The overall purpose of this PhD thesis is to investigate different aspects of fishermen’s behaviour using production economic models at the individual and industry levels. Three parts make up this thesis. The first part provides an overview of the thesis. The second part consists of four papers analysing efficiency at the vessel level and factors influencing this. The third part consists of two papers and presents industry level analyses and focuses in particular on the likely impacts of implementing individual transferable quotas. The models are able to allow for changes in fishermen’s behaviour via individual learning and adjustments in output mix. All the papers included in Part II: Modelling and Evaluating Fishermen’s Behaviour consider factors influencing fishermen’s behaviour. Knowledge about these factors is important to give a correct description of fishermen’s behaviour. However, including all relevant factors in specific analyses is impossible, and it is therefore important to be aware of the most essential ones. As demonstrated in the literature review of Paper 1, a large number of factors may significantly influence fishermen’s short run behaviour, i.e. choice of gear type or fishing location. Behaviour can be viewed as being determined by the fishermen’s objectives subject to different restrictions, given by physical resources, time, mental capacity and information, and institutions. The review of the extensive literature gives reasonable support to the neoclassical assumption of utility maximisation through profit maximisation. Furthermore, the literature has demonstrated the importance of physical restrictions and time. The emphasis on these may of course be a consequence of the relatively easy access to such data. In the following three papers, specific aspects of fishermen’s behaviour are analysed. In Paper 2, technical efficiency and reasons for inefficiency are estimated using the Stochastic Production Frontier Approach. The results suggest that the level of technical efficiency is not influenced by the choice of revenue or weight as the output measure. Also, it has no profound impact whether inputs are measured by including fishing power and fishing time separately or as a composite measure. However, the output elasticities are influenced by these choices. Furthermore, vessel size, employment status and experience is found to influence inefficiency. Paper 3 considers how to include fish stocks in efficiency analyses. The biological developments are important in relation to fisheries, because fish stocks are one of the primary components in the production process. It is worthwhile to evaluate whether different methods of including fish stocks give rise to different conclusions. Three methods are investigated as possible ways to include fish stocks. The first method is based on catch data, while the two other methods are based on independent stock measures. It is shown that estimations based on the former give different results from the ones based on the latter. This conclusion is independent of the choice of time horizon and choice of other input/output measures. Paper 4 considers fishermen’s behaviour to counteract uncertainty. When performing efficiency evaluations, this is done on ex post data. However, in relation to fisheries such an approach may be too harsh, because the fishermen are operating in an uncertain environment with variations in fish stocks, weather, etc. Fishermen therefore seek in their ex ante decisions to cope with uncertainty. If the conditions are better than expected, this may result in some inputs not being used. In ex post efficiency evaluation, this is interpreted as inefficiency, although it was - in an ex ante perception - rational to bring the inputs along. This type of inefficiency can be denoted rational inefficiency. By further developing the method from Bogetoft and Hougaard (2003), an evaluation of 308 Danish fishing vessels is performed. The results indicate that these vessels seek to insure themselves against uncertainty by allowing for the highest flexibility in crew payments, followed by fuel costs, sales costs and costs for ice/provisions respectively. Accounting for changes in fishermen’s behaviour at the industry level is investigated in Part III: Industry Models of Fishermen’s Behaviour and Individual Transferable Quotas. Many bioeconomic models have been set-up through time to evaluate such changes, but none of these have to the author’s knowledge allowed for the behavioural flexibilities, as included in the modelling framework presented here. The starting point for Paper 5 is the fact that management changes will most likely result in different behaviour by the fishermen. It is necessary to account for these changes, when evaluating the expected gains to be derived. Based on the Data Envelopment Analysis approach, a framework to calculate these gains is provided. The gains are calculated by comparing industry profits in the initial management system with industry profits in a management system based on Individual Transferable Quotas (ITQs). Two types of behavioural flexibility are allowed in the system. They concern the ability to learn best-practice (catch-up) and the ability to change the input and output composition (mix). The framework is then adapted to a dataset from the Danish fishery. Not surprisingly, the gains rise with increased behavioural flexibility. Under the most restrictive assumptions, reallocation of the ITQs will alone result in a 50% increase in gross profits, while this level increases to 87% in the most flexible situation. The final Paper 6 provides an extension of the framework developed in Paper 5. A complex of restrictions is included to obtain more realistic estimations of the potential reallocations gains, when applied to specific fisheries. The restrictions relate to the determination of best-practice, possible levels of individual learning and changes in output composition. By using a dataset covering the entire Danish commercial fishery, we obtain estimates of the plausible tradability gains if Danish fisheries had been regulated by ITQs in 2002. In the most flexible (optimistic) situation, a 92% increase in gross profits can be expected, but this level is significantly reduced if vessel behaviour is restricted. Furthermore, a series of policy implica-tions are considered in relation to an ITQ system, including concentration, specialisation, market activity and price changes. Finally, plausible consequences of exogenous chocks and changes in management practice in the form of mesh size increases are considered.