The Bio-Economic Model of European Fleets (BEMEF) analyses the socio-economic impacts of alternative states of the EU fishing fleet. It is a multi-period static equilibrium model, using fish stock assessments and total allowable catches/yields, along with updated datasets on key drivers as exogenous variables to estimate the performance of fleets in future time periods for which data is not wholly available (2015, 2016, 2017) and alternative static scenarios, including a state of long-term MSY.
The purpose of BEMEF, as explained in full in the background section is to:
The appropriate level of complexity is a major issue that bio-economic modelling struggle with. If modelling is too simple it may not be representative and is unlikely to produce relevant or practical results for management. Alternatively, if modelling is too complex it may not produce a workable representation and is difficult to calibrate. In between these extremes is an area of intermediate complexity where a model can address real problems but remain grounded in reality. This is referred to as the Medawar Zone after Sir Peter Medawar who hypothesised that there is a period in time when investigating a scientific question is particularly productive. Reviews of existing models confirm this relationship (Costanza & Sklar, 1985); (Grimm et al., 2005).
BEMEF is not the only bio-economic model of EU fisheries and each model comes with its own advantages and disadvantages. Many of these models have been summarised in overview reports (Prellezo et al., 2009) and are the subject of the ICES Working Group on Integrated Ecological and Economic Models (WGIMM), of which BEMEF is a part of. Whereas BEMEF has a wide scope to give a general picture, some models of European fisheries, in particular Fishrent are much better at modelling dynamic behavioural change and transitional dynamics. In this sense, BEMEF can be characterised as an economic simulation model analysing an exogenous change on a stable fleet structure, rather than an optimisation model, which can answer questions about “best”. Regarding the time dimension, an elaborated behavioural component and stock growth dynamics (such as FLR) may be integrated with BEMEF future stage, as explained in Works in progress at the end of this documentation.
A bio-economic model is not the only way to analyse socio-economic outcomes in fisheries management. A triple bottom line approach (Voss et al., 2014) or multi-criteria decision analysis (Rossetto et al., 2014) seem particularly well placed to compare trade-offs in fisheries management. Simplified economic analysis at a macro level can illustrate the problem of overfishing in economic terms and attempt to quantify the potential gains from sustainable management (Quaas et al., 2012; Salz, 2012).
Thank you to the following researchers for their important contributions to BEMEF: Thomas Thøgersen, Hans Frost, Jordi Guillen, Gorka Merino, Rainer Froese, Sandra Bernick, and to the many stakeholders and experts who have volunteered their time and comments over the past three years.
Thank you to the following conferences for hosting presentations on BEMEF: the International Council for Exploration of the Sea (ICES) Annual Science Conference 2014, the UK Network of Environmental Economists (UKNEE) Applied Environmental Economics Conference 2015, the 2015 Conference of the European Association of Fisheries Economists (EAFE), and the European Society for Ecological Economics (ESEE) 2015 Conference.
All errors and responsibilities belong to NEF.
This section on model documentation begins with a description of how landings are calculated, followed by revenues, then costs, and finally socio-economic indicators. Using this diagrammatic representation of BEMEF, the documentation proceeds through the middle branch, left branch, right branch, and then overlap. Work in progress and BEMEF-powered tools external to the model are noted at the end of the documentation.
First, it’s important to recognise the existing data and work BEMEF has been built upon. In particular, the Annual Economic Report and the EIAA model.
The Annual Economic Report of the EU fishing fleet and the associated database are produced from data calls to EU Member States. This information covers landings, effort, and economic variables at the fleet segment level. These fleet segments, defined through Data Collection Framework, do not always align with fleet segmentation that Member States use in their own national reporting, which is often at a more disaggregated level. BEMEF can be seen as a model built specifically around the available information in the AER database.
The AER is updated annually, but has a two-year time lag in terms of what data is available, with partial data for some Member States with a one-year time lag. The associated database contains hundreds of thousands of rows of data but is still incomplete in its coverage. As data is collected at the national level, there is some variance in the coverage of the information provided. As the model is largely driven by changes in TAC, and for simply clarity, only fleets in the Northeast Atlantic (Area 27) are included (which also cover the Azores). In total this covers 235 fleets from 15 Member States. For comparison, the impact assessment carried out for the Common Fisheries Policy analysed 50 fleets due to data limitations.
A modelling feature of BEMEF is that data from the AER database is automatically referenced in the model. This enables relatively quick calculation between fleets and the ability to calculate an ad hoc change in a model parameter across all fleets. Running the model in full across XXX fleets typically takes 30 minutes. The referencing system in the model also allows for easy year-on-year updates when the new AER dataset becomes available.
In the downloadable version of BEMEF, the AER database is provided as a worksheet and can be filtered to find a fleet of interest or to list and rank fleets by a certain parameter.
In order to avoid duplication of existing work, the EIAA bio-economic model (Economic Interpretation of ACFM Advice), originally designed and implemented under the “Promotion of Common Methods for Economic Assessment of EU Fisheries”, has formed the model backbone for BEMEF. The EIAA has been used historically in STECF working groups for the Annual Economic Report of the EU fishing fleet (Anderson & Carvalho, 2013) and in measuring the economic implications of TAC advice ( STECF, 2004 ). A modified version of the EIAA model was also used for the socio-economic Impact Assessment of recent reforms to the Common Fisheries Policy (Agnew et al, 2010).
The EIAA model has been well-documented, and advanced users of BEMEF will want to read The EIAA Model: Methodology, Definitions and Model Outline (pdf) by Hans Frost, Jesper Levring Andersen, Ayoe Hoff and Thomas Thøgersen.
The fishing industry is fairly unique in that there is a hard cap on the output of production through the management tool of total allowable catches (TACs or quotas). As such, the BEMEF model structure is ‘backwards’ by first setting the TAC, then determining landings, determining the amount of effort needed to harvest those landings, and finally the associated costs and revenues.
Using TACs as an exogenous variable, the following equation is used to determine landings for a fleet:
Lj,t,k = FSSj,k * RUj,k * RSc,k * Qt,k + NQLt,j
The components of this equation are explained below.
Fleet segment share and quota allocation
The default assumption in BEMEF is that quota is allocated by Member States to their fleets in a way that largely aligns with historic landings reported in the base period. At a policy level, many Member States do use historic track records in quota allocation, but there are alternative methods that could be used. Article 17 of the reformed Common Fisheries Policy requires Member States to consider socio-economics in the distribution of fishing opportunities, including quota.
To illustrate alternative states of the EU fleet under different quota allocation systems, BEMEF allows quota allocation along other considerations. As the Annual Economic Report dataset reports a number of metrics by fleet segment, four alternative criteria were derived from this dataset:
In these alternative quota allocation scenarios, fleets have their performance on the scenario criterion over the 2012–14 base period. Then, within a country and for a certain species and area, quota is allocated proportionately based on this recorded performance.
As fleet performance can not be broken down by species or TAC (ie. how much fuel was used to fish a particular species) the fleet’s whole performance is used. This makes measuring performance on a specific TAC impossible, but is a simple fact of fleets catching more than one species.
It is also important to note that these quota allocation scenarios are a mix of current fleets (vessel numbers, cost structure) but alternative quota allocation levels. In reality, there were be a process of dynamic adaptation to a policy change in quota allocation. Results are therefore only illustrative as a demonstration.
A couple constraints on quota allocation are used to more appropriately deal with changes in capacity. This is accomplished by calculating where fleets historically fish, and what species they land at port. Only fleets that currently fish within the same ICES area are compared on the different criteria used. Fleets that land less than 100 kg of a species were not considered for this comparison. The consequence of these constraints is that quotas in distant waters are only available to vessels that currently operate in this fishery, although the ICES areas are very large and analysis could be improved by comparing fleets at the level of ICES rectangles. Lastly, only the quota from fleets that are covered in the model is available for allocation to prevent the model results over and underestimating due to the fleets covered.
The outcomes of these quota allocation scenarios formed part of the report Managing EU Fisheries in the Public Interest (Carpenter & Esteban., 2015).
Not all of the quota that is allocated to a fleet will be landed in a given year and bio-economic modelling should account for this likely outcome. If the model is run where all quota is caught (a parameter in the model allows for this) then the results will overestimate the expected landings and other outcomes in future periods - a potential issue with previous studies on fishing at a state of MSY.
As such, the default parameter value is to use historical uptake. This level of uptake is calculated for each TAC and for each Member State using reported landings from the AER database and comparing this to the amount of quota allocated to a country through relative stability. The calculation occurs at the country level rather than at the fleet level because fleet shares are calculated based on landings adding no further detail to the results.
RUj,k = L0,j,k / FSS0,j,k * Q0,k
This calculation is complicated by the fact that there is quota trading that takes place. There is also some inevitable double counting that takes place (adjusted in landings but not elsewhere)
The alternative is to calculate uptake at the species level for a country, rather than for each TAC. This will remove the boundary issue as the overlap between TACs occurs for a species. The drawback of this approach is that one uptake percentage will be applied across different TACs while there may be a large degree of variance between levels of uptake. A trial of this alternative approach shows that the modelled results for the base period were further from the reported data. One additional issue is that there are areas where quota species can be caught without quota. This issue can be resolved once there is a complete matching between the AER database (for landings) and quota areas. Currently BEMEF estimates total landings without estimates for specific quota areas.
One key element as to how quota is allocated is the relative stability matrix. This matrix lists the fixed percentage of quota species that goes to each country as a way of avoiding conflict between countries and ensuring some stability within the industry. The main criteria used in determining the fixed percentages is landings data over the reference period of 1973-1978.
Similar to realised uptake, the model calculates how actual landings occurred over the base period so that the user can set landings using relative stability or the measured landings.
TACs and other fishing opportunities are set through a multi-stage process involving a number of several different actors institutions . Every year, Member States gather and submit data and research on fish stocks which is pooled in an international dataset. The International Council for the Exploration of the Seas (ICES), an intergovernmental scientific body, subsequently carries out annual stock assessments which it formulates into scientific advice: recommended levels of fishing to achieve MSY fisheries and the CFP’s objectives. This reviewed by an advisory committee before being passed on the Commission. On advice of the Scientific Technical and Economic Committee (STECF), the Commission formulates proposals on TAC levels in early the early Autumn. In October through November and December of each year, the Council, which has full legal discretion, sets the TACs and other fishing opportunities for the following year in a Regulation. The TACs are set for 36 commercially fished species in different fishing areas around EU waters - over 200 TACs in total.
The agreed TACs set at Council are the default parameter values in BEMEF. There are 25 species included that cover 150 TACs. Alternative scenarios for 2015, 2016 or 2017 can be modelled to estimate the in-year economic impacts of the ICES advice on TACs or the TACs proposed by the Commission. In instances where quota advice is not provided by ICES or the Commission, the actual Council TACs are used to fill those gaps. It is important to note that these are in-year impacts, and as such there is always by definition a higher economic benefit the higher the TAC (and vice versa).
There has been interest in using this quota-input function of BEMEF to complete an economic analysis of a quota proposal (STECF, 2014).
Gross revenue (total revenue) in future periods is calculated based on the computed future landings and prices.
Non-quota landed value is assumed to remain equal to va
NQLVt,j = NQLt,j * P t,j
Income from landings
Income from landings scales proportionately to landed value. For most fleet segments the values are almost identical.
LIt,j = LI0,j * (LVt,j/LV0,j)
TRt,j = (∑i Pt,i,j * Lt,i,j + Kt,j) * [GR0,j / ∑i P0,i,j * L0,i,j + K0,j)]
Where the landed value of non-quota species for fleet segment j in year t is defined as:
Kt,j = TR0,j - ∑i P0,i,j * L0,i,j
And where gross revenue including non-fisheries specific income of segment j is defined as:
GR0,j = TR0,j + O0,j
Fish prices & fish price flexibility
As the quantity of fish supplied to a market changes (whether local, national or international), so too will the price of fish. This price flexibility (the inverse of a price flexibility) tend to follow an inverse demand model with a decrease in supply leading to an increase in price.
The reference rates for a species’ price flexibility come from the bio-economic modelling completed for the CFP Reform (Agnew et al., 2012) with updates made to reflect more recent studies on EU wide price movements (Nielsen et al., 2012). For the adjustable price flexibility parameter on the model dashboard, the high and low price flexibilities are set at 200% and 50% of the reference value respectively. This scale is in line with the range of estimates from empirical studies (Nielsen et al., 2012).
There are few studies, either empirical or theoretical, on how long-term changes in the quantity of fish landed will impact first sale prices, yet this understanding is necessary for estimated the potential landed value at a long-term state of MSY. Any value chosen should be done so with caution, as the value used should be low as fish price flexibilities are specifically for the short-term and are often derived from small changes in quantities. In modelling MSY, some of the changes in quantity are so large that the price would fall to near zero, which is not expected in reality (Crilly & Esteban, 2012), especially given the more gradual rate of change where markets and supply chains can adjust.
There is also possibility that while landings at MSY are larger (pushing down the price), they are also in a sustainable state and there is evidence of a small price premium to products with sustainable certification (Roheim et al, 2011;Blomquist et al., 2014). The question remains whether this demand for sustainable fish is limited to a market niche or could cover an entire sector as news stories provide a more positive frame of healthy fish stocks.
In BEMEF, the long term price flexibility of a species is set at 25% of the short term reference value as markets and consumer tastes will have longer to respond to relative changes in quantities. This approach is conservative, although many bio-economic models have perfectly stable price (no price flexibility) in long term modelling (Guillen et al., 2016). This approach also recognises a declining price flexibility with time, which something that can be expanded upon in possible future developments of BEMEF with an intermediate transition period. The high and low range factors for the long-term price flexibilities are kept the same as short-term ranges (200% and 50%).
The consequence of fish price flexibility in the context of MSY fisheries management is that gross earnings at long-term MSY are reduced as a lower sales price is received due to higher landings (greater supply). Conversely, price flexibility also implies larger prices and revenues than would otherwise be expected during the transition to MSY as fishing mortality and landings are reduced.
While the AER historical price modelling is at the fleet level, there are sources of real market data for period t+1 and t+2 (2015 and 2016) from the European Market Observatory for Fisheries and Aquaculture Products (EUMOFA). These reported prices are at the Member State level rather than the fleet level, so the average market price for each species in a Member State acts as a reference point to scale all fleet segments. Integrating timely market data was suggested in a review of the EIAA model (STECF, 2004). Where market data does not exist in EUMOFA, the fish price flexibility is used as an alternative.
No exogenous shocks or market trends are used in the prices for the short-term forecasts or other scenarios. This omission is due to the complicated nature of global trade patterns and large changes underway in aquaculture production. This approach is reasonable given that the most recent OECD-FAO Agricultural Outlook forecasts stable prices in real terms.
First baseline price is calculated by fleet and species:
P0,i,j = V0,i,j,/L0,i,j,
Future prices where price flexibility is required:
Pt,i,j = P0,i,j * ∑Qet,i,j/∑Qe0,i,j
OIt,j = OI0,j * (Vt,j/Vt,j)
Production function and effort change
Following from the EIAA model design, an activity variable is calculated and used in the model to adjust variable costs. These changes are calculated within a fleet segment, rather than between fleets.
This calculation takes the form of an inverse Cobb-Douglas production function to isolate for the the effort change variable.
At,j = ∑ (L0,i,j * Pt,i,j * θt,i,j) * (SSBt,i,j/SSB0,i,j)ϒi,j * (Qt,i,j/Q0,i,j)χi,j
This activity coefficient is used as the driver of all variable costs (labour, energy, and other variable costs). When effort exerted on a certain stock, is reduced due to a lower TAC, the total variable costs of a fleet segment are reduced relative to the weight of the reduced species in the fleet segment's landings composition.
Other variable costs
Other variable costs in future periods are calculated by energy costs in the base period scaled by the change in effort.
OVCt,j = RC0,j * At,j
Energy costs in future periods are calculated by energy costs in the base period scaled by the change in effort and estimated changes in prices based on reported fuel prices or, if not yet available, fuel price forecasts. Recent fuel prices are reported in the EU weekly fuel price bulletin and fuel price forecasts are available from the EIA.
ECt,j = EC0,j * At,j * (EIAt/EIA0)
The effect of an increase in the fuel price is to raise fuel costs and thereby reduce the amount to other variable costs, including crew share. If the parameter for job estimation through wages is selected, this decrease in wages leads to a decline in employment as work is sought elsewhere. This relationship is similar to previous work on the relationship between fuel, wages and employment (Salz & Smit, 2006)
LCt,j = LC0,j * At,j
Fixed costs (repair and maintenance, unpaid labour value and other non-variable costs) are linked to vessels much more than activity. As such, they are scaled in BEMEF by the number of vessels rather than the activity variable.
Repair and maintenance costs
RCt,j = RC0,j * (Vt,j/Vt,j)
Other non-variable costs
NVCt,j = NVC0,j * (Vt,j/Vt,j)
Unpaid labour value
ULCt,j = ULC0,j * (Vt,j/Vt,j)
A number of economic indicators are calculated as shown by the subsequent expressions.
GFt,j = TRt,j - (RC t,j + CCt,j + FCt,j)
NPt,j = TRt,j - (RC t,j + CCt,j + FCt,j + DCt,j)
Operating profit margin
OPMt,j = [TRt,j - (RC t,j + CCt,j + FCt,j + DCt,j)]/ [TRt,j]
Gross value added
GVt,j = NPt,j + CCt,j+DCt,j)
There are a couple of methodological approaches one could take to calculate jobs. BEMEF uses an effort-based approach by calculating the labour by demanded. First the number of days at sea are calculated using data on landings and days at sea over the reference period. Then, the employment (FTE) required to work those days at sea is calculated using data over the reference period and a fixed relationship. Like many aspects of the model these relationships assume no changes to labour productivity or adjustments due to technological innovations.
This effort-based approach to employment differs from a wage-based approach that takes a more supply side view of labour markets and employment. Under this approach, employment is based on the number of people that are attracted to the industry through the wage being offered. This wage could be determined as a function of revenues, profits, or set at a specified level. In other models this wage-based approach has been used to verify the feasibility of results as a certain wage level may be required to attract younger workers to the fishing industry.
The effort-based approach is used as it is a better reflection of how hiring decisions are made by vessel owners and operators. In addition, the current state of labour markets across Europe is such that the demand for labour, rather than the supply of labour, is the constraint and the important dynamic to model.
There is no distinction in BEMEF between whether these jobs accrue to citizens of a specific member state, the EU or a non-EU country. Government data on the Scottish fleet shows that 20% of all positions are held by non-EU citizens - the vast majority from the Philippines. The data also shows a significant variation between fleet types, with pots and traps employing 100% EU citizens and <24 metre demersal trawlers employing 57% EU citizens (Marine Scotland Science, 2013).
First landings ability (CPUE) in future periods is calculated as an adjustment to the landings ability from the three year base period using stock-landings and effort-landings flexibility rates in a Cobb-Douglas production function. Landings-stock is set to 0.1 for pelagic stocks and 0.8 for demersal stocks. Effort-landings is set to 1 unless there is a TAC access effect.
CPUE0,i,j = L0,i,j/SD0,j
Then future catchability:
CPUEt,i,j = CPUE0,i,j * (Lt,i,j/L0,i,j)(1-(1/αi,j)) * (SSBt,i,j/SSB0,i,j)(βi/αi,j)
And future sea days:
SDt,i,j = L * CPUEt,i,j
To calculate future jobs in fishing:
JFt,j = (JF0,j/SD0,j) * SDt,i,j
The calculation of industry or fleet effort is an area of modelling where a few key assumptions play a major role in bio-economic models and should be drawn out explicitly.
Merino et al (2015) analyse a series of four stock MSY scenarios using the Mefisto bio-economic model with the results indicating a decrease in effort (and highly increased profitability) at a state of MSY. It is important to note that the model covers four Mediterranean stocks that do not have a significantly higher landings at MSY than is currently fished, so no increase in effort would be required.
In their assessment of the Common Fisheries Policy for the European Commission, Agnew et al (2010) assume a reduction in fleets by 2% a year, which is the general trend over the past few decades. While landings have been fairly stable during this period of capacity reductions, it is assumed that this trend will continue as stocks improve to MSY and landings increase. There is an implicit assumption then that a reduction in capacity can simultaneously handle an increase in landings through increased recruitment as stocks through catchability or through technological improvements.
Salz (2012) takes a similar approach but is more explicit about the underlying assumptions. Effort, again measured as the number of vessels, decreases for a three year period before flatlining as fish stocks and landings increase under two MSY projections. While scenario analysis is used to test four different theoretical possibilities (fast recovery, slow recovery, fast deterioration, slow deterioration), an underlying assumption across the four scenarios is that the number of vessels will fall 10% in the first three years and stay below this level over the course of the projection. The assumption, as explained in the paper, is that higher profitability from stock recovery and increased earnings will go towards investments in labour-saving technology and not towards employing additional labour.
What is clear from all of these effort-based scenarios of EU fisheries, as well as in the body of literature on fisheries management as a whole, is that there are economic benefits to be had from fishing at long-term sustainable levels. The balance of these benefits between jobs, wages, profitability, and other indicators is shifted with the model assumptions that are used. The question then is not whether there are economic benefits from sustainable fishing, but how they will manifest in the fishing economy, and as BEMEF highlights, how the benefits will impacts different fleets in particular.
Yet to some extent the balance of benefits is not a question confined to model assumptions, but a question at a policy and management level about what outomes, economic or otherwise, that we would like to target and optimise (Kell et al, 2007). New frameworks have shown that it is likely impossible to simultaneously maximise conflicting environmental, social, and economic objectives (Hilborn, 2007).
From a modelling perspective there is also ambiguity. Two effects determine the overall relationship between effort and population in the long run.
The stock effect: If a fishery were somehow completely duplicated in some other area of the world then the number of fishermen required would instantaneously double. More effort is required to catch more fish, if everything else stays the same (in particular, stock density). This could more accurately be named the “available catch effect” since it depends not just on the population but on how much of that is available for harvest (presumably determined by regulation).
The catchability effect: If a fishery’s stock population were somehow to duplicate but remain within the same geographical area as the original stock population then fishing would instantaneously become much easier. Less effort is required to catch each fish when stock density increases, if everything else stays the same (in particular, the geographical scope of the fishery).
These effects do not necessarily work in the same direction or have a comparable relationship with the level of stock population. In other words, the direction of change of total effort as population size increases is ambiguous until the population size associated with MSY is reached (since beyond that both sustainable yield and effort per unit of yield decreases).
The implication is that, for fisheries that are over-exploited (i.e. have a population lower than that which sustains MSY), rehabilitation of the stock to MSY may be associated with either an increase or decrease in total effort. Each fishery is likely to differ with respect to the exact shape of the population-effort function – strictly, this is an empirical question.
A rule is used in BEMEF whereby the share of gross earnings going to labour is determined by vessel size. This reflects the fact that some vessels, typically large ones, will cover energy costs and variable costs before wages are paid, whereas smaller vessels tend to operate on a more strict earnings share system. Although these different wage arrangements have been documented, the methodological split based on vessel length is an assumption as most studies conclude that the decision on crew share is made by the individual skipper (LaRiviere, 2008; HM Revenue & Customs, 2014).
For fleets of smaller vessels (<12 metres) the labour share is calculated as:
CWsj,t = [(Ej,t - FCj,t - VCj,t) * (LCj,0 / Ej,b)] / Jj
For fleets of larger vessels (>12 metres) the labour share is calculated as:
CWlj,t = [Ej * (LCj,0 / Ej,0)] / Jj
Fleets that could not be grouped into one of the two categorised were grouped with smaller vessels and a more strict earnings share.
This practice of crew wages as a share of earnings rather than a fixed wage is well established in fisheries (Sutinen, 1979) and is therefore the default scenario. However, if employment is calculated based on the wage rate (described in the fishing employment section) then clearly this amount is used instead of a crew share system.
Processing jobs are calculated through the use of country-based multipliers for primary processing employment and an EU-based multiplier for secondary processing employment. The multipliers were calculated by estimating the relationship between landings and processing jobs generated in primary and secondary industries (landings corresponding to one EU-based FTE), after making necessary adjustments to isolate only those jobs generated by fish landed (and remaining) in the EU.
For primary processing the number of jobs that are supported by aquaculture were removed. Landings likely entering the market unprocessed (based on FAO data) were also removed from the calculation before the multipliers. At the second stage we remove jobs generated by fish imported from outside the EU and fish exported extra-EU after the first processing stage subsequent to landing. Due the number of worksheets this calculation of the multiplier is held outside of the BEMEF excel workbook but is available on request.
Total EU-based processing multiplier (FTE job/tonne of landing)
JPo = JPp + JPs
Primary EU-based processing multiplier (FTE job/tonne of landing)
JPp = [(P1j + P2j) * (Lt/(AQ+Lt))] / [Lt - (Ld*Lt)]
Or in text: FTE jobs in primary processing * Marine landings as % of fish processed / Landings - Fish directly sold
Secondary EU-based processing multiplier (FTE job/tonne of landing)
JPs = [(P3j + P4j) * ((Lt/(AQ+Lt)) - ((P1i + P2i)/(P1 + P2))] / [Lt - (Ld*Lt) - P1e - P2e]
Or in text: FTE jobs in secondary employment * Marine landings as % of fish processing - Imported fish as % of primary processed fish / Landings - Fish directly sold - Exported fish after primary processing
Carbon emissions (all greenhouse gas emissions measured in CO2 equivalents are estimated based on reported fuel use. The emission factor used are for fuel oil (Defra, 2013). While there may be variance in the emission filtering of engines, this data is not reported by fleet and is expected to be small.
Changes in fuel use follow the same effort-based approach as jobs. Comparisons of effort change methodologies are underway to ensure consistency within BEMEF.
First future fuel use is estimated:
Ft,j = F0,j/SD0,j
And the carbon emissions using an emissions factor:
Ct,j = Ft,j * EF
Most of BEMEF operates through the use simulations to illustrate a concept, rather than predictive forecasts. Still, as it can be useful to attempt to forecast, especially in the short-term periods, time-dependent variables have been created for technological change and fuel cost change to fuel prices.
Technological change is often modelled as a capacity reduction. In the case of BEMEF this works by an adjustment to catchability and to cost reduction in the production function.
For the technological change parameter, the high end is set at the 4.4% rate of improvement in technological efficiency observed in the European fishing fleet (Villasante & Sumaila, 2010). As these technological efficiencies may not be passed on to capacity reductions, a low end of 2% is an also an option. This is in line with the current capacity reduction trend (Paulrud et al., 2014). Like fuel prices, the long-term state is assumed to be in 2020 as it aligns with the latest date for reaching MSY in EU waters.
This impacts the effort change and catchability equations.
The forecasted effort change takes the following form:
At,j = ∑ (L0,i,j * Pt,i,j * θt,i,j) * (SSBt,i,j/SSB0,i,j)ϒi,j * (Qt,i,j/Q0,i,j)χi,j * Tt,j
One of the main features of BEMEF, particularly the online version, is the economic assessment of MSY.
Idea of the task
The outcomes of this MSY scenario formed part of the report Managing EU Fisheries in the Public Interest (Carpenter & Esteban., 2015).
MSY estimates There is a continuing debate among fisheries biologists as to the most reliable method to calculate MSY. In BEMEF, a hierarchy of data sources is used as there are multiple estimates for the same stock. The first set of estimates are from multispecies studies where ecosystem dynamics have been taken into account, predator-prey relationships in particular. The next set of MSY estimates comes from single stock analysis in the academic literature, where analytical assessments are prioritised. The third classification is where no MSY estimate is available and quota levels from the baseline are used.
In general, where there are multiple sources for an MSY estimate the alignment is good, although the multispecies estimates tend to be lower. That there is general alignment is partly due to the fact that single species values are estimated from a multispecies context - the current one. As such, assessments already involve interaction with other species and with the environment through natural mortality, somatic growth and population growth over the years.
The main data source of single species estimates (Froese et al., 2016) states that a 90% rule can be applied to the single species estimates which would imply stock sizes of ⅔ the unexploited size, which should be large enough for predators and preys to fulfill their respective ecosystem roles.
There are concordance issues between the fish stock areas (for the MSY estimates) and TAC areas. This issue has been detailed but remains unresolved. It also presents issues with the use of ICES data on current stock biomass in BEMEF. Some matching had to be undertaken to fit the two areas as closely as possible.
As new MSY estimates with improved methodology become available they will be incorporated into BEMEF. Currently, the following sources are used in the model:
ICES. (2013) Report of the Benchmark Workshop on Baltic Multispecies Assessments (WKBALT), 4–8 February 2013, Copenhagen, Denmark. ICES CM 2013/ACOM:43. (Link)
ICES. (2013) Multispecies considerations for the North Sea stocks, Advice June 2013. (Link)
Guillen, J., Macher, C., Merzéréaud, M., Bertignac, M., Fifas, S. & Guyager, O. (2013) Estimating MSY and MEY in multi-species and multi-fleet fisheries, consequences and limits: an application to the Bay of Biscay mixed fishery. Marine Policy 40, 64-74. (Link)
Froese, R., Garilao, C., Winker, H., Coro, G., Demirel, N., Tsikliras, A., Dimarchopoulou, D., Scarcella, G., Sampang‑Reyes, A. (2016) Exploitation and status of European stocks. (Link)
Merino, G., Barange, M., Fernandes, J., Mullon, C., Cheung, W., Trenkel, V. & Lam, V. (2014) Estimating the economic loss of recent North Atlantic fisheries management. Progress in Oceanography 129, 314-323. (Link)
For any remaining stocks without MSY estimates from the available literature, MSY values are set at the base period TAC. This approach has been used in previous studies of MSY potential where data is limited (Cunningham et al., 2010).
Even in its most simple form, there exists uncertainty about these estimates. Some of this is due to the lack of historical experience with low fishing pressure (below Fmsy). There are also issues about the density-dependence effect (Ohlbeger et al., 2014) and how this interacts with low fishing pressure (Svedäng & Hornborg, 2014). Ultimately these are uncertain waters, but the modelling in BEMEF takes the approach of dozens of previous studies (incl. Cunningham et al., 2010; Quaas et al., 2012; Crilly & Esteban, 2012; Salz, 2012; Merino et al., 2015; Guillen et al., 2016) in forging ahead and attempting to analyse MSY as one of the fundamental objectives of the Common Fisheries Policy.
The following features have been developed alongside BEMEF but have been removed from the core file due to their size and now exist as separate files.
As the species landings recorded by fleet in the AER dataset is also recordedr by geographic (ICES areas), a visual display is available for a fleet under analysis. On the map tab a fleet is selected and then a heatmap shows where that fleet is catching a certain species (or all species together). The darker the shade the greater the catch is relative to the other areas.
There also exists landing data by fleet type at the smaller resolution of (ICES rectanlges) and a similar heatmap has been produced where a user can select a fleet and species to see where in EU waters the catch is taking place. Due to the processing power of the mapping visualisation, this feature exists as a stand alone file to BEMEF but is available upon request.
As employment is a central concern of fisheries management, particularly in coastal communities, an extended component is available in the excel version of BEMEF to analyse how MSY or other variables will impact on port employment. This is accomplished using the EU Fleet Register which records the length of a vessel and where it is registered. The distribution of registered vessels can also be viewed on the European Atlas of the Seas by vessels number or size at ports around the EU.
With this information on where vessels are registered, the economic outputs of a fleet can then be distributed across ports in the relevant country where vessels of that fleet category are registered. This approach follows in line with practiced methodology (Natale et al., 2013) for finding fisheries-related employment in coastal communities.
There are some major assumptions in this approach due to the limited information that can be concluded from the Fleet Register. One, since there is no component for species, landings are treated on aggregate. Two, fleet landings are assumed to take place within its home country. And three, vessel landings are assumed to be made in its home port.
It is clear from these assumptions that this approach to port analysis is only the best approach when there is no alternative data to the Fleet Register. For some countries this is not the case. In the UK, the Marine Management Organisation (MMO) records landings by species, gear, length, and port. This dataset removes the need for the fleet register as fleet landings can be more allocated to the ports where they (historically) occur.
Due to the size of the port databases, this analysis exists as a stand alone file to BEMEF but is available upon request. The local economic impacts in the UK of rebuilding fish stocks was modelled using the port analysis function of BEMEF for the Blue New Deal report (Balata et al., 2016).
Transition to MSY
An important aspect of the economics of MSY fisheries management in the EU is not only what fishing at a long-term state of MSY would mean for the fishing fleets, but also what the economics of a transition to this state would look like. This transition is much more difficult to model as it involves the continuous interaction between the biological components (stock growth) and the economic components (landings from fleets) of an economic model. An added difficulty is modelling the interactions between predator and prey species as the stocks grow in number and at different rates. Incorporating a transition period to MSY using stock growth rates associated with a series of TACs would change BEMEF from a comparative-static model to a recursive-dynamic model over multiple, linked years.
While recent work in bio-economic modelling has estimated the economic benefits of a faster transition to MSY (Guillen et al., 2016), this is through a standardised one-fleet bio-economic model.
The development of uncertainty bands follows from best practice in economic modelling and the result tables and graphs in BEMEF would benefit from this addition. Currently the range of values for key parameters is being gathered in order to run multiple scenarios and calculate the corresponding uncertainty bands.
As noted, most of BEMEF operates through the use simulations to illustrate a concept, rather than forecasts, although some developments such as technological change and changes to fuel prices are designed to allow for additional forecasting power.
Following on best practice in economic modelling, a good test of model accuracy would be to set the base period at earlier dates so that the t+1, t+2 and t+3 periods have recorded data to compare with the modelled data. This test would show the reliability of effort-landings dynamics and an indication of the uncertainty in prediction (which may also help with the uncertainty bands). However, as many of the variables in BEMEF are estimated by scaling historical results it is unlikely that the findings will show significant departures between estimated and actual data.
Fleet inclusion for quota allocation
As few studies have attempted an economic analysis of alternative quota allocations, there is a degree of experimentation within BEMEF regarding what fleets should realistically be assessed against each other in quota allocation comparisons. There are three adjustments currently under review.
The first potential adjustment is the current landings threshold under which a fleet is included. This is to strike a balance between removing fleets that are only landing a particular species as by-catch without removing fleets that would have the capacity to target that species but have low historic landings only due to the current allocation arrangement.
A second potential adjustment is to look at data reporting issues for various fleets. This problem has arisen as some fleets reporting in the AER Database are fishing species that are not associated with the gear type recorded for that fleet. This is likely a fleet classification/segmentation issue due to primary and secondary gears used onboard vessels.
Adjustable stock-catch coefficients There is little information on stock-catch coefficients, but a larger range of values from 0.1 for pelagic stocks and 0.8 for demersal stocks could be applied. This change is important given the relative significance of the coefficient in calculating effort change.Back to top