to a static measure that does not take into consideration the development of the effects in different periods. Studies in the fishery management literature are primarily related to reference points such as ICES working groups, which have a measure for a precautionary approach to determining the spawning stock biomass for several species (ICES, 1998) but do not include

9

Called effects or consequences.

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

problems of weighting different achievements into a lumped effect measure. Furthermore, these effect measures are static in nature because they are concerned with the stock in a given year. An evaluation of the implementation of a more selective gear type is dynamic in the sense that it imposes costs and benefits due to changes in harvest over several years, but it also changes the stocks in all these years compared with a baseline. What should be included in the effect measure, however, is not obvious. For example, should it be an increase in the stocks in all years compared with a baseline scenario, or only the equilibrium increase in the stock? Furthermore, a more robust stock may also be related to the age composition of the stock, and a broader approach may be to consider all species in question. The effects should be clearly defined in the environmental cost-effectiveness analysis because effects across alternatives have to be comparable. The one-shot equilibrium effect, traditionally applied in the environmental literature (Schou et al., 2000), has the advantage of being easy to calculate. It takes the simplest approach by including the effects from the increase in stock once equilibrium is reached. If, however, the measure is to include the intangible values of the improved stocks, then part of the improvement will also be seen in all the years leading up to the equilibrium. The one-shot equilibrium effect does not capture these improvements. Thus, the approach seems too simplistic regarding biomass changes over the years. Different alternative effect measures are proposed here to address this problem. These measures can be applied in an evaluation of rebuilding policies. 1) Single species stock effect over time This effect measure captures the change in biomass due to a stock rebuilding policy until equilibrium in the stock is reached and is compared with the baseline continuously in each year under consideration. This measure includes the total present value of the yearly effects measured as an increase in biomass per year:

DE ¼

G X SWP  SWOP t

t¼0

t

(4)

ð1 þ rÞt

where G refers to the year when the stock reached equilibrium with the implementation of the rebuilding policy, SWP refers to the total t biomass at time t with the rebuilding policy (WP), SWOP refers to t the total biomass at time t without the rebuilding policy (WOP) and r is the discount rate. Adding the value of the increased biomass beyond year G,10 when the stock reaches equilibrium, is a way to extend this measure. An intuitively clear formulation of this approach is to apply the formula for perpetuity, assuming that the biomass stays at the equilibrium level. Formally, this approach assumes there is no WOP growth in the difference of biomass ðSWP Gþ1  SGþ1 Þ after the equilibrium is reached in year G, once the rebuilding policy has been implemented. The present value of the potential effects of the rebuilding policy is calculated by applying the formula to calculate the effects in the year when equilibrium is reached and discounting them to present value terms:

DE ¼

G X SWP t t¼0

 SWOP t

ð1 þ rÞt

þ

 . WOP SWP r Gþ1  SGþ1 ð1 þ rÞG

(5)

10 The equilibrium may not be reached in the same year for all policies; in such cases, it is suggested that the same G be applied for all policies, where G is defined as the largest equilibrium year for the different policies.

273

This effect measure has an advantage over the traditional one-shot measure in that it includes the effects in each period under consideration. 2) Age-weighted single species stock effect An alternative proposal is to include aggregated effect measures in ECEA, in which effects on different age classes of the stock are included. This measure takes the stance that some age-classes in a stock are more important than are others. From an economic perspective, one way to define weights for different age-classes in a stock is to consider the economic value of the different age-classes, reflected in the ex-vessel market price. For this purpose, an index based on prices is defined that allows the different effects on the age composition in the biomass to be lumped into a weighted effect measure. The following formula demonstrates the method for a limited age-weighted single stock effect:

DE ¼

G X

Pn

i¼1 wi

t¼0



WOP sWP i;t  si;t

ð1 þ rÞt

 (6)

where n is the total number of age-classes for the species in question; si,t is the total weight of the age class i at time t; WP and WOP represent cases with and without the rebuilding policy, respectively; and wi is the weight year class i has in the aggregate effect measure. This effect measure has the additional advantage of evaluating a change in the age composition. This measure can be extended to an infinite measure similarly to that above, and it can be easily extended to a multispecies effect measure. The above measures implicitly assume that stocks reach an equilibrium stock size at a certain point in time. In fisheries where equilibrium with constant stock sizes may not exist (for example, in predator-prey systems, or if stochastic shocks affect the system), the extension of the first measure, the single species stock effect over time, would not be well-defined. For the other measures, the point in time G could, in principle, refer to any point in time the politicians would like to include and does not necessarily have to refer to the time when equilibrium has been reached. In addition, all the measures implicitly assume that all the effects are discounted at the same rate as used for discounting monetary units. The arguments for discounting effects are based on the logic that discounting only costs, and not effects at the same rate, can result in inconsistencies in reasoning. The decisionmaking process can break down completely if effects are not discounted (Brent, 2003). The measures have increased complexity compared with the traditional one-shot measure, and they allow for the evaluation of effects in a broader time frame. Further, the second measure allows for evaluating the consequences for the age composition or, alternatively, species composition for alternative policies. 3.1. Decision rules We apply two commonly used decision rules in costeffectiveness analysis (Johannesson, 1996). These two rules depend on whether (1) effects are to be maximized given a fixed budget within which most effects are to be achieved or (2) there is a maximum price per unit of effect achieved. For decision rule one, in the case of an ECEA for fisheries there may not be a physically fixed budget available to pay for the implementation of selective gear but, rather, a maximum acceptable loss measured as the present value of net benefits (DB  DC), which we will refer to as a budget limitation. For decision rule 2, the maximum price (e.g., cost of

274

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

If, on the other hand, the policies are mutually exclusive, the incremental cost-effectiveness ratio is calculated first by the slope from the origin (e.g., the baseline) to policy 1 (the least effective alternative) and then by the slope from policy 1 to policy 2, illustrated by the dotted line, which is the ICER. Based on the C/E-ratio or the ICER, the decision now depends upon one of the two decision rules. This situation reveals 4 different decision scenarios dependent on, first, the type of decision rule and, second, whether the policies are independent. The decision scenarios are summarized below.

Fig. 1. Illustration of C/E-ratio (slope of the solid lines) and ICER (slope of the dotted line).

implementing the gear) in the case of ECEA in fisheries is interpreted as an acceptable political wtp per unit of effect achieved. To apply both decision rules, it is necessary to calculate the appropriate cost-effectiveness ratio. However, the appropriate costeffectiveness ratio depends on whether the different alternatives are independent or mutually exclusive. If the policies are independent, one policy will not affect the implementation of another policy on the cost side or the effect side, and all policies will be compared with the baseline scenario. This approach allows for calculating the traditional Cost-Effectiveness Ratio (C/E-ratio) for each policy as the cost per unit of effect. If different alternatives for obtaining the achievements are mutually exclusive, two alternatives cannot be implemented simultaneously because they either exclude each other or affect the costs/effects of each other. An example of exclusivity is a vessel that cannot fish with two gear types simultaneously but that can fish part time with each gear type. Having the exclusivity or independency, and the goal of the most cost-effective policy given one of the two decision rules, the policy recommendation can be formulated. Policies should be implemented based on the increase in effect because another policy with a greater effect can replace a policy with a lesser effect. For this purpose, the incremental changes in effects and costs can be calculated as the marginal, or incremental, cost-effectiveness ratio (ICER), which is defined as the incremental costs per incremental effect. This calculation gives a measure of the cost of achieving additional effects compared with the effects that can be achieved by the less effective alternative, and it is calculated, after ordering alternatives according to increasing effects, as follows: ICER ¼ ðDðDNBÞ=DðDEÞÞ ¼ ðDNB1  DNB2 Þ=ðDE2  DE1 Þ.11 The ICER measure applies the changes in net benefits rather than the changes in costs because the approach is developed for those cases where the net benefits of the tangible benefits and costs are negative. The reason for applying the ICER, and not the traditional C/E-ratio, is the exclusivity of the policies, which implies that if one policy is already in place then the policy must be removed before an alternative policy can be implemented. Fig. 1 illustrates two policies and the corresponding C/E-ratio and ICER. If policies 1 and 2, are independent, the cost-effectiveness ratio is calculated by the slope of the line from the origin12 to the policy.

a) With a maximum acceptable loss and independent policies, the most cost-effective solution is to start by implementing the policy with the lowest cost-effectiveness ratio and then implement the policies in the order of increasing cost-effectiveness ratios until the limit of acceptable loss is achieved.13 b) With a political wtp and independent policies, the most costeffective solution is achieved by implementing all the policies with a cost-effectiveness ratio below or equal to the political wtp. With mutually exclusive policies, the situation is slightly more complex because one policy replaces another, and therefore, the incremental increase in the cost-effectiveness ratio is important.14 c) With a maximum acceptable loss and mutually exclusive policies, the cost-effective solution is achieved by implementing the policy with the highest ICER but still within the limit, which is the only policy that can be fully implemented. If the policies are perfectly divisible, then it may be worthwhile to replace only part of this policy with the next-most effective policy, which can then fully meet the limit for acceptable loss. If the policies are not divisible, a partial implementation is not possible. d) With a political wtp and mutually exclusive policies, the policy with the highest ICER just below or equal to the political wtp is recommended. 4. An empirical example An example is given in which an empirical investigation of the proposed ECEA framework is applied to the mixed fishery in Kattegat and Skagerrak. The Kattegat and Skagerrak waters, located north of Denmark, are shared among Norway, Sweden and Denmark. These waters correspond to the ICES area IIIa. Our paper extends the bio-economic model setup in Kronbak et al. (2009), where standard cohort models with no biological interaction inbetween species are formulated for each species.15 The bioeconomic model was applied to calculate the effects of changed landings, discards and biomass by introducing more-selective gear in the Danish mixed trawl fishery in Kattegat and Skagerrak. There are other gear types and nations participating in the fishery, which are grouped and called other fleets in this example. The implementation of more-selective gear in this mixed fishery aims to improve cod stocks. With no clear signs of improvement in the cod stocks in Kattegat and Skagerrak since the cod recovery plans of

13

This assumes that programs cannot be repeated. If the assumptions about constant returns to scale and perfectly divisible policies are met, then any dominated policy can be eliminated; otherwise, they cannot (Johannesson, 1996). In such a case, instead of using ICER, policies should be treated by using the C/E, remembering that policies cannot be implemented simultaneously. The elimination of a dominated program implies that the same effects can be achieved at lower costs with a combination of programs. 15 For a description of the model, please refer to the Appendix. 14

11

It may be worth noting that for the policy with lowest effect-level, it will be compared with the baseline, and thus, the ICER will be equal to the C/E for this policy. The ICER is used in the empirical example in Section 4. 12 Comparison with a baseline scenario and only including changes in net benefits and in effects allows us to apply the origin in the diagram instead of the baseline scenario.

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

2004 and 2008 (Council Regulation (EC) 1342/2008), there is a clear interest in improving the conditions for this species. Atlantic cod are categorized as vulnerable (Sobel, 1996) on the IUCN (International Union for Conservation of Nature and Natural Resources) Red List of Threatened Species. A species categorized as vulnerable by the IUCN faces a high risk of extinction in the wild in the mediumterm (ICUN website 16). The cod in Skagerrak and Kattegat are considered a species at a low or critical level (Eggert and Ulmestrand, 2000; ICES, 2006a,b). In Kattegat and Skagerrak, cod are primarily caught in the mixed trawl fishery, which harvests Norway lobster (Nephrops norvegicus), Atlantic cod (Gadus morhua), Common sole (Solea solea), European plaice (Pleuronectes platessa) and other less economically important species. In 2004, Denmark had 91 trawl vessels in the fishery (Kronbak et al., 2007). The Danish fishery in Kattegat and Skagerrak falls within Denmark’s exclusive economic zone (EEZ) and is subject to the Common Fishery Policy (CFP). The fishery is regulated by a set of different management tools, which include output restrictions (quotas and licenses), input restrictions (days at sea) and technical measures (minimum mesh size and minimum carapace length/landing size). With respect to selectivity, the main problem in this fishery is, together with the very high discard level of undersized Norway lobster, the by catch of cod that is discarded. For a further description of the fishery, see Kronbak et al. (2007), Kronbak et al. (2009) and ICES (2006a,b). Some of the discard problems may be solved by implementing more selective gear, which would improve the level of cod stock in particular. Another aspect is the change in the age composition of the stock; the quantities of mature fish should be increased, while catches of younger cod are to be avoided. The bio-economic model in Kronbak et al. (2009) was designed to evaluate the changes in biomass levels discards and landings under constant recruitment, which allows for consideration of the changes in age and species compositions in the fishery.17 Assuming constant recruitment allows us the calibrate the model to the actual fishery in the base year and by changing the gear the stock will over a 10 year period reach a new equilibrium. In addition to the biological consequences, Kronbak et al. (2009) evaluate the economic consequences of implementing a more selective gear type in the trawl fishery compared with the baseline defined as the actual fishery with 90 mm cod-end in 2004. The fleets apply the same intensity of effort, but the structure in fishing mortality on the species and year classes is changed, which isolates the effects of changing the gear type.18 Kronbak et al. (2009) show that two different gear types implemented in the trawl fishery can improve the size and composition of the cod stock: a 90 mm cod-end with a 120 mm mesh panel (referred to as 90/120 mm) and a 90 mm codend with 35 mm between the bars in the lower part and 80 mm

16

http://www.iucnredlist.org/ (last accessed March 2013). As a reviewer noticed, with higher biomass levels, other things equal, it is likely that recruitment will increase. Constant recruitment is, therefore, likely to be a conservative measure, and would be an obvious area for future expansion of the model. However, the relationship between recruitment and biomass level for these species in Kattegat and Skagerrak is not well understood (ICES, 2006a,b). Further, we are not convinced about empirical evidence for Kattegat cod about the relationship between recruitment and stock size. E.g. Cardinale and Svedang (2004) concluded that fishing effort intensity was the main driver of the dynamics of the cod stock and that the population size and recruitment was substantially uncoupled. In our model the intensity of fishing does not change only the effort composition. 18 The full bio-economic model is presented in Kronbak et al. (2009). Because the purpose of this paper is not to introduce a more comprehensive bio-economic model but to develop the economic evaluation framework and demonstrate it by example, we refer to the description of the bio-economic model available in Kronbak et al. (2009). 17

275

Fig. 2. Percentage change in biomass for the four most important species comparing 90/120 mm to baseline (90 mm) e upper panel and comparing grid to baseline e lower panel.

between the bars in the upper part (referred to as a grid). Fig. 2 illustrates the changes in biomass for cod compared with the baseline (90 mm cod-end). The upper panel illustrates the improvement in stocks for 90/120 mm compared with 90 mm, and the lower panel illustrates the improvement in the stocks for the grid compared with 90 mm. The results are based on the bioeconomic model calibrated to the actual fishery in Kattegat and Skagerrak. The upper panel in Fig. 2 demonstrates how the selectivity in the 90/120 mm gear is primarily designed to protect the cod stock, while it has negative consequences for the biomass of the plaice, sole and Norway lobster species compared with the baseline. The insertion of a 120 mm panel reduces catches of undersized cod and haddock, while catches of undersized Norway lobsters increase.19 The lower panel in Fig. 2 illustrates that the selectivity of the grid has a large positive impact on the biomass level of cod, plaice and sole. The gear is more or less neutral concerning the biomass of Norway lobster compared with the baseline.20 The change in landings from the whole fleet determines the producer’s total net benefits of different selective fishing gear. This empirical example further contributes to the literature by not only considering the changes in the landings but also relating the achieved effects on the

19 The panel is likely to have an effect on the flow in the gear, but it is uncertain whether this can explain the increased catches of Norway lobster. 20 The modification of the model by Kronbak et al. (2009) is made by calculating the fishing mortality (including landing mortality and discard mortality) imposed from the other fleet. Based on the determined fishing mortality for this fleet, which is assumed to be constant over the 10-year period, and the biomass level determined in Kronbak et al. (2009), the landings for this fleet and the value from these landings can be determined and included in the economic analysis. The model in the current paper is extended to also include the positive net benefits to other fleets due to increased biomass levels. The other fleets comprise the Danish gillnet and international (Swedish) fleets.

276

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279 Table 1 Change in present value of future net benefits (2004e2013) for all vessels at different discount rates (1000 DKK).

Fig. 3. The net benefit for all vessels in the fishery applying 90/120 mm vs grid, respectively. Compared to the baseline (90 mm).

recovery of species. In particular, this paper relates the negative net benefits of more selective gear to the effects achieved, which is defined as the recovery of endangered species by the different stock effect measures and the approach introduced in Sections 2 and 3. 4.1. Results The changes in biomasses are the result of changes in landings of the different species due to the implementation of the moreselective gear in the trawl fishery, which has an effect on the net benefit, not only for the trawlers but also for the other fleet. The changes in the net benefit for all vessels in this fishery are summarized in Fig. 3. Implementing the 90/120 mm gear in the trawl fishery results in increased catches of lobster, plaice and sole in the trawl fishery and, thus, a positive net benefit in the year of implementation. This increase in catches from the trawlers will result in a future reduction in all stocks, apart from cod, as illustrated in Fig. 2. The lower stock levels of the three species imply a decrease in the net benefit for both trawlers and the other fleet over the years, which is insufficient to outweigh the increase in revenues from the improved cod stock available to the other fleet. Implementing the grid in the trawl fishery reduces catches for the trawlers for most species (place, sole and cod to a larger degree and lobster to a smaller degree), which implies an improved stock available for the remaining fleet in the following years, as can be observed in Fig. 2. The large reductions in the landings and, hence, revenues for the trawlers due to the far more selective gear cannot be outweighed by the increased landings due to stock improvements for the other fleet. The net benefits of this gear for all vessels are, therefore, negative, as illustrated in Fig. 3. The total present value of net benefits for the whole fleet is calculated and summarized in Table 1 for a set of discount rates. Because the two different gear types have different effects on the biomass of the cod, it is not reasonable to compare the costs of the different gear without adjusting for the effects of the different gear types. Therefore, the effects are calculated according to the measures defined in Section 3. Using methodology 1) single species stock effect over time, the changes in biomass for the cod stock each year, compared with the baseline scenario, are calculated. The model uses the equilibrium level of stocks after 10 years and the time horizon, G, is therefore set to 2013. These changes are summarized in Table 2. From the above table, the differences in biomass with the moreselective gear seemingly reach an equilibrium, such that the biomass from year 2014 and for 90/120 mm increases by approximately 254,500 kg/year and with the grid increases by approximately 698,000 kg/year if the two selective types of gear are compared with continued use of the baseline 90 mm gear. To calculate the horizon value for the single species stock effect over

Discount rate

90/120 mm

Grid

3% 5% 7%

176 181 154 272 135 529

499 090 455 291 417 139

time, the effect is assumed to be an infinite constant increase in biomass compared with the baseline. Applying Eqs. (4) and (5), the present value of the effects can be calculated. Depending on the discount rate for 90/120 mm, the value ranges between 1480e 1848 tons and 4073e5080 tons with the grid for the period 2004e 2013. The value is 3328e8161 tons for 90/120 mm and 9142e 22 393 tons with the grid for the period 2004 to infinity. By relating the present value of effects to the present value of net benefits, the appropriate cost-effectiveness ratio can be calculated. The alternatives are mutually exclusive because the use of one gear implies that the other gear cannot be used at the same time. In principle, therefore, the appropriate measure for calculation is the incremental cost-effectiveness for the more effective of the two alternatives and the average cost-effectiveness ratio for the least effective alternative. Thus, a measure is given of the cost of achieving additional effects in terms of improved cod stock compared with the those that can be achieved by the less effective gear. It is, however, not simple to argue for perfectly divisible policies and constant returns to scale in a real-world setting. For the alternatives to be divisible would require that some trawl vessels have different gear restrictions than others. In addition to the policy resistance against such different regulations for the same types of vessels, such a policy also would be extremely difficult to enforce. Therefore, it is questionable whether these assumptions are fulfilled, and for the remainder of the paper both the ICER and the C/Eratio for the different alternatives are retained for the decisionmaking analysis. Table 3 summarizes the average and incremental costeffectiveness ratios for the two alternatives, with the single species stock effect measure over time according to three different discount rates. If policies are accepted to fulfill the two assumptions (perfect divisibility and CRS), we compare columns 2&3 and 5&6; otherwise, columns 2&4 and 5&7 should be compared. With the fulfillment of the two assumptions, if the ICER is lower for the grid compared with the 90/120 mm gear, then 90/120 mm is dominated by the grid and should, in principle, be eliminated from the analysis. This is the case of the single species stock effect measure over time from 2004 to infinity. This is an interesting conclusion because it demonstrates the importance of finding an appropriate measure of effects. Without the fulfillment of the two assumptions, it also becomes clear from Table 3 that a willingness to pay for improved cod stocks in the Kattegat and Skagerrak of approximately 100 DKK per tons of cod in the sea is sufficient for recommending the two alternatives. Additionally, the grid is more cost-effective than the 90/120 mm gear if the infinite horizon is included. This conclusion may be controversial because, at first glance, it seems to be very costly in terms of the reduction of harvests. Using methodology 2) the age-weighted single species stock effect, the change in biomass for the cod stock is calculated for each year by including an index for the year classes. The effect measure thereby takes into account that it is not only the quantity but also the quality (the age composition) of the stock that matters. This is a weighted CEA, where the weights are defined as an index based on the economic value of the species. Using the revenue shares as

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

277

Table 2 Change in biomass for cod compared with baseline (90 mm) (tons).

90/120 mm Grid

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

0.00 0.00

186.68 571.38

243.65 649.63

252.99 676.17

254.31 688.75

254.48 694.32

254.50 696.54

254.50 697.34

254.50 697.64

254.50 697.84

weights is a common way of aggregating species into species groups (see, e.g., Squires, 1987). This approach has also been used to create an aggregate stock index; see Andersen (2005). Because the recovery plans focus on the cod stock, we apply the price of the cod to illustrate the relative economic importance of the different age classes; these indices can be found in appendix Table A.1. Due to the pricing of the different year classes of cod, it becomes more valuable to society to have one ton of cod at year class 7e10 compared with having one ton of cod at a lower year class. Thus, the age composition of the stock is important when determining the effects of recovery plans. Based on the index, the present value of the weighted net effects are calculated and related to the net costs of the alternatives. Table 4 summarizes the calculations of the costeffectiveness and incremental cost-effectiveness ratios with the age-weighted single species effect measure applied. Applying this age-weighted single species stock effect implies that a one-ton increase in biomass of younger cod has a smaller effect than a one-ton increase in biomass of older cod. Therefore, the cost-effectiveness ratio is higher compared with the single species effect measures over time as presented in Table 3. This methodology has the particular advantage of allowing for different ecological consequences for the composition of a single stock between the two selective gear types. This advantage is apparent for the effects of the 90/120 mm gear, which decrease significantly because it rebuilds the stock biomass by a higher share of younger year classes than does the grid. Without the two assumptions, there is now a significant difference between the cost-effectiveness ratios (e.g., 173.1 >> 118.2 for a 3% discount rate). With these assumptions (comparing columns 2&4 in Table 4), it is also concluded that the 90/120 mm gear is dominated by the more cost-effective grid. In principal, therefore, the 90/120 mm gear should be excluded from the analysis.

5. Discussion The different effect measures affect the cost-effectiveness-ratio differently and therefore lead to different results. The infinite time horizon of the single species stock effect over time and the weighting of year classes inside the cod stock (age-weighted single species stock effect) demonstrate that the grid is more costeffective than the 90/120 mm gear. The single species stock effect over time from 2004e2013 indicates that the grid is less costeffective than the 90/120 mm gear. The 90/120 mm gear can be interpreted as demonstrating the minimum acceptable political wtp, or the minimum acceptable loss in the fishery, because it is already implemented in the fishery. Because the infinite time

Table 4 The cost-effectiveness and incremental ratios of 90/120 mm and grid, respectively (DKK/tons). Discount rate

90/120 mm C/E

Grid C/E

Grid ICER

3% 5% 7%

173.1 170.0 166.7

118.2 120.6 123.1

100.7 105.0 109.3

horizon of the single species stock effect over time and the ageweighted single species stock effect measures show a significant improvement in cost-effectiveness compared with the 90/120 mm gear, it is natural to explore the reasons for not implementing the grid in the fishery. The answer to this question is likely to be found in the size of the acceptable loss in the fishery. The grid demonstrates a substantial loss in the fishery, paid by the Danish trawlers, while the biomass shows greater improvement than in the case of the 90/120 mm gear. In a ten-year time horizon, the loss is more than double that with the 90/120 mm gear and may exceed the acceptable loss. The analysis in this paper does, however, demonstrate that accepting a loss in the fishery corresponding to the loss from the 90/120 mm gear can be more effective by implementing the grid in parts of the fishery. An example of this point is illustrated in Fig. 4, where the limit for loss is set to the loss from applying 90/120 mm and the age-weighted single species stock effect and a discount rate of 3% is applied. Fig. 4 demonstrates that implementing the grid in 35.3% of the fishery corresponds to the same loss in net benefits as the 90/ 120 mm gear, but it can improve the biomass with 473 tons of cod. For this case to be implemented in a real-world fishery would require a rule stipulating who should implement the gear type: should it be based, for example, on vessel type (e.g., vessel length), specific harvest areas or the season? It is not obvious which of the proposed measures should be recommended; there are pros and cons to all of them. The time perspective matters in the analysis. The infinite version of the single species stock effect over time captures all future improvements in the stock. It is based on the ceteris paribus assumption that nothing else changes in the future. This might be an unlikely assumption, and the infinite perspective may be out of scope from a political perspective. On the other hand, it provides the only full picture of all future net benefits and effects from the policy. Including only stock

Table 3 Average and incremental cost-effectiveness ratios of 90/120 mm and grid, respectively, using a single species effect measure over time (DKK/tons). 2004e2013

2004einfinity

90/120 mm C/E Grid ICER Grid C/E 90/120 mm C/E Grid ICER Grid C/E 3 % 95.3 5 % 93.5 7 % 91.6

99.9 104.2 108.6

98.2 100.3 102.4

158.4 148.0 138.9

66.3 73.0 79.7

99.8 100.3 101.2

Note: The horizon value of costs has been calculated similar to the horizon value of benefits (assuming a yearly loss of 45 mill and 70 mill DKK, respectively).

Fig. 4. Illustration of improved effect by accepting Grid in part of the fishery, applying the effect measure, Age-weighted single species stock effect.

278

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

effects up until the equilibrium year provides only a partial measure of effects and net benefits. The main advantage is that it includes a (shorter) time horizon in the evaluation. The disadvantage is the focus on a single species and that this species is lumped into biomass independent of the age-structure of the species. The age-weighted single species stock effect overcomes this problem by including the age composition of the stock in the effect. This measure might be extended to include several stocks. This effect measure and that including the infinite time horizon are the most comprehensive measure, and hence, from an economic evaluation perspective, the most relevant for application. A limitation of the empirical model is that it applies constant recruitment and, therefore, only identifies part of the effect from the improved gear selectivity, namely, the consequences on age and species compositions in the fishery. The gains via an improved spawning stock and recruitment are thus excluded from the analysis by construction of the model and could be an area of future research. The model includes a sensitivity analysis of the discount rate, which evaluates the importance regarding the flow of costs and benefits. The model also demonstrates that the results are rather robust to changes in the discount rate. The model is, however, entirely deterministic. Thus, a second limitation of the model is that is does not take into consideration any environmental and economic uncertainty. 6. Conclusions This paper develops a framework for the application of a modified cost-effectiveness analysis as an analytical approach e environmental cost-effectiveness analysise to evaluate the policy implications when some of the benefits are intertemporal and difficult to measure in monetary units but are quantifiable. The advantage of the methodology is that it contributes the formulation of management actions as solutions in cases where objectives are predetermined, and even non-economic, and it is an alternative approach to the CBA in evaluating policies, including non-market valuation. The need for such a methodology is present in many natural and environmental resource management problems, for example, air pollution and the protection of endangered species. Furthermore, the paper discusses how to measure effects in natural renewable resource problems, where it is not only the stock in a given future/target year that is important but also the size of the stocks in each of the years in the considered period. The paper suggests different ways of measuring the effects in cases of prioritizing a single natural resource stock, which is often the case in fishery recovery plans. The effect measures vary from having an infinite time horizon to being weighted by the age composition of the target species, in contrast to the traditional single static measure in a given target year. The main problem in evaluating the consequences of natural resource policies is balancing the certain short-run loss due to a policy with uncertain long-run gains. What complicates matters further is that the certain short-run loss is often measured in monetary units, while some of the long-run gain is uncertain and intangible. An example is the implementation of more-selective gear in fisheries, which results in lost revenues due to harvest reductions and stock improvements in the future and is connected to intangible benefits. The shorter the evaluation time horizon, the higher the cost-effectiveness ratios of the alternatives and the less likely they are to be implemented. The outcome of the analysis is based on the willingness-to-pay, or the acceptable loss, for a given achievement. Therefore, the intangible benefits are included in the analysis. The paper applies the methodology to an empirical example in the Kattegat and Skagerrak mixed fishery, where the cod stock is in

poor shape. Politically, the recovery plans for this species are ratified, and to achieve the goals of these recovery plans, among other measures, different selective gear types are introduced in the trawl fishery. The environmental cost-effectiveness analysis is applied to evaluate the consequences of two different more-selective gears: a 90/120 mm cod-end and a grid. When implementing the different types of gear, stocks will adjust and eventually reach an equilibrium level. The model does not evaluate which level of equilibrium might be better but rather evaluates the cost per unit of biomass at the equilibrium level. The analysis applies proposed effect measures for calculating the units of biomass in the cod stock. The analysis concludes that what seems like an overly selective gear type (grid) e because harvests are significantly reduced for most species in question, leading to large negative net benefits e may be an equally good or more cost-effective way of improving the cod biomass than the 90/120 mm gear, depending on the definition of effects (i.e., intangible benefits). However, this gear type may not be implemented due to large negative net benefits, which can have significant and severe consequences for fishing vessels. However, the analysis provides the new insight that, if scaling is possible, the same effect can be achieved at lesser cost by implementing the grid in part of the fleets. This proposal is interesting and has not been considered by policy makers. Our approach can be observed to be a part of recent developments in age-structured bio-economic modeling and optimal policy analysis (e.g., Diekert et al., 2010; Quaas et al., 2013; Tahvonen et al., 2013) because we formulated a methodology that can, within the framework of an age-structured model, rank different policy proposals. Our analysis only includes different effect measures on a single species. An obvious extension would be a mixed effect measure. Implementing selective gear affects several species in a multispecies fishery, which is reflected in the changes in present net value as a result of changes in harvest. Therefore, more than just a single stock is changed by the policy, and it would be obvious to include these changes in the effect measure. One way to do this might be to extend the age-weighted single species stock effect measure to include more species (i.e., to construct an index). The problem in defining the relative importance of the different species is that the political importance of the different species might be significantly different from their economic value. For example, cod is a species of focus because of the poor shape of the stock, even though the economic ex-vessel price of cod is only approximately 1/3 of the exvessel price of sole.

Appendix. Model description (based on Kronbak et al., 2009) A.1. The biological model The model used is a population dynamic equilibrium and forecasting model of stock landings and discards with the calculation of total and partial fishing mortality to assess the relative differences between different scenarios for changes to the technical mesh and gear selection. The model builds upon the models in TEMAS (TEchnical MAnagement MeasureS)21 and can be seen as an extension and expansion of a traditional projection model (forecast model) as used by ICES.22 Based on the input of initial stock size, recruitment and fishing mortality for fish fully recruited to the

21 The full model is described in Ulrich-Rescan, C., Andersen, B.S., Sparre, P.J., and Nielsen, J.R. (2007). TEMAS: fleet-based bio-economic simulation software to evaluate management strategies accounting for fleet behavior. ICES Journal of Marine Sciences. 64: 647e651. 22 An overview of the biological model is available upon request.

L.G. Kronbak, N. Vestergaard / Journal of Environmental Management 131 (2013) 270e279

fishery, the model calculates stock size (biomass and number) and total fishing mortality per age group in the following year. The model also calculates fleet-specific partial fishing mortalities and landings and discards per age group by fleet. The growth of all species in the analysis follows the von Bertalanffy growth equation:

L ¼ LN *ð1  expð  K*ðAge  t0 ÞÞÞ; where L ¼ length; LN ¼ maximal fish length (L-infinity); K ¼ the curvature parameter, which determines how fast the fish approaches LN; and t0 ¼ the point in time the fish has zero length. Fishing mortality in the model is estimated for a given stock and age in the stock assessment. Fishing mortality is split into partial fishing mortality for Danish trawlers and partial fishing mortality for “Others”. A constant yearly recruitment is a precondition for a population dynamic equilibrium model. The recruitment at age 0 is estimated by calculating “backwards” from the recruitment estimated from the analytical assessments. Other model input parameters are gear mesh size; a selection ogive resulting from the input selectivity parameters L50; and selection range, SR, where L50 is the length of the fish at which 50% of the fish escape the net and 50% are retained by the net (SR is L75eL25). An estimated sorting (or discard) ogive obtained from the results of at-sea sampling programs is also included in the model. The length where 50% of the fish are discarded and 50% retained on board (L50) is estimated from a linear regression between the length where all specimens are discarded and the length where all specimens are kept on board. The estimation of L75 (length where 75% of the fish are retained on board) is based on the same regression. Appendix A.2 Table A.1 summarizes the price for cod based on different age classes and defines an index based on normalization to the highest price. Table A.1 Market price for cod landings in 2004 connected to different year classes and the defined index. Age class

Price (DKK/kg)

Index

0 1 2 3 4 5 6 7 8 9 10

12.23 12.23 12.23 19.59 23.20 25.67 25.67 28.18 28.18 28.18 28.18

0.43 0.43 0.43 0.70 0.82 0.91 0.91 1.00 1.00 1.00 1.00

References Andersen, J.L., 2005. The inclusion of stocks in multi-species fisheries: the case of Danish Seiners. Mar. Resour. Econ. 20, 163e184. Armstrong, C., 2007. A note on the ecological-economic modelling of marine reserves in fisheries. Ecol. Econ. 62, 242e250. Berndes, G., Hansson, J., 2007. Bioenergy expansion in the EU: cost-effective climate change mitigation, employment creation and reduced dependency on imported fuels. Energy Policy 35, 5965e5979. Brent, R.J., 2003. Cost-Benefit Analysis and Health Care Evaluations. Edward Elgar, Cheltenham, UK. Cardinale, M., Svedang, H., 2004. Modelling recruitment and abundance of Atlantic cod, Gadusmorhua, in the eastern SkagerrakeKattegat (North Sea): evidence of

279

severe depletion due to a prolonged period of high fishing pressure. Fish. Res. 69, 263e282. Clark, C.W., 1976. Mathematical Bioeconomics: The Optimal Management of Renewable Resources. John Wiley & Sons Inc., USA. Commission of the European Communities, 2009. Brussels, 22.4.2009, COM(2009) 163 final GREEN PAPER, Reform of the Common Fisheries Policy. Council Regulation (EC) 1342/2008Establishing a long-term plan for cod stocks and the fisheries exploiting those stocks and repealing Regulation (EC) No 423/ 2004. Council Regulation (EC) No 388/2006 of 23 February 2006 establishing a multiannual plan for the sustainable exploitation of the stock of sole in the Bay of Biscay. Council Regulation (EC) No 2166/2005 of 20 December 2005 establishing measures for the recovery of the Southern hake and Norway lobster stocks in the Cantabrian Sea and Western Iberian peninsula and amending Regulation (EC) No 850/98 for the conservation of fishery resources through technical measures for the protection of juveniles of marine organisms. Diekert, F., Hjermann, D., Nævdal, E., Stenseth, N.C., 2010. Spare the young fish: optimal harvesting policies for North-East Arctic Cod. Environ. Resour. Econ. 47, 455e475. Drummond, M., Sculpher, M., Torrance, G., O’Brian, B., Stoddart, G., 2005. Methods for the Economic Evaluation of Health Care Programmes, third ed. Oxford Medical Publications. Eggert, H., Ulmestrand, M., 2000. A bioeconomic analysis of the Swedish Fishery for Norway Lobster. Mar. Resour. Econ. 14, 225e244. Goulder, L.H., Parry, I.W.H., Williams III, R.C., 1999. The cost-effectiveness of alternative instruments for environmental protection in a second-best setting. J. Pub. Econ. 72, 329e360. Grafton, R.Q., Kompas, T., Hilborn, R., 2007. The economics of overexploitation revisited. Science 318, 1601. Hajkowicz, S., Spencer, R., Higgins, A., Marinoni, O., 2008. Evaluating water quality investments using cost utility analysis. J. Environ. Manage. 88, 1601e1610. Hoff, A., Frost, H., 2008. Modelling combined harvest and effort regulations: the case of the Dutch beam trawl fishery for plaice and sole in the North Sea. ICES J. Mar. Sci. 65, 822e831. ICES, 1998. Report of the Study Group on the Precautionary Approach to Fisheries Management. ICES CM 1998/ACFM:10. ICES, 2006a. Report of the Working Group on the assessment of demersal stocks in the North Sea and Skagerrak (WGNSSK). ICES CM 2006/ACFM: 35 http://www. ices.dk. ICES, 2006b. Report of the Baltic Fisheries Assessments Working Group (WGBFAS). ICES CM 2006/ACFM: 24 http://www.ices.dk. Innes, J.P., Pascoe, S., 2010. A multi-criteria assessment of fishing gear impacts in demersal fisheries. J. Environ. Manage. 91 (4), 932e939. IUCN website. http://www.iucnredlist.org/ (accessed September 2009). Johannesson, M., 1996. Theory and Methods of Economic Evaluation of Health Care. Kluwer Academic Publishers, The Netherlands. Kraak, S.B.M., Buisman, F.C., Dickey-Collas, M., Poos, J.J., Pastoors, M.A., Smit, J.G.P., van Oostenbrugge, J.A.E., Daan, N., 2008. The effect of management choices on the sustainability and economic performance of a mixed fishery: a simulation study. ICES J. Mar. Sci. 65, 697e712. Kronbak, L.G., Andersen, S.E., Boom, J.T., Eliasen, S., Frost, H., Hoff, A., Jørgensen, O., Krog, C., Mathiesen, C., Nielsen, J.R., Munch Petersen, S., S Sverdrup-Jensen, Vestergaard, N., 2007. Implementering af mere selektive og skånsomme fiskerier e Konsekvenser for ressource, fiskere og samfund. IME Report 11/07. Department of Environmental and Business Economics (IME), University of Southern Denmark (in Danish). Kronbak, L.G., Nielsen, J.R., Jørgensen, O.A., Vestergaard, N., 2009. Bio-economic evaluation of implementing trawl fishing gear with different selectivity. J. Environ. Manage. 09 (11), 3665e3674. Perman, R., Ma, Y., McGilvray, J., Common, M., 2003. Natural Resources and Environmental Economics, Pearson Higher Education. Addison Wesley. Quaas, M.F., Ruckes, K., Requate, T., Skonhoft, A., Vestergaard, N., Voss, R., 2013. Incentives for optimal management of age-structured fish populations. Resour. Energy Econ. 35 (2), 113e134. Schou, J.S., Skop, E., Jensen, J.D., 2000. Integrated agri-environmental modelling: a cost-effectiveness analysis of two nitrogen tax instruments in the Vejle Fjord watershed, Denmark. J. Environ. Manage. 58, 199e212. Sobel, J., 1996. Gadus morhua. In: IUCN 2008. 2008 IUCN Red List of Threatened Species. http://www.iucnredlist.org. Downloaded April 2009. Squires, D., 1987. Public regulation and the structure of production in multiproduct industries. Rand J. Econ. 18, 232e247. Tahvonen, O., Quaas, M.F., Schmidt, J.O., Voss, R., 2013. Optimal harvesting of an agestructured schooling fishery. Environ. Resour. Econ. 54 (1), 21e39. Van Beukering, H., Bartelings, V.G.M., Linderhof, Oosterhuis, F.H., 2009. Effectiveness of unit-based pricing of waste in the Netherlands: applying a general equilibrium model. Waste Manage. 29, 2892e2901. Van Kooten, G.C., Bulte, E.H., 2000. The Economics of Nature: Managing Biological Assets. Blackwell Publishers Ltd., Oxford.