The use of individual transferrable quotas (ITQ) has been seen by many as the best solution to combine economic success while avoiding over-exploitation. Since the economic value of quota shares increases when fish stocks are well managed, ITQ shares create an economic incentive for stewardship [1,14,18,19]. Specifically, while most fishermen are directly concerned with cost-benefit in terms of catch per unit effort (CPUE), ITQ programs alter the equation to a combination of short-term CPUE and long term share value. Thus, the success of ITQ programs is based on the notion of increased incentives for long-term management due to a form of property rights or virtual ownership. ITQ programs have shown evidence of success toward many goals including biological sustainability (long-term stock management) and economic profits and output [16,19-22]. Both of these goals are strongly linked with the reduction in fishing capacity that typically results from ITQ program adoptions. In addition, Costello et al. [17] examined over 11,000 global fisheries and found empirical evidence that ITQ managed fisheries that were collapsed was about half that of non-ITQ fisheries.

ITQ program successes in Canada, New Zealand, Australia, Iceland, Norway and other locations have prompted changes in the United States systems and currently six of the eight federal regions are working to develop catch share or ITQ programs [18]. In 2005, approximately 10% of the total marine harvest was managed by some form of ITQ, and the numbers are increasing [19]. Recent adoptions include the establishment of catch share programs for 19 stocks of New England groundfish, including cod [23].

A major criticism or weakness of ITQ systems is that they do not create true property rights. Holding a quota share only provides a right to harvest, but confers no real control over the resource itself [16,24,25]. For example, the cost of any short-term sacrifice is born by the individual. However, long-term benefits that accrue from stock rebuilding are shared by all fishery participants. Such a structure provides short-term benefits from actions such as high grading, quota busting, and misreporting catch. In fact, ITQs systems are often criticized because of the potential for cheating at the individual level [1,9,15,21].

Perhaps the biggest issue with ITQ adoption has been the concentration of quota shares to fewer and generally larger-scale fishing operations and the resultant impact on communities [1,12,19,21,26,27]. This concentration of shares is a natural consequence of a market based approach where shares are traded. While less efficient operators are pushed out of the market this results in: (1) A reduction of fishing capacity, (2) a reduction in the number of fishermen with a corresponding large social impact on community structure and employment; and (3) the potential for shares to be controlled by non-local owners. While the first factor is a desirable impact since it relieves pressure on over-fishing, the second factor is undesirable because it both results in social inequity and it places pressure on communities and fishermen to remain profitable. In addition, the transfer of shares to more profitable operators can result in an increase in non-local owners who may not have a long-term interest in the resource. Essentially this may mean a decoupling of ownership from resource stewardship [28].

Finally, it must be noted that ITQ systems are not appropriate for all contexts. ITQ systems are primarily designed for single, target species management and may be more difficult to manage in multi-species fisheries [19,22]. In addition, ITQ systems require substantial investments in enforcement, monitoring and scientific assessment [5].

Figure 1 highlights areas where high leverage policy may be effective, specifically the management control points or areas of potential management influence. However, the development of a stock and flow model and computer simulation is necessary to more precisely understand the long-run behavior of the ITQ fishery over time.

Understanding and managing stocks and flows, which includes resources that accumulate or decay and the flows that govern their change is a fundamental process in social and natural systems. All stock and flow problems share the same underlying structure; the resource level or stock accumulates its inflows less its outflows [33]. Many real world problems can be modeled with stock and flow simulations. Example stock and flows include: Inventory (stock) that increases through production (inflow) and decreases through sales (outflow); and population (stock) that increases through births (inflow) and decreases through deaths (outflow). Underlying equations represent the flows and determine how the stocks change through time. System dynamics modeling with stock and flow is extremely useful since many real world systems involve many interconnections, non-linear dynamics and feedbacks. The complexity of systems means that humans are unable to infer the long-run consequences of our actions without the use of computer simulations [33].

The stock and flow model is summarized in four different views (Figures 2 through 5) and a number of select formulas are shown in Appendix A (The complete model, developed in Vensim Professional, can be downloaded by contacting the author).

In our model of an ITQ fishery, we are primarily interested in the behavior of the fishery as a whole and not on comparing the success of one fishing organization relative to another. Overall, each fishing organization will be compelled to add fishing units or vessels as vessel efficiency increases. Specifically, if the current efficiency, as measured by catch per unit effort (CPUE) is above the break-even CPUE, then it will be economically justifiable to add vessels. In a similar fashion, as the fishery becomes overcapitalized, CPUE values will fall and fishing units (and organizations) will leave the fishery (see Figure 2). This mirrors real-world observations as ITQ programs are designed to promote consolidation and remove excess capacity. Fishing organizations that are inefficient with high costs are driven to drop out. Our model uses the average CPUE and average profit to determine the entry and exit of vessels for the fishery as a whole.

In contrasting ITQ with non-ITQ TAC fisheries, of central concern is the incentive structure or the system induced influence that leads to different fishermen behavior (especially with respect to capacity decisions). In non-ITQ TAC fisheries, investment behavior is likely more aggressive than ITQ programs as there are no individual restrictions on catch. This leads to the “race to fish” and overcapacity. In ITQ systems however, capacity is constrained when individual quotas constrain catch. The model implements this behavior using information feedback on the gap between TAC and fishermen catches.

Figure 3 models the TAC setting process for the entire fishery. In an ITQ system, the TAC would then be divided-up according to the pre-defined ITQ shares for each of the fishing units or fishing organizations. Our model simply assumes equal shares among fishing units.

The model first determines a scientific proposed TAC that is used as input to management's proposed TAC. The scientific proposed TAC is purposely mismatched with the fish biomass model in Figure 4 to reflect the often real-world situation that scientific fishery models do not match the complex realities of actual ecosystems with perfect precision. The model is intended to be more descriptive than prescriptive since the goal is to identify areas for improved management decision making in the context of stakeholder interference in TAC setting and imperfect scientific information.

It is beyond the scope of this paper to develop a sophisticated TAC-setting or harvest control strategy. Rather the purpose is to examine the difficulties found in the human-based components of the system that can steer TAC management away from scientific recommendations.

To model real-world interference in TAC setting, the model then depicts two major avenues for TAC changes: (1) Direct pressure on management's TAC by the community, and (2) lobbying pressure. Lobbying pressure is a function of the proposed TAC's impact on the community, relative to capacity.

ITQ design and management influence can then negate or counter attempts at TAC change through strength of management mandate (strong central management), management's community bias (degree of concern for community employment and short-term economics), and ownership.

The variable ownership is used to address the difference in incentives that are provided by ITQ programs relative to non-ITQ fisheries. The ownership variable can be used as a switch variable; setting it to 0 or 1 allows us to distinguish between non-ITQ TAC fisheries, where there is strong interference in TAC setting, and ITQ programs where share values influence users to consider their long-term stakes in the fishery.

After the management TAC has been recommended and potential lobbying pressure determined, the TAC setting process then takes a conflict resolution, or debate oriented approach, where the political pressure from lobbying is weighed against management's bargaining power. We adapt the approach used by Dudley [34] whereby management's bargaining power is primarily a function of: (1) the strength of the original management mandate, as designed by policy makers; and (2) the degree of political support for management. As the fishery becomes overfished and declines in overall ecosystem health, the political authority provides strengthened authority to management.

The TAC setting process variable is a negotiated value determined by a weighted average of the two competing interests. The ratio of recent average catch relative to the past (historical average) is used to determine the political support for management; the amount of short-term economic pressure is a function of the ratio of catch to fishing capacity.

We adopt a simple approach to fish population dynamics, similar to the approach suggested by Dudley ([34], pp. 3-4) who states “There is a need for models that allow us to examine complex fishery issues in a transparent and understandable manner without becoming overly focused on details of population dynamics.”

The model views fish population as an aggregate for ease of communication and avoids a more complicated model with cohorts. The model is based on the biomass dynamic model [35]: (1) dB / dt = rB - rB ( B / k ) - qEwhere, an increase in fish biomass is equal to the biomass fractional growth rate, r, times the current biomass B. The natural decrease in biomass is given by -rB times the ratio of B to k, where k is the maximum, possible biomass of the fish population for that ecosystem. The outflow of fish biomass due to the harvest is equal to the catch per units of fishing gear, q, times the number of gear units, E (see Figure 4).

The model is based on the alternate model developed by Dudley [34]. Separate inflows to biomass come from both growth additions and recruitment. The percentage of additions from recruitment relative to growth varies by average biomass age. Recruitment increases biomass based on the intrinsic rate of growth, r, base rate, however, above a certain biomass stock size constant recruitment is used. This mirrors many real-world fish stocks [7]. Growth additions to biomass are computed based on g rate growth, which is a function of fish density. Catch per unit effort (CPUE) is also a function of fish density and in addition, CPUE is affected by technology investments in fishing gear (max CPUE/yr, see Appendix). Since the impact of fish density on CPUE may also be a function of the particular fish species [36], a switch variable—hyperstability is used for sensitivity tests in this context. Both types of density impacts on CPUE are modeled using lookup functions.

There are three major biomass outflows: Natural death, fish harvest and discards. Notably, the total catch or fish harvest is also based on illegal catch or quota busting. Quota busting is computed based on fishermen's incentives to cheat and the countervailing level of enforcement. Incentive to cheat is computed in the TAC Management section of the model. The fish stock is also reduced by discards. While exact equations to compute cheating and fish dumping are generally unknown, empirical evidence suggests a rate of about 8% to 10% worldwide [37,38], (we conservatively use 5% in the base case). Price dumping is computed as a non-linear function using the ratio of average catch (the 5-year smoothed average) relative to the current TAC. As catches are likely to become limited due to quota, fishermen have more and more incentive to maximize catch revenue. In other words, why turn-in lower valued fish to count against quota?

Market price is computed based on recent market prices and a supply/demand curve (See Figure 5). Average profit per unit is computed based on various cost dimensions including new investments in fishing technology or gear that are necessary to keep competitive (See Appendix for key equations and parameters).

As identified in Figure 1, the rational design of ITQ fisheries and TAC setting is not adequate (in itself) to control sustainability and guarantee the entire system will behave in consistent and mutually supportive ways. We could test if TAC setting policy is based on intended rationality, by running a partial model test [39]. We could re-write Figure 1 to reflect only the ‘rational intention’ of TAC setting (and thereby remove policy resistance from the model).

However, by setting a range of management controllable variables (enforcement, management control, and ownership), we can also effectively “turn off” the policy resistance from the fishing community (the right-hand part of Figure 1). There are essentially two ways to apply outside management control to thwart or cancel policy resistance: (1) Set variables or switches reflecting strong, centralized fishery management control (strength of management mandate to cancel lobbying influence; management's community bias to cancel direct TAC interference), or (2) set the ownership variable to reflect the idealized state of ITQ shareholder and community responsibility, stewardship, and ownership. In either case, strong management control or high ownership, there is a resulting high level of enforcement. Both cases reflect successful fishery strategies as seen in real world observations [2].

In the base case or ideal state, TAC changing and other systemic cheating is prevented. Figure 6 shows fish biomass (tons), catch (tons/year), fishing units (or vessels, units) and realized catch per unit effort (CPUE, tons/(year × unit). Figure 6 shows that the fish biomass converges to, approximately, an “optimal 20,000 tons” or a level of k/2. At around year 35 the catch peaks and is constrained by TAC.

The ITQ model was developed for the primary purposes of: (1) comparing TAC effects from other ITQ design effects; (2) examining the range of human influence points involved in Individual Transferrable Quota systems; and (3) developing insight into influential management control points. As a starting point, previously published models and an extensive literature review were used to formulate the initial model description. In particular, Dudley's [34] model provided a starting point for both the biological model and the initial political context.

Using the general approach to model validation recommended by Barlas [40], the model was examined using the following general tests: (1) Direct structure tests; (2) structure-oriented behavior tests; and (3) behavior pattern tests. An extensive literature review was performed to provide insight into the workings of ITQ systems including the behavior of fishermen and other stakeholders. Equations were developed and examined with respect to their correspondence with the knowledge reviewed in the literature (theoretical structure test). Parameter confirmation was performed by examining prices, costs, and corresponding values found in prototypical fisheries [41-43].

In order to examine for possible structural flaws in the system, sensitivity analysis of key parameters and extreme condition tests were performed. Model structure was also examined and re-examined to determine if stock and flow behavior would realistically match real system behavior. For example, vessel entry and exit to the fishery were re-modeled to reflect the more realistic flow of vessels using separate flows for: (1) Vessel entry; (2) vessel exit; (3) vessel retirement; and (4) vessel re-entry. In addition, comparison of non-ITQ, TAC fishery investment behavior to ITQ systems required additional modifications to reflect observed and reported real-world behavior.

Behavior pattern tests were performed to examine model outcome patterns to changes in parameters and key management leverage points. For example, by varying the levels of enforcement the model produced expected amounts of quota busting and ultimately resulted in stock crashes.

Using the same initial conditions as the base case, Figure 7 displays the simulation results for an extreme condition test where biomass is suddenly increased by 25% in year 50. We show the system response with two possible reactions by the human component of the system: (1) “business as usual” where the TAC management component continues to manage the fish biomass for an optimal, long-run equilibrium, and (2) with fishery management being open to community pressure. That is, the management's community bias variable is set to 1 to allow for changes in TAC when the short-term economic pressure increases (measured using catch/fishing capacity). The sudden increase in biomass results in improved productivity in terms of CPUE which then leads to an increase in fishing capacity and to an increase in short-term economic pressure. In this second case, with management responding to community pressure, TAC and thus catch are increased in year 50 (graph line 1). We allow this openness to community pressure to persist for 5 years and then it is switched off. During this 5-year period “avg profit/unit” increases (graph line 3) and is higher than “avg profit/unit” for the “business as usual” case (graph line 4). However, with higher profits there is also a corresponding increase in vessels added to the fishery. Thus, after year 55 when TAC and catches are again reduced by a return to normal policy, the higher capacity results in a lower long-run equilibrium “avg profit/unit” (graph line 3) compared to “business as usual” (graph line 4).

Figure 8 displays the simulation results for an extreme condition test where TAC is suddenly reduced by 25% in year 50. We again show the system response with two possible reactions by the human component of the system: (1) “business as usual” with strict enforcement to prevent quota busting, and (2) enforcement turned off for years 50 to 55. Catch is reduced in both cases; however, with quota busting allowed (no enforcement) catch is comparatively higher (graph line 4 > graph line 3). In addition, lack of enforcement means that fish biomass is stabilized at a lower, long-run level and “avg profit/unit,” while higher in the short-run, is also lower in the long-run compared to strict enforcement (graph line 2 < graph line 1). In this scenario, enforcement was only relaxed for a short 5 year period. Naturally, if quota busting persisted for a longer period there would be correspondingly larger impacts on biomass and average profit/unit.

Overall, the ITQ model responded as expected to extreme condition tests. In both test cases, when the human component of the system influenced TAC policy due to short-term economic pressures the end result was higher short-run but lower long-run average profit/unit. However, it should be noted that all simulations have assumed a small variation in recruitment. It is not uncommon for fisheries to experience large fluctuations in recruitment [34].

In Figure 9 we show the results of increasing random variation in recruitment with the stock set to 10,000 tons biomass (k = 40,000 tons) and 200 fishing units or vessels (an overcapitalized fishery). Random changes cause large boom and bust cycles and make fishery management and stock rebuilding more difficult. As seen in Figure 9 periods of high biomass lead to increases in catch (via increases in TAC) and since capacity is more than adequate, average profit/unit rises. Reductions in capacity are difficult (via TAC decreases) since biomass levels are high. In years 20 through 60, capacity remains at an equilibrium level of approximately 140 vessels (units). In years 60 through 80, average profit/unit remains steady despite declines in biomass because the fishery finally loses capacity. Overall, large fluctuations in biomass may mean increased short-run financial pressures in some years and may make community and management cooperation more difficult.

As pointed out by Chu [1] in eight of the 20 stocks examined in her study, ITQ programs were not able to rebuild fish stocks. We explore the possibility of rebuilding stocks with the ITQ model. We start the model with the stock set to 10,000 tons biomass (k = 40,000 tons) and 200 fishing units or vessels (an overcapitalized fishery). The stock rebuilds, but only to 17,016 (k/2 = 20,000) in year 100 with CPUE rising from 12.5 tons/unit/year to 16.8 tons/unit/year.

We next readjust the model to implement scientific TAC recommendations using a more precautionary approach (the scientific TAC is decreased by 5%). The stock rebuilds more quickly to 11,435 tons in 10 years and to 20,647 (approximately k/2) in 100 years. While stock rebuilding appears to be better accomplished with a precautionary policy, we must consider whether such policy is easily implemented.

Toward that end, we compare and contrast TAC setting approaches and examine their impact on total fishery profits. Figure 10 contrasts run 1, a precautionary approach where TAC is set 10% lower than normal scientific advice, with run 2, a liberal approach where TAC is set 10% higher than scientific advice. In the short-run, the liberal TAC approach dominates with total fishery profit in year 5 ($1.88 million compared to $1.6 million) and in year 10 total fishery profits are higher ($1.98 million compared to $1.87 million). However, in the long-run the precautionary approach dominates the liberal approach. For example, in year 50 the precautionary approach yields $2.5 million compared to $1.9 million for the liberal approach and by year 100 the profits are $2.5 million to $1.6 million, respectively (or 50% higher than the liberal policy).

Although the precautionary approach appears to be the superior TAC setting policy for long-run sustainability and fishery profit, in Figure 11 we compare accumulated discounted fishery profit over time for these two approaches. Using this approach for comparison, liberal TAC setting dominates since discounted profits or losses means that short-run profit will count much more than later profits.

Figures 10 and 11 contrast a precautionary approach with liberal TAC setting. Figure 10 shows the obvious superiority, from the public's perspective, of precautionary TAC setting for long-run profitability and sustainability. Unfortunately, the precautionary approach generates a “worse before better” long-run result. Figure 11 shows discounted cash flows (based on operating profit) that reflect fishermen's necessary concerns for immediate results.

As noted by Forrester [44], many of the actual practical working mechanisms of a system tend to create conflict between short-run and long-run results. The free-market design of ITQ systems means that fishermen may be in conflict with the long-run, public sustainability goals of fishery management.

While long-run sustainability is a primary concern of the public, ITQ programs however give incentives for fishermen to compete and remain profitable, and therefore short-run goals will dominate their decision making. Thus, it becomes very difficult for fishermen to agree on conservative TAC levels. In practice, using the precautionary principle (to build-up stocks) is very difficult, even when it is incorporated into law [7,45].

In order to achieve a proper balance between short-run and long-run goals that will lead to better outcomes for both local communities and the public, we propose a public transfer payment. That is, if fisheries are able to make progress toward long-run sustainability targets then they will qualify for short-run off-set payments. Payments would be contingent on achieving target goals and would provide incentive for local communities to adjust their short-run versus long-run tradeoffs. Since everyone in the community benefits if targets are met, there are additional incentives for communities to monitor and enforce quota compliance.

Ecosystem-based management offers many obvious advantages to traditional fishery management approaches. EBM is necessary to provide valuable information to fishery management since ecosystem structure and function impact ecosystem productivity, ecosystem diversity can help to prevent species collapse [4,17], and finally EBM is necessary to achieve conservation objectives.

While ITQ systems are clearly not appropriate for all fishery management contexts, they can be effective in reducing overcapacity in many fisheries. Reductions in overcapacity may be an important and practical first step toward achieving the goals of EBM [9]. In fact, Mace recommends three fundamental prerequisites to help achieve EBM: (1) Reduce exploitation rates on individual target species; (2) reduce overcapacity; and (3) conduct adequate baseline monitoring of marine species and their environment.

There is no “silver bullet” with respect to effective fisheries management. The use of rights-based incentives (including ITQs) does not automatically account for ecosystem damage because fishermen have little incentive to minimize by-catch or habitat damage since it does not immediately impact profits from their target species [5,46]. On the other hand, while the EBM approach is necessary to account for fishery impacts on ecosystems, it is insufficient to address two major factors impacting unsustainable fisheries: (1) The inappropriate incentives imposed on fishers (that lead to policy resistance); and (2) the ineffective governance often found in commercial, developed fisheries managed by total harvest limits (non-ITQ TAC fisheries) and input controls [13,47]. Indeed, traditional commercial fisheries that impose input controls result in increased costs of fishing and often fail to reduce fishing effort due to the substitution of un-regulated inputs for controlled ones [13]. Similarly the use of competitive TACs in input controlled fisheries leads to the “race to fish” and ultimately harm the sustainability of fisheries [2].

Reducing the impact of fishing on ecosystems requires the ability to control fishing effort [10,47]. Use of ITQ systems, where appropriate, can be beneficial in reducing fishing capacity, effort and exploitation rates of target species. Other incentive-based approaches have been recommended to effectively control ecosystem impacts, including: Closed areas (marine protected areas); incentives to deploy by-catch reduction devices; on-board observers to reduce by-catch; restrictions on fishing gear; and vessel monitoring systems, among others [10,13].

Due to the increased costs required to supply the appropriate scientific information for successful EBM, appropriate intermediate steps might include reducing fishing mortality and eliminating overcapacity [9,10]. These two steps could reduce the frequency with which stock assessments need to be made and free up resources for scientific and ecosystem research [9].