The Benefits of Responsible Fishing

BENEFITS OF RESPONSIBLE FISHING: THE IMPACT OF AN INNOVATIVE TRIAL OF VOLUNTARY RESTRAINT

Early in 2002, members of the North Sea Fishermen’s Organisation (NSFO), in the United Kingdom, and of the Cooperative Producentenorganisatie Oost Nederland UA (CPO), in the Netherlands, realised that quotas of plaice and sole on their North Sea fishing grounds were insufficient to ensure a fishery until the end of the quota period and would be quickly exhausted.

In response to this realisation, NSFO members agreed voluntarily to tie up their vessels for one week in four during the first three months of the year. Similarly, members of CPO limited their fishing to 40 days at sea during the first quarter, with at least 3 weeks without any fishing activity at all.

The intention of this action was several-fold. First, it would ensure a more steady supply of plaice and sole to processors and consumers. Secondly, it could have been expected slightly to raise the price received at auction while the tie-ups were in place, but to have prevented a sharp rise in prices when the quota was exhausted early as had happened in 2001.

There may also have been benefits to the fleet in slightly increased catches arising from fishing more-abundant (less fished) stocks later in the year, as well as a reduction in costs, due to fewer days being spent at sea.

The temptation to fish illegally is also reduced owing to the consequent benefits of stock conservation and future earnings.

As such there appears to have been the chance of gains to all participants in the supply chain right through to consumers, and also to the crew share of sales, from this socially responsible approach.

Normally, these would be impossible to achieve without appropriate government intervention; intervention which has been successful hardly anywhere in the world. It represents in a sense the ideal of fishery managers – fishing enterprises acting collectively to constrain their action for their own and others' benefit.

The action was repeated under the same principle but slightly different terms in 2004 and 2005 but not in 2003.

This project responds to Criteria 1(3) of the Call for Proposals in that if it can be demonstrated empirically that this innovative socially responsible behaviour and good practice in the fisheries sector can result in gains for the participants then there may be an incentive for others to follow, with the consequent positive impact on illegal fishing, stock conservation and future earnings.

This is particularly pertinent given the continued apparent failure of measures, imposed on the fishing industry as part of the Common Fisheries Policy (CFP), to reduce fishing effort in Northern Europe, whilst maintaining fleet viability.

The results will be of general interest in fisheries management, serve to improve the image of fishing, and have wider benefits for the community, preserving a steady flow of raw material to processors and stable prices for consumers.

The objective of this study has therefore been to identify whether, following the voluntary restraint on fishing imposed by the NSFO and CPO during the early part of 2002, 2004 and 2005 the expected improved price stability for plaice and sole on the Urk auction market was achieved, and to estimate the amount of the improvement in revenues, crew share, profits and cost-savings for the fleet.

Section 2 Background

The membership of CPO is made up of beam trawlers fishing for flatfish in the North Sea. They land almost exclusively to the markets at Urk and Harlingen. CPO membership in 2004 consisted of more than 100 vessels of widely varying sizes. The annual turnover of CPO members is approximately €115 million.

The NSFO is a Producer Organisation (PO) with a membership of some 25 active vessels representing mainly British and Anglo-Dutch beam trawlers, most of which originate in the Dutch port of Urk. They constitute the majority of the UK’s beam trawlers fishing in the North Sea. The beam trawlers in the NSFO average 35m in length; only four of the beam trawlers in membership are the smaller class of ‘Eurocutter’ which work mainly from ports in the south-west of England. Total income of the 23 North Sea beam trawlers in the NSFO, in 2004 was approximately €30 million.

A further 7 Urk beam trawlers with similar fishing patterns work within other UK Producer Organisations.

The Anglo-Dutch vessels also land mostly into the fish auction at Urk, or Harlingen, also in the Netherlands and the next most important market.

Urk vessels also work under the flags of Belgium, Denmark and Germany, representing a significant fleet tonnage in both Belgium where they comprise approximately 25% of the beam trawl fleet and Germany, 80%.

The Dutch vessels operate under an Individual Transferable Quota (ITQ) management system where each fishing company originally had a quota given according to previous fishing practice and which could subsequently be traded. Under the UK’s Fixed Quota Allocation (FQA) system each vessel has been allocated a share of the UK fish quotas also based on track record. The vessel quotas are allocated to a vessel’s PO and administered by the PO. It is effectively very similar to an ITQ system since quota may be, and is, traded, both between PO members and between POs.

The mainstay of both Urk fishermen and the Urk fish auction has historically been flatfish, in particular plaice and sole, and it is changes to the availability of plaice quota that appears to have driven the move by fishermen from a single port to be so adventurous in flagging vessels elsewhere, combined with the financial strength induced by the Individual Transferable Quota (ITQ) regime under which they work.

Figures 2.1 and 2.2 trace the changes in plaice and sole landings and quota since 1987. Much of the behaviour of Urk fishermen can be seen as a response to these declines, in particular to the dramatic cut in the Total Allowable Catch (TAC) in 1995 and again in 1996. Flagging out tonnage to other Member States has allowed total plaice quota available to Urk fishermen to remain relatively stable during a period of significant overall decline in the TAC.

Figure 2.1: Landings of North Sea Plaice and the Total Allowable Catch 1987 to 2003

Source: ICES

Figure 2.2: Landings of North Sea Sole and the Total Allowable Catch 1987 to 2003

Source: ICES

Fleet ownership remains remarkably similar to the patterns seen in Urk 20 years ago, with the majority of vessels in single ownership, often with more than one family member on board, and others associated with the shore side aspects of running the vessel.

A trend towards rationalisation is discernable, but family-owned vessels remain the mainstay of the Urk fleet. Technological changes within the beam trawl fleet in the Netherlands have been well documented by the European Commission.

Whilst the flagging out of the Urk fleet has continued since the first tentative approaches to the UK around the time of the Factortame case in the 1990s, numbers of vessels and quota available have not increased. The changes in ownership, with the associated costs of operating away from Urk, have been in effect moves to maintain the status quo within the port.

Under the system of Relative Stability within the CFP, the United Kingdom has 27% of the North Sea plaice TAC and Netherlands 38%, a combined holding of two-thirds the stock available to be fished.

Quota prices in the United Kingdom have historically been well below those of the Netherlands, and the quota market developed several years later than in the Netherlands, allowing the first Dutch operators to flag into the United Kingdom to gain access to quota at much cheaper rates than were available on their own flag.

Vessels operating in Belgium, Germany and Denmark are allocated quotas centrally, usually on a monthly basis.

The average British North Sea beam trawler has access to significantly more annual plaice quota than average Dutch flagged vessels (approximately 400 tonnes compared to 175), and significantly less sole quota (15 tonnes compared to 50), leading to different fishing patterns evolving between what are physically very similar vessels.

Vessels each had (and still have) significantly less quota of sole and plaice than they are capable of catching; they have in the order of 70-80% of their needs for sole but only 50-60% of the plaice they could catch. [1]

Dutch vessels are likely to fish in areas where sole are found, typically the muddy grounds in the southern North Sea, close to the Plaice Box, an area off the coasts of Belgium, Denmark, and the Netherlands where plaice fishing is restricted.

Vessels flying the United Kingdom flag spend longer on grounds further north, where sole are less abundant.

Virtually the entire Urk-based fleet leaves for the fishing grounds early on a Monday morning, returning to land at the end of the week. The reason for this is mainly cultural, but there has been an increasing trend for British-flagged vessels, fishing further afield, to fish longer trips, though many still choose not to fish on a Sunday. The preference for fishing patterns of this type leads to very large auctions on a Friday and Monday, with much smaller volumes of fish available for sale midweek. Urk processors appear to be well adapted to these peaks and troughs.

The fish action at Urk is the leading fish auction in the Netherlands, taking a third of all Dutch sales, with a throughput of 35,500 tonnes in 2004. The majority of this is flatfish, with 20,000 tonnes of plaice, 3,800 tonnes of sole and 2,750 tonnes of dabs.

United Kingdom vessels consigned 9,000 tonnes of fish to Urk for auction in 2004, the bulk being from vessels owned and operated from Urk itself.

Urk vessels flagged out to Belgium, Denmark and Germany consigned a further 5,000 tonnes to the auction. 94% of the plaice sold on the market is purchased by local processors. A third of this is sold as fresh fish, mainly within the Netherlands, the remainder being processed for freezing and subsequent export.

Section 3 A Literature Review of Demand and Inverse Demand Functions in the Quayside Market for Fish

Introduction

The large number of studies that have been made of the landings market for fish have confirmed the responsiveness of the price of fish to the quantity landed for sale as economic theory predicts.

Not surprisingly, the development of understanding of demand and inverse demand functions in general has been considerably enhanced by the emergence of information technology and the refinement of econometric techniques that powerful desk-top computers have made possible. These have made it possible to test the theory empirically.

The purpose of this literature review is:

to appraise the economic theory relating to the relationship between the auction price of fish and the quantity landed and offered for sale,

to consider the results of empirical analysis undertaken, and

to discuss how other economists have dealt with the myriad of theoretical and practical problems that empirical investigation of fish auction markets presents.

This will serve to give a sound theoretical foundation to the work to be reported later in this study.

The Inverse Relationship between Quantity and Price

It has already been obvious for many years that the demand for fish and the price it obtains are responsive to supply, especially on a daily basis. For example, during the heyday of the seasonal East Anglian Herring Fishery in the early 1930s the fleet would land its catch each morning before bunkering and returning to sea.

The fish processing companies in Great Yarmouth would accept all the fish for which they had capacity. When they were full the international code flag “N” would be hoisted on the fishwharf and vessels too late to land to the market would turn and dump their catch at sea. Even now the trade press carries occasional stories of perfectly good fish being withdrawn and sent for reduction because of excess supply on a particular day.

These anecdotes suggest that there are limitations to the amount of fish the market can absorb on a daily basis governed perhaps by processors throughput and chilling capacity.

In the famous case of the pacific halibut in the USA an attempt was made to control output by limiting fishing effort. Instead, the year-round fishery of the early 1920s became one of only 48 hours and less in the 1990s. The economic distortion of investment on-shore resulting from the failure of the policy was the development of vast chilled storage capacity to provide steady supplies for the downstream consumer market.

Adam Smith (1776) demonstrated that the price of goods will vary according to the amount becoming available for the market. In developing his rudimentary theory he was particularly influenced in his thinking by the markets for agricultural produce. He noted that a glut was accompanied by a fall in price, whereas a poor harvest would lead to high prices for the produce. This Classical view of the inverse relationship between the market-clearing price and quantity sold suggested that the determination of price was driven by supply.

The Keynesians suggested that the direction of causality might be in the opposite direction. They felt that the price set by the market determined what producers could supply (Keynes 1936).

This difference in view, however, is in large part due to the fact that the Keynesians were writing in a time when industrial production was in the ascendancy in the world economy. Industrialists have much greater control over the level of output than crop-growers. They are generally not subject to the responsiveness of production-in-hand to the natural factors that cause gluts and shortages in crops. An industrialist can plan to use given amounts of the factors of production and expect a certain level of output, within the variability caused only by reasons such as breakdowns and human error. This means that industry can adjust its output to the going price.

The relationship between the market-clearing price of fish and the quantity sold is closer to the Classical view of markets than that of the Keynesians, because fishing enterprises do not have the tight control of output that is available to industrialists.

The supply of fish is governed by the same kind of variability faced by agricultural output. Environmental factors may play a considerable part in deciding the amount of fish that is available to a market. Annual recruitment of juvenile fish to a fishery is highly variable, with the maximum being 7 times the minimum for plaice and 45 times the minimum for sole since stock assessments began in 1957 (ICES 2003).

The existence of TACs has served only to cap the amount of fish that may be landed in any one year, allowing for misreported and illegal landings. However, this has not removed the natural variability of landings on a daily basis.

There are several reasons for this variability. First, stock estimates may be inaccurate, and the TAC set incorrectly or in the face of a stock assessment for political reasons. Sometimes, in spite of the availability and use of fish-finders, the fish are just not about. So, whatever the stock level, there are days when fish landings are scarce or abundant. Bad weather may prevent a fleet from fishing or force it to move to a second-best fishing ground. This variability of supply to the market may be expected to drive the price.

Early work in estimating the relationship between the quantity of fish supplied and the price achieved at fish auctions concentrated on the market for fish in the USA, where Bell (1968), Nash and Bell (1969), and Waugh and Norton (1969) established the empirical linkage between the price obtained for fish and the quantity landed.

Gates (1974) confirmed the relationship and showed the importance of including the size of fish as an explanatory variable in the inverse demand function.

This relationship between changes in the quantity supplied and variations in the market price is called the price-elasticity of demand but is determined in a function where the quantity supplied is the dependent variable. The price-elasticity of demand gives the percentage change in quantity-sold brought about by a 1% change in the price.

When the price is the dependent variable the relationship is called a flexibility. The price-flexibility of demand is the percentage change in price brought about by 1% change in quantity sold. The price flexibility is the inverse of the price elasticity.

Other elasticities and flexibilities may reflect the responsiveness of demand or price to changes in another variable such as landings of another species.

In a case study of the yellow-tail flounder Gates found a price flexibility of demand of -0.63, equivalent to an elasticity of -1.59. The price flexibility of the mean size of fish landed was 0.43.

Previous Estimates of the Price Elasticity of Demand and of the Price Flexibility of Demand for Sole and Plaice in the Northern European Arena

In Europe, Ioannidis and Whitmarsh (1987) estimated the demand relationship for plaice in the United Kingdom and found the market to be highly competitive. Their models offered high explanatory power for the effect of landings on the price level.

Lagged effects explained the price achieved by landings quite sharply but the impact disappeared after a month or two. An important finding was the cross-price effect, where the price obtained for plaice was influenced by landings of other species, notably haddock.

Rodgers (1987) also estimated the demand relationship for plaice in the United Kingdom, concluding that the long-run price elasticity of demand using monthly observations over the period from January 1981 to December 1986 was -3.05. The short-run elasticity was more volatile at -6.72, reaffirming the belief that the market reacts sharply to short-term over-supply.

Jorgensen et al (1989) suggested that the responsiveness of price to changes in supply of sole in the Dutch market was very sluggish, taking longer than a month to adjust perhaps as a result of frictions faced by processors in adjusting capacity.

They found a short-run flexibility of -0.19 which rose in the long-run to -0.57, but both were sensitive to the degree of change of supply. The mean-size flexibility was 0.12 in the short-run and 0.36 in the long-run.

In the Danish market the own-price flexibility of plaice was found to be -0.074, implying that a 10% change in the volume of landings will bring about a 0.74% change in price.

Table 3.1a Price Elasticities of Demand for Sole in the North European Arena

Study	Market	Short Run Elasticity	Long-Run Elasticity
Jorgensen et al (1989)	Netherlands	-5.26	-1.75
Jaffry et al (1997)	United Kingdom	-	-5.00
MOR (2001)	United Kingdom	-	∞

Table 3.1b Price Flexibilities of Demand for Sole in the North European Arena

Study	Market	Short Run Flexibility	Long-Run Flexibility
Jorgensen et al (1989)	Netherlands	-0.19	-0.57
Jaffry et al (1997)	United Kingdom	-	-0.20
MOR (2001)	United Kingdom	-	0

Table 3.2a Price Elasticities of Demand for Plaice in the North European Arena

Study	Market	Short Run Elasticity	Long-Run Elasticity
Rodgers (1987)	United Kingdom	-6.72	-3.05
Ioannidis and Whitmarsh (1987)	United Kingdom	-55.80	-14.00
Jorgensen et al (1989)	Denmark	-	-13.51
Jorgensen et al (1991)	Netherlands	-8.33	-3.70
MOR (2001)	United Kingdom	-	-5.21 to -9.17

Table 3.2b Price Flexibilities of Demand for Plaice in the North European Arena

Study	Market	Short Run Flexibility	Long-Run Flexibility
Rodgers (1987)	United Kingdom	-0.15	-0.33
Ioannidis and Whitmarsh (1987)	United Kingdom	-0.018	-0.07
Jorgensen et al (1989)	Denmark	-	-0.074
Jorgensen et al (1991)	Netherlands	-0.12	-0.27
MOR (2001)	United Kingdom	-	-0.11 to -0.19

Jorgensen et al (1991) found that the own-price flexibilities for plaice in the Netherlands reflected the same structure as found in their earlier study for sole with a short-run flexibility of -0.12 and a higher long-run flexibility of -0.27. However, they were unable to find a statistically significant mean-size flexibility.

In contrast, Jaffry et al (1997) using a systems approach and correcting for co-integration found a high elasticity of -5.00 for sole in the British market.

MOR Research Ltd (2001) found high flexibilities for plaice in the British market ranging from 1.58 to 7.75 and again adjustment was noted to be rather slow. The estimates for sole proved insignificant suggesting the hypothesis that the estimates were not significantly different from zero cannot be rejected.

Tables 3.1a-b and 3.2a-b summarise the elasticities and flexibilities estimated in European studies of plaice and sole.

It is noticeable that the elasticities estimated for plaice and sole are very much lower in the Dutch market than in the markets in Denmark and the United Kingdom. This bears out Nielsen's proposition (Nielsen 1999) that in smaller, even partially integrated markets where the Law of One Price (Stigler 1979) is in operation, elasticities will be over-estimated. The Dutch market is clearly the price-setter and the others are price-takers.

Problems of Estimation

Identification

One of the difficulties of estimating demand and inverse demand relationships is of identifying the demand (or inverse demand) curve. The observations in the data each represent a point where it is assumed that market clearing has taken place and so demand and supply are in equilibrium.

In moving from point to point, there must be at least a movement along one curve and a shift of the other, or a shift of both curves. It is not possible easily to decide which of these the observations show unless some further piece of information is available.

The most satisfactory extra information would be explanatory variables for supply which would provide information identifying the supply curve and by default also the demand curve, in which case the demand and supply curves may be estimated simultaneously.

Alternatively, if some evidence can be brought forward that supply is perfectly inelastic then it may be assumed credibly that the demand or inverse demand function is identified and a single equation estimation performed.

Ioannidis and Whitmarsh (1987) supplied a theoretical underpinning for assuming that the supply of fisheries products at the quayside is indeed perfectly inelastic and that the demand function is therefore identified. A conclusion of their study was that the direction of causality ran from quantity to price, confirming that the supply of fish is exogenous in the short-run.

This is enormously helpful as it serves to permit single functions of demand to be adequately identified for econometric estimation, without the need for a supply curve to be estimated simultaneously or for instrumental variables to act as a proxy for a missing supply side.

At a daily level, it is difficult for skippers to know what is likely to be landed to the markets and even if they have some idea that the market may be over-supplied, they may be unwilling to remain at sea because of the deterioration in the quality of the fish that they already have caught.

In models of greater periodicity, it has been shown that skippers give more consideration to landing their quota than maximising profits (Valatin 1995).

In reality, the market-clearing price and quantity sold are unlikely exactly to match supply and demand in each time period (especially in this study, where the observations are daily). Instead, the market is likely to be a little out of equilibrium, but facing economic forces which make it hunt for the equilibrium.

However, as it moves towards it, the equilibrium may itself shift so that the market has to turn towards a new equilibrium. This of course is a continual process and the market will never settle absolutely. If it does so momentarily a disturbance can be expected which will set it moving again.

The action by the CPO and NSFO members in limiting the activity of their vessels was taken with some awareness in mind of this relationship between market-clearing price and quantity sold.

If the quota available to each enterprise individually is insufficient to keep their vessels active the whole time it seems apparent that they could benefit from ensuring a steady regular supply of fish to the market by spreading their activity more evenly between enterprises.

If all the vessels fished and landed during the same limited period within each month the price received could be expected to be lower owing to the limited throughput and chilling capacity of processors.

There may be some potential mutual benefit, therefore, to be gained from collusive behaviour and additional benefits to processors and consumers from a stable market and prices.

While the cost of investing in chilled storage might appear not to be a problem of the fleets, in fact it does impinge upon them because it raises processors’ costs and therefore lowers their demand (via their ability to pay) for fish.

While some of the extra cost will be passed on to the consumer, elementary economic theory shows that some will be borne by the processors and some passed back to the fleets (Archibald and Lipsey 1973).

The supply schedule for the processing sector can be expected to assume the normal upward gradient of most industries, reflecting the fact that the higher the market price achievable for output the more processors are willing and able to supply to the market.

With the demand curve having a downward gradient – the lower the market price, the more output purchasers are willing and able to acquire – the market-clearing quantity of output and the market price are given by the unique point of intersection of the curves representing the two sides of the market.

If we now assume that further chilled storage is required processors face both increased fixed cost in having to invest in the plant and increased variable costs in having to operate and maintain it. Hence the new supply curve has both a higher intercept and gradient.

The immediate result is lower output and a higher market price in the market for processors output.

However, that is not the end of the story, because a given supply is coming through the market and the processor market must return to that. Hence, something must happen to return the supply curve to its previous position.

That something is a fall in the auction price of fish. Indeed, the whole of the cost of the cold storage will ultimately be passed to the fleet in a rare case brought about by the short-run perfect elasticity of supply. Since some vessels will be unable to remain viable with the lower fish prices, they will leave the fishery and the remainder will have better quotas offering some improvement in the situation for those which remain.

Cointegration of Price and Quantity

Cointegration of price and quantity variables in a market occurs where the two variables are not independent over time. It causes a relationship between them simply to be re-established at a new level if some shock occurs to the market, rather than a new relationship developing.

The problem was first discovered in the 1970s when economists sought to discover what would be the impact on world economies of the shock to economic growth caused by large increases in the price of crude oil.

They wished to know whether there would be a permanent shift to a new level in the relationship of related variables or whether the world economy would simply return slowly, possibly by an oscillatory trajectory, to the trend that it had been following (Dickey and Fuller 1981, Perron 1989, Hylleberg and Mizon 1989).

Several papers have now established that fish prices and quantities sold at auction are often cointegrated. Ioannidis and Matthews (1995) showed that the prices and quantities of both cod and haddock in the British market were cointegrated and proceeded to estimate demand functions for those species accordingly.

However, there is a question mark over the use of methods to correct for co-integration where the period covered by time series is relatively short.

In this study observations are on a daily basis giving more than five hundred observations in a very efficient market. Though only two years and three months are covered, cointegrated variables for price and quantity may well demonstrate the characteristics of dependence over time in response to changes in the Total Allowable Catch of the species under consideration. Such shocks may have occurred twice over the period investigated, at the start of each year following the initial year 2002.

However, the time period is too short for the impact of shocks of the significance of, for example, 1970 oil prices, to be identified. Since these types of shock were intended to be accommodated by analysis for co-integration and the use of error correction mechanisms, there is no justification for pursuing this approach and testing and correction for co-integration has not been carried out.

Integration of Markets

One notable feature of empirical work on demand for fish products at the quayside is the disparity between American and European estimates.

The American estimates are consistently lower than those estimated for the European market until the most recent. Stigler (1969) established the Law of One Price and Nielsen has shown that to ignore the Law of One Price may invite estimates of price elasticity of demand which are excessive (Nielsen 1999).

Thus the continental differences may be due to the fact that whereas both the American and European markets are each single integrated markets, empirical work has often treated the USA as a single market but European Union member states as a set of independent markets. Compare, for example, Gates (1974) in the US with Rodgers (1987) in the United Kingdom which is possibly a sub-market of a wider European market. This is important for this study since excessive estimates will over-estimate gains.

Neilsen (1999) is especially valuable in that not only does he show the importance of market integration, he also offers a simple method for correcting estimates made where an apparent market is in reality part of a sub-market.

Guillotreau et al (1998) concluded that the markets are integrated for roundfish in Belgium, Denmark, France, Germany and United Kingdom.

Section 4 Econometric Investigation of the Tie Up Scheme using the Inverse Demand Function

The 2002 and 2004 tie-up schemes were introduced by CPO in the Netherlands and NSFO in the United Kingdom whereby vessel-owners agreed to restrict their fishing activity in order to preserve quota and to ensure a steady supply of their output to the market.

The purpose was to stabilise the market and ensure that by acting together all enterprises benefited.

Normally collusive behaviour by firms to influence prices is regarded as undesirable because it results in higher prices for the consumer, lower output, and excess profits.

In this case, however, if the scheme succeeded there should be greater profits for fishing enterprises, a steadier supply of raw material and stable prices for processors, and regular supplies of fish for caterers, retailers and consumers at steady prices. This is an example of a Pareto improvement, a benefit where no-one loses. Thus, rather than being a monopolistic exercise the intention was to manage scarce resources of fish to the benefit of all levels of the market.

The vessels of the CPO and NSFO are beam trawlers and their target species are dover sole and plaice. There is a small by-catch, principally of other flatfish, including lemon sole. In 2001, quota had been exhausted early and the resulting shortages in the market produced high prices contrasting with relatively depressed prices earlier in the year when large quantities of fish were being landed. The annual plaice quota held by CPO vessels is generally between 150 and 175 tonnes but because of large cuts in the TAC this is sufficient only for about three months fishing. Thus there is more than the usual pressure on the fleet to obtain the best possible prices.

For a tie-up scheme to succeed and be repeated required not only the cooperation of virtually all the vessels, something often considered infeasible in fisheries, but also for them at least not to lose financially. Remarkably, the members of both organisations agreed to and have apparently adhered to the collective behaviour necessary. The purpose of this section is to examine whether they might have improved the revenue from sales of fish resulting from a more stable market.

In 2002 the scheme covered the first nine weeks of the year and ended 28^th February. It applied to the whole Dutch fleet including the beam trawlers and larger Eurocutters with a beam greater than 4.5m. However, later in the year bad weather reduced the catch sufficiently for the fleet to receive compensation from the Dutch government.

In 2003 no scheme was implemented, probably as a reaction to the misfortune of the previous year, but in 2004 a new scheme was introduced. It allowed vessels two weeks fishing per month in January and February with a maximum of 24 days and was applied to all beam trawlers of greater than 300 kiloWatts engine power. It was extended at the request of processors who asked for an extra week tied up in March.

To determine the effect on earnings it was necessary first to calculate the relationship between the quantities offered for sale of the principal species landed by the NSFO and CPO vessels and the average price each achieved at auction.

The data used cover the quantities and values of sales of plaice and sole at Urk on a daily basis. Data on daily landings and the prices achieved at Urk have been provided by Visafslag Urk – the Urk fish auction market. They include landings and revenues achieved for lemon sole, dover sole, and plaice by EU market size category (Council Regulation 2406/96) from 1^st January 2002 to 31^st March 2004, a total of 584 observations. Sales were also disaggregated by nationality of the vessel landing the fish and this covered seven nations; Belgium, Denmark, Germany, Ireland, Netherlands, Norway, and the United Kingdom.

Since this covers two periods when a collective agreement to tie-up in order to preserve market stability was in force, the data have proved adequate for the needs of the report. They are unaffected by any misreporting as they concern sales through the market rather than landings officially reported via log-book data. In any case, misreporting is not thought to be a serious problem in this fishery.

Papers as far back as Ioannidis and Whitmarsh (1987) have reported co-integration as a problem in fish auction market data and applied corrections in the econometric investigations. However, tests for co-integration were not carried out in this study as the time period over which the model was estimated is less than three years and it is not possible for the influence of a shock to be identified in such a short time-span.

Similarly, the data were not corrected for inflation. Obtaining an index for daily inflation is impossible, and to interpolate rates on inflation on a daily basis between monthly statistics would risk introducing more errors than it removed.

Early investigations concentrated on finding an inverse demand curve for each of the species with the mean price obtained as the dependent variable and explanatory variables consisting of the own-quantities offered for sale, the mean size of fish offered, the cross-quantities within the three species, and seasonal dummies to represent both the day of the week and the month of the year. The price, quantity and size variables were in logs.

Data were aggregated to market level, requiring the national quantities to be summed and using the aggregated international mean prices and sizes. This was done because the intention was to understand and estimate the effect on the British and Dutch vessels, but the prices they receive will be determined by the action of the whole market and not their own action alone. Should they, for example, run out of quota and be unable themselves to land any fish the market will still continue but with supplies provided only by the vessels of the other nations which sell their catch though Urk. Thus while they are an important influence on the market they do not entirely control it.

Normally, to estimate a demand or inverse demand function in isolation runs into the problem of identification.

The observations represent market clearing points given both by the demand side and the supply schedule. Since each relates a relationship between price and quantity, in order to model the demand or inverse demand function it is necessary also to estimate the supply curve as part of a system or at least that instrumental variables representing the supply side are used.

However, in the case of fisheries such as the North Sea beam trawl fishery, the price is governed by supplies which are pre-determined in the short run. The market is more akin to Adam Smith’s market for primary produce with price determined by supply than the Keynesian notion of manufacturers facing a pre-determined price and adjusting supply accordingly.

Hence, it is possible to assume that in the short-run supply is perfectly inelastic. This identifies demand and ensures that all the observations represent points on the inverse demand curve. This practice is well established in estimating fisheries demand curves (see for example Ioannidis and Whitmarsh 1987, Jorgensen et al 1989, 1991).

To reduce the number of dependent variables and therefore the number of equations in the system, it was decided to adopt a weighting system for the quantities of each size category for each species and to sum them to produce a single independent variable for the quantities. This reduced the number of equations to be estimated from twelve to three.

The weights were obtained from regressing a curve through the points given by the definitions of each size category and estimating a mid-point. The weights used were as set out in Table 4.1.

Table 4.1: Weightings used in determining mean size of fish landed

	Size 1	Size 2	Size 3	Size 4	Size 5
Lemon Sole	0.984	0.457	0.258	-	-
Dover Sole	0.734	0.408	0.285	0.205	0.146
Plaice	0.921	0.490	0.327	0.221	-

The units are implicit kilograms.

To determine the effect on earnings it has been necessary first to calculate the relationship between the quantities offered for sale of the three species and the average price each achieved. The econometric modelling has achieved this and the results generally satisfy the statistical tests for reliability, though there appear to be some problems with outliers and with heteroskedasticity.

Initial investigation took place using ordinary least squares with the price and quantity variables in logs. Then, working from the general to the specific, insignificant variables were eliminated. However, before more complex attempts to find the coefficients could begin, the first simple attempts to estimate an inverse demand function in the form

proved inadequate.

The presence of the size variable proved destructive to the relationship between price and quantity in the lemon sole equation, giving a right-signed and significant coefficient for size, but a wrong-signed if significant coefficient for the quantity. The size variable was wrong-signed in the dover sole function and wrong-signed and insignificant in the plaice function.

The inverted demand curve for each of the three species confirmed the expected downward gradient of the curve in establishing the market price. However, tests on the data suggest the presence of heteroskedasticity, and outliers.

It became clear early on that the market has no memory. There are no lagged variables influencing the prices determined. This confirmed the belief that the use of a method which corrected for co-integration would have been inappropriate.

The use of lagged dependent variables to represent seasonality on a weekly and monthly basis was not possible as days when there were no sales disrupt the series. Therefore dummy variables were used instead but they offered no indication of seasonality in prices for any of the three species and although they gave a markedly improved coefficient of correlation corrected for degrees of freedom, the temptation to accept them in preference to a model showing the cross-elasticities was resisted.

At this point Zellner's Seemingly Unrelated Regression (SUR), a collapsed version of Three-Stage Least Squares, was used. The value of SUR is that, while the functions appear independent with no dependent variables appearing as explanatory variables in other equations within the whole system, the cross-correlations of the residuals, containing the unobserved influences on the markets, are preserved and this extra information is used to obtain more reliable estimates.

A problem with a systems method of estimation is that a complete set of observations is required for each time element. However, not all species were available everyday when at least one was available. This presents the researcher with the choice either of discarding observations when there were no supplies of one or more of the species with the consequent loss of information or of setting an arbitrary price for an observation of zero supplies. Despite reservations, the latter was chosen as the loss of information would have been considerable. Further, the arbitrary price selected and the double-log form of the function estimated suggest that the effect on the estimates would not be unacceptable.

Functions of the form

were estimated, where the P are the prices of the species, f=1 to F, the Q are the quantities offered for sale.

Various formulations to proxy for the influence of the mean size of fish offered were tried but none proved significant. The final selection of the cross-price elasticities was made using the method of the general to the specific, with insignificant relationships being dropped.

The use of a dummy variable to model the presence or not of the tie up scheme is not appropriate in this case. The reason is that the presence of the tie up scheme did not suggest a shift in the demand curve in a single direction. It can have been expected to lower landings early in each month within the period and then raise them towards the end when otherwise the fleet would have run out of quota. This complex reaction to the effect of the scheme is not readily amenable to being modelled by binary variables. It is not believed that the bad weather in 2002, which disrupted the landings to some extent, affected the result in that by tying up when they did supplies were restricted even if the weather frustrated them taking their quota.

The estimated equations for the three species are set out in Tables 4.2a-c.

Table 4.2a: Estimated inverse demand function for lemon sole

Lemon Sole	Parameter Estimate	t-Statistic and Significance
Constant Term	9.0590
Lemon Sole Own-Quantity	-0.9584	22.38**
Dover Sole Cross-Quantity	-0.0967	2.94**
R-bar-squared	0.763

Table 4.2b: Estimated inverse demand function for dover sole

Dover Sole	Parameter Estimate	t-Statistic and Significance
Constant Term	9.2750
Dover Sole Own-Quantity	-0.4586	16.70**
Plaice Cross-Quantity	-0.2426	9.79**
R-bar-squared	0.845

Table 4.2c: Estimated inverse demand function for plaice

Plaice	Parameter Estimate	t-Statistic and Significance
Constant Term	8.3960
Plaice Own-Quantity	-0.6998	52.37**
	0.826

** indicates significant at the 1% level.

The explanatory power of the estimates indicated by the is good, higher than is often the case for landings of fish. This is believed to be due to the quality of the data provided by Visafslag Urk and the fact that the data relate to auctions of fish through the market. There are no official statistics from any of the six countries whose vessels sold their catch through Urk which might have been corrupted by misreporting.

The results are interesting in that they suggest fairly isolated product markets. The price of each species reflects the availability of supplies with the expected inverse relationship duly appearing.

The markets for lemon sole and dover sole are affected by the availability of dover sole and plaice respectively but in the case of lemon sole the influence of dover sole is weak and the price of plaice is unaffected by the availability of the other species. Lemon sole has emerged almost as own-price inelastic which confirms its special position as destined for the catering trade.

The price elasticities implied by the estimates are set out in Table 4.3.

Table 4.3: Estimated own-price elasticities

Product	Own-Price Elasticity
Lemon Sole	1.043
Dover Sole	2.181
Plaice	1.429

These are long-run estimates and the absence of any memory in the market suggests that as such there are no short-run estimates or that the short-run elasticities are identical.

The Simulation Tests

The next stage was to compare, by simulation, what the model calculates were the earnings with the tie-up scheme in place to what it suggests would have been the earnings if it had not been. (It is necessary to let the model estimate the earnings under the scenarios of the tie up scheme being in place and not in place in order to be able to compare like-for-like when it is asked to estimate an alternative scenario.)

The precise details of the alternative scenario are, of course, speculative. The one chosen for this study has been to assume that the monthly sales of fish through the market by British and Dutch vessels in each month of the tie-up scheme’s operation would have been put through in the first three weeks of each month. This was achieved by raising the level of actual landings proportionately. Supplies from other countries through the market were left unchanged.

The alternative scenario is in fact probably best regarded as a limiting case in that it is unlikely that there would have been such a complete stop to fishing. Any of a myriad of causes could have prevented a vessel from fishing early in each month and to that extent the results found are exaggerated. Nevertheless, the results indicate that, except perhaps for lemon sole, there was a gain in earnings from the voluntary restraint and that the maximum it could have been was as in the Table 4.4 below.

Table 4.4: Aggregate Gains in Earnings Arising from the Tie-Up Scheme

	CPO	NSFO
2002	16%	15%
2004	12%	18%

The prices for each of the three species estimated by the model both with and without the tie-up scheme in place are set out in Table 4.5a-c.

Figure 4.5a: Estimated Prices of lemon sole with and without the tie-up scheme operating

Lemon Sole

Figure 4.5b: Estimated Prices of dover sole with and without the tie-up scheme operating

Dover Sole

Figure 4.5c: Estimated Prices of plaice with and without the tie-up scheme operating

Plaice

In the case of all three species examined, it can be seen that prices are more volatile when the tie up scheme was not operating.

The full results of the econometric investigation are set out in Appendix 2

Section 5 Statistical Analysis of the Impact of the Tie-Up Scheme using Costs, and Earnings Data

The econometric modelling sought to estimate the increased revenue from sales of fish that occurred as a result of the self-restraint imposed by the tie-up scheme compared to what might have happened otherwise.

However, revenue is only one side of the equation that produces profitability – which is the real objective of fishing; the other is costs.

To test whether profitability might have improved as a result of the scheme, it was decided to assemble costs and earnings data for a sample of vessels and to test whether a number of variables were significantly different when the scheme was in operation.

The variables tested were:

revenue,

costs,

profits

profits as a percentage of revenue,

crew share, and

crew share as a percentage of revenue

Costs and earnings data were obtained for a sample of nine vessels for the years 1999 to 2003, the latest available. Eight vessels were members of CPO and the ninth a member of NSFO, all fishing from Urk by beam trawling for flatfish in the North Sea.

In 1999 one vessel was omitted because of missing data and in 2002 one set of observations was omitted as the vessel was replaced and there had clearly been an unusual impact on the costs and earnings for the year. This appeared to have been partly due to a substantial period of inactivity with neither the old vessel nor the replacement vessel being available for fishing, resulting in a substantial amount of quota being leased, and partly because of the learning effect experienced with a new vessel.

Thus in the period 1999 to 2001 and 2003 there was a total sample of 35 observations and in 2002, the year when the tie-up scheme was in place, a sample of 8 observations.

The data were adjusted to correct for inflation using the EU Harmonised annual average consumer price index set out in Appendix 1.

One-tailed t-tests under the assumption that the variances were unequal were conducted on the means of the sample statistics, revenue, profits and profits as a percentage of revenue.

The alternative hypothesis, H₁, was that each was significantly higher at the 5% level in 2002 when the tie-up scheme was in place than in the remainder of the period 1999 to 2003 when it was not. The results are set out in Appendix 3.

For revenue, the t-statistic is not significant and we cannot reject the null hypothesis, H₀, that there is no difference between the revenue obtained in 2002 when the tie-up scheme was in force and in the other years.

This is not in conflict with the findings of the econometrics which estimated the impact on revenue had the tie-up scheme not been in force. Combined the two different methods of analysis suggest that if the tie-up scheme had not been in force in 2002 the level of revenue would have been noticeably lower than usual. This concurs with the impact of the bad weather experienced.

While the primary effect of the tie up scheme was expected to be on the stability of prices, a secondary effect may have occurred on the costs of the fleet. The impact could have been that the congestion cost experienced by the fleet was lowered.

Another possible effect was that the stock had grown by being allowed to remain in the sea a little longer and so the quota could be taken with less fishing. However, the t-tests suggest that costs although they were on average 12% lower in 2002 could not be considered statistically to be significantly below normal.

Profits are a sharper test as they represent the difference between the revenue and costs and are thus more volatile in the face of commercial pressures and, indeed, the t-test of profits confirms that the null hypothesis, that the level of profit in 2002 was the same as in other years, can be rejected. The evidence is quite strong with a likelihood of only about 1 in 30 of being incorrect.

A similar result was obtained for profits as a percentage of revenue.

It should be noted that it does not necessarily follow that the finding, of significantly higher profits during the year in which the tie-up scheme was in place, was brought about by the tie-up scheme itself. However, we are not readily able to offer any coincidental cause.

A possibility might have been abundant stocks but neither plaice nor sole stocks were out of the 1999 to 2003 range in 2002, as shown in Figure 5.1. There was however, a slight upward trend in stocks of plaice, but no discernible trend in the biomass of sole.

Figure 5.1: Stocks of North Sea plaice and sole 1999 to 2003

Source: ICES

It is possible that the bad weather in late 2002 had more influence in holding up prices but plaice and sole have not been reported as having elasticities which would increase revenue as a result of generally lower landings. The evidence from the econometrics is that improved stability of prices resulting from smoother landings has made the difference.

While these results suggest that profits were higher in 2002 than the other years tested, they do appear to support the view that the Dutch fleet was prevented from taking its quota owing to bad weather in the later part of the year. This conclusion is supported by the observation that revenue to the fleets was not found to be higher during 2002.

Section 6 Discussion and Inferences

The results found by the econometric testing corroborate the work of earlier studies that have found an inverse relationship between the amount of sole and plaice available and the prices achieved at auction. They also verify the finding that the price of both is fairly elastic but in this study it is relatively low. This confirms the importance of Urk as the centre of the market for sole and plaice in the North Sea.

The value of the elasticities estimated and the subsequent simulation of an alternative landings pattern indicate that the fleet did indeed benefit from the collective restraint in supplying fish to the market in a more orderly fashion than might otherwise have occurred. However, there are some weaknesses indicated by the statistical tests of the data. It was disappointing that a relationship between the price achieved and the average size of fish being offered for sale was not found, and likewise that there appears to be no seasonal influence on prices at either a daily or monthly level.

The statistical test on the costs and earnings data also indicate some benefit from the collective action to smooth the pattern of landings. While the tests give quite a strong indication that profits in the year of the tie-up scheme were better than they would have been in a free-for-all, the question of causation is not proven.

Individually the evidence of the econometrics and statistical tests is not strong. Indeed, that of the statistical tests need not necessarily be attributable to the tie-up scheme at all. However, combined, the two pieces of evidence do suggest that the tie-up scheme did result in higher revenues and profits than would have been available to the fleets if the vessels had not collaborated in the tie-up scheme.

Further testing with additional data for 2004 and 2005 when the data become available would serve to add further credence to the results.

Nevertheless, despite some statistical problems with the data, it is the conclusion of this study that the collective behaviour of the two fleets of beam trawlers belonging to CPO and NSFO members did achieve the objective of making gains from more stable markets, and that the processors, caterers, retailers and consumers of flatfish will have benefited accordingly.

Plaice in the North Sea spawn from January to mid-March, with the peak of the spawning season being in January and February. Deferring a proportion of fishing effort in these months to a period later in the year should allow a greater number of plaice to spawn successfully.

However, the stock/recruitment relationship for plaice is relatively flat as shown in Figure 6.1, and benefits to future recruitment through deferral of fishing effort would be extremely hard to demonstrate.

Figure 6.1: The stock/recruitment relationship for North Sea plaice

Source: ICES

However, the loss of condition of plaice during spawning season, with a consequent reduction in both meat yield and quality during the first three months of the year, should also mean that deferring fishing effort from this quarter leads to reduced numbers of fish being taken to fulfil a particular quota, as well as a greater unit value per fish caught. The possible benefits of this to the stock overall may well be worth more detailed investigation.

This collective behaviour of the beam trawl fleets is unusual in fisheries and has wider implications. Beam trawler owners have often been accused of being among the least responsible of fishing enterprises. This study suggests that if a good case can be made to them they are capable of leading the way if necessary towards responsible fishing.

There is a further inference that can be drawn that is of some importance throughout the fishing industry in Europe. This study suggests that in many fish markets, fleets could improve their own incomes by collective behaviour to stabilise markets through providing a steady supply rather than by racing to fish and then racing to the auction.

Whilst the European legislation relating to the common organisation of the market has given Producer Organisations the ability to impose certain marketing regulations on all vessels landing to that market, this ability has been seriously compromised by the huge number of additional blanket restrictions placed on the fishing industry in recent years.

It would be particularly unfortunate if the extreme measures imposed on the beam trawl fleet under the auspices of the Cod Recovery Plan prevent a similar tie up scheme from operating in future years.

Many of these serve only to deter Producer Organisations from taking on the responsibilities that they have been given. This is particularly the case when regulations are seen by fishermen to bear no reality to their understanding of their particular stock or market.

In many industries such collective behaviour might prove detrimental to the interests of the consumer, but in the particular circumstances in which fisheries are to be found operating, it appears to offer benefits to the fleets, processors, caterers, retailers and consumers alike, with no-one losing.

References

Archibald G.C. and R.G. Lipsey (1973) An Introduction to a Mathematical Treatment of Economics, 2^nd Edition, Weidenfeld and Nicholson, London.

Bell, F.W. (1968) The Pope and the Price of Fish, American Economic Review, December, pp 1346-1350.

Council Regulation (EC) No 2406/96 of 26 November 1996 laying down common marketing standards for certain fishery products, EU Official Journal, L 334, 23.12.1996, p 1, amended by Commission Regulation (EC) No 323/97 of 21 February 1997, L 52, 22.02.1997, p 8, Council Regulation (EC) No 2578/2000 of 17 November 2000, L 298, 25.11.2000, p1, and, Commision Regulation (EC) No 2495/2001 of 19 December 2001, L337, 20.12.2001, p23.

Dickey, D.A. and W.A. Fuller (1981) Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root, Econometrica, 49, 1057-1072.

Gates, J.M. (1974) Demand Price, Fish Size and the Price of Fish, Canadian Journal of Agricultural Economics, 22(3).

Guillotreau P, R. Hannesson, A. Hatcher and R. Lostado (1998) Foreign Trade and Seafood Prices: Implications for the CFP, Report for the European Commission DGXIV, LEN-CORRAIL, University of Nantes, July.

Hylleberg S, and G.E. Mizon (1989) Cointegration and Error Correction Mechanisms, Supplement to the Economic Journal, 1989.

ICES (2003) Working Group on the Assessment of Demersal Stocks in the North Sea and Skagerrak, Report for the International Council for the Exploration of the Sea, Copenhagen, October.

Ioannidis C, and D. Whitmarsh (1987) An Econometric Model of the Price Determination of Fresh Fish Landings in the United Kingdom, Economics Research Paper No 4, SFIA, Edinburgh.

Ioannidis C, and K. Matthews (1995) Determination of fresh cod and haddock margins in the UK 1988-1992, Seafish Report 3007, SFIA, Edinburgh.

Jaffry S, S. Pascoe and C. Robinson (1997) Long-run price flexibilities for high-valued species in the UK: a co-integration systems approach, CEMARE Research Paper No 111, Portsmouth.

Jorgensen H.P, P.E. Rodgers and J. Smit (1989) The Economic Impact of Fisheries Management in the North Sea: Cod, Haddock, Sole, and Plaice, Report for the EU Fisheries Directorate DGXIV, SFIA, Edinburgh.

Jorgensen H.P, P.E. Rodgers, J. Smit and G. Valatin (1991) The Economic Impact of Fisheries Management in the North Sea: Cod, Saithe, Whiting, and Plaice (Vol II), Report for EC Fisheries Directorate, SFIA, Edinburgh.

J M Keynes, J.M. (1936) General Theory of Employment, Interest and Money, Macmillan, London.

Marshall A. (1936) Principles of Economics, 8th edition, London.

MOR Research Limited (2001) Study of Quayside Demand for Demersal Fin Fish 1980-2000, SFIA, Edinburgh.

Nash, D.A, and F.W. Bell (1969) An Inventory of Demand Equations for Fishery Products, Working Paper No 10, US Dept of the Interior, Fish and Wildlife Service, Bureau of Commercial Fisheries.

Nielsen M, (1999) EU Seafood Markets - Integration and Demand, Working Paper No13/1999, SJFI, Copenhagen.

Perron P. (1989) The Great Crash, the Oil Price Shock, and the Unit Root Hypothesis, Econometrica, 57, 6 (November), pp 1361-1401.

Rodgers, P.E. (1987) The Imposition of a Plaice Box in the Eastern North Sea: An Econometric Appraisal of the Effects on the UK Market for Plaice, Economics Research Paper No 7, Sea Fish Industry Authority, Edinburgh.

Smith, A. (1776) An Inquiry into the Nature and Causes of the Wealth of Nations, The Adam Smith Institute, London, 2001, www.adamsmith.org/smith/won-intro.htm.

Stigler G.J. (1969) The Theory of Price, Macmillan, London.

Valatin G.B. (1995) A Survey of Skippers' Motivations, SFIA, Edinburgh.

Waugh, F.W, and V.J. Norton (1969) Some Analysis of Fish Prices, Bulletin No 401, University of Rhode Island Agricultural Experiment Station.

[1] This should not be taken as evidence of excessive over-capacity in the fleets and even less a measure of it. It is perfectly normal for an industry to work at less than full capacity. Operating at 35% less than full capacity is not at all uncommon in other industries and since these fleets both exist under efficient management regimes historic overcapacity is being shed and they can be expected to be approaching an industry optimum by economic means. That having been said, it is equally clear that the sharp falls in quota of recent years have left fleets needing a better quota to use their capital efficiently.