Efficiency, together with sustainable utilization and equity in access to resources, is seen as one of the main aspects to be considered in the management of fisheries (Cochrane 2000). According to Grafton et al. (2006), efficiency in fisheries is considered to be doing the best possible with the available resources in terms of fish stock and fishing inputs. To measure firm-level efficiency, one technique that can be applied is data envelopment analysis (DEA), which was first introduced in the literature in 1957 (Farrell) for application in the public sector and then in agriculture. It was then expanded on and further developed until today, when it has become a fundamental tool for efficiency analysis in fisheries. The technique extracts information about the production process of each vessel [herein called a decision making unit (DMU)] through linear calculations used to construct a non-parametric piece-wise frontier of the data. Efficiency measures for each DMU are then calculated relative to this frontier. Each DMU’s performance measure traces out a best-practice frontier (Coelli et al. 2005), so all DMUs lie either on or below the frontier, and this is how performance, in the form of efficiency, is tracked. Technical efficiency (TE) and scale efficiency (SE) are measures of efficiency calculated using DEA. According to Farrell (1957) and Kumbhakar and Lovell (2000), TE measures the relationship between a vessel’s inputs to the fishing process and its outputs. Full efficiency is achieved when outputs are maximized from a given set of inputs. By contrast, technical inefficiency can result from a fishing strategy with inputs, such as energy, labour or other running costs, that are too high or a fishing strategy that has excessive invested capital in relation to the potential revenue. Reasons for inefficiency have been associated with shifts in management (e.g. Pascoe et al. 2001), vessel characteristics (e.g. Pascoe and Coglan 2002) and skipper characteristics (e.g. Squires and Kirkley 1999, Tingley et al. 2005, Vázquez-Rowe and Tyedmers 2013). According to Grafton et al. (2006) SE expresses whether a firm is operating at its “optimal” size so that any modifications to its size will render the unit less efficient, and it is measured by dividing the aggregate efficiency by the TE of each DMU. In the literature, TE has been generally estimated through an output-oriented approach (e.g. Kirkley et al. 2002, Reid et al. 2003, Vestergaard et al. 2003, Kirkley et al. 2004). In the studies found in the literature, an output-orientation has been generally used to investigate the factors affecting the TE of fishing vessels under a quota system (e.g. Tingley et al. 2005). In the Mediterranean, an application of output-oriented DEA was found in Maravelias and Tsitsika (2008), who assessed the efficiency in purse-seine fishing in the eastern Mediterranean. Another example of estimating TE by applying DEA in small-scale fisheries in the Mediterranean was provided by Idda et al. (2009). Both of these studies used variations of output-oriented approaches. Few studies in fisheries have utilized an input-oriented approach, even in the Mediterranean where fisheries are largely managed under input control systems. However, the application of input-oriented DEA has been considered particularly meaningful for fisheries managed under an input control system (e.g. Pinello et al. 2016).

Input-oriented DEA models have also been applied to estimate input-based measures of overcapacity, where overcapacity is synonymous with overcapitalization (Ward et al. 2004): usually, when fishing vessels are not technically efficient it is because they have running costs that are too high and they are operating in a state of overcapacity in the sense that a lower capacity could be used to achieve the same harvest. A fleet in a status of overcapacity is characterized by low TE and is then not properly using its input in order to produce outputs. The link between overcapacity and efficiency of fishing fleets has been investigated by Pascoe et al. (2003), Grafton et al. (2006) and Maravelias and Tsitsika (2008). Overcapacity has been a key issue in the EU fleets, with most of the fisheries in a status of inefficiency and in most cases heavily subsidized, as was highlighted by Villasante (2010), Villasante and Sumaila (2010) and Maravelias and Tsitsika (2008). In the Mediterranean Sea, the overfishing of most of the fishery resources and the overcapacity of fleet have been clearly diagnosed (Colloca et al. 2013, 2017). To better balance the equilibrium between the stock status and the fleet capacity, the European Commission has adopted programmes for reducing the fleet capacity. Since 2002, this reduction programme has placed particular emphasis on monitoring fleet exit and entrances (EU 2371/2002) and supported the scrapping of fishing vessels (EU No 508/2014). This has led to a consistent decrease in the Italian fleet in terms of gross tonnage and engine power, which has particularly impacted the Mazara del Vallo fleet in Sicily. This fleet is composed of distant-water trawlers, mainly operating in the Strait of Sicily, one of the most important fishing area of the Mediterranean, with many demersal stocks—including deep-water shrimp—shared between European and third countries (Fiorentino et al. 2008, Gristina et al. 2006, Knittweiss et al. 2013). Shared stocks require a particularly effective management strategy, as their management is outside the responsibility of the individual states and consequently of the General Fisheries Commission for the Mediterranean, and the European Commission are undertaking several common actions to enforce rational management and improve utilization of these marine resources. Therefore, a broad analysis of the fleets exploiting shared stocks, including the efficiency component, is critical for effective management to be exercised.

To contribute to this process, this paper aims to investigate the efficiency of the strategies employed by the Mazara del Vallo fleet: pesca a fresco, trawling mainly in shallow-water and pesca a congelato, mainly targeting deep-water shrimp. The analysis was made through an input-oriented DEA comparing TE and SE performance. Differences in efficiency were found between the different fishing strategies, and these are discussed in light of the use of LED lights, along with the implications for discard reductions.

Differences in fishing grounds of shallow and deep-water trawling were reflected in the differences in the crew size of the vessels as well as in the activity costs and the market modality. The pesca a fresco trawling was more labour-intensive (on average 8-9 crew per vessel) than the pesca a congelato (on average 6-7 crew per vessel). The pesca a fresco trawls had an average tow duration of approximately two hours with a tow speed of 3-4 knots, while the pesca a congelato had an average tow duration of four to five hours with a tow speed of 2-3 knots and a slower engine speed. These factors result in lower fuel consumption. Different crew sizes for the two strategies did not result in significant increases in personnel costs because under the crew-share system the proportion of revenues assigned to the crew-share is a fixed ratio, no matter the number of crew members (Gee et al. 2017, Pinello et al. 2017).

During a trawl operation, one of the most labour-intensive components is the process of sorting the commercial catch from the discards. The reduced labour required for this activity with the pesca a congelato strategy is shown by the lower number of crew members required on these vessels and by evidence in the literature (e.g. Milisenda et al. 2017).

In Table 4 the main findings of the focus group discussion are reported. The discussion focused on eight main points, including the social component, the fishing activity and the use of the LED lights. Of the 23 individuals interviewed, more than 80% were owners who worked onboard their vessels as the skipper. Although demographic information such as age and educational levels were collected, no specific pattern related to this information was identified, so it was not reported here. When the focus group members were queried on the main problems related to the fishing activity, the majority of respondents highlighted the shortage of specialized crew, in particular skippers and engineers, and the need for more technological innovation aimed at further reducing discards and increasing the quality of the landings.

The pesca a congelato strategy showed the highest average levels of efficiency, with a TE of 0.96 and an SE of 0.99 (Table 5). In terms of TE, all of the pesca a congelato vessels were very close to the frontier and they were working at nearly the maximum level of efficiency. In fact, 15 vessels were found to be on the frontier of efficiency. The very high SE score highlighted that the vessels were operating close to the optimal scale of production. On average, only one percent of inefficiency can be attributed to the scale size. CRS were detected for 2 vessels, while the remaining 30 vessels were equally distributed between those operating at IRS and those operating at DRS.

On average, the mixed strategy fell between the two other strategies, as measured by TE and SE (Table 6). The vessels following the mixed strategy had a TE level equal to 0.91 and there was only one vessel on the frontier, while the SE was 0.97. Three vessels were operating at CRS, while the other thirteen vessels were at IRS.

On average, the pesca a fresco was the less efficient fishing strategy, as measured by TE (Table 7). The vessels following this strategy had the lowest TE level, equal to 0.74, and could therefore proportionally decrease their inputs by 36% and still produce the same amount of output. No vessels were on the frontier for the TE, while the SE was 0.97. IRS was found for seven vessels, while the optimal scale was detected for one vessel.

There is an evident positive correlation between the occurrence of shrimp in the landings and the higher efficiency scores found for the three categories of vessels (Table 8). The higher the presence of shrimp in the landings, the higher the TE for the trawlers. The pesca a congelato showed the highest TE performance, with the greatest number of vessels on the frontier of efficiency while the pesca a fresco showed the lowest TE performance, with no vessels on the frontier of efficiency. The mixed-strategy vessels had a TE performance that was a mix of the two other categories and accordingly had no vessels on the frontier. The range of TE scores for the three categories corroborated the correlation between shrimp in landings and efficiency.

According to the focus group discussion, the fleet targeting shrimp started equipping their nets with LED lights in 2013. The fishers reported that the lights attract mainly ARA/ARS and DPS, therefore increasing the catchability of the gear, and seemed to be particularly effective at night. Although the effectiveness of the lights during daylight hours is probably lower because the shrimp would already be near the bottom, the fishers usually prefer to leave the lights operating. The preferred colours are white, green and blue, in that order, and the choice depends mostly on experiences gathered by fishers. In general, the trawl nets are equipped with about 100 flashing lights installed in groups of ten spaced two or three metres from each other. The LED lights are attached to the headrope of the net and to the bridles. They cost about €3 to €5 each, and to be kept constantly in operation they require two new batteries every 15 days, which cost about €0.15 each. The annual cost for the lights is therefore about €300 to €500 per net, while their maintenance costs are about €20 to €30 per fishing trip (20 days). Under the pesca a fresco strategy, the main targets were bony fish and cephalopods, which were landed fresh because the fishing trips lasted three days or less, with each tow lasting two hours or less. The short duration of the tow was an intentional strategy to reduce the total amount of catch in the cod end. Catches of unwanted species were a matter of concern to the fishers on two operational fronts. First, too high a quantity of catch in the codend increases the mechanical damage, which reduces the product quality; second, the sorting of the catch on the deck takes longer with higher quantities of non-targeted species. The vessels mainly operating under the a pesca a fresco strategy employed 8 to 9 fishers as a result of the increased workload due to more tows per fishing day and increased quantities of catch (landings plus discards) to handle. The pesca a congelato strategies targeting DPS (250-400 m) and ARA and ARS (>400 m) had fishing trips that lasted between one week and 20 days, and the catch was landed frozen. The tow lasted more than four hours at lower speeds as the lower amount of biomass allowed for a greater duration. The crews were smaller, with an average of 6 to 7 fishers because the smaller number of daily tows and the lower catch quantities reduce the required effort, while the lower speeds reduce the fuel consumption. The entry costs of using LED lights in the pesca a congelato strategy was so low (<1% of the overall turnover) that the entire fleet has started to use them, and no correlation could be found with age of the skippers/owners who participated in the focus group discussions. Even fishers who were not convinced of the effectiveness of using the lights used them, just in case. Further, the fishers reported that they did not believe that the use of the lights during the daytime was effective, but the maintenance costs were so low and there did not seem to be any negative consequences of using the lights during daytime fishing, so the fishers left the lights on for all fishing activity. The use of lights has become a structural component of the gear, and it has become the latest technological innovation for this fleet.

Between the two main strategies followed by the Mazara del Vallo trawl fleet, pesca a congelato (at depths >250 m) showed higher TE and SE scores than pesca a fresco (at depths <250 m). Considering these positive efficiency performances, it would appear that the negative effects of overcapacity are not evident for the portion of the fleet which targets deep-water shrimp. Further, from the perspective of the DEA analysis, it would appear that the pesca a congelato fleet is close to a point of equilibrium, with a good balance between costs and revenue. Over the years, an increased use of LED lights has been found in the pesca a congelato fleet.

According to the fishers, the adoption of the use of LED lights has increased the catchability of the gear towards the deep-water shrimp, in particular during the night-time. For this reason, the use of lights extended the potential fishing time, and this allowed the fishing activity to become more specialized while further increasing its productivity rate. The adoption of the lights for the Mazara del Vallo fleet was done without any scientific inputs—the fishers adopted and spread this technological improvement within their own community. It has now spread across the entire fleet and has become a structural element of the fishery operations.

Further to the focus group discussion, it is well established that during the day shrimp feed mainly on detritus on the bottom while at night they migrate off the seafloor to predate on zooplankton (Shumway et al. 1985, Sobrino et al. 2005, Savenkoff et al. 2006). Light could lead shrimp to swim towards it, or, additionally, attract shrimp searching for prey towards the bottom, favouring their capture in the fishing gear.

Regarding the discard production, the focus group reported that there were reduced levels of discards in the pesca a congelato strategy, and this was further corroborated by the longer tow duration and lower crew sizes required for sorting the commercial catch from the discards (Table 4). This reporting was supported by findings in the literature, in which deep-water fishing resulted in lower discard levels. The available literature from the Mediterranean suggested that trawlers targeting deep-water crustaceans produced fewer discards than shallow-water trawlers. An average overall discard rate for bottom trawlers in the Strait of Sicily during the early 2000s was reported by Vitale et al. (2006) to be 52%. For the shrimp fisheries of the Strait of Sicily, Castriota et al. (2001) reported a mean discarded biomass value of 49% of the total catches, similar to the discard ratio produced by other shrimp trawls in the Mediterranean Sea (43.2%±27.2%), but higher than mean values reported for other Mediterranean trawl fisheries (32.9%±2.8%) (Tsagarakis et al. 2013). More recently, Milisenda et al. (2017) analysed catches of bottom trawlers targeting deep-water rose shrimp (DPS) in the Strait of Sicily over the period 2009 to 2013, and reported a mean discard rate of 36% of the total catch. Finally, information reported in the review of Tsagarakis et al. (2013) suggested that there is a clear gradient of decreasing discard levels with increasing depths of fishing grounds and that deep-water (>400 m) trawling targeting shrimp generates fewer discards than shallow trawling that targets mid-water depths (Table 9).

The combination of these factors and the shortage of specialized crew has also had some influence on the shift of activities from the pesca a fresco towards the pesca a congelato strategy. The change of fishing strategy occurring in distant trawlers could have had a positive effect on discards. According to Gristina et al. (2013), the amount of immature individuals routinely discarded in demersal assemblages exploited by trawlers in the Strait of Sicily amounted to more than 60% at depths of less than 200 m and less than 30% at depths between 200 and 800 m. Moving from shallow-water fisheries targeting fish and cephalopods to deep-water fisheries targeting crustaceans could be a good strategy for more sustainable trawling in the Mediterranean.

In summary, a clear correlation was found between the occurrence of shrimp in the landings and the higher efficiency scores of vessels with the pesca a congelato strategy. The lower levels of discards caught through pesca a congelato than pesca a fresco and the higher catchability of the gear through the regular utilization of lights can be considered further important points of reflection for the management of shared stocks exploited by trawling in the Mediterranean. Under an effective management framework, the evolution from classical multispecies fisheries targeting the shelf bottoms to a fishery targeting deep-water crustaceans on the slope could shift trawling in other Mediterranean fleets towards a more sustainable pattern from both an economic and from an ecological point of view.