Previous Page Table of Contents Next Page


6. ADDITIONAL CONSIDERATIONS


6.1 Distinction between technical and economic measures

The methods of output-based capacity measurement outlined in Section 4 were developed from a technological perspective. Although based on production theory, the measures have little direct economic content. Formally, technological-engineering measures are defined according to the maximum possible output that could be produced per unit of time (a year or season), with existing plant (the vessel and its characteristics) and equipment (gear or nets), and under customary and usual practices, provided the availability of variable factors of production (e.g. labour and fuel) are not restricted. Such measures thus represent maximum physical output levels that a vessel or operating unit could produce, regardless of input and output prices. Some economic content is implicitly accommodated in what have been called technological-economic measures that do not impute capacity output measures beyond the scope of the observed data, and thus are restricted by observed behaviour. To some extent, the technological-economic measures implicitly reflect economic responses (i.e. landings were actually determined by fishers in accordance with underlying behavioural objectives). However, technological measures are only rough approximations of an economic measure of capacity, since they are not explicitly linked to any behaviour from economic incentives.

Different cost structures within a fishery could result in measured potential (technical) output levels for some vessels being inconsistent with an operator’s objective of profit maximization. Whether this is a problem, however, depends on whether profit maximization may be considered a relevant goal for the operator. For many fisheries it may not be reasonable to think that such a goal is the main driving factor behind behaviour. It remains true, however, that some measured potential levels of output might not be economically feasible, as the added cost of harvesting the catch would exceed the additional revenue gained. This may not be well reflected by the technological-economic approach. Alternatively, cost minimization may not be the behavioural objective that best characterizes decision-making behaviour, and approaches for determining capacity based on cost minimization may be inappropriate to use estimating capacity output. Also, even if cost minimization or profit maximization may not be perceived as a relevant goal, in a multiple species setting with possible choice among species, revenue maximization may be a relevant economic goal to build into the analysis. Färe, Grosskopf and Kirkley (2000) demonstrate how a cost- or revenue-based approach may be used to estimate capacity when production involves multiple outputs.

Economic indicators, or at least imputation of the implied costs of production decisions, are even more important to represent at the fleet level, since generation of rents from reduced capacity is one of the goals of fishery management. Therefore, even if individual fishers do not have explicit economic incentives, such goals may be relevant to managers. Also, if regulatory schemes in place may impose upon property rights, thus causing fishers to operate more clearly in response to economic motivations, such behaviour is important to impute for appropriate measurement of capacity and capacity utilization.

The economic concept of capacity output is the output level (nominal catch or landings) determined in accordance with a given behavioural objective (e.g. profit maximization, cost minimization, or revenue maximization) by a fishing unit operating under customary and normal operating conditions. The economic measure is distinguished from the technological-economic measure in that it explicitly determines the economically optimal output or input levels consistent with optimizing behaviour of fishing units or operators. Provided adequate data on costs and earnings are available, such economic capacity measures may be calculated in several ways. For example, one very crude measure sometimes suggested is to determine the output level corresponding to minimum average cost. This would require a sufficiently long-time series of costs and landings data to compare unit costs across boats and time.

Although such a measure could provide useful information for fisheries managers, it does not directly address the issue of associated capacity utilization. One type of cost-oriented approach, however, may be drawn from the technological economic framework. If an input-oriented approach is used, the implied reduction of capital from existing K levels that would still support production of observed catch levels is the focus of the analysis. If the associated capital costs associated with this contraction in the fleet can be estimated, the implied reduction in costs for a given output (and thus revenue) level may be imputed. This is not directly related to economic optimization but allows implicit consideration of potentially reduced costs associated with contractions of the fleet. With heterogeneous capital stocks, however, and thus potential ambiguities associated with what form a capacity reduction programme would take, construction of such measures is difficult. This suggests, in turn, that an appropriate “price” of capital to use for cost-based economic models might be similarly difficult to compute.

A more direct approach involves the use of economic optimization models based on cost or profit functions, using DEA or SPF methods. With information on input and output prices, an economic measurement consistent with cost minimization, revenue maximization, or profit maximization may be calculated by imputing the least cost, fixed input level for production of observed output levels. A capacity utilization measurement then can be constructed in terms of the additional costs that are unnecessarily incurred in the fishery by non-optimal fixed input levels. And the deviation of the fixed input levels may in turn be imputed from this cost gap. Although this method has rarely been applied to DEA or SPF methods, such models in a standard economic framework have been specified and implemented in studies such as Morrison (1985) and Färe, Grosskopf and Kirkley (2000). It should again be emphasized, however, that using this type of modelling framework requires one to assume that the relevant behavioural objective of boat operators is the economic one of, say, cost minimization.

An advantage of using an economic-based approach to capacity measurement is that potential economic waste in fisheries may be identified. Excess capacity can be measured not just in terms of changes in the quantity of catch, or more relevantly in the level of inputs, but also in terms of foregone economic profits. Difficulties that persist, however, involve the existence (and appropriateness) of cost and price data; the relevance of economic behavioural assumptions in fisheries that remain subject to common property motivations; and the nature of existing management and regulation.

Also, estimation of economic measures of capacity and capacity utilization requires significant economic data, and these data are generally not available.[43] It is therefore unrealistic to require states to produce such estimates for the purpose of international comparisons. Ultimately, it will be important for states to develop such measures for managing their fishing capacity, if the economic ramifications of excess capacity are driving management decisions.

6.2 Aggregation across species and fleet segments

Estimation of capacity output or catch for fisheries is best carried out at low levels of aggregation - for example, the boat level - for a particularly fishery. Once capacity utilization indicators have been estimated for the boats and species in individual fisheries, however, this information must be aggregated by various dimensions - such as boats, gear, species, fisheries and regions - to provide useful information about excess capacity over entire fisheries, or even countries. Such aggregation does not, however, have a strong theoretical basis unless production is fully and linearly additive (Daal and Markies, 1984). That is to say, the individual components of the overall aggregate are essentially independent from one another, and thus can simply be added together.

The basic problem is this: How does one use estimates at the firm or operating unit level to obtain estimates of the fleet, fishery, or industry. Daal and Merkies (1984) suggest that realistic and consistent aggregation is nearly impossible. Moreover, in the presence of technological externalities, consistent aggregation is not possible. This is likely to be the case for many fisheries. Kirkley et al. (2001), Färe and Zelenyuk (2001) and Färe, Grosskopf and Zelenyuk (2001) provide a comprehensive theoretical framework and discuss aggregation over firms to obtain a measure of industry efficiency or capacity. They demonstrate that industry capacity is greater than or equal to the sum of firm level capacity. In contrast, Blackorby and Russell (1999, p. 7-8) state “...there does not exist a technology set such that the widely used Debreu (1951)/Farrell (1957) measure of technical efficiency can be aggregated.” The three previously cited works, however, adopt Koopmans (1957) concept of efficiency and propose the use of directional distance functions to examine technical efficiency and capacity.

Aggregation of output-based measures of capacity becomes increasingly less definitive at higher levels, since comparability is lost. For example, it is less problematic across fisheries and between countries that harvest a shared stock for a given species, such as the cod stock in the North Sea. In this scenario, capacity output can be derived from the addition of such output of cod from each country participating in the common-pool fishery. This, however, provides only a rough approximation, which would underestimate total capacity. The sum of individual capacity estimates would be underestimated, because it does not allow for allocation of inputs among different operating units (e.g. allocating labour or days from one vessel in a given fishery to another vessel in a different fishery). This will be even more true when adding across species, particularly if capacity output measures impute the potential output from latent capacity, and thus, possibly double-count boats that are currently operating in different fisheries. Also, with diminishing returns, increased exploitation of a shared stock by all participants would result in a less than proportional increase in output because the stock is limited. However, since output measures are typically used indirectly to impute required capacity or capital contraction to produce desired catch levels, rather than as an indicator of what would happen if capacity were actually unleashed on the fishery, this is unlikely to be a binding constraint in practice.

With several fleet segments catching different combinations of species, the problem of aggregation becomes even more complex. One possibility is to use techniques such as DEA or SPF to estimate the capacity output of each species per fleet segment separately in a multiple species fishery. These can be aggregated across fleet segments for individual species as indicated in Table 3, where the X’s represent the capacity output of a given species of a given fleet segment. An example of aggregation of species across fleet segments is also provided in Appendix C. It is preferable, however, to recognize multispecies issues more directly, at least within a particular fishery, by using DEA or SPF models that recognize technical and economic interactions among the various outputs produced (e.g. how the catch of one species increases or decreases as the catch of another species increases or decreases). That is, the estimation may be performed for a multi-output, or multispecies, production technology that accommodates at least some forms of jointness (more than one species or product is produced for a given level of fishing effort) that are ignored when potential output from separate estimations are simply added together. Also, if revenue rather than quantity is the focus of the analysis, and estimation of multispecies fishery capacity output is carried out directly rather than simply added, it also may be useful to recognize the economic motivations underlying different catch compositions. This can be accomplished by postulating revenue rather than output maximization as a basis for capacity output measure.

Table 3 - Interactions between fleet segments and species in multispecies/fleet fisheries

Fleet segment

Species


1

2

3

4

A

X

X



B


X

X


C

X


X


D




X

Total fishery

A1+C1

A2+B2

B3+C3

D4

Deriving overall output-based measures of capacity utilization and excess capacity at higher aggregation levels, such as for a country will inevitably require some form of aggregation across species, gear and region, since estimation cannot justifiably be carried out at such an aggregated level. The simplest approach is to add up the quantities of different fish stocks. However, for most purposes it will be more informative to weight this sum in some manner, such as weighing the output of each species by its price to produce a total value of output (Gross Value of Production). Note also that in order to impute capacity utilization measures from such an aggregation process, target output measures of capacity used as comparison points for CU measures also must be aggregated using price weights.

A potential interpretation difficulty for measures added according to their value is that excess capacity measures can vary with a change in relative prices, all other things being equal. This is particularly problematic when examining trends in capacity and comparing capacity measures between years. For example, an apparent decline in total excess capacity over time may be a result of a decrease in price for a species that is subject to equivalent or even greater levels of excess capacity than in the previous year. To limit this, a constant set of prices could be applied to a given time series of output values for purposes of international comparisons. However, this raises questions about how market mechanisms and true capacity output are linked and how to distinguish their effects. This, in turn, suggests that evaluation of capacity output in such terms is questionable. In summary, aggregation or even comparison of capacity output measures across fisheries is difficult to accomplish effectively and should be undertaken with care.

Aggregation of many input-based measures of capacity also can be undertaken, although it again raises difficulties of interpretation. For example, total gross tonnage or total kW days fished can be aggregated across all fleet segments, as can their target levels. However, the more variation there is across boats, the more this measure is questionable, since it implies that a “representative” boat can be defined and the relationship between inputs and output harvesting capacity is linear. Similarly, inaccuracies in the aggregate measure may increase at higher levels of aggregation due to incompatibility of effort units. For example, the importance of kW days is greater for fleet segments using mobile gear (e.g. trawl gear) than for fleet segments using static gear (e.g. pots). It also precludes the incorporation of activity in fisheries that are not based on readily measurable physical inputs (e.g. labour rather than capital intensive fisheries). And when only measures such as boat numbers are available, a mixture of large and small boats in the population will create a bias in the estimate of total capacity (most likely an overestimate, because one large boat may be equivalent to several small boats in harvesting capacity). Consequently, any aggregation of input-based capacity measures should be viewed with substantial caution.

Despite the problems associated with aggregation, such information is important for providing a general indication of the order of magnitude of capacity utilization in a fisheries sector. Computing an indicator of total capacity utilization for all fleet segments that are harvesting a given species or stock provides a useful, albeit approximate, indication of the magnitude of balance or imbalance that exists between fishing capacity and the overall resource.

At the international level, aggregation could potentially be undertaken between those countries harvesting shared international, transboundary, highly migratory and straddling fish stocks, although again the aggregation is problematic and should be undertaken with care. The purpose of this exercise would be to provide information to the appropriate, regional fisheries management organization (RFMO) about the potential risks that a combined national fleet capacity prosecuting these shared fish resources may present for the short- and long-term conservation of such stocks. In this context, RFMO officials would have an opportunity to consider the implications of the mobility of certain countries and/or fleet segments across species and/or national lines and to discuss any policies or measures that may eventually be considered to manage such fleet mobility (FAO, 2000). For such aggregation to work, the countries involved will need to coordinate their data collection and capacity estimation approaches to ensure that compatible measures are developed.

6.3 Artisanal fleets

In many countries, the artisanal sector is often not adequately incorporated into fisheries management plans and measures, despite its importance. In many developing countries, attention is focused on the development of mechanized and/or commercial fisheries, with traditional and subsistence fisheries often incorrectly regarded as being insignificant. Even in countries that have relatively advanced fisheries management systems, such as the United Kingdom, the level of information collected on the small boat sector (under ten s in length) is negligible, even though these boats comprise almost two thirds of the entire fleet.

Three main types of artisanal fleets/fishers can be defined as: pure subsistence fishers, part-time commercial fishers, and full-time commercial fishers. The capital used by these fishers may not involve a vessel but, instead, may take the form of fishing gear or even labour. In such cases, the most appropriate inputs should be used to define fishing units in subsequent analyses.

For pure subsistence fisheries, the concepts of capacity utilization and excess capacity as defined in previous sections of the guidelines are not necessarily meaningful. This sector catches only what is needed and, while they could catch more, by definition they do not catch more than is required for food or subsistence purposes. As a result, it is not clear that they behave in the same optimizing manner as commercial fishers (e.g. who maximize their outputs given fixed inputs, or minimize their costs to achieve a desired catch). They may, of course, have other optimizing behaviour), and hence, the analytical methods such as SPF and DEA may not be appropriate. For this user group, rather than operating according to a strict, firm level objective, individuals may be more concerned about satisfying or maximizing utility subject to various constraints. Similar problems are likely to exist when attempting to assess capacity in recreational fisheries.

For purposes of defining and measuring capacity in subsistence fisheries, the current catch levels can be considered to be the current output capacity, because, by definition, this is the maximum catch that will be taken under normal operating conditions. Furthermore, because most subsistence fisheries interact with commercial or industrial fisheries to some extent, the ability of their fishing activity to expand is limited.

Small-scale fishing in many countries is also associated with part-time farming (or other activities). Hence, when conditions are not favourable for farming, fishing activity may increase. In such cases, the potential capacity of this group should be considered in the same manner as full-time fishing units. This will result in these fleet segments demonstrating substantial latent effort and capacity underutilization. This needs to be considered when assessing the overall level of overcapacity in the fishery.

Small-scale commercial fishing units operating on a full-time basis need to be assessed in the same manner as their larger counterparts in the measurement of fishing capacity. However, data related to this sector are often poor or non-existent. As a result, the available approaches may be limited and resulting estimates, subject to some uncertainty. This may present problems when aggregating capacity measures at the national or regional level, particularly if output-based measures of capacity by species are not available.

6.4 Pelagic and highly variable fisheries

Many pelagic fisheries are subject to large inter-annual variations in catch, because stock size is highly dependent upon spawning success and subsequent recruitment, both of which are highly susceptible to variations in environmental conditions (e.g. food availability and water temperature). This represents an extreme example of the general issue of short run fluctuations in stocks that generate output changes that should not be attributed to capacity changes.

Without some measure of stock that can be used to control for these fluctuations, estimates of capacity output derived through peak-to-peak, SPF, or DEA analysis with a single series of fleet level data will be largely influenced by the years in which the fish stocks were either highly abundant or very dense.[44] Where panel data (time series of individual vessel level data) are available, dummy variables can be used in the SPF approach to try to capture the effects of such stock fluctuations on output. Similarly, treating time (and, implicitly, stock size) as a categorical variable and estimating capacity output in each separate time period will reduce potential distortions when using DEA.

Ideally, some measure of stock or resource density can be directly incorporated into the analysis as a fixed input into the production process. In such a case, the resulting estimates of capacity output would be more representative of the real value. When using DEA, the stock needs to be treated as a non-discretionary input. (See Cooper, Seiford and Tone, 2000.) In actuality, however, stock or resource conditions represent disembodied technical change (i.e. technical progress that is generally beyond the control of the vessel operator).

The issue of short-run fluctuations is particularly a problem when imputing long-run measures from short run evidence; for example, when comparing current capacity output measures with target catch estimates such as MSY. As noted above, target measures are based on long run equilibrium values of output, and implicitly a stable (or average) stock size, whereas usual capacity output estimates are based on current stock size. If comparison is carried out using these types of measures, it is particularly likely that a fishery may be perceived as not having overcapacity in a “poor” recruitment year, because capacity output is less than (average) target capacity. However, the level of inputs employed may be greater than that which would be expected to produce the target capacity under “normal” or average conditions. Conversely, a fishery may be perceived as having substantial overcapacity in a “good” year when capacity output exceeds target capacity, but the level of inputs may be less than or equal to the level associated with target capacity under average conditions.

In such highly fluctuating fisheries, controlling for stock levels and for long-run comparisons that impute capacity output levels at target rather than current biomass stock levels is key to constructing interpretable and useful measures. For short run comparisons, if a bio-economic model of the fishery is available, optimal yields given current stock conditions can be estimated to provide a short run measure of target capacity for comparative purposes. Also, directly constructing input-oriented measures of capacity could bypass some of these issues if an estimate of optimal input use at (average) target capacity output can be derived.

6.5 Processing and hold capacity

Both onshore and onboard processing can affect the measurement of fishing capacity. Onboard processing can act as a constraint to vessel production. That is, some of the input base is used for processing purposes rather than catch, so if variations in output composition are not taken into account, the link between measured input and output is misrepresented. Also, processing facilities are only able to process a given quantity of catch-per-unit of time, and thus, onboard processing activities may actually determine harvesting capacity levels. While boats could potentially catch more, they are unable to process this catch, so there is no economic incentive to continue fishing beyond the ability of the boat to process the catch. This can be incorporated into DEA analysis directly as a technical constraint, provided information is available. Alternatively, vessel capacity could be estimated by examining onboard processing capacity. Where information on output composition or the processing constraint is not available, measures of capacity output are likely to be over-estimated.

Similarly, constraints in onboard freezer and storage capacity also will restrict the level of potential catch. Once full, the boat must return to port to unload, even though higher catches could be achieved by continuing to fish. In most cases, these constraints will be implicitly incorporated in the analysis as boats of similar sizes would be expected to have similar hold or freezer constraints, and hence, returning to port would be part of the normal working practice reflected in their effort data (e.g. days at sea). This issue can be accommodated if data on this capital characteristic are available and used as part of the measure of K.

In addition, in some cases, onshore processing may impose limits on the quantity of catch that can be utilized upon landing. In such cases, processing acts as a constraint to the total capacity of the fleet. The maximum output for all boats may be negatively impacted, and hence, the resulting estimate of capacity reflects the processing rather than the harvesting capacity. Given that an objective of capacity management is to ensure that harvesting capacity is commensurate with reproductive capacity of the resource, an unconstrained estimate of output capacity is required. Where such a constraint can be identified and the carrying capacity measured, it can be incorporated into the analysis, much like onboard processing capacity. The constraint can then be relaxed to provide a more appropriate measure of fleet capacity.

It is also the case that when fishing vessels and processors are vertically integrated, production decisions are made on the basis of the value of the final product. As a result, cross-subsidization may occur between the processing activity and the harvesting activity in order to maximize overall profits. Consequently, fleet activity may not be consistent with the assumptions underlying the main techniques used to assess capacity. In such a case, estimated capacity output may not reflect the potential output of the fishery, unless output compositional and values are taken into account. This requires both the separate measurement of different types of product, and the assumption of maximum revenue rather than output as the retained assumption about behavioural motivations. If this is not possible, the bias in the measure cannot readily be identified; the measures may be either over- or under-estimated depending on the extent of cross-subsidization in the fishery.

6.6 Other factors that may affect the measure of capacity

Other factors that also may distort the measurement of capacity include quality and discards. Haul reduction may increase the quality of the landed fish and result in higher market prices. Where a fishery consists of a mix of fishers, some of whom aim to land a lesser quantity of high-quality catch and others, to land a higher quantity catch, the capacity output will be defined by the latter rather than the former. As a result, boats landing the higher quality catch will be perceived as operating at less than full capacity. One way to deal with this issue, as alluded to above for other types of output compositional issues, is to weight the measure of the catch by their prices and construct the estimating framework according to revenue instead of output maximization. Alternatively, if boats that aim to land higher quality but small quantity catches can be identified, constraints on their capacity output can be incorporated directly into the analysis, or the boats can be analyzed separately (so the quality aspect is treated as a categorical variable). This may only be accomplished, however, if individual vessel level data are available, and the harvesting strategies of the individual vessels can be identified. The latter concern, in particular, requires information on individual boats that is not likely to be readily available.

Estimates of output capacity are generally based on estimates of the landed catch rather than the total catch, as often only the former is recorded. In many fisheries, particularly those subject to quota controls, part of the catch of some species is discarded. The effect of this is that both actual and capacity outputs are under-estimated.

Where discard data are available, they can be incorporated into the analysis (i.e. added to the landed catch to provide an estimate of total catch). It may also be possible in this case to take discards or by-catch into account as a negative output, particularly if discards stem from restrictions on catch for controlled species. However, it is unlikely that such information is available at the individual boat level for all boats in the fishery. If estimates of discards are available at the fleet level (e.g. derived from a discard sampling programme), some adaptation to estimated capacity output levels may be possible. For example, if it is assumed that discards are proportional to catch, then discarding at the capacity output level can be estimated by dividing the current discard estimate by the measure of capacity utilization.


[43] Several European Union Nations and the United States have begun collecting detailed economic data on costs and earnings. The eventual availability of detailed economic data, thus, warrants that various economic concepts of capacity be estimated. Methods and procedures for these approaches are available in Morrison (1985a, b), Fousekis and Stefanou (1996), Keeler and Ying (1996), Fagnart, Licandro and Portier (1999), Färe, Grosslopfand Kirkley (2000), and Coelli, Grifell-Tatje and Perelman (2001). There are also several problems with estimating a stochastic cost or profit function (see Kumbhakar and Lovell (2000) for a comprehensive discussion on estimating stochastic cost, revenue and profit functions).
[44] In addition, some pelagic species, such as tuna, are often prosecuted by complex fishing gear and technology. For example, purse vessels are typically deployed from a mother ship in response to aerial-based descriptions of stocks. In other cases, vessels use dolphin feeding behavior to identify schools of tuna, speedboats to help herd the dolphin, and divers to free the dolphin. In addition, the inputs typically considered for many fisheries may not be appropriate indicators of capacity. For such fisheries, it will be a challenge to estimate capacity output. Squires (pers. comm.) is presently estimating capacity for several Pacific tuna fisheries. The problem may become more complicated because of the need to treat undesirable outputs (e.g. dolphin).

Previous Page Top of Page Next Page