© 2004 by ICES/CIEM International Council for the Exploration of the Sea/Conseil International pour l'Exploration de la Mer
Risk and strategy of fishers alternatively exploiting sea bream and tuna in the Gibraltar Strait from an efficiency perspective
Department Economía y Empresa, University Pablo de Olavide Ctra. de Utrera, Km 1, 41013 Sevilla, Spain
*Correspondence to I. Herrero: fax: +34 954 349 339. e-mail: ihercha{at}dee.upo.es.
While in the last few years interest in efficiency analyses has increased among fisheries economists, efficiency analyses have not been used so far to compare the profitability of alternative fisheries. We compared the profitability of the sea bream fishery and the tuna fishery, both operating in the South of Spain (Gibraltar Strait), based on an efficiency analysis and on a comparison of the efficient frontiers of the two fisheries. Results for 2002 show that, even if the sea bream is a higher-value species than tuna, the profitability of the latter is higher. However, sea bream still appears to be the target species of the fleet whereas the tuna fishery appears only as an alternative fishery. The key reason appears to be that the latter involves more risk than the former. The management of each fishing vessel seems related to the management strategy of the captain and to his assessments of risk.
Keywords: data envelopment analysis, efficiency, fisheries, risk
Received 15 May 2003; accepted 9 December 2003.
 |
1 Introduction |
|---|
 Top 1 Introduction 2 Methods3 Results4 DiscussionReferences |
|---|
In the last few years, interest in efficiency has increased among fisheries economists (e.g. Kirkley et al., 1995, 1998; Pascoe et al., 2001a). These studies have not just been focused on efficiency in itself but also on the factors affecting efficiency (Pascoe and Coglan, 2002) or on estimates of capacity utilization (Pascoe et al., 2001b; Vestergaard et al., 2002) or technical change. However, efficiency analysis has not been used so far to compare the profitability of alternative fisheries. We compared the profitability of two hook and line fisheries, the sea bream fishery (SBR), and the tuna fishery (BFT), operating in the Gibraltar Strait (these two FAO acronyms will be used throughout this article). Vessels of the same fleet may alternate their participation in the two fisheries, using essentially the same gear (although the tuna may be hauled manually rather than by using the device used for sea bream) and thus employing the same fixed inputs for both. During the period when the tuna may be caught, fishers may opt on a daily basis for participating in either SBR or BFT. A particular fisher may choose for a potentially high profit even if it is associated with a high risk, whereas another may go for a lower profit with less risk. This may show different fishing strategies and different assessments of risk among fishermen, and we will analyse the possible reasons affecting their decisions.
1.1 Description of the fisheries
The fleet investigated operates in the Gibraltar Strait. Most of the vessels participating in this coastal fishery are based in the port of Tarifa. We used daily landings data, aggregated by month, for 91 vessels operating in 2000, considering the monthly number of trips as a variable input. Tuna is a migratory species and consequently the fishery is a seasonal one. In 2002, the tuna fishery operated only during June–November (excluding August), though June was excluded because observations were few. Therefore, only trips during 4 months were finally considered in the analysis. Only vessels that have access to the SBR were considered as they use the same fixed inputs (GRT, HP, and number of crew) in both fisheries.
The target species is either the high-valued sea bream (with other species taken as by-catch) or bluefin tuna. While sometimes tuna appears in the landings together with sea bream, most trips are dedicated exclusively to either species. During 2002, the amount of tuna in landings represented 43% of the total weight of the catch (39% of the value), while the amount of sea bream represented only 25% in weight but 43% of the total value. Sea bream and tuna appeared together in the landings of a single vessel only during 18 trips. In these cases, the vessel must have changed its target species during the same trip as the two species cannot be caught in the same fishing grounds or at the same depth. We selected those trips dedicated exclusively to either species (although vessels may have a by-catch of other species) among those vessels that are based in Tarifa to avoid the inclusion of larger vessels with different characteristics from the North of Spain that participate only during the tuna season. On average, vessels participating more often in BFT have more engine power than the rest (Table 1). This seems reasonable as tuna is generally larger than sea bream and it is easier for more powerful vessels to fish this species. Different regulations exist for the management of this fishery. Among them, a seasonal closure was imposed during February and March in the year under study.
|
The most commonly used gear when fishing sea bream is a long line of special design ("voracera") and the gear used by this fleet for catching tuna is a much stronger line with hooks. Some years ago, the lines were hauled in manually. However, nowadays most vessels have machines to haul in the lines. This implies that the fishing capacity of the fleet is now much higher than it used to be and as vessels are investing to buy more of these machines, the fishing power is increasing rapidly.
The average price per kilogram (in euros) of sea bream is more than double the average price of tuna, while the coefficient of variation (CV) is roughly the same in the two species (Table 1). Even though the average price of sea bream is higher, the BFT has a much higher average value per trip (VPUE) than the SBR (Table 2). However, in this case, the CV is much higher. In economic analyses, risk is often related to the coefficient of variation of the profit or other income variables. Effectively trying to catch tuna involves a higher risk regarding income than the catch of sea bream. More risk-averse fishers may often choose the security of the sea bream fishery at the cost of lower profitability, whereas others may prefer to choose the tuna fishery even if it implies more risk.
|
) with coefficient of variation (CV) in the sea bream (SBR) and the tuna (BFT) fisheries in 2002 by month.1.2 Variables included in the analysis
The dependent variable used was the total value of the catch of the target species over the month. As stated before, we only considered the trips in which landings were composed of either sea bream or tuna (irrespective of other by-catch species). We used the value of the total catch as this measure is effectively an aggregated index of output, with each component of the catch weighted by its price (reflecting its relative importance to the overall objective of profit maximization). Prices vary from one catch to another, and within landings by a single vessel, as they depend on quality (there are around eight categories for each of the two species, depending on size and freshness), each agreement, total landings per day, the time the vessel got to port, and the demand.
The key inputs used in the model were the size of the boat (represented by its gross registered tonnage, or GRT), engine power (in hp), number of crew, and effort (number of trips made each month).
|
2 Methods |
|---|
|
Top1 Introduction2 Methods3 Results4 DiscussionReferences |
|---|
2.1 DEA background
Data Envelopment Analysis (DEA) is a non-parametric method to estimate technical efficiency based on optimization techniques (Charnes et al., 1978). DEA frontiers are based on optimal units, and the frontier in the output (input) oriented case represents the maximum level of output (or minimum level of inputs) that might be obtained given a certain level of inputs (outputs) if the units considered were efficient. The selection of an input- or output-oriented model depends on the case and interest of the study. While the emphasis of the former is on input reduction, the emphasis of the latter is on output augmentation. The units cannot reach a value above the frontier, because the major drawback is that the technique is deterministic and does not consider random variation in the data. Consequently, no assumptions on error distributions have to be made, as is the case for evaluating stochastic production frontiers.
The original (CCR) model of Charnes et al. (1978) assumed constant returns to scale. Banker et al. (1984) extended the model and constructed a new one, known as BCC, that allows for variable returns to scale. Figure 1 illustrates the two models for the case of one input and one output. In the case of constant returns to scale (Figure 1a), the feasible production set (or production possibility set, PPS) would be represented by points between the X-axis and the ray starting at the origin and passing through point C (the efficient frontier). The virtual unit associated with unit D would be D' in an input-oriented model (unit D' produces the same output as unit D by using less resources) or alternatively point D'' in an output-oriented model (keeping the same inputs, D should expand its outputs to the same level as unit D'' to be efficient).
|
In the case of varying returns to scale (Figure 1b), the PPS corresponds to the points enclosed between the X-axis and the segmented line joining G, A, C, B, and E and continuing parallel to the X-axis through H. The efficient frontier is given by points on the segmented line AE. Inefficient decision making units (or DMU) again can be transformed into efficient ones by contracting its inputs (D') or expanding its outputs (D'') for input and output orientation, respectively.
If the DMU under evaluation is excluded from the constraints (i.e., not included in the reference set), then inefficient units get the same score as in the original model whereas efficient ones get what has been called a "super-efficiency score" that can be >1. The higher this score, the more efficient is the associated DMU. This modified BBC model (Andersen and Petersen, 1993) has been commonly used to rank efficient units.
2.2 Models used
When different fishers (or firms) operate differently in respect of adjusting expected output (e.g. target species) relative to chosen input (e.g. investment in vessel and gear), the interest of decision makers may be focused on knowing who among them is able to develop the more efficient system. Similarly, in a fisheries framework, managers might be interested to know which of two alternative fisheries is the more profitable one. The criterion for profitability might be based on observations of the efficiency of the units using different fishing grounds or different target species. To avoid the influence of the inefficiency inherent to all vessels, it is useful to distinguish between what is known as "technical efficiency" within alternative fisheries and the "programme efficiency" among alternative fisheries (Charnes et al., 1981). This is achieved by first obtaining a set of efficiency measures for just those vessels operating in the same fishery (the reference group being either SBR or BFT). Secondly, all vessels are made efficient (Figure 2) by projecting the data points for each vessel on the efficient frontier (multiplying its output by the expansion ratio 
). Thus, observation A of an SBR vessel is projected on the efficient frontier associated with SBR (A*SBR) and observation B of a BFT vessel onto the BFT frontier (B*SBR).
|
The last step consists of measuring the distance between the two frontiers to investigate whether one lies consistently above the other (implying a higher programme efficiency). For example, the projected point A*SBR would appear as inefficient when compared to efficient BFT vessels (A*BFT), while the projected point B*BFT would appear as super-efficient when compared to the SBR frontier (B*SBR). This situation would imply that BFT vessels generally perform better, even if individual vessels may be less efficient than some SBR vessels.
To compare the means of the efficiency rates for vessels participating in each of the two fisheries, we first used a standard BCC output-oriented model for the SBR fishery. For each SBR observation in period t (vessel j0), this model calculates the proportional increase in the catch of sea bream that would make this vessel efficient (radial expansion ratio ; 1<<
) by comparison with all SBR observations (j) in the same period:
|
|
|
| (1) |
|
|
|
|
is the weight associated with each unit in the reference set (i.e., the optimal efficient unit associated with the unit under analysis is the convex combination of the units in the reference set, given by the positive 's). The technical efficiency (TE) score associated with vessel j0 is 1/1 which varies between zero and one ("radially efficient"). Subsequently, all catch observations (y) are transformed to their efficient equivalent (y*) by
|
| (2) |
A completely analogous DEA analysis is carried out to estimate the efficiency of the vessels operating in the BFT, and again the observations are transformed to their efficient equivalent. Based on these values for efficient vessels, a new set of two analogous DEA analyses is solved. First, each unit operating in SBR is compared with all units operating in BFT during the same period by substituting y by y* (j0 referring to SBR units) and nSBR by nBFT (j referring to BFT units) in Equation (1). Next, the same is repeated for the BFT fleet (j0 referring to BFT and j to SBR). In both cases, the parameter to be maximized is 2, corresponding to a programme efficiency (PE) of 1/2. The scores of PE per observation were averaged for the two fleets separately.
All models were programmed in GAMS, a multi-purpose software package that is mainly used for solving optimization problems.
|
3 Results |
|---|
|
Top1 Introduction2 Methods3 Results4 DiscussionReferences |
|---|
The average scores per month for the technical efficiency (Table 3) were generally slightly higher but also more variable for the BFT (range: 0.63–0.96) than for the SBR (range: 0.72–0.79). The standard deviations were also more variable (range 0.08–0.25 vs. 0.19–0.24), but the overall standard deviation was higher for the BFT (0.3) than for the SBR (0.21). The monthly differences appear not to be related to differences in the number of observations in each month or the percentage of the total fleet participating in each fishery (Table 4). Also, the technical characteristics of both fleets (GRT and engine HP) showed similar levels of variability (Table 1), suggesting that the variability in TE scores of the two fleets cannot be caused by one fleet being more heterogeneous than the other. The high variability of the TE in the BFT might also be influenced by the randomness, or risk, in the profitability of the different trips. During months when profits are more variable, observations associated with vessels having had a "bad day" appear as inefficient, resulting in lower average TE and higher standard deviations. Catching tuna involves a higher degree of uncertainty regarding potential income than catching sea bream as the average TE scores vary greatly over the months (Table 3).
|
|
Table 3 also shows that the average programme efficiency scores differed considerably between the two fisheries. Vessels were more efficient when operating in the BFT (average PE=3.42) than when operating in the SBR (average PE=0.32) during all the months analysed, indicating that the BFT is the most productive fishery.
A relationship appears to exist between the programme efficiency scores and the relative effort employed in each fishery (Tables 3 and 4). The differences in PE between BFT and SBR were largest in September, when 80% of the fishers joined the BFT, and they were smallest in November, when 86% joined the SBR. This seems consistent with general fisherman behaviour. In contrast, the standard deviation of PE in the BFT by month does not seem to be related to the percentage of vessels involved in the fishery (Tables 3 and 4).
To explore these differences further, vessels were classified into four groups (G1–G4) depending on their relative involvement in each fishery. G1 was composed of 25% of the vessels having the highest percentage of trips in the BFT, G2 was composed of the next 25%, and so on (G4 actually representing full-time SBR vessels during the 4 months analysed). Table 5 provides the average TE and average PE for the four groups. While the efficiency scores of the SBR observations just fluctuate among groups, both technical and programme efficiency associated with trips in the BFT were highest for G1 and steadily decreased for vessels participating less often in this fishery.
|
The reason for the declining trend in programme efficiency associated with trips made by vessels participating less often in the BFT might be that vessels become more experienced and therefore more efficient as they join the fishery more often. Another possible incentive for joining the BFT more often might be that for these vessels, the difference in efficiency experienced when alternating between BFT and SBR is even higher than for the rest of the group. Whatever the causes are, the results indicate that vessels in G1 are characterized by a certain degree of specialization in the BFT.
A Wilcoxon test (Table 6) showed that the programme efficiency scores of the observations in the two fisheries were significantly different (p<<0.01) for all the months considered, vessels in the BFT always outperforming those in the SBR.
|
|
4 Discussion |
|---|
|
Top1 Introduction2 Methods3 Results4 DiscussionReferences |
|---|
The study of fishers' behaviour and their reaction to fisheries management has become a hot issue among fisheries researchers. Hilborn (1985) concluded that most management failures were due to a misunderstanding of fishers' behaviour, and not due to a poor understanding of fish dynamics. According to Babcock and Pikitch (2000), fisheries managers manage fishers rather than fish, and their dynamic behaviour should therefore be studied in greater depth. Many studies have appeared in the literature (Hilborn, 1985; Hilborn and Walters, 1987), investigating characteristics ranging from information flow among vessels (Mangel and Clark, 1983; Little et al., 2004) to responses to stock collapse (Mackinson et al., 1997). Zetina-Rejón et al. (2004) analysed fishers' behaviour based on ecological, economic, and social criteria. Recent papers have focused on behaviour in relation to regulation compliance (Eggert and Ellegard, 2003; Nielsen and Mathiensen, 2003), while others have investigated different strategies in multispecies fisheries (Babcock and Pikitch, 2000; Pech et al., 2001). However, so far they have not included the potential risk involved in choosing alternative target species.
While most studies have used traditional modelling, we estimated the profitability of two alternative fisheries from an efficiency-theory perspective. Even if the results may have been slightly distorted because a few observations for trips that landed both sea bream and tuna were excluded (19 out of 1176) or because of possible errors in the data (fisheries data are not always reliable), the methodology presented for the analysis of the profitability of alternative fisheries and its potential relation to fishermen's strategies and assessment of risk seems promising. While sea bream is the target species of the fleet operating in the Gibraltar Strait and the tuna fishery represents only an alternative fishery, the latter appears to have a higher potential for profit. Using an efficiency perspective has the advantage of being able to carry out the analysis without knowing the prices of all productive inputs. The efficiency is calculated relative to the given inputs and allowing a multi-input multi-output context when required (Charnes et al., 1981). This methodology has never been applied to the field of fisheries.
The average technical efficiency score did not differ much between the two fisheries. This TE score is just an indicator of how efficient vessels are relative to vessels operating in the same fishery. A vessel may be operating very inefficiently in spite of participating in a very profitable fishery and vice versa. The scores may also be influenced by the random variations (or risk involved) in the profitability of each fishery. Traditionally, the concept of risk has been related to variability in the value-per-unit-of-effort (VPUE). However, VPUE may be heavily affected by differences in the fishing power of vessels belonging to the same fleet. Thus, the standard deviation of the VPUE will be much higher for a heterogeneous fleet than for a homogeneous fleet, even if the amount of risk involved in the fishing activity is low. In contrast, TE scores take such differences into account as a vessel with less fishing power may be efficient even if the value of the landings is low as long as it does not use a large amount of resources. Thus, if risk is low then even a heterogeneous fleet may show a low variability in the TE scores. Of course, variations in TE scores are due not only to risk but also to differences in efficiency among vessels but the same bias affects variations in VPUE. Therefore, variability in the average TE, and its standard deviation, over the fishing season is supposedly a better indicator of the level of risk involved in the fishing activity than the variability in the VPUE.
One possible explanation for the high variability of the TE scores in the BFT over time might be that sea bream move easily from one fishing ground to another. The strong winds often blowing through the Strait of Gibraltar may cause shifts in the banks favoured by this species to areas closer to Tarifa or to areas farther away around Morocco depending on wind direction.
The relationship observed between the programme efficiency scores and the relative effort employed in each fishery may imply that, regardless of the monthly risk involved in the tuna fishery, the number of vessels that decide to join the tuna fishery is strongly related to the absolute difference in profitability between the two fisheries. Hence, when the difference in expected average profitability reaches a certain level (depending on the attitude of individual fishers to risk and on the utility function of profitability), most fishers should take the option of joining the tuna fishery even if it involves more risk. However, even if fishermen may be generally considered as risk seekers, the Tarifa fishers require a huge difference in profitability before joining the tuna fishery.
The standard deviation per month of the programme efficiency is not related to the percentage of vessels joining each fishery. This may be due to fishermen not being able to assess the variability in the benefits within each month. They may know that there is a higher risk involved in one fishery than in the other, but guessing the risk in each month may be too difficult. In any case, most fishers seem to join the more risky BFT only when its profitability is much higher than the profitability of the SBR in which they usually operate.
In the light of the results, the conclusion must be that the sea bream fishery is less profitable than the tuna fishery, although the latter involves more risk (as profits are more variable). Variations in risk do not seem to be closely related to variations in prices as coefficients of variation in the prices of the two species are roughly the same (40 and 38% for SBR and BFT, respectively) but to the variations in the total catch by trip. Tuna shoals often move in this area of high winds and the amount of fish present at the fishing ground where a vessel had decided to go fishing varies from one day to another (Garcia et al., 2001). This could be one reason why the BFT is highly variable.
While vessels basically employ the same fixed inputs on both fisheries, captains should, based on expected catches, prefer the BFT to the SBR. Although the BFT is the preferred option in July and September, overall, most fishers (62%; Table 4) prefer the SBR. This seems to be highly related to the fisher's strategy and assessment of risk. The variability in the VPUE and in the TE scores associated with the tuna fishery is higher than that associated with the sea bream fishery. These measures may be indicative of a higher amount of risk involved in the tuna fishery. Given this, most fishers seem to prefer the security of the all-year-round SBR to the highly profitable but more uncertain BFT. Only some fishers with a higher tendency to take risks may prefer the tuna fishery.
While sea bream fishery is nowadays mechanized, the devices used to haul in lines cannot always be used to haul in the much larger tuna, implying that more human labour is required in the BFT than in the SBR. This may be another key reason for captains to prefer the latter, even at the cost of a lower profit.
|
Acknowledgements |
|---|
The author thanks the editor, two anonymous referees, and Cristina García for their valuable comments, and the Dirección General de Pesca de la Junta de Andalucía (Andalusian Local Government) for providing the data. The Centro de Estudios Andaluces (CentrA) has partially financed this work.
|
References |
|---|
|
Top1 Introduction2 Methods3 Results4 DiscussionReferences |
|---|
-
Andersen P. and Petersen N.C. (1993) A procedure for ranking efficient units in DEA. Management Science 39:1261–1264.[Web of Science]
Babcock E.A. and Pikitch E.K. (2000) A dynamic programming model of fishing strategy choice in a multispecies trawl fishery with trips limits. Canadian Journal of Fisheries and Aquatic Science 57:357–370.[CrossRef]
Banker R.D., Charnes A., Cooper W.W. (1984) Some models for estimating technical and scale inefficiencies in DEA. Management Science 30:91078–1092.[Web of Science]
Charnes A., Cooper W.W., Rhodes E. (1978) Measuring the efficiency of decision making units. European Journal of Operational Research 2:429–444.[CrossRef][Web of Science]
Charnes A., Cooper W.W., Rhodes E. (1981) Evaluating program and managerial efficiency: an application of data envelopment analysis to program follow through. Management Science 27:6668–696.[Web of Science]
Eggert H. and Ellegard A. (2003) Fishery control and regulation compliance: a case for co-management in Swedish commercial fisheries. Marine Policy 27:525–533.[CrossRef][Web of Science]
García J. J., Garcia F., Gonzalez M. D., Vilches M. (2001) Políticas de Gestión en Pesquerías Artesanales (in Spanish). Servicio de Publicaciones de la Universidad de Huelva.
Hilborn R. (1985) Fleet dynamics and individual variation: why some people catch more fish than others. Canadian Journal of Fisheries and Aquatic Science 42:2–13.
Hilborn R. and Walters C.J. (1987) A general model for simulation of stock and fleet dynamics in spatially heterogeneous fisheries. Canadian Journal of Fisheries and Aquatic Science 44:1366–1369.
Kirkley J.E., Squires D., Strand I.E. (1995) Assessing technical efficiency in commercial fisheries: the Mid-Atlantic Sea scallop fishery. American Journal of Agricultural Economics 77:3686–697.[CrossRef][Web of Science]
Kirkley J.E., Squires D., Strand I.E. (1998) Characterizing managerial skill and technical efficiency in a fishery. Journal of Productivity Analysis 9:145–160.[CrossRef][Web of Science]
Little L.R., Kuikka S., Punt A.E., Pantus F., Davies C.R., Mapstone B.D. (2004) Information flow among vessels modelled using a Bayesian network. Environmental Modelling and Software 19:27–34.[CrossRef]
Mackinson S., Sumaila U.R., Pitcher T.J. (1997) Bioeconomics and catchability: fish and fishers behaviour during stock collapse. Fisheries Research 31:11–17.[CrossRef][Web of Science]
Mangel M. and Clark C.W. (1983) Uncertainty, search and information in fisheries. Journal du Conseil International pour l'Exploration de la Mer 41:93–103.
Nielsen J.R. and Mathiesen C. (2003) Important factors influencing rule compliance in fisheries lessons from Denmark. Marine Policy 27:409–416.[CrossRef][Web of Science]
Pascoe S., Andersen J.L., de Wilde J.W. (2001) The impact of management regulation on the technical efficiency of vessels in the Dutch beam trawl fishery. European Review of Agricultural Economics 28:2187–206.[Abstract]
Pascoe S. and Coglan L. (2002) Contribution of unmeasurable factors to the efficiency of fishing vessels: an analysis of technical efficiency of fishing vessels in the English Channel. American Journal of Agricultural Economics 84:345–57.
Pascoe S., Coglan L., Mardle S. (2001) Physical vs. harvest measures of capacity: the case of the UK vessel capacity unit. ICES Journal of Marine Science 58:61243–1252.
Pech N., Samba A., Drapeau L., Sabatier R., Laloe F. (2001) Fitting a model of flexible multifleet–multispecies fisheries to Senegalese artisanal fishery data. Aquatic Living Resource 14:81–98.[CrossRef]
Vestergaard N., Hoff A., Andersen J., Lindebo E., Gronbaek L., Pascoe S., Tigley D., Mardle S., Guyader O., Daures F., van Hoof L., de Wilde J. W., Smith J. (2002) Measuring capacity in fishing industries using the Data Envelopment Analysis Approach. Research project report. (coordinator) Department of Environment and Business Economics. University of Southern Denmark. ISBN 87-90872-05-3.
Zetina-Rejón M. J., Arreguín-Sánchez F., Chávez E. A. (2004) Exploitation of harvesting strategies for the management of a Mexican coastal lagoon fishery. Ecological Modelling (in press).
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||