- Email: [email protected]
International shark fin markets and shark management: an integrated market preference–cohort analysis of the blacktip shark (Carcharhinus limbatus)
Ecological Economics 40 (2002) 117– 130 www.elsevier.com/locate/ecolecon
ANALYSIS
International shark fin markets and shark management: an integrated market preference –cohort analysis of the blacktip shark (Carcharhinus limbatus) Quentin S.W. Fong a,*, James L. Anderson b,1 a
Marine Ad6isory Program/Fishery Industrial Technology Center, School of Fisheries and Ocean Sciences, Uni6ersity of Alaska Fairbanks, 118 Trident Way, Kodiak, AK 99615, USA b Department of En6ironmental and Natural Resource Economics, Coastal Institute, Uni6ersity of Rhode Island, 1 Greenhouse Road, Kingston, RI 02881, USA Received 15 November 2000; received in revised form 10 October 2001; accepted 31 October 2001
Abstract The increasing demand for shark fins in Asia, and the publicity resulting from finning and discarding live sharks, has generated concern regarding the sustainability of the world’s shark populations. These concerns can be attributed to the shark’s life history, which is characterized by a pattern of slow growth, late maturity, few offspring, and long life, making populations vulnerable to overexploitation. Once overexploited, shark stocks will be slow to recover due to these constraints. Despite an increase in consumption and trade of shark fins and other shark products, and the vulnerability of shark populations once overexploited, little effort has been expended to understand the biology and economics of sharks and shark fisheries until recently. This study adds to the understanding of linkages between shark product markets, specifically shark fins, and the biology of shark populations by explicitly incorporating multi-attribute market information into bioeconomic modeling. Results from conjoint analysis of the Hong Kong dried, processed end-user markets is incorporated into a blacktip shark (Carcharhinus limbatus) cohort model to estimate the optimal harvest size and age that maximize economic value. Results show that optimal harvest sizes and ages for all mortality and discount factor scenarios are greater than the maturation sizes and ages for both male and female blacktip. Policy implications for this study are also discussed. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Shark fin market; Shark management; Bioeconomic model
1. Introduction
* Corresponding author. Tel./fax: + 1-907-486-1516. E-mail addresses: [email protected] (Q.S.W. Fong), [email protected] (J.L. Anderson). 1 Tel.: + 1-401-874-4568.
Bioeconomic models utilize an integrated economic and biological systems approach to evaluate the performance of fishery resources using different management strategies. Traditionally, most bioeconomic analysis in the fisheries man-
0921-8009/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 0 0 9 ( 0 1 ) 0 0 2 7 3 - 7
118
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
agement literature examines issues such as extraction rates and fleet size and/or capacity with simplistic assumptions about market behavior. Demand functions are estimated using highly aggregated data to generate price estimates. Little attention has been given to using realistic market information, especially when considering the multi-attribute nature of fishery products (e.g. Carroll et al., 2001). This is important, because in the economics of the international food marketing system, which includes seafood, it is consumers who are the driving force for product selection and consumption (Schaffner et al., 1998). While lack of understanding of consumer tastes and preferences may prevent successful marketing of fishery products by producers, the misunderstanding of consumer product markets by policymakers may promote ineffective fishery management schemes. This can result in welfare losses to all resource users (e.g. Homans and Wilen, 1997). The incorporation of product characteristics in bioeconomic analysis for fishery management was investigated by Gates (1974), who illustrated the importance of product size as a function of market price for ex-vessel demand analysis. Subsequently, several bioeconomic studies have incorporated size-dependent market prices in examining optimal resource management strategies (e.g. Thunberg et al., 1998). The consideration of size as a product characteristic that affects market price has certainly added a realistic dimension to fishery management analysis. However, in certain fisheries, particularly those with foreign ethnic markets such as bluefin tuna and shark fins, the use of product size alone may not be sufficient for demand and bioeconomic analysis. Indeed, recent marketing studies of seafood products using various multi-attribute utility approaches have demonstrated that size is only one of several product attributes that end-users evaluate (e.g. Zucker and Anderson, 1998). For example, results from hedonic price analysis of fresh North Atlantic bluefin tuna (Thunnus thynnus), show that fat content, color, shape, freshness, and size are significantly correlated to ex-vessel price (Carroll et al., 2001). McConnell and Strand (2000) found similar results from tuna in Hawaii.
Until recently, none of the aforementioned seafood marketing studies has explicitly incorporated multi-attribute market values into capturedbased fishery bioeconomic analysis. Larkin and Sylvia (1999) explicitly incorporate intrinsic fish quality into a standard bioeconomic fisheries model for Pacific whiting that includes the harvesting and processing sectors. Intraseasonal price for Pacific whiting is estimated by a multi-attribute, seemingly unrelated regression model that incorporates flesh composition, product form, and hedonic price function for surimi products. This research also explicitly incorporates multiattribute market information to bioeconomic modeling. Here, a framework that incorporates market information of a shark product, shark fin, is merged with the biological growth function of a shark is presented. Specifically, the fishery management objective of harvesting shark fins that are of the most preferred quality to maximize economic return to society is investigated. This is achieved by incorporating the results of a conjoint analysis of dried processed shark fin in Hong Kong into a bioeconomic model of the blacktip shark. The optimal harvest size and age that maximizes economic value of the shark fin set (caudal, dorsal, and two pectoral fins) for a single cohort of blacktip shark under different biological and economic scenarios is estimated.
2. Background Shark fin, a product that is traditionally consumed in Hong Kong, Singapore, Macao, China, and other countries with large ethnic Chinese populations, is one of the most valuable food items in the world. For instance, in 1998, the average price for dried processed caudal fins 25.4 cm (10 in.) in length was US$ 415.00 retail in Hong Kong (Fong and Anderson, 2000). As a consequence of liberalization and increasing spending power of the Asian middle class, the demand for shark fins has increased significantly. For instance, Hong Kong, a trader, processor, and consumer of shark fins, and the most important market for shark fins in the world, increased shark fin imports more than 214% from 2648 mt
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
in 1985 to 8323 mt in 1998 (Vannuccini, 1999; Hong Kong Census and Statistics Department, 2001). Similarly, shark fin imports by Thailand increased 42% from 97 to 138 mt (Food and Agriculture Organization, 2001). The increasing demand for shark fins in Asia, and the publicity resulting from finning and discarding live sharks, have generated concern regarding the sustainability of the world’s shark populations. These concerns are due to the nature of the shark’s life cycle, which makes them vulnerable to overexploitation (Holden, 1977). Once overexploited, shark stocks are slow to recover. Other biological factors, such as schooling by age, sex, and reproductive state, also make some shark species (e.g. blue shark, Prionacae glauca) highly vulnerable to overfishing. High fishing mortality may deplete certain segments of the age class, which may significantly affect the reproductive dynamics of shark populations (Anonymous, 1996). Despite an increase in consumption and trade of shark fins and other shark products, and the vulnerability of shark populations once overexploited, relatively little effort has been spent to understand the biology and economics of sharks and shark fisheries until recently (e.g. Pascoe et al., 1992). This lack of research in the biology and economics of sharks may stem from the traditionally small scale of shark fisheries relative to other fisheries, a lack of understanding of ethnic markets for shark products, and the incidental bycatch nature of many shark fisheries. Further, little attention has been paid by domestic and international fishery management institutions to the management of shark stocks. Bonfil (1994) found that only three (Australia, New Zealand, and the US) out of 26 countries with reported annual shark landings of over 10 000 mt have domestic shark research programs and management plans. Few existing economic studies concerning sharks address the linkage between shark products and shark harvest management. Pascoe et al. (1992) developed a bioeconomic model to estimate the effects of different management options for the southern shark fishery in Australia. They measured the price flexibilities of shark meat in
119
the Melbourne market to get an indication of how prices change with changing quantity. However, they could not make any conclusive statements from this portion of the study since imports and domestically caught sharks, which do not enter the Melbourne market, were not included in the analysis. Thus, the authors treat the price of shark meat as exogenous and assume that price remains constant over the simulated 30-year period. This is a restrictive and unrealistic assumption. This study adds to the understanding of linkages between shark product markets, specifically shark fins and shark biology.
3. Bioeconomic model The overall structure of the model is presented in Fig. 1. First, the biological growth of an individual blacktip shark is modeled with respect to the length and weight of three fin types— caudal, dorsal, and pectoral. Results of a shark fin preference analysis, conjoint analysis, is applied to calculate the utility index of the dried, processed fin set as a function of blacktip shark growth. Finally, the harvest size (age) for a blacktip shark
Fig. 1. Blacktip shark market – cohort model.
120
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
cohort that maximizes the Hong Kong shark fin buyer’s utility for dried, processed fins under different mortality rates and discount rates is estimated.
3.1. Biological component Age and growth estimates for blacktip shark of both sexes are represented by a von Bertalanffy growth function: TLt = L [1− e − K(t − t0)]
(1)
where TL is the total length of the blacktip in centimeters; L the attainable maximum size, is 176 cm total length; K the rate that approaches L is 0.27; and t0 is the age at which the fish would have been zero size is − 1.20 year (Branstetter, 1987). The subscript t represents age in quarters; t=1.0, 1.25, 1.50, …, 30, assuming the blacktip has an average life expectancy of 30 years. The total length equation is then converted to pre-caudal length by using the following equation estimated by Castro (1996) in millimeters: PCLt = − 23.1+0.74TLt
(2)
where PCL is the pre-caudal length of a blacktip (mm) and TL is the total length. To estimate the functional relationships between shark and fin growth in terms of fin size and weight, data from Al-Quasmi (1994) and measurements from a commercial sample of blacktip shark fins are utilized. These relationships are estimated by a series of regressions that linked dried, processed fin size to the von Bertalanfy growth function of blacktip shark. The first step of the relationship is estimated by the relationship between fresh dorsal fin and pre-caudal length of the blacktip shark as reported by AlQuasmi (1994):
3.1.1. Caudal fin 3.1.1.1. Fin size estimate. The relationship between fresh caudal fin and fresh dorsal fin is represented by: ln(MidCaut ) = 0.2670 (1.21)+0.8777 (12.81)ln(FDFt )
FDFt =0.5524 (0.33)+ 0.1608 (14.85)PCLt R 2 =0.83
ship between dried, processed shark fins and blacktip shark growth. First, fin length measurements are made from the tip of the fin to the middle base of the fin and along the anterior edge of the shark fin. These measurements are done because Hong Kong shark fin buyers use the anterior edge measurement as an indicator of fin size while Al-Quasmi (1994) uses Food and Agriculture Organization/World Health Organization’s standard (FAO/WHO, 1987). The weight, in grams, of the dried, unprocessed shark fin samples is also recorded. These measurements are used to estimate the relationship between the length and weight of dried, unprocessed shark fins. Dried, unprocessed shark fins of each type are processed into the end-user product form— dried, processed fins. Before final processing, dried, unprocessed fins are rehydrated until fin lengths became constant. Length measurements of the rehydrated fins are used as a proxy for fresh shark fins to estimate the relationship between fresh and dried unprocessed fin lengths. After processing the fin sets into final dried, processed form, length measurements were taken from the processed fins and used to calculate the conversion ratios of dry, unprocessed fins to dried, processed fins. Weight conversion ratios between dried, unprocessed and dried, processed shark fins were calculated following Al-Quasmi (1994) (Tables 18, 19). The following sections describe the estimated linkages by fin type.
(3)
where FDFt is the fresh dorsal fin size measured from the shark fin tip to the middle base of the fin. Numbers in parenthesis represents the t-ratio. A sample of 36 sets of dried, unprocessed blacktip shark fins was obtained from Guyana to complete the estimation of the functional relation-
R 2 = 0.79
(4)
where ln denotes natural logarithmic transformation; MidCau is the middle length measurement of the fresh caudal fin; FDF is the middle length measurement of the fresh dorsal fin. The relationship between the fresh and dried unprocessed fin is represented by:
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
DryMidCaut =0.7999 (0.81)
121
sented by: DryMidDort = 0.7486 (1.03)
+ 0.8768 (17.27)MidCaut R 2 =0.97
+0.8877 (20.94) FDFt
(5)
where DryMidCau is the middle length measurement of the dried caudal fin. The relationship between the two length measurements of the dried caudal fin is represented by: DryOutCaut =3.17596 (1.64)
DryOutDort = − 0.8986 (−0.28)
+ 1.009 (9.33)DryMidCaut R 2 =0.91
(6)
where DryOutCau is the anterior length measurement of the dried caudal fin. The relationship between dried caudal fin and dried processed caudal fin based on five samples is represented by: DryProcOutCaut = 0.96DryOutCaut
R 2 = 0.98 (10) where DryMidDor is the middle length measurement of the dried caudal fin; FDF is the middle length measurement of the fresh dorsal fin. The relationship between the two length measurements of the dried dorsal fin is represented by:
(7)
+1.4841 (7.51)DryMidDort R 2 = 0.90
(11)
where DryOutDor is the anterior length measurement of the dried dorsal fin. The relationship between dried caudal fin and dried processed caudal fins based on five samples is represented by: DryProcOutDort = 0.96DryOutDort
(12)
where DryProcOutCau represents the anterior length measurement of the dried processed caudal fin.
where DryProcOutDor represents the anterior length measurement of the dried processed dorsal fin.
3.1.1.2. Fin weight estimate. The relationship between the weight and length of the dried caudal fin is represented by:
3.1.2.2. Fin weight estimate. The relationship between the weight and length of the dried dorsal fin is represented by:
ln(DryCaugmt )
ln(DryDorgmt )
= − 4.5691 (− 3.87) + 2.8660 (7.445)ln(DryOutCaut )
= − 5.2522 (−5.96) 2
R =0.87 (8)
+3.0628 (10.82)ln(DryOutDort )
R 2 = 0.95 (13)
where DryCaugm is the dried caudal fin weight in grams. The weight relationship between dried caudal fins and processed dried caudal fin based on five samples is represented by:
where DryDorgm is the dried dorsal fin weight in grams. The weight relationship between dried caudal fins and processed dried caudal fin based on five samples is represented by:
DryCauKgt =0.74DryCauKgt
DryProcDorKgt = 0.530DryDorKgt
(9)
(14)
where DryCauKg is the dried processed caudal fin in kilograms.
where DryDorKg is the dried processed dorsal fin in kilograms.
3.1.2. Dorsal fin
3.1.3. Pectoral fin
3.1.2.1. Fin size estimate. The relationship between the fresh and dried unprocessed fin is repre-
3.1.3.1. Fin size estimate. The relationship between fresh pectoral fin and fresh dorsal fin is
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
122
represented by:
samples is represented by:
ln(MidPect )
DryProcPecKgt = 0.42DryPecKgt
=1.1582 (6.35)+ 0.7438 (13.09)ln(MidDort )
where DryPecKg is the dried processed pectoral fin in kilograms.
2
(15)
R =0.79
where MidPec is the middle length measurement of the fresh pectoral fin. The relationship between the fresh and dried unprocessed fin is represented by:
(20)
3.2. Utility index for an indi6idual shark
where DryOutPec is the anterior length measurement of the dried pectoral fin. The relationship between dried pectoral fin and dried processed pectoral fins based on five samples is represented by:
3.2.1. Conjoint analysis and consumer choice A market preference model, conjoint analysis, is used to determine the utility of the shark fin set to Hong Kong shark fin importers/processors as a function of blacktip shark growth. Conjoint analysis is a form of multi-attribute utility model, which all, or in part, link to the notion that utility is derived from the attributes that the good possesses (e.g. Lancaster, 1966). It is assumed that the utility Hong Kong shark fin importers/processors obtained from a specific shark fin product is a function of the utility derived directly from the product’s attributes and levels of those attributes and indirectly from the profits associated with the product’s attributes (Lancaster, 1966). For example, a Hong Kong shark fin buyer may prefer medium-sized dried, processed dorsal shark fin to large-sized dried, processed pectoral shark fin. The utility derived from a given product may then be expressed in general form as a quasi-concave, twice continuously differentiable utility function:
DryProcOutPect = 0.96DryOutPect
U(sh )= U{Xh ; y(Xh )}
DryMidPect = − 1.1327 (−1.37) + 1.0006 (25.80)MidPect R 2 =0.99
(16)
where DryMidPec is the middle length measurement of the dried pectoral fin. The relationship between the two length measurements of the dried pectoral fin is represented by: DryOutPect = 3.2687 (2.18) +1.1607 (15.61)DryMidPect R 2 =0.97
(17)
(18)
where DryProcOutPec represents the anterior length measurement of the dried processed pectoral fin.
3.1.3.2. Fin weight estimate. The relationship between the weight and length of the dried pectoral fin is represented by: ln(DryPecgmt ) = − 4.7353 (− 7.16) + 2.7462 (13.56)ln(DryOutPect )
R 2 =0.87 (19)
where DryPecgm is the dried pectoral fin weight in grams. The weight relationship between dried pectoral fins and processed dried pectoral fin based on five
(21)
where U(sh ) is the utility the buyer derives from the hth composite dried, processed shark fin product sh ; Xh is a vector of levels making up the composite product sh ; y(Xh ) is the profit function associated with the product’s attributes. Since a decision maker obtains some degree of satisfaction from each product, the alternative selected for consumption would be the one that provides the highest satisfaction. For example, a shark fin buyer would choose product s4 over product s2, only if U(s4) is greater than U(s2). However, the utility of the shark fin importer/processor is not directly observable and is unknown. The utilities, therefore, are treated as random variables, and the probability of choosing alternative dried, processed shark fin product s4 over s2 is equal to the probability that U(s4) is greater than U(s2) (Manski, 1977).
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
3.2.2. Conjoint analysis model specification and estimation Conjoint analysis of dried, processed shark fin was conducted with Hong Kong shark fin importer/processors. This method uses field experiments by asking respondents to rank or rate products with predetermined attributes and levels of attributes to measure the buyer’s preference or utility as the dependent variable (Green and Srinavasan, 1990). Here, the conjoint analysis evaluates the utility function of Hong Kong shark fin importers/processors directly by asking respondents to rate a set of stimuli from 0 to 10, with 0 being the least preferred, and 10 being the most preferred. In this case, a reduced design of 11 dried, processed shark fins was obtained by using an asymmetrical factorial orthogonal experimental plan (Addelman, 1962). The attributes included were fin size and type. The conjoint model employed in this research uses the traditional non-interaction-effect model, which is assumed to be additive in levels of the attributes (e.g. Green and Srinavasan, 1990): U(sh )= i%ij x (h) ij + mij
mij N(0, 1)
L(h, v) m
= % % Rq,y log(Lq,y (vj − 1 −i%X) q = 1y = 1
−Lq,y − 1(vj − i%X))
Table 1 Results of conjoint model estimation (ordered logit) Variable
Coefficient
S.E.
T-ratio
Constant Size Dorsal Pectoral v(1) v(2) v(3) v(4) v(5) v(6) v(7) v(8) v(9)
2.78 2.32 −8.36 −13.11 3.22 6.16 7.68 10.18 14.26 17.31 18.91 20.43 23.76
0.66 0.23 0.89 1.38 1.72 0.75 0.88 1.85 1.76 1.82 1.90 1.99 2.84
4.20** 10.19** −9.41** −9.47** 1.87* 8.26** 9.34** 5.50** 8.12** 9.49** 9.93** 10.26** 8.38**
Log likelihood function =−192.58; N =187; Restricted log likelihood=−448.41; 2 =511.66**; **, significant at the 0.01% level; *, significant at the 10% level.
individuals in the experiment, and m is the number of stimulus in the conjoint experiment. Maximizing L(h, v) provides estimates of the parameters h and v (McKelvey and Zvoina, 1975).
(22)
where U(sh ) is the random utility that an individual derives from hth product, iij is the parameter matrix that represents the relative importance of the levels, x (h) ij represents the deterministic independent variable matrix associated with attribute j and level i for product h, and mij is the random error term. An ordered logit model was used to analyze the rating data generated by the conjoint experiment. For an independent sample of n individuals, the log likelihood function, L(h, v), is: n
123
(23)
where L(·) is the logistic distribution function e(·)/1 +e(·); v’s are the unknown threshold variables to be estimated with h where j= 0, …, 9; i is the matrix for the coefficients h; and X is the matrix for the independent variables constant, fin size, dorsal fin, and pectoral fin; n is the number of
3.2.3. Utility index formulation Maximizing the log likelihood function in Eq. (23) provides estimates for h, the coefficients of the independent variables and v, the threshold levels between ratings. The estimated equation is: Usc = 2.8+ 2.4Sz− 8.3Dor− 13.1Pec (24) where Usc is the utility score for the three fin types, caudal, dorsal, and pectoral, at various fin sizes; Sz is fin size; Dor is dorsal fin, and Pec is pectoral fin. Both Dor and Pec are coded in dummy variables. All estimated coefficients are significant at the 0.01% level (Table 1). The estimated utility score Eq. (24) and the estimated threshold level (v) are then used to calculate the probability of a dried, processed shark fin being rated in a certain category (e.g. rating= 10) for a given fin size and fin type. The ordered logit model specification captures the preference structure for dried, processed shark fin by a representative shark fin processor/importer in Hong Kong. It is assumed that the shark fin processor/importer assigned ratings to product profiles in the conjoint experiment relative to his/ her most desired product profile. Thus the specific
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
124
utility score for an index is calculated from the logistic probability function for the most preferred rating (rating=10), the estimated utility score (Usc) from Eq. (24), and the estimated v9 (23.759; significant at 0.01%) that represents the lower bound threshold level for the most preferred rating. The formula for the utility index, which represents the probability of being the most preferred dried, processed shark fin product is: UWi,Sz =1−
e[23.7 − (2.8 + 2.4Sz − 8.3Dor − 13.1Pec)] 1+e[23.7 − (2.7 + 2.4Sz − 8.3Dor − 13.1Pec)] (25)
where UWi is the utility per unit weight for the three fin types; i is fin type, and 23.7 is the estimated lower bound threshold level for the most preferred rating from the estimated ordered logit model. The utility index for fin type, i, is calculated as the product of the utility per unit weight, UWi, and the dried, processed weight, DPWi, for fin type i: UIi =UWi ×DPWi
(27)
where TUI is the total utility index for an individual blacktip shark.
3.3. Utility index for cohort To calculate the utility index for the blacktip shark, the initial population of the cohort is assumed to be 10 000, with both sexes combined. The quarterly numbers-at-age for the blacktip shark cohort is: Nt + 1 = Nt · e − M/4
Mt =
1.92 − 0.25 Wt 4
(28)
where Nt is the number of sharks at age t, expressed in quarters; and M is the natural mortality rate. Three quarterly natural mortality rates, 0.025, 0.050, 0.075, and a natural mortality function
(29)
where Mt is the quarterly natural mortality, and Wt is the dry weight of individual shark in grams, assuming dry weight is 0.2 of wet weight. This mortality function simulates the decrease in natural mortality as the size of a shark increases with age in a cohort. The weight of an individual blacktip shark is determined by: WKGt = (2.51×10 − 9)TLM3.12 t
(30)
where WKGt is the wet weight of an individual blacktip shark (kg), and TLMt is total shark length (mm) (Castro, 1996). The total utility of a cohort using the utility index approach by conjoint analysis is represented by:
(26)
where UIi is the utility index for fin type i. The total utility index for an individual blacktip is the sum of the utility indexes of the three fin types, taking into account that sharks have one caudal, one dorsal, and two pectoral fins: TUI =% UIi
proposed by Peterson and Wroblewski (1984) are used for sensitivity analysis. The Peterson and Wroblewski (1984) natural mortality function is:
TUCt =
TUIt × Nt (1+r)t
(31)
where TUCt is the total utility index for the cohort at age t; TUIt the total utility index for an individual shark; Nt the number of sharks in the cohort; and r is the discount rate, which is set at 0, 0.02, 0.03, 0.05, 0.07, 0.1, and 0.2, respectively.
4. Optimal harvest Results from a multi-attribute marketing analysis were incorporated into a market preference–cohort model of the blacktip shark. The optimal harvest size of the blacktip shark was investigated for the conjoint preference–cohort model under four natural mortality scenarios. Within each natural mortality scenario, the effects of seven discount factors were also simulated. These results are presented in Table 2 and Figs. 2–5. Three quarterly natural mortality parameters, 0.025, 0.05, and 0.075, are used to determine the optimal harvest size/age of the blacktip shark. Results show that as quarterly natural mortality increases from 0.025 to 0.075 at any given discount
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
rate, the optimal harvest size/age for the blacktip shark decreases. For example, at a discount rate of 0.03, the optimal harvest size estimated with the conjoint market– cohort model decreases from 171.88 (12.50 years of age) to 169.34 cm (10.75 years), then to 166.62 cm (9.50 years) as the quarterly mortality rate increases from 0.025 to 0.075 (Table 2 and Figs. 2– 4). The performance of the conjoint market– cohort model using a size-dependent natural mortality function is also investigated (Peterson and Wroblewski, 1984). This function assumes that as the size of an individual shark increases with age (expressed in weight), the natural mortality rate for the cohort decreases. This assumption is an improvement in realism over the constant mortal-
ity scenarios, since a shark cohort of a small-size class (i.e. younger age) would be more vulnerable to predation than a cohort of a large-size class. Results show that the size-dependent mortality conjoint market–cohort model provides the least conservative optimal harvest sizes/ages of all mortality scenarios (Table 2 and Fig. 5). For example, at zero discount rate, the optimal harvest size/age for the size-dependent mortality scenario is 169.34 cm (10.75 years), as opposed to 172.86 cm (13.50 years), 170.19 cm (11.25 years), 167.24 cm (9.75 years) for 0.025, 0.050, and 0.075 constant quarterly mortality rates, respectively. Seven discount rates, ranging from 0 to 20%, are used to examine optimal harvest size and age of the blacktip shark. These rates are used to
Table 2 Optimal harvest for conjoint market–cohort model Natural mortalitya
Discount rate (%)
Total utility index
Optimal harvest sizeb (cm)/age (years)
0.025
0 2 3 5 7 10 20 0 2 3 5 7 10 20 0 2 3 5 7 10 20 0 2 3 5 7 10 20
86.01 66.29 58.56 46.18 36.89 26.87 10.71 25.49 20.51 18.47 15.08 12.42 9.41 4.15 8.99 7.42 6.76 5.64 4.74 3.69 1.76 10.77 8.72 7.84 6.47 5.36 4.09 1.84
172.86/13.50 172.15/12.75 171.88/12.50 171.27/12.00 170.94/11.75 170.19/11.25 167.82/10.00 170.19/11.25 169.34/10.75 169.34/10.75 168.87/10.25 168.36/10.25 167.24/9.75 165.24/9.00 167.24/9.75 166.62/9.50 166.62/9.50 165.95/9.25 165.24/9.00 164.48/8.75 162.79/8.25 169.34/10.75 168.87/10.50 167.82/10.00 167.82/10.00 167.24/9.75 166.62/9.50 164.48/8.75
0.050
0.075
P & Wc
a
In quarters. Total length. c Peterson and Wroblewski (1984). b
125
126
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
Fig. 2. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality=0.025 per quarter).
simulate the divergence between the social and private opportunity cost of capital, time reference, and risk premium. Real discount rates between 0 and 5% have been suggested as an appropriate social discount rate for the 30-year horizon (Clark, 1990). The differences can be attributed to the differences in risk premium perceptions. Results show that in all scenarios, size (age) of optimal shark harvest decreases as discount rate increases (Table 2). For example, given a discount rate of 3%, the optimal harvest size and age for blacktip under a size-dependent natural mortality conjoint market–cohort simulation is 167.82 cm (10.00 years of age). Alternately, the optimal harvest size and age for size-dependent natural mortality given a 20% discount rate is 164.48 cm (8.75 years of age), 2 years younger than the case with a lower discount rate (Fig. 5).
5. Summary and conclusions The increasing demand for shark products, including shark fins, and the life-history pattern of long living, late maturity, and low-reproductive potential of sharks have generated concerns regarding the health of the world’s shark stocks. These concerns are generated not only by regulatory agencies and non-governmental organizations but also from resource users. A survey of shark fin importers/processors in Hong Kong, the center for shark fin trade and consumption in the world, has shown that more than 41% of the respondents expressed concerns of overharvesting of sharks (Fong and Anderson, 2000). Responding to these concerns, an International Plan of Action for Conservation and Management of Sharks (IPOA-SHARKS) was developed by the FAO Technical Working on the Conservation and Management of Sharks (Food
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
and Agriculture Organization, 2000). The IPOASHARKS calls and provides guidelines to countries with directed and non-directed shark catches to adopt a national plan of action for conservation and management of shark stocks. Countries considering or adopting the guidelines are encouraged to ‘implement harvesting strategies consistent with the principles of biological sustainability and rational long-term economic use’ (Food and Agriculture Organization, 2000). The objective of our work is to develop an analytical framework for shark management in the context of using market information to develop and incorporate economic incentives to help ensure biological sustainability and rational economic use of shark populations. This is achieved by assembling a bioeconomic model by examining the linkages between the shark fin market and biological parameters of sharks based on best scientifically available information in conjunction with original data (shark fin market and fin length/weight conversion). Specifically, the preference structure of Hong Kong shark fin importers/wholesalers is
127
elicited by conducting a conjoint analysis using real dried processed shark fins, a common product form at the retail/wholesale level. The results of the conjoint analysis in the form of a utility index is then integrated with the biological growth function of the blacktip shark. The objective of determining the optimal harvest size/age of a cohort of blacktip shark under different discount factors and mortality scenarios are investigated. Results from the preference–cohort analysis show that given the reproductive maturation size of 145.00 cm (5.25 years) for males and 158.00 cm (7.25 years) for females, optimal harvest sizes and ages for all scenarios are greater than the maturation sizes/ ages for both sexes (Castro, 1996). For example, the optimal harvest size based on Hong Kong buyer’s preference and age for the size-dependent mortality function is 167.82 cm (10.00 years)— 2.75 years beyond the maturation age for the male and female blacktip, respectively. Shark stocks are currently being managed by management measures such as reduce harvest levels or effort, use of alternate gears, reduce adverse
Fig. 3. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality=0.05 per quarter).
128
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
Fig. 4. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality=0.075 per quarter).
effects on essential fish habitats, implement minimum sizes, and time-area closures (Shotton, 1999). Results from our work show that Hong Kong shark fin end-users prefer larger sized fins. Moreover, our results also show that the optimal harvest age/size of the shark is beyond the maturation of the representative shark. These results can be utilized by policy makers to internalize economic incentives and strengthen existing management policies to manage shark stocks in a biologically and economically sustainable manner. For instance, size limits can be imposed such that harvested sharks have had the opportunity to reproduce to help ensure biological sustainability. While this measure helps biological sustainability, it also ensures harvesting of high quality (preferred larger sized) shark fins. This harvest strategy would then also be consistent with the ‘rational long-term economic use’ guideline set forth by IPOA— sharks (Food and Agriculture Organization, 2000). Second, gear types that are size selective with low mortality should be used to harvest sharks.
Many of the targeted and non-targeted shark fisheries employ gear types that produce high mortality. The use of gear types (e.g. trawl, gill-nets etc.) that harvest and kill sharks non-selectively would render size limit management measures ineffective. Alternate gear-types such as hook and line, which under-sized sharks or sharks that have low quality fins can be released live, may also be used as a management measure. Third, rights-based fishing should be introduced as a management measure. Rights-based systems, whether in the form of individual transferable quota, individual fishing quota, cooperatives or community development quota, allocate property rights of specific fish stocks to resource users usually on a percentage of the total allowable catch (TAC). This management system gives resource users the flexibility and motivation to maximize economic gain in a sustainable fashion. For example, shark fishers for fins would be able to decide when and how to fish in the most economically efficient manner given the TAC and other manage-
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
129
Fig. 5. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality= Peterson and Wroblewski, 1984).
ment measures. This flexibility will generate economic incentives for shark fishers to selectively harvest sharks that correspond to market signals—in this case, larger sized sharks for their fins. Further, ‘ownership’ by resource users also generates the incentive to biologically conserve shark stocks to ensure economic sustainability. This work has developed a framework to merge economics and ecosystems, in this case a market system and an apex predator in the ecosystem, to optimally manage scarce resources using the best available scientific information. The paucity of biological and economic data prevented us from conducting a full model incorporating intertemporal species and fleet dynamics. Future work includes continuing work on determining the population dynamics of specific shark populations, obtaining biological data that is pertinent to ‘rational economic use’, such as shark fin length and weight, incorporate other uses of sharks (e.g. meat and non-use values) in analysis, and obtaining costs structures for processing and fleet operations. This work is only the first step.
References Addelman, S., 1962. Orthogonal main-effects plans for asymmetrical factorial experiments. Technometrics 4, 21 – 46. Al-Quasmi, A.M., 1994. Physical – chemical characterization of shark fins. M.S. thesis, Food Science and Nutrition Department, University of Rhode Island, 164 pp. Anonymous, 1996. The implications of biology for the conservation and management of sharks. Paper prepared by the Shark Specialist Group of the Species Survival Commission of the IUCN — The World Conservation Union for the 13th Meeting of the CITES Animals Committee, 23 – 27 September 1996, Pruhonice, Czech Republic. Bonfil, R., 1994. Overview of world elasmobranch fisheries. FAO Fisheries Technical Paper No. 341, FAO, Rome, Italy, 119 pp. Branstetter, S., 1987. Age and growth estimates for blacktip, Carcharhinus limbatus, and Spinner, C. bre6ipinna, sharks from the Northwestern Gulf of Mexico. Copeia 1987, 964 – 974. Carroll, M., Anderson, J.L., Martinez-Garmendia, J., 2001. Pricing of US North Atlantic bluefin tuna and implications for management. Agribus 17, 243 – 254. Castro, J.I., 1996. Biology of the blacktip shark, Carcharhinus limbatus, off the Southeastern United States. Bull. Mar. Sci. 59, 508 – 522.
130
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
Clark, C.W., 1990. Mathematical Bioeconomics: The Optimal Management of Renewable Resources. Wiley, New York, p. 384. FAO/WHO, 1987. First draft of codex worldwide standard for shark fins. Codex Alimentarius Commission, Joint FAO/ WHO Food Standards Program, Document CX/LA 87/9, Rome, 8 pp. Food and Agriculture Organization, 2000. The International Plan of Action for the Conservation and Management of Sharks. FAO, Rome. Food and Agriculture Organization, 2001. FAO Yearbook, Fishery Statistics, Commodities. FAO, Rome. Fong, Q.S.W., Anderson, J.L., 2000. Assessment of the Hong Kong shark fin trade. INFOFISH Int. 1/2000, 28 – 32. Gates, J.M., 1974. Demand price, fish size and the price of fish. Can. J. Agric. Econ. 22, 1 –12. Green, P.E., Srinavasan, V., 1990. Conjoint analysis in marketing: new development with implications for research and practice. J. Market. 54, 3 –19. Holden, M.J., 1977. Elasmobranchs. In: Gulland, J.A. (Ed.), Fish Population Dynamics. Wiley, London, pp. 187 – 215. Homans, F.R., Wilen, J.E., 1997. A model of regulated open access resource use. J. Environ. Econ. Manage. 32, 1 –21. Hong Kong Census and Statistics Department, 2001. Hong Kong Trade Statistics: Imports. Hong Kong Special Administrative Region, People’s Republic of China. Lancaster, K., 1966. A new approach to consumer theory. J. Pol. Econ. 74, 132 – 157. Larkin, S.L., Sylvia, G., 1999. Intrinsic fish characteristics and intraseason production efficiency: a management-level bioe-
conomic analysis of a commercial fishery. Am. J. Agric. Econ. 81, 29 – 43. Manski, C., 1977. The structure of random utility models. Theor. Dec. 8, 229 – 254. McConnell, K., Strand, I., 2000. Hedonic price for fish: tuna prices in Hawaii. Am. J. Agric. Econ. 82, 133 – 144. McKelvey, R.D., Zvoina, W., 1975. A statistical model for the analysis of ordinal level dependent variables. J. Math. Soc. 4, 103 – 120. Pascoe, S., Battaglene, T., Campbell, D., 1992. A bioeconomic model of the southern shark fishery. ABARE Research Report 92.1, Project 9343.103, Australia, 60 pp. Peterson, I., Wroblewski, J.S., 1984. Mortality rate of fishes in the pelagic ecosystem. Can. J. Fish. Aquat. Sci. 41, 1117 – 1120. Schaffner, D.J., Schroder, W.R., Earle, M.D., 1998. Food Marketing: An International Perspective. WCB/McGrawHill, Boston, MA, p. 487. Shotton, R. (Ed.), 1999. Case Studies of the Management of Elasmobranch Fisheries. FAO Fisheries Technical Paper No. 378, FAO, Rome, 920 pp. Thunberg, E.M., Helser, T.E., Mayo, R.K., 1998. Bioeconomic analysis of alternative selection patterns in the United States Atlantic silver hake fishery. Mar. Resour. Econ. 13, 51 – 74. Vannuccini, S., 1999. Shark utilization, marketing and trade. FAO Fisheries Technical Paper No. 389, FAO, United Nations, Rome, 470 pp. Zucker, D.A., Anderson, J.L., 1998. Implications of choice behavior and preferences in niche markets. Aquacult. Econ. Manage. 2, 61 – 70.
ANALYSIS
International shark fin markets and shark management: an integrated market preference –cohort analysis of the blacktip shark (Carcharhinus limbatus) Quentin S.W. Fong a,*, James L. Anderson b,1 a
Marine Ad6isory Program/Fishery Industrial Technology Center, School of Fisheries and Ocean Sciences, Uni6ersity of Alaska Fairbanks, 118 Trident Way, Kodiak, AK 99615, USA b Department of En6ironmental and Natural Resource Economics, Coastal Institute, Uni6ersity of Rhode Island, 1 Greenhouse Road, Kingston, RI 02881, USA Received 15 November 2000; received in revised form 10 October 2001; accepted 31 October 2001
Abstract The increasing demand for shark fins in Asia, and the publicity resulting from finning and discarding live sharks, has generated concern regarding the sustainability of the world’s shark populations. These concerns can be attributed to the shark’s life history, which is characterized by a pattern of slow growth, late maturity, few offspring, and long life, making populations vulnerable to overexploitation. Once overexploited, shark stocks will be slow to recover due to these constraints. Despite an increase in consumption and trade of shark fins and other shark products, and the vulnerability of shark populations once overexploited, little effort has been expended to understand the biology and economics of sharks and shark fisheries until recently. This study adds to the understanding of linkages between shark product markets, specifically shark fins, and the biology of shark populations by explicitly incorporating multi-attribute market information into bioeconomic modeling. Results from conjoint analysis of the Hong Kong dried, processed end-user markets is incorporated into a blacktip shark (Carcharhinus limbatus) cohort model to estimate the optimal harvest size and age that maximize economic value. Results show that optimal harvest sizes and ages for all mortality and discount factor scenarios are greater than the maturation sizes and ages for both male and female blacktip. Policy implications for this study are also discussed. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Shark fin market; Shark management; Bioeconomic model
1. Introduction
* Corresponding author. Tel./fax: + 1-907-486-1516. E-mail addresses: [email protected] (Q.S.W. Fong), [email protected] (J.L. Anderson). 1 Tel.: + 1-401-874-4568.
Bioeconomic models utilize an integrated economic and biological systems approach to evaluate the performance of fishery resources using different management strategies. Traditionally, most bioeconomic analysis in the fisheries man-
0921-8009/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 0 0 9 ( 0 1 ) 0 0 2 7 3 - 7
118
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
agement literature examines issues such as extraction rates and fleet size and/or capacity with simplistic assumptions about market behavior. Demand functions are estimated using highly aggregated data to generate price estimates. Little attention has been given to using realistic market information, especially when considering the multi-attribute nature of fishery products (e.g. Carroll et al., 2001). This is important, because in the economics of the international food marketing system, which includes seafood, it is consumers who are the driving force for product selection and consumption (Schaffner et al., 1998). While lack of understanding of consumer tastes and preferences may prevent successful marketing of fishery products by producers, the misunderstanding of consumer product markets by policymakers may promote ineffective fishery management schemes. This can result in welfare losses to all resource users (e.g. Homans and Wilen, 1997). The incorporation of product characteristics in bioeconomic analysis for fishery management was investigated by Gates (1974), who illustrated the importance of product size as a function of market price for ex-vessel demand analysis. Subsequently, several bioeconomic studies have incorporated size-dependent market prices in examining optimal resource management strategies (e.g. Thunberg et al., 1998). The consideration of size as a product characteristic that affects market price has certainly added a realistic dimension to fishery management analysis. However, in certain fisheries, particularly those with foreign ethnic markets such as bluefin tuna and shark fins, the use of product size alone may not be sufficient for demand and bioeconomic analysis. Indeed, recent marketing studies of seafood products using various multi-attribute utility approaches have demonstrated that size is only one of several product attributes that end-users evaluate (e.g. Zucker and Anderson, 1998). For example, results from hedonic price analysis of fresh North Atlantic bluefin tuna (Thunnus thynnus), show that fat content, color, shape, freshness, and size are significantly correlated to ex-vessel price (Carroll et al., 2001). McConnell and Strand (2000) found similar results from tuna in Hawaii.
Until recently, none of the aforementioned seafood marketing studies has explicitly incorporated multi-attribute market values into capturedbased fishery bioeconomic analysis. Larkin and Sylvia (1999) explicitly incorporate intrinsic fish quality into a standard bioeconomic fisheries model for Pacific whiting that includes the harvesting and processing sectors. Intraseasonal price for Pacific whiting is estimated by a multi-attribute, seemingly unrelated regression model that incorporates flesh composition, product form, and hedonic price function for surimi products. This research also explicitly incorporates multiattribute market information to bioeconomic modeling. Here, a framework that incorporates market information of a shark product, shark fin, is merged with the biological growth function of a shark is presented. Specifically, the fishery management objective of harvesting shark fins that are of the most preferred quality to maximize economic return to society is investigated. This is achieved by incorporating the results of a conjoint analysis of dried processed shark fin in Hong Kong into a bioeconomic model of the blacktip shark. The optimal harvest size and age that maximizes economic value of the shark fin set (caudal, dorsal, and two pectoral fins) for a single cohort of blacktip shark under different biological and economic scenarios is estimated.
2. Background Shark fin, a product that is traditionally consumed in Hong Kong, Singapore, Macao, China, and other countries with large ethnic Chinese populations, is one of the most valuable food items in the world. For instance, in 1998, the average price for dried processed caudal fins 25.4 cm (10 in.) in length was US$ 415.00 retail in Hong Kong (Fong and Anderson, 2000). As a consequence of liberalization and increasing spending power of the Asian middle class, the demand for shark fins has increased significantly. For instance, Hong Kong, a trader, processor, and consumer of shark fins, and the most important market for shark fins in the world, increased shark fin imports more than 214% from 2648 mt
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
in 1985 to 8323 mt in 1998 (Vannuccini, 1999; Hong Kong Census and Statistics Department, 2001). Similarly, shark fin imports by Thailand increased 42% from 97 to 138 mt (Food and Agriculture Organization, 2001). The increasing demand for shark fins in Asia, and the publicity resulting from finning and discarding live sharks, have generated concern regarding the sustainability of the world’s shark populations. These concerns are due to the nature of the shark’s life cycle, which makes them vulnerable to overexploitation (Holden, 1977). Once overexploited, shark stocks are slow to recover. Other biological factors, such as schooling by age, sex, and reproductive state, also make some shark species (e.g. blue shark, Prionacae glauca) highly vulnerable to overfishing. High fishing mortality may deplete certain segments of the age class, which may significantly affect the reproductive dynamics of shark populations (Anonymous, 1996). Despite an increase in consumption and trade of shark fins and other shark products, and the vulnerability of shark populations once overexploited, relatively little effort has been spent to understand the biology and economics of sharks and shark fisheries until recently (e.g. Pascoe et al., 1992). This lack of research in the biology and economics of sharks may stem from the traditionally small scale of shark fisheries relative to other fisheries, a lack of understanding of ethnic markets for shark products, and the incidental bycatch nature of many shark fisheries. Further, little attention has been paid by domestic and international fishery management institutions to the management of shark stocks. Bonfil (1994) found that only three (Australia, New Zealand, and the US) out of 26 countries with reported annual shark landings of over 10 000 mt have domestic shark research programs and management plans. Few existing economic studies concerning sharks address the linkage between shark products and shark harvest management. Pascoe et al. (1992) developed a bioeconomic model to estimate the effects of different management options for the southern shark fishery in Australia. They measured the price flexibilities of shark meat in
119
the Melbourne market to get an indication of how prices change with changing quantity. However, they could not make any conclusive statements from this portion of the study since imports and domestically caught sharks, which do not enter the Melbourne market, were not included in the analysis. Thus, the authors treat the price of shark meat as exogenous and assume that price remains constant over the simulated 30-year period. This is a restrictive and unrealistic assumption. This study adds to the understanding of linkages between shark product markets, specifically shark fins and shark biology.
3. Bioeconomic model The overall structure of the model is presented in Fig. 1. First, the biological growth of an individual blacktip shark is modeled with respect to the length and weight of three fin types— caudal, dorsal, and pectoral. Results of a shark fin preference analysis, conjoint analysis, is applied to calculate the utility index of the dried, processed fin set as a function of blacktip shark growth. Finally, the harvest size (age) for a blacktip shark
Fig. 1. Blacktip shark market – cohort model.
120
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
cohort that maximizes the Hong Kong shark fin buyer’s utility for dried, processed fins under different mortality rates and discount rates is estimated.
3.1. Biological component Age and growth estimates for blacktip shark of both sexes are represented by a von Bertalanffy growth function: TLt = L [1− e − K(t − t0)]
(1)
where TL is the total length of the blacktip in centimeters; L the attainable maximum size, is 176 cm total length; K the rate that approaches L is 0.27; and t0 is the age at which the fish would have been zero size is − 1.20 year (Branstetter, 1987). The subscript t represents age in quarters; t=1.0, 1.25, 1.50, …, 30, assuming the blacktip has an average life expectancy of 30 years. The total length equation is then converted to pre-caudal length by using the following equation estimated by Castro (1996) in millimeters: PCLt = − 23.1+0.74TLt
(2)
where PCL is the pre-caudal length of a blacktip (mm) and TL is the total length. To estimate the functional relationships between shark and fin growth in terms of fin size and weight, data from Al-Quasmi (1994) and measurements from a commercial sample of blacktip shark fins are utilized. These relationships are estimated by a series of regressions that linked dried, processed fin size to the von Bertalanfy growth function of blacktip shark. The first step of the relationship is estimated by the relationship between fresh dorsal fin and pre-caudal length of the blacktip shark as reported by AlQuasmi (1994):
3.1.1. Caudal fin 3.1.1.1. Fin size estimate. The relationship between fresh caudal fin and fresh dorsal fin is represented by: ln(MidCaut ) = 0.2670 (1.21)+0.8777 (12.81)ln(FDFt )
FDFt =0.5524 (0.33)+ 0.1608 (14.85)PCLt R 2 =0.83
ship between dried, processed shark fins and blacktip shark growth. First, fin length measurements are made from the tip of the fin to the middle base of the fin and along the anterior edge of the shark fin. These measurements are done because Hong Kong shark fin buyers use the anterior edge measurement as an indicator of fin size while Al-Quasmi (1994) uses Food and Agriculture Organization/World Health Organization’s standard (FAO/WHO, 1987). The weight, in grams, of the dried, unprocessed shark fin samples is also recorded. These measurements are used to estimate the relationship between the length and weight of dried, unprocessed shark fins. Dried, unprocessed shark fins of each type are processed into the end-user product form— dried, processed fins. Before final processing, dried, unprocessed fins are rehydrated until fin lengths became constant. Length measurements of the rehydrated fins are used as a proxy for fresh shark fins to estimate the relationship between fresh and dried unprocessed fin lengths. After processing the fin sets into final dried, processed form, length measurements were taken from the processed fins and used to calculate the conversion ratios of dry, unprocessed fins to dried, processed fins. Weight conversion ratios between dried, unprocessed and dried, processed shark fins were calculated following Al-Quasmi (1994) (Tables 18, 19). The following sections describe the estimated linkages by fin type.
(3)
where FDFt is the fresh dorsal fin size measured from the shark fin tip to the middle base of the fin. Numbers in parenthesis represents the t-ratio. A sample of 36 sets of dried, unprocessed blacktip shark fins was obtained from Guyana to complete the estimation of the functional relation-
R 2 = 0.79
(4)
where ln denotes natural logarithmic transformation; MidCau is the middle length measurement of the fresh caudal fin; FDF is the middle length measurement of the fresh dorsal fin. The relationship between the fresh and dried unprocessed fin is represented by:
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
DryMidCaut =0.7999 (0.81)
121
sented by: DryMidDort = 0.7486 (1.03)
+ 0.8768 (17.27)MidCaut R 2 =0.97
+0.8877 (20.94) FDFt
(5)
where DryMidCau is the middle length measurement of the dried caudal fin. The relationship between the two length measurements of the dried caudal fin is represented by: DryOutCaut =3.17596 (1.64)
DryOutDort = − 0.8986 (−0.28)
+ 1.009 (9.33)DryMidCaut R 2 =0.91
(6)
where DryOutCau is the anterior length measurement of the dried caudal fin. The relationship between dried caudal fin and dried processed caudal fin based on five samples is represented by: DryProcOutCaut = 0.96DryOutCaut
R 2 = 0.98 (10) where DryMidDor is the middle length measurement of the dried caudal fin; FDF is the middle length measurement of the fresh dorsal fin. The relationship between the two length measurements of the dried dorsal fin is represented by:
(7)
+1.4841 (7.51)DryMidDort R 2 = 0.90
(11)
where DryOutDor is the anterior length measurement of the dried dorsal fin. The relationship between dried caudal fin and dried processed caudal fins based on five samples is represented by: DryProcOutDort = 0.96DryOutDort
(12)
where DryProcOutCau represents the anterior length measurement of the dried processed caudal fin.
where DryProcOutDor represents the anterior length measurement of the dried processed dorsal fin.
3.1.1.2. Fin weight estimate. The relationship between the weight and length of the dried caudal fin is represented by:
3.1.2.2. Fin weight estimate. The relationship between the weight and length of the dried dorsal fin is represented by:
ln(DryCaugmt )
ln(DryDorgmt )
= − 4.5691 (− 3.87) + 2.8660 (7.445)ln(DryOutCaut )
= − 5.2522 (−5.96) 2
R =0.87 (8)
+3.0628 (10.82)ln(DryOutDort )
R 2 = 0.95 (13)
where DryCaugm is the dried caudal fin weight in grams. The weight relationship between dried caudal fins and processed dried caudal fin based on five samples is represented by:
where DryDorgm is the dried dorsal fin weight in grams. The weight relationship between dried caudal fins and processed dried caudal fin based on five samples is represented by:
DryCauKgt =0.74DryCauKgt
DryProcDorKgt = 0.530DryDorKgt
(9)
(14)
where DryCauKg is the dried processed caudal fin in kilograms.
where DryDorKg is the dried processed dorsal fin in kilograms.
3.1.2. Dorsal fin
3.1.3. Pectoral fin
3.1.2.1. Fin size estimate. The relationship between the fresh and dried unprocessed fin is repre-
3.1.3.1. Fin size estimate. The relationship between fresh pectoral fin and fresh dorsal fin is
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
122
represented by:
samples is represented by:
ln(MidPect )
DryProcPecKgt = 0.42DryPecKgt
=1.1582 (6.35)+ 0.7438 (13.09)ln(MidDort )
where DryPecKg is the dried processed pectoral fin in kilograms.
2
(15)
R =0.79
where MidPec is the middle length measurement of the fresh pectoral fin. The relationship between the fresh and dried unprocessed fin is represented by:
(20)
3.2. Utility index for an indi6idual shark
where DryOutPec is the anterior length measurement of the dried pectoral fin. The relationship between dried pectoral fin and dried processed pectoral fins based on five samples is represented by:
3.2.1. Conjoint analysis and consumer choice A market preference model, conjoint analysis, is used to determine the utility of the shark fin set to Hong Kong shark fin importers/processors as a function of blacktip shark growth. Conjoint analysis is a form of multi-attribute utility model, which all, or in part, link to the notion that utility is derived from the attributes that the good possesses (e.g. Lancaster, 1966). It is assumed that the utility Hong Kong shark fin importers/processors obtained from a specific shark fin product is a function of the utility derived directly from the product’s attributes and levels of those attributes and indirectly from the profits associated with the product’s attributes (Lancaster, 1966). For example, a Hong Kong shark fin buyer may prefer medium-sized dried, processed dorsal shark fin to large-sized dried, processed pectoral shark fin. The utility derived from a given product may then be expressed in general form as a quasi-concave, twice continuously differentiable utility function:
DryProcOutPect = 0.96DryOutPect
U(sh )= U{Xh ; y(Xh )}
DryMidPect = − 1.1327 (−1.37) + 1.0006 (25.80)MidPect R 2 =0.99
(16)
where DryMidPec is the middle length measurement of the dried pectoral fin. The relationship between the two length measurements of the dried pectoral fin is represented by: DryOutPect = 3.2687 (2.18) +1.1607 (15.61)DryMidPect R 2 =0.97
(17)
(18)
where DryProcOutPec represents the anterior length measurement of the dried processed pectoral fin.
3.1.3.2. Fin weight estimate. The relationship between the weight and length of the dried pectoral fin is represented by: ln(DryPecgmt ) = − 4.7353 (− 7.16) + 2.7462 (13.56)ln(DryOutPect )
R 2 =0.87 (19)
where DryPecgm is the dried pectoral fin weight in grams. The weight relationship between dried pectoral fins and processed dried pectoral fin based on five
(21)
where U(sh ) is the utility the buyer derives from the hth composite dried, processed shark fin product sh ; Xh is a vector of levels making up the composite product sh ; y(Xh ) is the profit function associated with the product’s attributes. Since a decision maker obtains some degree of satisfaction from each product, the alternative selected for consumption would be the one that provides the highest satisfaction. For example, a shark fin buyer would choose product s4 over product s2, only if U(s4) is greater than U(s2). However, the utility of the shark fin importer/processor is not directly observable and is unknown. The utilities, therefore, are treated as random variables, and the probability of choosing alternative dried, processed shark fin product s4 over s2 is equal to the probability that U(s4) is greater than U(s2) (Manski, 1977).
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
3.2.2. Conjoint analysis model specification and estimation Conjoint analysis of dried, processed shark fin was conducted with Hong Kong shark fin importer/processors. This method uses field experiments by asking respondents to rank or rate products with predetermined attributes and levels of attributes to measure the buyer’s preference or utility as the dependent variable (Green and Srinavasan, 1990). Here, the conjoint analysis evaluates the utility function of Hong Kong shark fin importers/processors directly by asking respondents to rate a set of stimuli from 0 to 10, with 0 being the least preferred, and 10 being the most preferred. In this case, a reduced design of 11 dried, processed shark fins was obtained by using an asymmetrical factorial orthogonal experimental plan (Addelman, 1962). The attributes included were fin size and type. The conjoint model employed in this research uses the traditional non-interaction-effect model, which is assumed to be additive in levels of the attributes (e.g. Green and Srinavasan, 1990): U(sh )= i%ij x (h) ij + mij
mij N(0, 1)
L(h, v) m
= % % Rq,y log(Lq,y (vj − 1 −i%X) q = 1y = 1
−Lq,y − 1(vj − i%X))
Table 1 Results of conjoint model estimation (ordered logit) Variable
Coefficient
S.E.
T-ratio
Constant Size Dorsal Pectoral v(1) v(2) v(3) v(4) v(5) v(6) v(7) v(8) v(9)
2.78 2.32 −8.36 −13.11 3.22 6.16 7.68 10.18 14.26 17.31 18.91 20.43 23.76
0.66 0.23 0.89 1.38 1.72 0.75 0.88 1.85 1.76 1.82 1.90 1.99 2.84
4.20** 10.19** −9.41** −9.47** 1.87* 8.26** 9.34** 5.50** 8.12** 9.49** 9.93** 10.26** 8.38**
Log likelihood function =−192.58; N =187; Restricted log likelihood=−448.41; 2 =511.66**; **, significant at the 0.01% level; *, significant at the 10% level.
individuals in the experiment, and m is the number of stimulus in the conjoint experiment. Maximizing L(h, v) provides estimates of the parameters h and v (McKelvey and Zvoina, 1975).
(22)
where U(sh ) is the random utility that an individual derives from hth product, iij is the parameter matrix that represents the relative importance of the levels, x (h) ij represents the deterministic independent variable matrix associated with attribute j and level i for product h, and mij is the random error term. An ordered logit model was used to analyze the rating data generated by the conjoint experiment. For an independent sample of n individuals, the log likelihood function, L(h, v), is: n
123
(23)
where L(·) is the logistic distribution function e(·)/1 +e(·); v’s are the unknown threshold variables to be estimated with h where j= 0, …, 9; i is the matrix for the coefficients h; and X is the matrix for the independent variables constant, fin size, dorsal fin, and pectoral fin; n is the number of
3.2.3. Utility index formulation Maximizing the log likelihood function in Eq. (23) provides estimates for h, the coefficients of the independent variables and v, the threshold levels between ratings. The estimated equation is: Usc = 2.8+ 2.4Sz− 8.3Dor− 13.1Pec (24) where Usc is the utility score for the three fin types, caudal, dorsal, and pectoral, at various fin sizes; Sz is fin size; Dor is dorsal fin, and Pec is pectoral fin. Both Dor and Pec are coded in dummy variables. All estimated coefficients are significant at the 0.01% level (Table 1). The estimated utility score Eq. (24) and the estimated threshold level (v) are then used to calculate the probability of a dried, processed shark fin being rated in a certain category (e.g. rating= 10) for a given fin size and fin type. The ordered logit model specification captures the preference structure for dried, processed shark fin by a representative shark fin processor/importer in Hong Kong. It is assumed that the shark fin processor/importer assigned ratings to product profiles in the conjoint experiment relative to his/ her most desired product profile. Thus the specific
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
124
utility score for an index is calculated from the logistic probability function for the most preferred rating (rating=10), the estimated utility score (Usc) from Eq. (24), and the estimated v9 (23.759; significant at 0.01%) that represents the lower bound threshold level for the most preferred rating. The formula for the utility index, which represents the probability of being the most preferred dried, processed shark fin product is: UWi,Sz =1−
e[23.7 − (2.8 + 2.4Sz − 8.3Dor − 13.1Pec)] 1+e[23.7 − (2.7 + 2.4Sz − 8.3Dor − 13.1Pec)] (25)
where UWi is the utility per unit weight for the three fin types; i is fin type, and 23.7 is the estimated lower bound threshold level for the most preferred rating from the estimated ordered logit model. The utility index for fin type, i, is calculated as the product of the utility per unit weight, UWi, and the dried, processed weight, DPWi, for fin type i: UIi =UWi ×DPWi
(27)
where TUI is the total utility index for an individual blacktip shark.
3.3. Utility index for cohort To calculate the utility index for the blacktip shark, the initial population of the cohort is assumed to be 10 000, with both sexes combined. The quarterly numbers-at-age for the blacktip shark cohort is: Nt + 1 = Nt · e − M/4
Mt =
1.92 − 0.25 Wt 4
(28)
where Nt is the number of sharks at age t, expressed in quarters; and M is the natural mortality rate. Three quarterly natural mortality rates, 0.025, 0.050, 0.075, and a natural mortality function
(29)
where Mt is the quarterly natural mortality, and Wt is the dry weight of individual shark in grams, assuming dry weight is 0.2 of wet weight. This mortality function simulates the decrease in natural mortality as the size of a shark increases with age in a cohort. The weight of an individual blacktip shark is determined by: WKGt = (2.51×10 − 9)TLM3.12 t
(30)
where WKGt is the wet weight of an individual blacktip shark (kg), and TLMt is total shark length (mm) (Castro, 1996). The total utility of a cohort using the utility index approach by conjoint analysis is represented by:
(26)
where UIi is the utility index for fin type i. The total utility index for an individual blacktip is the sum of the utility indexes of the three fin types, taking into account that sharks have one caudal, one dorsal, and two pectoral fins: TUI =% UIi
proposed by Peterson and Wroblewski (1984) are used for sensitivity analysis. The Peterson and Wroblewski (1984) natural mortality function is:
TUCt =
TUIt × Nt (1+r)t
(31)
where TUCt is the total utility index for the cohort at age t; TUIt the total utility index for an individual shark; Nt the number of sharks in the cohort; and r is the discount rate, which is set at 0, 0.02, 0.03, 0.05, 0.07, 0.1, and 0.2, respectively.
4. Optimal harvest Results from a multi-attribute marketing analysis were incorporated into a market preference–cohort model of the blacktip shark. The optimal harvest size of the blacktip shark was investigated for the conjoint preference–cohort model under four natural mortality scenarios. Within each natural mortality scenario, the effects of seven discount factors were also simulated. These results are presented in Table 2 and Figs. 2–5. Three quarterly natural mortality parameters, 0.025, 0.05, and 0.075, are used to determine the optimal harvest size/age of the blacktip shark. Results show that as quarterly natural mortality increases from 0.025 to 0.075 at any given discount
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
rate, the optimal harvest size/age for the blacktip shark decreases. For example, at a discount rate of 0.03, the optimal harvest size estimated with the conjoint market– cohort model decreases from 171.88 (12.50 years of age) to 169.34 cm (10.75 years), then to 166.62 cm (9.50 years) as the quarterly mortality rate increases from 0.025 to 0.075 (Table 2 and Figs. 2– 4). The performance of the conjoint market– cohort model using a size-dependent natural mortality function is also investigated (Peterson and Wroblewski, 1984). This function assumes that as the size of an individual shark increases with age (expressed in weight), the natural mortality rate for the cohort decreases. This assumption is an improvement in realism over the constant mortal-
ity scenarios, since a shark cohort of a small-size class (i.e. younger age) would be more vulnerable to predation than a cohort of a large-size class. Results show that the size-dependent mortality conjoint market–cohort model provides the least conservative optimal harvest sizes/ages of all mortality scenarios (Table 2 and Fig. 5). For example, at zero discount rate, the optimal harvest size/age for the size-dependent mortality scenario is 169.34 cm (10.75 years), as opposed to 172.86 cm (13.50 years), 170.19 cm (11.25 years), 167.24 cm (9.75 years) for 0.025, 0.050, and 0.075 constant quarterly mortality rates, respectively. Seven discount rates, ranging from 0 to 20%, are used to examine optimal harvest size and age of the blacktip shark. These rates are used to
Table 2 Optimal harvest for conjoint market–cohort model Natural mortalitya
Discount rate (%)
Total utility index
Optimal harvest sizeb (cm)/age (years)
0.025
0 2 3 5 7 10 20 0 2 3 5 7 10 20 0 2 3 5 7 10 20 0 2 3 5 7 10 20
86.01 66.29 58.56 46.18 36.89 26.87 10.71 25.49 20.51 18.47 15.08 12.42 9.41 4.15 8.99 7.42 6.76 5.64 4.74 3.69 1.76 10.77 8.72 7.84 6.47 5.36 4.09 1.84
172.86/13.50 172.15/12.75 171.88/12.50 171.27/12.00 170.94/11.75 170.19/11.25 167.82/10.00 170.19/11.25 169.34/10.75 169.34/10.75 168.87/10.25 168.36/10.25 167.24/9.75 165.24/9.00 167.24/9.75 166.62/9.50 166.62/9.50 165.95/9.25 165.24/9.00 164.48/8.75 162.79/8.25 169.34/10.75 168.87/10.50 167.82/10.00 167.82/10.00 167.24/9.75 166.62/9.50 164.48/8.75
0.050
0.075
P & Wc
a
In quarters. Total length. c Peterson and Wroblewski (1984). b
125
126
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
Fig. 2. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality=0.025 per quarter).
simulate the divergence between the social and private opportunity cost of capital, time reference, and risk premium. Real discount rates between 0 and 5% have been suggested as an appropriate social discount rate for the 30-year horizon (Clark, 1990). The differences can be attributed to the differences in risk premium perceptions. Results show that in all scenarios, size (age) of optimal shark harvest decreases as discount rate increases (Table 2). For example, given a discount rate of 3%, the optimal harvest size and age for blacktip under a size-dependent natural mortality conjoint market–cohort simulation is 167.82 cm (10.00 years of age). Alternately, the optimal harvest size and age for size-dependent natural mortality given a 20% discount rate is 164.48 cm (8.75 years of age), 2 years younger than the case with a lower discount rate (Fig. 5).
5. Summary and conclusions The increasing demand for shark products, including shark fins, and the life-history pattern of long living, late maturity, and low-reproductive potential of sharks have generated concerns regarding the health of the world’s shark stocks. These concerns are generated not only by regulatory agencies and non-governmental organizations but also from resource users. A survey of shark fin importers/processors in Hong Kong, the center for shark fin trade and consumption in the world, has shown that more than 41% of the respondents expressed concerns of overharvesting of sharks (Fong and Anderson, 2000). Responding to these concerns, an International Plan of Action for Conservation and Management of Sharks (IPOA-SHARKS) was developed by the FAO Technical Working on the Conservation and Management of Sharks (Food
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
and Agriculture Organization, 2000). The IPOASHARKS calls and provides guidelines to countries with directed and non-directed shark catches to adopt a national plan of action for conservation and management of shark stocks. Countries considering or adopting the guidelines are encouraged to ‘implement harvesting strategies consistent with the principles of biological sustainability and rational long-term economic use’ (Food and Agriculture Organization, 2000). The objective of our work is to develop an analytical framework for shark management in the context of using market information to develop and incorporate economic incentives to help ensure biological sustainability and rational economic use of shark populations. This is achieved by assembling a bioeconomic model by examining the linkages between the shark fin market and biological parameters of sharks based on best scientifically available information in conjunction with original data (shark fin market and fin length/weight conversion). Specifically, the preference structure of Hong Kong shark fin importers/wholesalers is
127
elicited by conducting a conjoint analysis using real dried processed shark fins, a common product form at the retail/wholesale level. The results of the conjoint analysis in the form of a utility index is then integrated with the biological growth function of the blacktip shark. The objective of determining the optimal harvest size/age of a cohort of blacktip shark under different discount factors and mortality scenarios are investigated. Results from the preference–cohort analysis show that given the reproductive maturation size of 145.00 cm (5.25 years) for males and 158.00 cm (7.25 years) for females, optimal harvest sizes and ages for all scenarios are greater than the maturation sizes/ ages for both sexes (Castro, 1996). For example, the optimal harvest size based on Hong Kong buyer’s preference and age for the size-dependent mortality function is 167.82 cm (10.00 years)— 2.75 years beyond the maturation age for the male and female blacktip, respectively. Shark stocks are currently being managed by management measures such as reduce harvest levels or effort, use of alternate gears, reduce adverse
Fig. 3. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality=0.05 per quarter).
128
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
Fig. 4. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality=0.075 per quarter).
effects on essential fish habitats, implement minimum sizes, and time-area closures (Shotton, 1999). Results from our work show that Hong Kong shark fin end-users prefer larger sized fins. Moreover, our results also show that the optimal harvest age/size of the shark is beyond the maturation of the representative shark. These results can be utilized by policy makers to internalize economic incentives and strengthen existing management policies to manage shark stocks in a biologically and economically sustainable manner. For instance, size limits can be imposed such that harvested sharks have had the opportunity to reproduce to help ensure biological sustainability. While this measure helps biological sustainability, it also ensures harvesting of high quality (preferred larger sized) shark fins. This harvest strategy would then also be consistent with the ‘rational long-term economic use’ guideline set forth by IPOA— sharks (Food and Agriculture Organization, 2000). Second, gear types that are size selective with low mortality should be used to harvest sharks.
Many of the targeted and non-targeted shark fisheries employ gear types that produce high mortality. The use of gear types (e.g. trawl, gill-nets etc.) that harvest and kill sharks non-selectively would render size limit management measures ineffective. Alternate gear-types such as hook and line, which under-sized sharks or sharks that have low quality fins can be released live, may also be used as a management measure. Third, rights-based fishing should be introduced as a management measure. Rights-based systems, whether in the form of individual transferable quota, individual fishing quota, cooperatives or community development quota, allocate property rights of specific fish stocks to resource users usually on a percentage of the total allowable catch (TAC). This management system gives resource users the flexibility and motivation to maximize economic gain in a sustainable fashion. For example, shark fishers for fins would be able to decide when and how to fish in the most economically efficient manner given the TAC and other manage-
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
129
Fig. 5. Total utility index from dried processed shark fins for the blacktip shark cohort (initial population = 10 000; natural mortality= Peterson and Wroblewski, 1984).
ment measures. This flexibility will generate economic incentives for shark fishers to selectively harvest sharks that correspond to market signals—in this case, larger sized sharks for their fins. Further, ‘ownership’ by resource users also generates the incentive to biologically conserve shark stocks to ensure economic sustainability. This work has developed a framework to merge economics and ecosystems, in this case a market system and an apex predator in the ecosystem, to optimally manage scarce resources using the best available scientific information. The paucity of biological and economic data prevented us from conducting a full model incorporating intertemporal species and fleet dynamics. Future work includes continuing work on determining the population dynamics of specific shark populations, obtaining biological data that is pertinent to ‘rational economic use’, such as shark fin length and weight, incorporate other uses of sharks (e.g. meat and non-use values) in analysis, and obtaining costs structures for processing and fleet operations. This work is only the first step.
References Addelman, S., 1962. Orthogonal main-effects plans for asymmetrical factorial experiments. Technometrics 4, 21 – 46. Al-Quasmi, A.M., 1994. Physical – chemical characterization of shark fins. M.S. thesis, Food Science and Nutrition Department, University of Rhode Island, 164 pp. Anonymous, 1996. The implications of biology for the conservation and management of sharks. Paper prepared by the Shark Specialist Group of the Species Survival Commission of the IUCN — The World Conservation Union for the 13th Meeting of the CITES Animals Committee, 23 – 27 September 1996, Pruhonice, Czech Republic. Bonfil, R., 1994. Overview of world elasmobranch fisheries. FAO Fisheries Technical Paper No. 341, FAO, Rome, Italy, 119 pp. Branstetter, S., 1987. Age and growth estimates for blacktip, Carcharhinus limbatus, and Spinner, C. bre6ipinna, sharks from the Northwestern Gulf of Mexico. Copeia 1987, 964 – 974. Carroll, M., Anderson, J.L., Martinez-Garmendia, J., 2001. Pricing of US North Atlantic bluefin tuna and implications for management. Agribus 17, 243 – 254. Castro, J.I., 1996. Biology of the blacktip shark, Carcharhinus limbatus, off the Southeastern United States. Bull. Mar. Sci. 59, 508 – 522.
130
Q.S.W. Fong, J.L. Anderson / Ecological Economics 40 (2002) 117–130
Clark, C.W., 1990. Mathematical Bioeconomics: The Optimal Management of Renewable Resources. Wiley, New York, p. 384. FAO/WHO, 1987. First draft of codex worldwide standard for shark fins. Codex Alimentarius Commission, Joint FAO/ WHO Food Standards Program, Document CX/LA 87/9, Rome, 8 pp. Food and Agriculture Organization, 2000. The International Plan of Action for the Conservation and Management of Sharks. FAO, Rome. Food and Agriculture Organization, 2001. FAO Yearbook, Fishery Statistics, Commodities. FAO, Rome. Fong, Q.S.W., Anderson, J.L., 2000. Assessment of the Hong Kong shark fin trade. INFOFISH Int. 1/2000, 28 – 32. Gates, J.M., 1974. Demand price, fish size and the price of fish. Can. J. Agric. Econ. 22, 1 –12. Green, P.E., Srinavasan, V., 1990. Conjoint analysis in marketing: new development with implications for research and practice. J. Market. 54, 3 –19. Holden, M.J., 1977. Elasmobranchs. In: Gulland, J.A. (Ed.), Fish Population Dynamics. Wiley, London, pp. 187 – 215. Homans, F.R., Wilen, J.E., 1997. A model of regulated open access resource use. J. Environ. Econ. Manage. 32, 1 –21. Hong Kong Census and Statistics Department, 2001. Hong Kong Trade Statistics: Imports. Hong Kong Special Administrative Region, People’s Republic of China. Lancaster, K., 1966. A new approach to consumer theory. J. Pol. Econ. 74, 132 – 157. Larkin, S.L., Sylvia, G., 1999. Intrinsic fish characteristics and intraseason production efficiency: a management-level bioe-
conomic analysis of a commercial fishery. Am. J. Agric. Econ. 81, 29 – 43. Manski, C., 1977. The structure of random utility models. Theor. Dec. 8, 229 – 254. McConnell, K., Strand, I., 2000. Hedonic price for fish: tuna prices in Hawaii. Am. J. Agric. Econ. 82, 133 – 144. McKelvey, R.D., Zvoina, W., 1975. A statistical model for the analysis of ordinal level dependent variables. J. Math. Soc. 4, 103 – 120. Pascoe, S., Battaglene, T., Campbell, D., 1992. A bioeconomic model of the southern shark fishery. ABARE Research Report 92.1, Project 9343.103, Australia, 60 pp. Peterson, I., Wroblewski, J.S., 1984. Mortality rate of fishes in the pelagic ecosystem. Can. J. Fish. Aquat. Sci. 41, 1117 – 1120. Schaffner, D.J., Schroder, W.R., Earle, M.D., 1998. Food Marketing: An International Perspective. WCB/McGrawHill, Boston, MA, p. 487. Shotton, R. (Ed.), 1999. Case Studies of the Management of Elasmobranch Fisheries. FAO Fisheries Technical Paper No. 378, FAO, Rome, 920 pp. Thunberg, E.M., Helser, T.E., Mayo, R.K., 1998. Bioeconomic analysis of alternative selection patterns in the United States Atlantic silver hake fishery. Mar. Resour. Econ. 13, 51 – 74. Vannuccini, S., 1999. Shark utilization, marketing and trade. FAO Fisheries Technical Paper No. 389, FAO, United Nations, Rome, 470 pp. Zucker, D.A., Anderson, J.L., 1998. Implications of choice behavior and preferences in niche markets. Aquacult. Econ. Manage. 2, 61 – 70.