3. The purse-seine fishery for tunas in the Eastern Pacific Ocean

In this section we focus attention on the purse-seine fishery for tunas in the EPO. We find that excess capacity exists for the EPO fishery with respect to yellowfin (YFT), skipjack (SKJ) and bigeye (BET) tunas caught in sets on dolphins, sets on floating objects and sets on unassociated schools.

3.1 Data and methodology

Capacity output, capacity output adjusted for TE and CU rates (observed output divided by CU or observed output divided by capacity output adjusted for TE) are estimated by DEA. We attempted to estimate the output-oriented non-radial method of Russell (1985), but the results were unsatisfactory. We instead estimated the output-oriented radial expansion approach, whereby all outputs were kept in fixed proportions as they were expanded, holding fixed factors constant and with full utilization of variable inputs. The CU rates are thus ray measures (Segerson and Squires, 1990).

The set- and vessel-level purse-seine data from the EPO tuna fishery were provided by the Inter-American Tropical Tuna Commission (IATTC) for 1980-2002. These data, by set and vessel, included landings of yellowfin, bigeye and skipjack tunas, vessel GRT and other measures of vessel size (cubic meters, net weight, or length, weight and depth in metres), trip lengths (days, arrival date minus departure date for trip), number of sets. Total is catch in tonnes, and is derived from observer data (or logbook data when observer date not available) raised to unloaded weight. All of these data were differentiated by mode of fishing, i.e. sets on fish associated with dolphins, sets on floating objects and sets on unassociated schools. The data were also differentiated by vessel size class (carrying capacity in tonnes) as follows: (1) 0-45 tonnes; (2) 45-91 tonnes; (3) 92-181 tonnes; (4) 182-272 tonnes; (5) 273-363 tonnes; (6) >363 tonnes. Biomass estimates for yellowfin, bigeye and skipjack tunas were provided by the IATTC (Maunder, 2003 personal communication; also see Maunder, 2002; Harley and Maunder, 2004 and Maunder and Harley, 2004).^[19] Monthly sea-surface temperature data were obtained from Rayner et al. (2003) for 5°N to 20°N between the coast of the Americas and 120°W to try to capture environmental influences.

Estimates of capacity outputs, allowing for variable returns to scale^[20], were made at the set and vessel level by mode of fishing (dolphin, unassociated or floating object). Data for yellowfin and bigeye tunas were combined to reduce the number of zero-valued observations of bigeye (which is troublesome to the operation of the DEA program). Output or retained catches in the analysis was specified by species and method of harvest per set as follows: (1) yellowfin and bigeye tuna caught in sets made on dolphins; (2) yellowfin and bigeye tuna caught in sets made on unassociated schools; (3) yellowfin and bigeye tuna caught in sets made on floating objects; (4) skipjack tuna caught in sets made on dolphins; (5) skipjack caught in sets made on unassociated schools; and (6) skipjack caught in sets made on floating objects. The retained catches of other fish were negligible, and hence not considered in the analysis. The analysis estimated capacity output for all six outputs and three types of fishing, specifying a common harvesting frontier (i.e. the DEA models were run with all six outputs at once, rather than separately for each of the three types of fishing). To be able to accurately estimate capacity output by individual vessel for each of the different types of fishing, each of the six outputs in the DEA model were specified as average landings per vessel per set per year.

Biomass estimates for yellowfin and skipjack were used to specify stock conditions, with sea-surface temperature used to account for environmental conditions. Both of these variables were specified as non-discretionary or fixed (constrained) inputs. The capital stock or capacity base of an individual vessel was captured by the GRT to allow for consistency with specifications for the other tuna purse-seine fisheries.

Although data were provided for 1980-2002, capacity output estimates were made only for 1998-2002. Limiting the analysis to the five most recent years captures more recent fleet configurations, cost conditions and fishing patterns, and also helps to control for the potential shifts in capacity output due to technical change. Limiting the number of years of analysis thus leaves differences in TE and variable input usage as the determinants of differences in observed output from capacity output (Färe, Grosskopf and Kokkelenberg, 1989). In addition, the technological-economic approach to capacity output is predicated on "normal practice" or "normal operating conditions" among the vessels, which is better given when the number of years is limited (cf. Corrado and Mattey, 1997).

Capacity output and TE were estimated separately for each of the following vessel size groupings: (i) classes 2 and 3 with 28 vessels; (ii) classes 4 and 5 with 43 vessels and (iii) class 6 with 188 vessels. There were no class-1 vessels in the data set. Classes 2 and 3 and classes 4 and 5 were combined to provide an acceptable minimum level of observations in each grouping.^[21] The full five years of data were available for only 50 vessels.

The technological-economic measure of capacity output specifies full utilization of variable inputs. However, estimates of TE by DEA were made using the number of sets per vessel by each type of fishing by year as the variable input.

Estimates of ray CU, in which deviations from full CU are due to either low variable input usage or technical inefficiency, are given by q in problem [1]. Estimates of ray CU purged for the effects of TE were given by the ratio q₂/q₁, where q₂ is derived from problem [1], allowing for variable inputs that are not necessarily fully utilized and q₁ is the q in problem [1] when variable inputs are fully utilized (Färe, Grosskopf and Kokkelenberg, 1989). Thus, estimates of ray CU purged for the effects of TE are due to low variable input usage. As noted above, we have attempted to control for deviations from full ray CU due to technical change in the later years by limiting the analysis to the last five years. We also attempted to control for deviations from full ray CU due to fluctuations in resource abundance and environmental conditions (which shift the capacity output frontier in or out) by specifying biomass and sea-surface temperature.

TABLE 3.1
Data used to estimate capacity for Class-2 and -3 vessels in the tuna purse-seine fishery of the EPO

Source: Inter-American Tropical Tuna Commission.
Note: There were no reported dolphin sets by Class-2 and -3 vessels in 1998, 1999, 2001 or 2002.

TABLE 3.2
Data used to estimate capacity for Class-4 and -5 vessels in the tuna purse-seine fishery of the EPO

Source: Inter-American Tropical Tuna Commission.
Note: There were no reported dolphin sets by Class-4 or -5 vessels between 1998 and 2001.

TABLE 3.3
Data used to estimate capacity for Class-6 vessels in the tuna purse-seine fishery of the EPO

Source: Inter-American Tropical Tuna Commission.

Annual capacity output on a per-set and per-vessel basis was estimated and subsequently converted to total annual fleet activity for each vessel size class by multiplying the per-vessel and per-set estimates of capacity output by the number of vessels and sets in each year for each vessel size class.

Technological change can also increase fishing capacity. To begin to evaluate the effects of technical change, we estimate a Malmquist index of technological change for the Class-6 vessels, which gives us balanced panel data set of nine years with total number of data for 128 vessels for all three set types. We estimate the Malmquist index DEA model with constant-returns-to-scale, which basically uses the same output-oriented DEA model as that in our capacity estimation, with number of sets as variable input and the interaction term of number of sets and gross weight of the vessel added as another input. This gives a flow measure of capital services for the vessel, engine and gear. Four CU rates (also called output distance in this methodology) are calculated. We provide annual year-to-year estimates and chain or cumulative indices over the nine years.

3.2 Results

3.2.1 Overall levels of capacity in the tuna purse-seine fishery of the eastern Pacific Ocean

The results of the analysis indicates that substantial excess fishing capacity, defined as fishing capacity output minus observed output (retained catches), when measured as: (1) potential catch minus actual catch or (2) potential catch, purged for TE, minus actual catch exists for:

• Skipjack for all vessel classes and set types utilised by the respective vessel class;

• Yellowfin and bigeye combined for all vessel classes and set types utilised by the respective vessel class.

In short, tuna purse-seine vessels had the capacity to catch substantially more of all species during 1998-2002 than they actually caught. The greatest contributor, by far, to excess capacity was Class-6 vessels, although there was excess capacity for Classes 2-3 and 4-5 vessels as well (Table 3.5). Excess capacity for all species combined, purged for TE, fluctuated from a minimum of 120 420 tonnes in 1998 to a maximum of 208 162 tonnes in 1999, dipping in 2000 and steadily rising to 193 199 tonnes in 2001 and to 196 178 tonnes in 2002 (Figures 3.5 and 3.6). Across all vessels it is estimated, after accounting for TE, that during 1998-2002 the combined catches of yellowfin and bigeye could have been 33 percent greater (Table 3.5, Figures 3.3 and 3.4) while those of skipjack could have been be 29 percent greater (Table 3.5, Figures 3.1 and 3.2).

The CU rates for all species combined also indicate substantial excess capacity, defined as capacity output minus observed output, regardless of whether TE is purged (Table 3.7). (CU is defined as observed output divided by capacity output. CU ranges from 0 to 1, where 0 indicates no observed output and 1 indicates that observed output equals capacity output.) The CU for Class 2-3 vessels, purging TE from capacity output, averaged 67 percent, i.e. on average a vessel caught about two-thirds of its potential catch. Across all Class 2-3 vessels it is estimated, after accounting for TE, that the combined catches of yellowfin and bigeye could have been 51 percent greater, while those of skipjack could have been 39 percent greater (Table 3.5). The CU for Class 4-5 vessels, purging TE from capacity output, averaged 72 percent; i.e. on average a vessel caught slightly less than three-quarters of its potential catch. Across all Class 4-5 vessels it is estimated, after accounting for TE combined, that the combined catches of yellowfin and bigeye could have been 10 percent greater, while those of skipjack could have been 28 percent greater (Table 3.5). The CU for Class-6 vessels, purging TE from capacity output, averaged 75 percent, i.e. on average a vessel caught about three-quarters of its potential catch. Across all Class-6 vessels it is estimated, after accounting for TE, that the combined catches of yellowfin and bigeye could have been 34 percent greater, while those of skipjack could have been 29 percent greater (Table 3.5).

Excess capacity exists for all vessel size classes combined for all set types for yellowfin and bigeye tuna when measured as either: (1) potential catch, purged for TE, minus actual catch (Table 3.5, Figure 3.3), or as (2) potential catch, purged for TE, minus the combined AMSYs for both yellowfin and bigeye (Tables 3.6, Figure 3.7). Excess capacity for yellowfin and bigeye tuna vis-à-vis their combined AMSY was relatively small in 1998 at 37 167 tonnes, i.e. capacity output, purged for TE, was 37 167 tonnes, or almost 11 percent more than the combined AMSYs. Capacity output, purged for TE, rose to 92 518 tonnes, or almost 27 percent more than the combined AMSY in 1999. In 2000, capacity output, purged for TE, decreased slightly to 89 704 tonnes, or almost 26 percent more than the combined AMSYs. By 2001, however, capacity output, purged for TE, rose to 210 915 tonnes, or almost 61 percent more than the combined AMSYs. In 2002, capacity output, purged for TE, rose to 241 835 tonnes, or almost 70 percent more than the combined AMSYs. In all cases, Class-6 vessels contributed the lion's share of the excess capacity.

In summary, by 2002 tuna purse-seine vessels had the capacity to harvest almost 70 percent more than the AMSYs for yellowfin and bigeye combined.

TABLE 3.4
Reported catch, estimated capacity and capacity purged for technical efficiency for the purse-seine fishery in the EPO.

Note: Actual output (retained catches in tonnes) from Inter-American Tropical Tuna Commission.

TABLE 3.5
Reported catch, estimated excess capacity and excess capacity purged for technical efficiency for the purse-seine fishery of the EPO

Notes: Excess capacity output is defined as capacity output less observed output (landings) in tonnes. Actual output (landings in tonnes) from Inter-American Tropical Tuna Commission.

3.2.2 The fishery by class-2 and -3 vessels

Potential catch exceeds actual catch for sets on unassociated schools and on floating objects for Class-2 and -3 vessels, i.e. there is excess capacity, regardless of whether capacity output is purged for TE (Table 3.4). (There were no dolphin sets for Class-2 or -3 vessels.) When TE is purged from capacity output for yellowfin and bigeye, this excess capacity is comparatively greater for sets on unassociated schools than for sets on floating objects, with an annual average about four times greater. Excess capacity for all set types for Class-2 and -3 vessels has been declined steadily during 1998-2002.

3.2.3 The fishery by class-4 and -5 vessels

Potential catch exceeds actual catch for sets on unassociated schools and on floating objects for Class-4 and -5 vessels, i.e. there is excess capacity, regardless of whether capacity output is purged for TE (Table 3.4). (There was a negligible number of dolphin sets for this size category.) When capacity output is purged for TE for yellowfin and bigeye, this excess capacity is comparatively greater for sets on unassociated schools than for sets on floating objects, with an annual average about three times greater. For skipjack, this excess capacity also averages three times greater for sets on unassociated schools than for sets on floating objects. Excess capacity over all set types averages about three times greater for sets on unassociated schools than for sets on floating objects. The trend for excess capacity for all set types has been roughly downward during 1998-2002, but with considerable variability.

3.2.4 The fishery by class-6 vessels

Potential catch exceeds actual catch for sets on unassociated schools and on floating objects for Class-6 vessels, i.e. there is excess capacity, regardless of whether capacity output is purged for TE (Table 3.4). When capacity output is purged for TE for yellowfin and bigeye, this excess capacity can be ranked by set type, from most to least excess capacity as: dolphin sets, sets on floating objects and unassociated school sets. For skipjack, this excess capacity is greatest for floating-object sets, intermediate for sets on unassociated schools sets and least for dolphin sets. Average excess capacity, purged for TE, is greatest for dolphin sets at 71 063 tonnes per year, intermediate for sets on floating objects at 61 462 tonnes per year and least for sets on unassociated schools at 23 009 tonnes per year. Excess capacity for all set types has been roughly upward over 1998-2002, but with considerable variability.

3.2.5 Technical change

Technical change on a cumulative basis for Class 6 vessels increased by about 60 percent for all set types, species during 1998-2002 (Figure 3.9). Thus "fishing power" or the state of technology increased considerably, and was an important factor in the exhibited increase in fishing capacity and excess capacity over this time period.

3.3 Summary and conclusions

Excess capacity for all species combined, defined as capacity output minus observed output (retained catches), exists for all vessel size classes individually and combined for all set types (dolphin, unassociated, floating objects) for yellowfin and bigeye tuna when measured as: (1) potential catch minus actual catch or (2) potential catch, purged for TE, minus actual catch. Excess capacity, purged for TE, for all vessel size classes has increased from about 120 000 tonnes in 1998 to close to 200 000 tonnes in 2002, an increase approaching 63 percent in five years. The largest contributor, by far, to excess capacity was Class-6 vessels, although there was excess capacity for Classes 2-3 and 4-5 vessels.

TABLE 3.6
Excess capacity for yellowfin and bigeye: capacity output purged for technical efficiency minus combined average maximum sustainable yield of yellowfin and bigeye for all vessels in the EPO

Notes: Excess capacity output is defined as capacity output, purged for technical efficiency, less combined AMSY for yellowfin and bigeye in tonnes. AMSYs from Inter-American Tropical Tuna Commission.

TABLE 3.7
Average vessel capacity utilisation and technical efficiency by vessel class

Notes: a. Output-oriented technical efficiency for a vessel size class is measured relative to that vessel size class's own vessels' best-practice production frontier. Vessel size, biomass and sea-surface temperature are held fixed.

Excess capacity exists for all vessel size classes combined for all set types for yellowfin and bigeye tuna when measured as either: (1) potential catch, purged for TE, minus actual catch, or as (2) potential catch, purged for TE, minus the combined AMSYs for both yellowfin and bigeye.

For yellowfin and bigeye, combining over all set types and vessel size classes, excess capacity (defined as capacity output, purged for TE, minus combined AMSY) climbed from an excess of about 11 percent in 1998 to an excess of almost 70 percent by 2002. In all cases, Class-6 vessels contributed the lion's share of the excess capacity.

Technical change on a cumulative basis increased by about 60 percent for all set types, species and vessel size classes during 1998-2002. Thus "fishing power" or the state of technology increased considerably, and was an important factor in the exhibited increase in fishing capacity and excess capacity over this period.

In short, there is considerable excess capacity, whether measured relative to existing catches or AMSY. There is also considerable technical inefficiency and considerable increases in "fishing power" or the state of technology due to technical change, which, in turn, is an important factor in increases in fishing capacity.