CONTENTS Executive Summary Introduction Global Haddock Supply Overview US-Canada Haddock Market Econometric Model Economic Forecast of Different Harvest Strategies Summary and Caveats References
Northeast Fisheries Science Center Reference Document 05-15
Economic Analysis of Alternative Harvest Strategies for Eastern Georges Bank Haddock [5Zjm; 551, 552, 561, 562]Eric M. Thunberg1, Doreen S.K. Liew2, Charles M. Fulcher1, and Jon K.T. Brodziak3
1National Marine Fisheries Serv., Woods Hole Lab., Social Sciences Branch, 166 Water St., Woods Hole, MA 02543
2Fisheries and Oceans Canada, Policy and Economics Branch, 6th Floor, 176 Portland Street, P.O. Box 1035, Dartmouth , Nova Scotia B2Y 4T3 Canada
3National Marine Fisheries Serv., Woods Hole Lab., Population Dynamics Branch, 166 Water St., Woods Hole, MA 02543
Web version posted October 31, 2005Citation: Thunberg EM, Liew DSK, Fulcher CM, Brodziak JKT. Economic analysis of alternative harvest strategies for eastern Georges Bank haddock [5Zjm; 551, 552, 561, 562]. US Dep Commer, Northeast Fish Sci Cent Ref Doc. 05-15; 42 p.
Information Quality Act Compliance: In accordance with section 515 of Public Law 106-554, the Northeast Fisheries Science Center completed both technical and policy reviews for this report. These predissemination reviews are on file at the NEFSC Editorial Office.EXECUTIVE SUMMARY -- This report examines the potential economic benefits of pursuing alternative harvest strategies for the Eastern Georges Bank haddock resource. Pursuit of this line of inquiry was prompted by the potential market impacts of the size of the 2003 year class of haddock; once thought to be 900 million fish but now assessed at 365 million age-1 fish. Even at this smaller size, the 2003 year class is large enough to have a material affect on haddock markets.
Estimation of economic effects on ex-vessel prices to United States and Canadian fishermen and effects on import quantities and prices was accomplished by applying an econometric model to projected landings under four alternative harvest strategies for Eastern Georges Bank haddock; constant catch of 30,000 mt, constant catch of 40,000 mt, constant catch of 50,000 mt, and constant fishing at FRef = 0.26. The econometric model was estimated using monthly data from 1989 to 2003 and was based on processor raw material demand for fresh whole haddock. Per guidance provided by the TRAC, projections of Eastern Georges Bank haddock landings were accomplished a using a stochastic approach. Specifically, realizations of recruitment were generated using two-stanza re-sampling from the empirical cumulative distribution functions below and above approximately 40,000 mt adult (ages 3+) biomass. Mean weights at age and partial recruitment were based on the most recent 3-year averages. Landings from the Western Georges Bank were calculated by subtracting projected Eastern Georges Bank haddock landings from landings projected for the entire stock area (Eastern and Western areas). The latter used stochastic projection methods based on advice from the Groundfish Assessment Review Meeting held in Woods Hole in August, 2005. Landings from non-Georges Bank sources of haddock were held constant at their most recent 3-year average.
The 30,000 mt constant catch strategy would be first achieved in 2007 and would be maintained through 2014. The 40,000 mt constant catch strategy would be first achieved in 2007 and would continue through 2012. The 50,000 mt constant catch strategy would be reached in 2007 and could be maintained through 2009. Of these strategies, harvesting at a constant FRef produced the highest present value of ex-vessel revenues in both the United States and in Canada and produced the highest present value of import sales of haddock from Canada to the United States. This finding was robust with respect to the choice of discount rate and assumed weights-at-age. Taking uncertainty over projected landings into account, also favored the constant FRef strategy as it more readily takes advantage of future recruitment events.
Of the alternative constant catch strategies evaluated herein the 50,000 mt constant catch strategy came closest to the revenue streams predicted for the constant FRef strategy. In fact, the projected landings between the two harvest strategies were quite similar since harvest at 50,000 mt would be sustained for only three years and there were only minor differences in accumulated benefit streams.1. INTRODUCTION
This report examines the potential economic benefits of pursuing alternative harvest strategies for the Eastern Georges Bank haddock resource. Pursuit of this line of inquiry was prompted by the potential market impacts of the size of the 2003 year class of haddock; once thought to be 900 million fish but now assessed at 365 million age-1 fish. Even at this smaller size, the 2003 year class is still considered to be very large. A harvest strategy that differs from that of current recommendations would have market effects on ex-vessel markets in both the United States and Canada and would affect trade in haddock products between the two countries.
In this report, we evaluate the impact of alternative harvest strategies for the Eastern Georges Bank haddock resource on ex-vessel prices in the United States and in Canada, and estimate how these changes would affect prices and import quantities of fresh whole haddock from Canada. The first section of the report provides an overview of global haddock supplies and the role of U.S. and Canadian haddock in domestic as well as international markets. The second section describes an econometric model developed to forecast how changes in haddock supplies affect import sales and revenues received by Canadian and U.S. vessels. The third section contains a discussion of several key assumptions and briefly describes how the econometric model was applied to a series of projected landings streams provided per guidance from the TRAC. The fourth section presents economic forecasts for four different harvest strategies; constant fishing mortality (FRef = FMSY = 0.26) and constant harvests of 30,000, 40,000, and 50,000 mt. Sensitivity analyses of the selected discount rate, the likelihood of achieving the highest benefit stream, and the effect of different average weights-at-age are presented in a fifth section. A final section provides conclusions, and discusses factors that were not possible to quantify yet may affect the economic impacts of a change in harvest strategy.
The landings streams produced under TRAC guidance for the purposes of this report were developed using a stochastic projection method. As such, the landings reported herein differ from those provided in the TRAC Status Report 2005/02. The TRAC used the stochastic projection method due to uncertainty over whether, and for how long, a given stream of constant harvest levels could be realized. A stochastic approach is better suited to this line of inquiry and has been used for exploratory purposes where a change in harvest strategy is being considered. The projected landings serve the purpose of providing input data only for the economic analysis (which itself is exploratory) and should not be used for purposes of quota setting. Additionally, where projections of landings were made (Eastern and Western Georges Bank resource areas) they were based on what would be allowable given the harvest strategy and prevailing resource conditions.2. Global Haddock Supply Overview
2.1 World Haddock Landings
There are two major haddock fishing regions in the world, the Northeast Atlantic/Arctic and the Northwest Atlantic (shaded red areas in Figure 1). In the Northeast Atlantic/Arctic, the main fishing areas are the Barents Sea, the Norwegian Sea, the North Sea, and the waters around Iceland, the Faroe Islands, and west of Scotland. In the Northwest Atlantic, the main fishing areas are on the Scotia Shelf (NAFO Areas 4TVW and 4X), Gulf of Maine (NAFO Area 5Y) Georges Bank (5ZE) and the Eastern United States.
The Northeast Atlantic/Arctic region has accounted for over 90% of the world haddock landings in recent years. In 2003, landings from the Northeast Atlantic/Arctic totalled about 252,000 mt (Table 1). This compares to about 23,000 mt caught from the Northwest Atlantic. World landings from 1950-2003 have ranged from a low of 190,000 mt (in 1992) to a high of 960,000 mt (in 1970), and averaged 427,000 mt (Figure 2). In 2003, landings of 275,200 mt were about 64% of the long term average. Until the mid- 1960’s, total landings from the Northwest Atlantic area were in the 100,000 to 200,000 mt range, and reached a high of 246,000 in 1965 (Figure 3). Since then, landings have been significantly lower and fell to a record-low of 7,300 mt in 1994. Landings in 2003 were about 23,000 mt.
2.2 United States – Canada Haddock MarketsCurrently, haddock landings from the Northwest Atlantic are mainly sold in the New England states and in Canada. In addition, landings from the Northeast Atlantic/Arctic are imported, making Canada and U.S. collectively a net importer of primary haddock products. Primary products in this context includes fresh or frozen whole or dressed fish, fillets or blocks and also salted or dried haddock.
CanadaSince 1994, Canada imports as much as it exports of primary haddock products, measured in dollar value (Figure 4). Imports are mostly frozen fillets and to a lesser extent, frozen whole fish. Exports are mainly fresh whole dressed fish. In 2003, Canada imported the equivalent of about 18,100 mt of haddock in round weight, while it exported the equivalent of about 12,500 mt, and thus was a net importer of primary haddock products. Estimated Canadian “consumption” of primary haddock products in 2003 was about 21,300 mt (landings of 15,700 mt plus net imports of 5,600 mt). Note that “consumption” here is of primary haddock product forms. Some of the primary products may have been sold directly to consumers or to processors to be further processed and exported. The flow of the secondary product is not traced here.
Canada imports mainly from the United Kingdom, Norway, Russia and China. Since 2000, China has emerged as a major player in imports of processed haddock products to Canada. In 1999, there were hardly any haddock imports from China. In 2000, it accounted for 14% of the haddock import value, and by 2003, that percentage had risen to 60%. These fish are predominantly imported in a frozen fillet form, but the country/countries of origin for the raw material supporting China’s exports are not known. Almost all of Canada’s haddock exports are to the U.S. Table 2 shows the imports and exports of haddock by country.
The United States is a net importer of primary haddock products and exports very little. Total imports in 2003 were about US$94.1 million, of which US$24.6 million were from Canada (Table 2).
Imports consist of fresh and frozen fillets, and also fresh and frozen whole fish. The U.S. imports the equivalent of about 52,000 mt of round fish while exports are negligible. U.S. landings of 6,800 mt plus the net imports of 52,000 mt puts the U.S. “consumption” at about 58,800 mt in 2003.
The U.S. imports mainly from Iceland, Canada and Norway. Canada supplies mostly fresh whole fish while Iceland supplies mainly fresh and frozen fillets. The 2003 value of U.S. imports by country and product form are tabulated in Table 3.
Overall size of the United States – Canada Market
The estimated overall size of the United States - Canada market was the equivalent of about 80,100 mt of round haddock in 2003. Of these, landings from the Northwest Atlantic were 22,556 mt (Canada, 15,771 mt and U.S., 6,785 mt). Landings from the Northwest Atlantic account for only 28% of the total Canadian and U.S. “consumption” of primary haddock products. This leaves net imports of about 72% or the equivalent of 57,600 mt of round fish from the landings of the Northeast Atlantic/Arctic to supply the overall United States - Canada market. Figure 5 shows the relative sizes of the Canadian and U.S. market. The total area shaded in yellow shows the estimated amount of haddock imported into the combined Canadian and U.S. market.
2.3 Resource Outlook
The major stock areas in the Northeast Atlantic/Arctic in 2003 are Iceland (comprising of 24% of total landings), Barents Sea (21%), North Sea (17%), Norwegian Sea (14%) and Faroe Islands (11%). Outlook for 2006 for some of the major stock areas are presented in Table 4. It appears that the only major stock area in the Northeast Atlantic/Arctic with a predicted increase in catches in 2006 is Iceland. This is somewhat offset by the slight declines in the Barents Sea, Norwegian Sea and Faroe Islands. Predicted allowable catches of haddock from the Western Georges Bank area in the U.S. are greater than the proposed TAC for the Eastern Georges Bank area. However, just as U.S. landings from the Eastern Georges Bank area have been well below allowable TACs due to management actions taken to protect other groundfish species in the U.S. EEZ, realized landings from the Western Georges Bank area are likely to be well below allowable levels.3. United States - Canada Haddock Market Econometric Model
3.1 Modeling Considerations
The economic implications of a change in haddock harvest strategy will depend on how international markets adjust to the change and how these adjustments are transmitted through the marketing chain to prices received in ex-vessel markets. Although no known studies of international markets for haddock have been conducted, available studies (Gordon and Hannesson 1996; Asche and Hannesson, 2002) of the cod market suggest that linkages between North American (United States and Canada) and European markets exist, but they are weak, while links within the two regional markets are strong. This means that the North American market for cod can be modeled separately from the European market and we assume that the same is likely to be true for haddock.
In developing a model of haddock markets between the United States and Canada, the own-price elasticity of demand measures how haddock price will respond to changes in quantities. If haddock demand is inelastic, then a proportional change in price will exceed an equi-proportional change in quantity and total revenues will go down. Conversely, total revenues will go up if haddock demand is elastic since a proportional change in price will be less than an equi-proportional change in quantity.
Modeling the United States and Canadian trade in groundfish products received considerable attention during the early to mid-1980’s as United States harvesters argued that Canadian imports were driving down ex-vessel prices. Given time constraints, an exhaustive review of this literature was not possible. Nevertheless, citations to many of these studies contained in both Felixson, Allen, and Storey (1987) and Hogan and Georgianna (1989) indicate that past modeling efforts have developed a system of equations to reflect supply and demand relationships at different market levels including imports, but focused on undifferentiated groundfish. Further, most of the cited econometric models focused only on processed products (fresh or frozen filets or frozen blocks).
Based on data presented in Section 2, the United States domestic market is assumed to be the primary market for domestic landings and for imported haddock from Canada, Iceland, and Norway. In evaluating the economic effect of different harvest strategies for haddock, the market of most interest is a raw material market. In effect, this market is a derived demand by United States processors for factor inputs. Within this context, it is important to determine the substitutability between domestic and imported whole haddock (fresh or frozen) and the impact that market substitution has on United States and Canadian ex-vessel prices. If United States processors substitute domestic landings for imported haddock, then an increase in United States landings would reduce the quantity demanded for Canadian raw material imports. Ex-vessel prices in both the United States and Canada would be expected to decline, but Canadian ex-vessel price may decline proportionally more due to the lowered demand for raw material imports.
The empirical model developed for this study was adapted from that of Hogan and Georgianna (1989). These authors estimated separate models for combined haddock and cod and for flatfish consisting of a three-equation system including import demand, import supply, and United States ex-vessel price. We made several modifications to Hogan and Georgianna’s original model to include a price equation for Canadian ex-vessel prices and to accommodate estimation issues encountered in developing a haddock-only model.
3.1 Econometric Model of the United States – Canada Haddock Market
Demand for Fresh Whole Canadian Imports
Processor demand for whole haddock imports was modeled as a function of the price of fresh whole haddock imports from Canada, the United States ex-vessel price of haddock lagged one period, the ex-vessel price of cod, United States domestic haddock landings, and a time trend.
The import price is expected to be negatively related to quantity demanded; as import prices go up, processor quantities demanded go down. The U.S. ex-vessel price of haddock is expected to be positively related to import demand; as domestic prices of haddock increase, processors substitute imports resulting in higher import demand. Note that our specification of the ex-vessel price variable in lagged form differs from that of Hogan and Georgianna. Although the demand equation was initially specified with the current U.S. ex-vessel price (as done by Hogan and Georgianna), it was not significant, whereas the lagged price was. The lagged price is likely reflective of some underlying adjustment process perhaps due to contractual obligations.
The ex-vessel price of cod is included to reflect demand for processed products and is a substitute for fresh whole haddock. The price of cod is expected to be positively related to import quantities; as the price of cod increases processors substitute away from cod to haddock increasing the demand for raw material imports.
Import demand is expected to be negatively related to the quantities of haddock landed by U.S. vessels; as available domestic landings go up, processors substitute away from imported haddock. Note that this variable also differs from that of Hogan and Georgianna. In their study, Hogan and Georgianna measure the substitution effect by constructing a measure of excess capacity. Implicit in their specification is the assumption that processors prefer domestic sources of haddock, importing only when domestic landings are insufficient to meet raw material requirements. In our study we make no assumptions as to the desirability of U.S. or imported haddock. Further, since the excess capacity variable in Hogan and Georgianna’s model was defined as the difference between capacity (measured using a modified peak-to-peak method) and U.S. landings, our specification should detect the same general effect. The expected sign for the time trend is indeterminate.
Supply of Fresh Whole Canadian Imports
Canadian fresh whole supply was specified as a function of fresh whole import price, Canadian haddock landings, Canadian haddock landings lagged one period, and the price of alternative product forms (frozen blocks and fresh and frozen filets). Import supply is expected to be positively related to import price as well as the quantity of Canadian landings. The sign of Canadian landings lagged one period is indeterminate but was included to reflect the possible presence of an adjustment process.
In Hogan and Georgianna’s model, Canadian importers were assumed to have a number of alternative markets for cod and haddock products. To reflect these alternatives they included a separate price series for each product form: frozen blocks, fresh filets, and frozen filets. Following Hogan and Georgianna, a haddock-only price series for each product was constructed from import data but none of these variables were found to be statistically significant. Further, there were a number of occasions where imported quantities were zero in a given month. Therefore, we estimated a price series based on a weighted average for all product forms and used this to capture the potential diversion of whole fresh haddock into products processed in Canada. The expected sign for this variable is negative; as the price of alternative products increases, Canadian exports of fresh whole haddock decreases.
United States Ex-Vessel Price
The U.S. ex-vessel price was specified as a function of the quantity of fresh whole imports, the quantity of U.S. landings, the ex-vessel price of cod, and the ex-vessel price of haddock lagged one period. Import quantity is expected to have a negative effect on U.S. ex-vessel price; as processors import more raw materials, the demand for U.S. raw material declines and ex-vessel prices decline. Ex-vessel price is expected to be negatively related to domestic landings; as landings increase, market-clearing prices decline. The expected sign of the ex-vessel price lagged one period is expected to be positive reflecting some stickiness or inertia in price determination. The ex-vessel price of cod is expected to be positively related to ex-vessel price; as cod prices increase, demand for haddock increases as processors substitute away from cod to haddock.
Canadian Ex-Vessel Price
Following Hogan and Georgiana, demand and supply of fresh whole imports of haddock were assumed to be simultaneously determined in a market clearing process. The Canadian ex-vessel price of haddock was assumed to be exogenously determined by the quantity of Canadian landings, the quantity of fresh whole exports to the United States, the Canadian ex-vessel price lagged one period, and the price of alternative products handled by Canadian processors. The quantity of Canadian landings is expected to be negatively related to ex-vessel price; as landings go up market-clearing prices go down. Ex-vessel price is expected to be positively related to export quantities; as export demand increases, ex-vessel prices increase. The expected sign of the Canadian ex-vessel price lagged one period is positive. As was the case for U.S. ex-vessel prices the lagged effect is intended to reflect some inertia or adjustment period in price determination. The expected sign for the price of alternative processed products is expected to be positive; as the value of alternative higher-valued processed product markets increases, Canadian ex-vessel prices increase.
The system of equations described above was estimated using monthly data obtained from several source (see Appendix B) for calendar years 1989 through 2003. These years were selected due to constraints on the ability to obtain reliable import quantities of haddock prior to the conversion in 1989 to a 10-digit harmonized code. In previous years most haddock was combined with a grouping of species including cod, pollock, and hakes and so the data were not useable. Data on monthly U.S. landings in live weight and value were obtained from the Northeast region dealer weighout data. Data on monthly Canadian landings in live weight and value were obtained from the Department of Fisheries and Oceans in Canada. Import quantities in product weight and values were obtained from NMFS headquarters Fisheries Statistics Division. These data are purchased by the Division from the Foreign Trade Division of the United States Census Bureau. All price data were expressed in nominal terms converted to U.S. dollars. Also, since both United States and Canada landings were measured in live weight, ex-vessel prices are expressed as dollars per pound live weight. By contrast, import quantities were measured in product weight so import prices are expressed in product weight.
The supply and demand system was estimated using two-stage least squares. All but one of the estimated parameters (the intercept in the Canadian ex-vessel price equation) were statistically significant (Table 5). The signs of all variables were consistent with theoretical expectations; all own-price relationships were negative in the import demand and ex-vessel price equations, and positive in the import supply equation. Similarly, all substitution effects were positive as were variables reflecting demand for United States processed products.
The F-test of all variables being simultaneously equal to zero was rejected for each of the four estimated equations. The adjusted R-square values for the import supply and Canadian ex-vessel price indicate that these equations fit the data reasonably well. However, the adjusted R-square values for the import demand and U.S. ex-vessel price indicate that though the signs of the model parameters are consistent with expectations, these models are estimated with considerably more error, perhaps due to some form of unaccounted for specification or measurement error. Model performance (see Appendix A for a more detailed discussion) over the time series suggests that haddock markets have undergone some structural changes that have not been completely captured, although the model does appear to reasonably capture contemporary market conditions. The reliability of model forecasts is uncertain as potential supplies of haddock may lie outside the range of observed data. However, even though the point estimates of model predictions are subject to uncertainty, underlying structural relationships capturing market behavior would be unaffected. This means that the econometric model is still likely to produce reasonably reliable relative or ordinal rankings of alternative harvest strategies affecting aggregate haddock markets.
The system of equations was specified in a double-log form so that the coefficients are interpretable as elasticities. The estimated import price elasticity of demand is quite high (-11.6) indicating that United States processor demand for imported whole fish from Canada is very responsive to the raw material price. By contrast, the substitution elasticity for United States domestic haddock landings is inelastic (-0.27) suggesting that processor demand for Canadian raw material imports is not particularly responsive to domestic landings.
The own-price elasticity in the supply equation is elastic indicating that Canadian exporters are responsive to changes in import price. Similarly, Canadian landings are responsive to the import price. The price of alternative products that may be processed in Canada is negatively related to supply of whole fresh haddock imports, but the proportional effect is less than unity.
The price equations for the United States and Canadian ex-vessel markets were specified as price-dependent demand which means that the estimated parameters should be interpreted as price flexibilities which under some conditions are theoretically equivalent to the inverse of the price elasticity. A price flexibility less than one is interpreted in the same manner as a price elasticity greater than one. The own-price flexibilities for both the Canadian and United States ex-vessel demand are less than one suggesting that prices respond proportionally less than quantities supplied, so that total ex-vessel revenues may be expected to increase even though prices decline. However, the Canadian own-price flexibility is larger (-0.23) than that of the United States (-0.06) suggesting that a proportional increase in United States landings will have a proportionally lower impact on ex-vessel prices than would be the case for an equi-proportional increase in Canadian landings. The negative substitution elasticity for imported haddock in the United States ex-vessel demand suggests that the availability of imports has a price dampening effect on prices received by United States harvesters.
Based on the Eastern Georges Bank haddock TAC of 23,000 mt for calendar year 2005, and assuming haddock from other U.S. and Canada sources remain at their recent average levels, the predicted ex-vessel price in Canada would be $0.56 per pound in U.S. dollars. Applying the current exchange rate the harvester price would be $0.67 per pound or $1,477 per metric ton. In the U.S. the predicted ex-vessel price would be $1.07 per pound and the import price would be $0.77 per pound in live weight.4. Economic Forecast of Different Harvest Strategies
4.1 Proposed Harvest Strategies for Eastern Georges Bank Haddock
Four alternative harvest strategies for
Eastern Georges Bankhaddock were evaluated. These alternatives were a constant harvest strategy of 30,000, 40,000 and 50,000 mt and a harvest strategy of constant fishing mortality set equal to FRef=0.26. Projected catches of eastern Georges Bankhaddock were based on the most recent stock assessment (van Eeckhaute and Brodziak 2005). Initial stock size at age in 2005 was estimated with a sequential population analysis. Variability in the 2005 estimate of stock was characterized using 1000 bias-corrected bootstrap realizations. The resulting bootstrap distribution of the initial stock size was used for projections to account for uncertainty in the initial stock size. Recruitment was simulated using a two-state spawning biomass-dependent recruitment function, similar to the recruitment model used for the entire Georges Bankhaddock stock (Brodziak et al. 2001, Brodziak et al. 2002). A cutoff of 40,000 mt was used to define the high and low spawning biomass states, as recommended by the Transboundary Resource Assessment Committee [TRAC]. This partitioned the set of observed recruitments into high and low spawning biomass states for simulating recruitment. As a result, projections of future catch streams accounted for both uncertainty in initial stock size and uncertainty in future recruitment (Brodziak et al. 1998).
The 3-year average of fishery selectivities and weights at age during 2002-2004 were used to characterize the projected catch under each harvesting scenario as recommended by the TRAC. The distribution of projected catches from 100 simulations for each bootstrap initial stock size were used to generate revenue streams for economic analyses. Since the 2003 year class is also a dominant feature in the
Western Georges Bankportion of the haddock resource, a set of projections from this source was also produced based on the harvest strategy adopted by the New England Fishery Management Council in 2004. Note that all projected landings are based on what could be landed at prescribed levels of fishing mortality rates and may not necessarily reflect realized landings given constraints imposed by management action taken to protect other stocks. For example, the U.S.portion of the Eastern Georges Bank area was closed effective August 26, 2005because the Georges Bankcod TAC had been taken. This means that the 2005 haddock TAC from this area will not be taken. Adjustments to management measures in the U.S. U.S.portion of the resource sharing area coupled with ongoing gear research may enable the to take its share of the TAC in the future. U.S.
A second set of projections were provided based on the possibility that environmental conditions would return to the longer term average. These projections were run using long term average weights-at-age (from 1987-2003), while retaining the same recent partial recruitment pattern. The economic effects of alternative harvest strategies under this environmental scenario were treated as a sensitivity analysis for purposes of this report. Procedures to evaluate the economic implications of the different harvest strategy are detailed below.
4.2 Procedures for Estimating Economic Effects of Different Harvest Strategies
The econometric model of haddock raw material market described previously generates an estimate of (a) the monthly haddock import price, (b) import quantity, (c) the Canadian ex-vessel price, and (d) the U.S. ex-vessel price. The economic model includes several exogenous variables that may affect any one of these endogenous but developing forecasts of these exogenous variables was outside the scope of analysis so they were held constant. These exogenous variables included the monthly pattern of landings, the U.S. ex-vessel price of cod, and the price of processed haddock imports (Table 6).
As projected landings were provided on an annual basis, the annual time-step was converted to a monthly time step to match the economic model. Since there was no reason to believe that the proposed harvest strategies would fundamentally alter the seasonal pattern of landings, annual projected landings were multiplied by the most recent 3-year average monthly share of total landings  . Further, since the reduced form equations (Table 7) in the market model include a one-month lag for U.S. haddock landings, Canadian haddock landings, and Canadian ex-vessel price, the 3-year December average for each of these variables was used for the first period of the economic forecasts. Predicted values for all subsequent periods were obtained by substitution of projected landings into the reduced form equations. These monthly values were then summed by year to obtain an annual estimate of import sales, import quantities, and ex-vessel revenues to Canadian and U.S. vessels.
Under the existing US/Canada resource sharing agreement, TACs for allocating the haddock resource between the two countries are established on an annual basis. This resource share was estimated to be 34/66, 33/67, and 34/66 percent for 2004, 2005 and 2006 respectively (Gavaris, Mayo, and O’Brien, 2005). Given the recent stability in the resource shares the most recent estimate of resource shares (66% Canadian and 34% U.S.) was assumed to remain constant for the period of analysis (2005 to 2014).
4.3 Results of Economic Projections
In addition to landing haddock from the Eastern Georges Bank area, U.S. vessels also land haddock from the Western Georges Bank area and the
Gulf of Maine. Canadian vessels also land haddock from non-Georges Bank stocks. Landings from these sources contribute to the overall haddock raw material market and affect import demand as well as ex-vessel prices in the U.S. and Canada. At this time, neither the U.S. nor Canada is considering a change in harvest strategy for any of these alternative haddock resources so the landings from these sources were invariant across all harvest alternatives for the Eastern Georges Bank area (Table 8)  . For this reason, reported results will focus on landings and revenue streams from the Eastern Georges Bank haddock resource, although landings and revenue totals for both countries will also be reported for completeness. Results include the median values of the projected landings by stock area, predicted prices, predicted total sales for imported Canadian whole fresh haddock, and predicted total ex-vessel revenues to Canadian and U.S. harvesters. Estimates of gross sales are reported in both nominal as well as present value terms by applying a discount rate of 7.0%. All quantities and prices are reported in live weight. These results provide an ordinal ranking of harvest strategies in terms of the expected present value of the benefit stream. Since calendar year 2005 is nearing completion, all forecasted results will be reported beginning in calendar year 2006 through 2014.
Constant Harvest Strategy 1: 30,000 mt
A constant harvest strategy of 30,000 mt would be first achieved in calendar year 2007 and would remain constant through 2014 (Table 9). This strategy would result in a potential annual catch of 43.7 million pounds by Canadian fishermen and 22.5 million pounds by U.S. fishermen. Given additional haddock supplies from other haddock resources total U.S. haddock supplies would be 74 million pounds in 2006, increase to 149 million pounds then gradually decline to 72 million pounds in 2014. This pattern of total landings is due to the projected allowable landings from the Western Georges Bank haddock resource. For reasons noted previously, realized landings from this resource are likely to be lower. In Canada, total landings would increase from 46.7 million pounds to 63 million pounds.
Given that the economic model predicts prices and quantities on a monthly time step, annual average prices were computed as the average of monthly predicted price and annual were computed by summing predicted values across months. Due to increased supplies, average annual live weight prices decline from 2006 to 2008 by $0.05 per pound in the U.S. ex-vessel market as aggregate haddock supplies peak in 2008 (Table 10). As aggregate haddock supplies decline, the U.S. ex-vessel price rises to $0.96 per pound in 2014. The average ex-vessel live weight price in Canada was predicted to initially decline from $0.53 per pound in 2006 to $0.52 per pound in 2008. Over the longer term, predicted ex-vessel price in Canada increases to $0.55 per pound due to the influence of both a decline in aggregate haddock supplies and to a small positive time trend. Similarly, the import price was predicted to decline initially from $0.75 to $0.74 per pound but recover to be above its 2006 level by 2014. Predicted imports from Canada increase from 38 million pounds in 2006 to 50 million pounds in 2014; again due to the effect of a positive time trend. Note that total import quantities may come from any source in Canada so the total presented in Table 10 does not represent imports only from the Eastern Georges Bank resource.
The total value of import sales increases from $28 million in 2006 to $38 million in 2014. Discounted at a rate of 7% the present value of the cumulative import sales was estimated to be $216 million (Table 11). Even though U.S. ex-vessel prices were predicted to go down, ex-vessel revenues from sales of haddock from the Eastern Georges Bank resource were predicted to increase from $19 million in 2006 to $22 million in 2014. This increase is due to the fact that haddock demand is elastic such that price goes down proportionally less than the increase in quantity. The present value of US ex-vessel revenue from the Eastern Georges Bank haddock resource was estimated to be $188 million and present value of total harvest revenue from all U.S. sources of haddock were estimated to be $641 million. Predicted returns from the Eastern Georges Bank resource to harvesters in Canada increased from $19 million in 2006 to $24 million in 2014. The present value of Canada revenues from the Eastern Georges Bank resource over the 9 years was estimated to be $147 million while discounted revenues in Canada from all sources of haddock were $214 million.
Constant Harvest Strategy 2: 40,000 mt
A 40,000 mt constant harvest strategy of haddock taken from the Eastern Georges Bank area is achieved in 2007 and is sustainable until 2012 (Table 12). In 2013 projected median landings decline to 38,000 mt and decline to 33,000 mt in 2014. Total Canadian catches from the Eastern Georges Bank haddock resource peak at 58.2 million pounds then decline to 48.5 million pounds in 2014. Potential catches to U.S vessels peak at 30 million pounds then decline to 25 million pounds in 2014.
From 2006 through 2014 average annual ex-vessel prices in the U.S. and Canada rise and fall with changes in total harvested supplies (Table 13). That is, predicted prices decline in 2006, 2007, and 2008 with the rise in landings from the Eastern Georges Bank resource. As landings decline in 2013 and 2014, predicted prices increase. Import prices follow a similar pattern as import quantities follow the same general pattern as landings from the Eastern Georges Bank resource even though not imports do come from other sources of haddock.
As noted previously the value of harvest revenues from the Eastern Georges Bank haddock resource increases in both countries in every year through 2012 even though average ex-vessel price is lower than the 2006 predicted price (Table 14). The present value of harvest revenue to vessels in Canada was estimated to be $182 million from the Eastern Georges Bank resource and a total of $248 million from all sources of haddock. In the U.S. discounted revenues from the Eastern Georges Bank resource were estimated to be $167 million and a cumulative total of $662 million from the entire Georges Bank and Gulf of
Maine stock areas.
Constant Harvest Strategy 3: 50,000 mt
A constant harvest of 50,000 mt of haddock from the Eastern Georges Bank area is achieved in 2007 and is sustainable for three consecutive years (2007 through 2009) before declining gradually from 47,000 mt in 2010 to 30,000 mt in 2014 (Table 15). Canadian landings from the Eastern Georges Bank resource peak at 72.7 million from 2007 through 2009 and decline to 43.5 million pounds in 2014. Similarly, U.S. haddock harvest from the resource peaks at 37 million pounds then declines to 22 million pounds in 2014. Total landings from all sources in Canada peak at 92 million pounds then decline to 63 million pounds. Aggregate U.S. harvest peaks at 164 million pounds then declines to 72 million pounds in 2014.
Predicted average ex-vessel prices and import price rise and fall with the change in harvested quantities (Table 16). As was the case for other harvest strategies, predicted prices in 2006 and 2007 decline as harvested quantities from the Eastern Georges Bank resource increase. Average prices then stabilize with constant harvests in 2007 through 2009. However, as landed quantities decline, predicted average prices rise through 2014.
As was the case for the 40,000 mt constant harvest strategy, decreases in prices are more than offset by increases in landings resulting in rising ex-vessel revenues from the Eastern Georges Bank resource in both Canada and the U.S. at least through 2009 (Table 17). Similarly the volume of import quantities more than offsets declines in import price resulting in higher value of sales through 2009. By contrast, lower landings from 2010 to 2014 are not offset by price increases so total value of import sales and harvest revenues declines. Applying a discount rate of 7% to annual sales, results in a present value of $262 million. The present value of Canadian and U.S. ex-vessel revenue from the Eastern Georges Bank resource was estimated to be $198 million and $249 million respectively.
Current Harvest Strategy: FRef = FMSY = 0.26
Employing a constant fishing mortality harvest strategy of FRef = FMSY = 0.26 for haddock in the Eastern Georges Bank area would result in an increase in potential landings from this area to 64,000 mt in 2008. Landings in 2009 would decline to 51,000 mt and would further decline in every year thereafter to 28,000 mt in 2014 (Table 18). The overall pattern of landings over time is similar to that of the 50,000 mt constant harvest strategy except that by landing fewer fish in 2008, the latter provides for elevated landings over the final five years, although by 2014 the difference between the two strategies is only 2,000 mt.
Predicted average prices rise and fall with projected haddock landings (Table 19). Predicted ex-vessel prices and import price are lowest in 2008 when landings peak. Given roughly equivalent landings in 2007 and 2009, predicted prices and estimates of import quantities are also similar in those two years although both import price and import quantity are higher in 2009 as compared to 2007. This small difference is due to the time trend which has a positive influence on both import price and import quantity. From 2010 onward predicted imports as well as ex-vessel prices increase as landed quantities decline.
Revenues to harvesters from the Eastern Georges Bank haddock resource in Canada peak at $45 million in 2008 then decline steadily with resource abundance to $22 million in 2014 (Table 20). The present value of these revenue streams was estimated to be $201 million. The present value of harvest revenue from all sources of haddock in Canada was estimated to be $266 million. Predicted commercial fishing revenue from Eastern Georges Bank haddock in the U.S. also peaked in 2008 at $41 million then declined over the remainder of the projection period to $20 million in 2014. The present value of Eastern Georges Bank haddock sales in the U.S. was estimated to be $184 million and was estimated to be $673 million in sales from all U.S. sources of haddock.
Harvest Strategy Comparison
Cumulative projected landings are highest (384,000 mt) for the FRef harvest strategy but by only 2,000 mt when compared to a constant catch harvest strategy of 50,000 mt (Table 21). Although the 30,000 mt constant harvest strategy would be sustainable from 2007 through 2014, it also results in the lowest cumulative yield (284,000 mt). The 40,000 mt constant harvest strategy has the third lowest cumulative yield (355,000 mt); a difference of 27,000 and 29,000 mt respectively as compared to the 50,000 mt constant harvest and FRef strategies.
On an annual basis, the 30,000 mt constant harvest strategy produces estimates of yield that exceed any other strategy in only the terminal year and only compared to the FRef strategy. The 40,000 mt harvest strategy produces annual yields lower than either the 50,000 mt or FRef harvest strategies through 2011 but produces higher yield in each of the last three years of the projection period. As noted earlier, the 50,000 mt and FRef harvest strategies produce annual yields that are not markedly different from one another with the exception of 2008 where FRef yield is greater by 16,000 mt.
Compared to other harvest strategies predicted cumulative import demand for fresh whole haddock from Canada is largest for the 50,000 mt and constant FRef harvest strategies (Table 22). Further, both cumulative import quantities and value of sales for these two harvest strategies are virtually identical. As noted above, the 30,000 mt harvest strategy does not produce higher import sales than other alternatives until the terminal year of the projection period. Similarly, the 40,000 mt harvest strategy produces lower value of import sales in years up to 2011 but does produce higher import sales from 2012 through 2014. The present value of import sales for the 50,000 mt harvest strategy exceeds that of the 40,000 mt strategy by $13 million and the constant FRef strategy exceeds the 40,000 mt strategy by $15 million.
Predicted total ex-vessel revenues (cumulative 2006 to 2014) to both U.S. and Canadian harvesters is greatest under a constant FRef fishing strategy although the cumulative difference between the constant FRef and constant harvest strategy of 50,000 mt is no more than $1 million in nominal terms or $3 million in present value (Table 23). On an annual basis, the comparative stream of harvest revenues follows the same pattern as noted previously for landings and for imports. Specifically, the 30,000 mt harvest strategy produces lowest catches in all years except 2014 and the 40,000 mt strategy produces lower revenues from 2007 through 2010 but higher revenues from 2011 onward.
4.4 Sensitivity Analysis
The present value of benefits from any one harvest strategy will depend on the selected discount rate, the variability or likelihood of achieving higher or lower benefits, and the effect on projected yields of assumed average weights-at-age.
The sensitivity of the ordinal ranking of the harvest strategies was examined by applying discount rates of 3%, 5%, 7%, and 9% to the annual nominal values for import sales, Canadian ex-vessel revenues, and U.S. ex-vessel revenues (Table 24). Although present values decline at higher discount rates, the ordinal ranking of harvest strategies was unaffected by the discount rate. That is, at all tested interest rates the constant FRef harvest strategy produced highest present value followed by the 50,000, 40,000, and 30,000 mt constant harvest strategies. Note that in all cases the difference between the constant FRef and 50,000 mt constant harvest strategies is slight.
Variability in Projected Yield
A stochastic projection allows for consideration of variability in predicted catches due to uncertainty in recruitment. That is, catch in any given year may be at or near some average level or could be well above or below average because recruitment in prior years can affect the sustainability of any given harvest level or harvest strategy. To examine how potential present value of harvest revenues may be affected by this uncertainty, the present value of gross harvest revenue was calculated for different percentiles of the realized landings streams for each harvest strategy (Figure 6). In Figure 6 combined harvest revenues for U.S. and Canadian vessels are reported for convenience because separate plots of each value displayed the same pattern and are interpreted the same way. The values shown at 50% probability are equivalent to the median values reported in the results section for each harvest strategy. However, the cumulative probability means that there is a 50% probability that the present value of harvest revenues will be equal to the median or less. For example, there is a 50% chance that the 30,000 mt harvest strategy will yield a present value of approximately $283 million (U.S.) or less. By contrast, there is roughly a 10% probability that any of the other harvest strategies will be produce less than $283 million (U.S.).
Figure 6 illustrates that at least up to the 50th percentile the cumulative probability distributions for the 50,000 mt constant harvest and constant FRef harvest strategies are virtually identical. The two distributions diverge at higher percentiles because the potential harvest revenue for the 50,000 mt constant harvest strategy is bounded by the TAC. This is also true of the 30,000 and 40,000 mt constant harvest strategies, which is why their cumulative probability distributions are nearly vertical. In essence, the constant harvest strategies may be well suited to take advantage of a particular recruitment event, but may not be as well suited to take advantage of future recruitment events.
Alternative Environmental Conditions
The TRAC provided a set of baseline projections using recent three-year average weights-at-age. These projections reflect current environmental conditions that have produced lower weight-at-age than the observed longer term average and are considered to be reflective of conditions that may be expected to prevail at least over the medium term. The TRAC also conducted a set of projections to examine how yields might differ if environmental conditions (and growth rates in particular) were to return to their long term average. The resulting projections indicate that the 30,000 mt constant harvest strategy would be achieved one year earlier than under the baseline condition and would be sustainable for all subsequent years of the projection period (Table 25). Compared to the baseline, the 40,000 mt harvest strategy would be achieved in the same year (2007) but would be sustainable through 2014 instead of 2012. Similarly, a 50,000 mt constant harvest would be achieved in the same year as that of the baseline projection but would be sustainable for three additional years. The constant FRef strategy would peak at 75,000 mt instead of 64,000 mt and the difference between in cumulative total landings between constant FRef and the 50,000 mt harvest strategy would be much larger.
As the projected landings in Table 25 indicate, should environmental conditions become more favorable, any given constant harvest strategy would be sustainable over a longer period of time as compared to the baseline projections. However, the favorable conditions would prevail for all harvest strategies and the ordinal ranking of harvest strategy under baseline conditions would be unaffected. Specifically, in terms of present value of discounted import sales and harvest revenues in both countries, the FRef strategy produces highest gross benefits regardless of the selected discount rate (Table 26). Further, the separation between the FRef strategy and its next best alternative (a 50,000 mt constant harvest strategy) is larger than was the case under baseline environmental conditions.
 Note that the most recent three-year average may not reflect the future monthly pattern of landings from the Eastern Georges Bank resource sharing area, because closures in both the U.S. and Canada would tend to distort seasonal landings patterns. However, unless seasonality is expected to continuously change over the time period of analysis, any alternative assumption of constant monthly landings shares would have no affect on the ordinal ranking across alternative harvest strategies.
 Landings from both the Gulf of Maine and Non-Georges Bank Canadian haddock resources were held constant at their most recent three-year average. Although landings from these resource areas are likely to change, any alternative assumed time stream of landings would still be applied across all harvest scenarios for the Eastern Georges Bank resource sharing area. This means that even though predicted prices and revenues would be different, the ordinal ranking across alternative harvest strategies would be unaffected.5. Summary and Caveats
The findings of this study demonstrate that haddock markets create interdependencies between the United States and Canada such that changes in management strategy will affect haddock trade and will have impacts on prices received by fishermen in both countries. Of the harvest strategies considered in this report, harvesting at a constant FRef produced the highest present value of ex-vessel revenues in both the United States and in Canada and produced the highest present value of import sales of haddock from Canada to the United States. This finding was robust with respect to the choice of discount rate and assumed weights-at-age. Taking uncertainty over projected landings into account, also favored the constant FRef strategy as it more readily takes advantage of future recruitment events.
Of the alternative constant catch strategies evaluated herein the 50,000 mt constant catch strategy comes closest to the revenue streams predicted for the constant FRef strategy. In fact, the projected landings between the two harvest strategies were quite similar and there were only minor differences in accumulated benefit streams. Throughout the report a number of caveats have been mentioned. These caveats are reiterated below.
As with any fitted statistical model, predictions will be more reliable when applied to conditions that are within the range of observed data. Even though the size of the 2003 haddock year class is less than once thought, landings are still projected to increase to levels that exceed the range of observed data used to estimate the econometric model. The potential directionality or magnitude of any forecast error is not known with certainty.
In this study, only haddock supplies from domestic fisheries in the United States and Canada were included. However, over the past several years both countries have been importing increasing supplies of haddock from Iceland and Norway, and in recent years Canada has imported growing amount of processed products from China. The role of these import supplies in U.S. and Canadian markets was not explicitly modeled because of data limitations principally due to missing observations over the time series used to develop the econometric model. With greater available domestic supplies, imports from these and other countries may decline as processors substitute away from imports and buy higher quantities of domestic haddock. Even if this is the case, the presence or opportunity to source haddock from other countries is likely to have some price dampening effect that would be transmitted down the marketing chain to processors and ex-vessel markets. Ultimately, the ability of domestic processors in the U.S. and Canada to compete with imported processed products will depend on cost and production efficiencies relative to import prices and the cost of shipping.
Where landings projections were made available (i.e. haddock from both Eastern and Western Georges Bank resource areas), they were based on what would be allowable under any of the four harvest strategies evaluated for this report. This does not necessarily mean that these landings or TAC levels would actually be realized. For example, the recent closure of the U.S. portion of the Eastern Georges Bank resource sharing area because of the U.S. cod TAC had been reached means that the U.S. TAC for Eastern Georges Bank haddock will not be taken. In the context of this study, 2005 import and ex-vessel prices would be underestimated as realized market supplies of haddock will be lower than projected. Given the comparatively low cod TAC, bycatch rates of cod in the haddock fishery may make any one of the allowable levels of constant catch or constant FRef harvest strategies difficult to achieve.
The difference between the ex-vessel “scrod” and “large” haddock market categories in the U.S. averaged about $0.14 per pound between 1990 and 2003; a premium of approximately 25% of the scrod price. Unfortunately, the majority of haddock sold to U.S. dealers does not identify the market category and these data were not available in Canada so a price premium for larger fish was not included in the estimated price models. Including a price premium could affect the choice of harvest strategy because at lower harvest rates the proportion of larger more valuable fish in the exploitable population would increase compared to a harvest strategy where harvest rates were higher. Further exploration of this issue is not possible at this time given present data limitations and modeling approach.REFERENCES
Asche, F. and R. Hannesson. 2002. Allocation of Fish Between Markets and Product Forms. Marine Resource Economics 17:225-238.
Brodziak, J. K. T., W. J. Overholtz, and P. J. Rago. 2001. Does spawning stock affect recruitment of
New England groundfish? Can.J. Fish. Aquat. Sci. 58:306-318.
Brodziak, J., M. Thompson, and R. Brown. 2002.
Georges Bank haddock. In NEFSC, Assessment of 20 northeast groundfish stocks through 2001, pp. 36-59. NEFSC Ref. Doc. 02-16, 509 p. Available at: http://nefsc.noaa.gov/publications/crd/crd0216/
Brodziak, J., P. Rago, and R. Conser. 1998. A general approach for making short-term stochastic projections from an age-structured fisheries assessment model. In F. Funk, T. Quinn II, J. Heifetz, J. Ianelli, J. Powers, J. Schweigert, P. Sullivan, and C.-I. Zhang (Eds.), Proceedings of the International Symposium on Fishery Stock Assessment Models for the 21st Century.
Alaska Sea Grant College Program, Univ. of Alaska, Fairbanks.
Eeckhaute, L.V., and J. Brodziak. 2005 in press. Assessment of eastern
Georges Bank haddock. Canadian Science Advisory Secretariat Research Document.
Felixson, T., P. G. Allen, D. A. Storey. 1987. An Econometric Model of the Market for Fresh
New England Groundfish with Emphasis on the Role of Canadian Imports. Northeast Journal of Agricultural and Resource Economics 16(1):24-34.
Gavaris, S., R. Mayo, and L. O’Brien. 2005. Update of Allocation Shares for
Canadaand the USAof the Transboundary Resources of Atlantic Cod, Haddock, and Yellowtail Flounder on Georges Bank Through Fishing Year 2006. TRAC Reference Document 2005/06, 24 p.
Gordon, D. V. and R. Hannesson. 1996. On Prices of Fresh and Frozen Cod Fish in European and
U.S.Markets. Marine Resource Economics 11(4):223-238
Hogan, W. and D. Georgianna. 1989.
U.S.Fish Processing Capacity and Imports of Whole Groundfish from Canada. Marine Resource Economics 6(3):213-225.