Economic essays on marine invasive species and international fisheries agreements

This thesis is divided into two parts, as explained in Chapter 1, which focus on different aspects of marine ecological change. Part A considers marine Invasive Alien Species (IAS), which are taxa introduced outside of their native range. The detrimental consequences of invasions for human welfare necessitate management of IAS. There are two types of IAS management. These are (i) management of the risks that an invasion will become established, termed “prevention”, and (ii) management of already established invasions, termed “control”. Chapter 2 considers prevention of invasive species with Ballast Water Management (BWM). Vessels transport invasive species in their ballast water. BWM involves treating ballast water to reduce the risk of successful invasion establishment. Chapter 2 studies the determinants of optimal ballast water treatment standards from a theoretical perspective. Chapter 3 considers control of already established invasions from a spatial and dynamic perspective. We model a non-native habitat divided into patches, where each patch may contain a population of the invasive species, and where spread of the invasion between patches is a stochastic process. In this context, we derive optimal management policies. The second part of this thesis: Part B, considers International Fisheries Agreements (IFAs). IFAs facilitate cooperation in the management of fish stocks. Cooperation is necessary to ensure sustainable management. Part B focuses on two issues which may affect the stability of cooperation within IFAs. These are; in Chapter 4, changes in stock location, which may occur due to climate change, and in Chapter 5, the risk of stock collapse, which may exist due to overfishing. Part B uses game theory to analyse the effects of these two issues on the stability of the Grand Coalition, which is the state of affairs where all parties cooperate to maximize their joint benefit from the fish stock. The methods and findings of the thesis are summarized as follows: in Chapter 2 (Part A), we construct a model to study optimal BWM standards. The model is built around the assumption that invasions arriving via ballast water are irreversible, i.e. once an invasion has arrived, it is not possible to reduce the size of the invasive population to zero. The hazard rate of invasion establishment can be reduced by setting a BWM standard. The hazard rate is also affected by the Minimum Viable Population (MVPs) of the species and the possibility of an Allee effect. An MVP exists if there is some population size below which there is an insufficient number of invasive individuals to sustain a population. An Allee effect exists if the probability that a population survives increases at an increasing rate in the size of the population. Our analysis focuses on the conditions under which a BWM standard which aims to reduce invasive populations in ballast water to below their MVPs (as is aimed for by the BWM convention) can be optimal. We find that the current aim of the BWM convention can only be optimal in the case that the hazard function (which determines the hazard rate) is not continuously differentiable around the MVP. We find that Allee effects are a requirement for a continuously differentiable hazard function. Therefore, we find that whether or not an Allee effect exists fundamentally affects whether it is optimal to aim to reduce an invasive population in ballast water to marginally below its MVP. In Chapter 3 (Part A), we combine aspects of previous modelling approaches to provide new generalized management insights for controlling established invasions. We employ a metapopulation network consisting of patches which are arranged one-dimensionally (i.e. in a line), which is relevant, among other cases, for invasive species spreading along coastlines. We allow for the population size of the invasion within patches to be reduced, which we term “removal”, and we allow for the probability of spread between patches to be reduced without affecting the population sizes directly, which we term “containment”. We employ numerical stochastic dynamic programming to explore how these two interventions (removal and containment) can be optimally applied to minimize the sum of damages from the invasion and the costs of removing and containing the invasion. We find that allowing for varying stock sizes within patches facilitates optimal timing of the application of containment. We also identify two novel optimal policies: the combination of containment and removal to stop spread between patches and the application of up to four distinct policies for a single patch depending on the size of the invasion in that patch. Chapter 4 (Part B) considers how Grand Coalitions can be stabilized in the face of changing stock location. To do so, we employ the Gordon-Schaefer fisheries model. We consider farsightedness as a mechanism by which stability of the Grand Coalition can be increased in the face of changing stock location. Farsightedness allows players to respond to deviations of other players by deviating themselves. This reduces the incentives to leave the Grand Coalition. This is in contrast to shortsightedness, whereby players cannot decide to leave the Grand Coalition in response to such a choice by another player. We begin by modifying the farsightedness concept such that it can be used in games with asymmetric players and transfer payments. We proceed to analyse the modified farsightedness concept in the case where players are symmetric (stock location does not change) in order to identify the properties of the concept in the base case. We find that farsightedness increases Grand Coalition stability with respect to shortsightedness. We proceed to analyse the extent to which farsightedness increases Grand Coalition stability, relative to shortsightedness, as fish stock location changes, using sensitivity analysis. We find that farsightedness increases the stability of the Grand Coalition, but also increases the sensitivity of stability to changes in fish stock location. Thus, for any fish stock location, a Grand Coalition is more likely to be stable if players are farsighted, but shifts between a stable and an unstable Grand Coalition will occur more frequently if players are farsighted. In Chapter 5 (Part B), we analyse how the stability of Grand Coalitions is affected by an endogenously determined risk of stock collapse. We do so using the Levhari and Mirman (LM) fisheries model, which is adapted such that there is a risk of stock collapse which increases as the fish stock size decreases. We numerically solve the model and calculate the stability of the Grand Coalition. We find that the effect of an endogenously determined risk of stock collapse depends heavily on the assumptions made regarding how payoffs are determined. A common assumption in the literature is that payoffs are determined at the steady state fish stock. Under this assumption, endogenous risk means that for specific discount and growth rates, a Grand Coalition is stable for any number of players. This is a very different result from the original LM model whereby Grand Coalitions can never be sustained. This is because players can essentially follow two strategies in response to the risk. Firstly, they can attempt to maintain the fish stock by fishing less. In doing so they are running the risk of collapse. Secondly, they can avoid the risk by pre-emptively depleting the fish stock, i.e. harvesting the stock to zero immediately to avoid the risk. Grand Coalitions of any number of players are stable for parameterizations for which a Grand Coalition attempts to maintain a non-zero fish stock and if a deviation from the Grand Coalition would result in pre-emptive depletion. We proceed by relaxing the assumption that payoffs are determined in the steady state by allowing for deviators to obtain payoffs in the transition between steady states. In this case, only Grand Coalitions of two players are stable, and then only for certain parameterizations. The reason is that players can now gain payoffs in the process of pre-emptively depleting the stock, i.e. payoffs are received from the process of fishing the stock down to zero. This increases the benefit of deviating from the Grand Coalition. In this case, Grand Coalitions are only stable for two players for specific parameterizations. Chapter 6 summarises the research questions formulated in Chapter 1 and evaluates the work of the thesis. Regarding Chapter 2, we justify our theoretical approach with the following two points. Firstly, BWM management is a global and complex problem, which means that the information required to formally calculate an optimal standard is prohibitively burdensome. Secondly, we argue that the complexity of BWM necessitates a sound theoretical understanding of the problem in order to evaluate the current BWM standard, and also to aid in future policy formulation. Similarly, in Chapter 3, we focus on deriving generalized management insights which are applicable to a variety of real-world cases, as opposed to deriving an optimal management strategy for a specific case. In addition to the data requirements necessary to derive such a management strategy, the complexity of such applied cases leads to potentially excessive computational burden. Chapter 3 analyses systems of two and three patches, which are likely to be too simple to analyse specific real world cases, but are sufficient to derive generalized management insights. The game theoretic methodologies in Part B are evaluated principally in terms of the assumptions about changes in stock location in Chapter 4 and the numerical method in Chapter 5. In Chapter 4, the fish stock is conceptualised as existing at a single point in space. The location of this point is determined in relation to fishing nations, which are also conceptualised as single points in space. Changes in stock location result from rises in ocean temperatures due to climate change. Such rises in temperature are likely to lead to other changes in the fish stock such as the size of the area where the fish stock can be found and increases in the maximum fish stock size which the ecosystem can support. These other aspects of changing stock location need to be considered in evaluating Chapter 4, as well as in formulating more applied models. In Chapter 5, a numerical method is adopted to analyse the effects of an endogenous risk of stock collapse. To do so, the utility function in the LM model is adapted such that it can be used in a numerical model. In order to isolate the effect of endogenous risk from changes in the utility function, a validation procedure is carried out by comparing analytically derived results in the deterministic case (without endogenous risk of stock collapse) to numerically derived results in the deterministic case. This reveals that changes to the utility function have a negligible effect and thus the results, in terms of the stability of Grand Coalitions can be attributed solely to endogenous risk of stock collapse. Overall, Part A of this thesis presents new insights into the determinants of optimal BWM standards. These insights demonstrate the conditions under which the current BWM standard, which aims to eliminate the risk of invasion establishment, may or may not be optimal. Part A therefore provides a novel theoretical framework which aids in the evaluation of current, and the determination of future standards. Part A also provides new insights into the control of established invasions, by extending existing spatially explicit optimal control models. Specifically, dividing space into patches and allowing for varying invasive population sizes within patches facilitates the optimal timing of management interventions and, in general, more detailed, and thus more efficient, management strategies. Part B provides a novel analysis of the effects of changing stock location on Grand Coalitions by explicitly introducing fish stock location in the analysis, and shows how farsightedness can stabilize Grand Coalitions in the face of such changes. Part B also shows how the effects of an endogenous risk of stock collapse on the stability of Grand Coalitions depends vitally on whether transition payoffs are included. These results can form the basis for more interdisciplinary analyses, analyses of different types of marine ecological change, and analyses of these changes in different settings, such as non-European countries.