1 JIMAR ANNUAL REPORT FOR FY 2001 (Project 653540) P.I. Name: PngSun Leung, Khem Sharma and Sam Pooley Project Research Assstant: Naresh Pradhan Project Ttle: Analyzng the Techncal and Economc Structure of Hawa s Pelagc Fshery Fundng Agency: NOAA/NMFS/Pelagc Fsheres Research Program Project start date: October 1, 1999 1. Purpose of Project The man objectve of ths study s to determne the technologcal and economc nterrelatonshps n Hawa-based longlne, troll, and handlne pelagc fsheres usng a multproduct dual revenue functon approach. A secondary objectve s to provde a prelmnary test of ncorporatng these estmated relatons nto the exstng allocaton model. The specfc objectves of the project nclude: 1. Comple exstng secondary trp-level nformaton on revenue, landngs and prces by speces as well as other trp characterstcs, ncludng trp lengths, and target for Hawa-based longlne, troll, and handlne pelagc fsheres contaned n NMFS logbook and HDAR (Hawa Dvson of Aquatc Resources) catch records; 2. Conceptualze and specfy mult-product dual revenue functon models for longlne and nonlonglne commercal (troll/handlne) fsheres and develop ther estmaton procedures; 3. Test for nput-output separablty and nonjontness-n-nputs of the harvestng technology of Hawa-based longlne, troll, and handlne pelagc fsheres; 4. Estmate own and cross-prce supply and effort elastctes; 5. Estmate the mult-speces economes of scope, speces-specfc economes of scale, mult-speces economes of scale, speces-specfc margnal costs and cost elastctes; 6. Assess the mpact of area-closure by conductng separate analyses of the data collected before and after the mplementaton of area-closure regme and comparng the estmates for the two perods; and 7. Incorporate the estmated relatons nto the exstng mathematcal programmng allocaton model as a demonstraton on how the estmated relatons can be used. 2. Progress Durng FY 2001 (July 1, 2000 to June 30, 2001) Most of project work durng FY 2000 was related to collecton, complaton and organzaton of necessary data for the longlne fleet. Work durng FY 2001 s concerned wth descrptve data analyses and the specfcaton and estmaton of the revenue functon model for the longlne fshery. The descrptve results nclude the contrbutons of ndvdual speces to total marketed landngs and ex-vessel revenues and tme trends of ex-vessel prces of major longlne speces. Informaton on contrbutons of each speces to total landngs and revenues was used n determnng the number of speces or speces groups n the revenue functon. The estmaton of output supply functons and related prce and effort elastctes for the longlne fshery has been completed. These tasks and related results are hghlghted below.
2 The HDAR data for the non-longlne commercal pelagc fsheres are also compled n FY 2001. These nclude troll, handlne, and aku boats. Snce aku boats harvest prmarly aku and one or two other mnor speces, they are not consdered n revenue functon analyses. Unlke the longlne fshery, effort (trp length, number of hooks) and vessel-specfc (tonnage, sze, etc.) nformaton do not exst for the nonlonglne fleets. Ths problem s beng overcome by aggregatng the trp level data to both monthly and quarterly levels and defnng effort n terms of number of trps per month or quarter. The descrptve analyses as well as the specfcaton and estmaton of the revenue functon models for troll and handlne fsheres are n progress. 2.1 Pounds Sold and Ex-Vessel Revenues Total annual pounds sold and ther speces composton for each trp type (swordfsh, mxed and tuna) as well as for entre longlne trps durng 1991 to 1998 are presented n Appendx Table 1. Total ex-vessel revenues and revenue shares by speces are presented n Appendx Table 2. Note that total pounds sold and total revenues presented here correspond to the matched observatons only (.e., 6,666 or 77% of HDAR total observatons) and hence can be dfferent from those found n longlne annual reports. For trps targetng swordfsh, as expected, swordfsh s the domnant speces, accountng for about 80% each of total pounds sold and total revenue. Bgeye s the second mportant speces for swordfsh trps. For mxed trps, swordfsh s domnant, followed by bgeye tuna and yellowfn tuna, and for tuna trps bgeye s domnant, followed by yellowfn and albacore. Comparng over tme, the mportance of swordfsh has declned whle that of bgeye tuna has ncreased. For example, based on matched observatons, the contrbuton of swordfsh to total longlne revenue has decreased from 52% n 1991 to 23% n 1998, whle the contrbuton of bgeye tuna has ncreased from 28% to 49%.
3 2.2 Ex-vessel prces The mean annual prces for the major longlne speces are presented n Appendx Table 3. On the average, prces have been farly stable durng 1991 to 1998, especally for the major speces targeted by longlners (bgeye tuna, broadbll swordfsh and yellowfn tuna). Ths suggests that although speces composton has changed substantally n recent years, relatve prces have remaned farly stable. However, as ndcated by the standard devatons, varatons n prces wthn a year are consderable. Comparng among the speces, bgeye tuna fetched consstently the hghest prce, followed by swordfsh and yellowfn tuna. In general, albacore tuna fetched the lowest prce. 2.3 Revenue functon specfcaton Based on pound and revenue shares, prces and bologcal characterstcs, for the revenue functon analyss, varous longlne speces are aggregated to sx (6) speces or speces groups. These nclude yellowfn tuna (Y), albacore (A), bgeye tuna (T), broadbll swordfsh (B), marln (M) and other pelagc speces (O). Marln (M) s an aggregate of black marln, blue marln and strpped marln. Smlarly, other pelagc speces nclude varous shark speces (mako, thresher, tger, and other sharks) and other pelagc speces (aku, barracuda, bluefn tuna, mahmah, monchong, ono, opah, papo, salfsh, short nose, walu and other unclassfed pelagcs). Varous vessel-specfc (vessel length, horsepower, gross regstered tonnage and net tonnage) and trp-level nputs (trp length, and number of hooks and number of sets) were consdered n dervng a measure of composte nput/effort n the revenue functon. Based on the correlaton results of these varables, trp length and vessel net tonnage were selected n order to compute the composte effort. A sngle composte nput (Z) s derved as the product of trp length (n days) and vessel net tonnage. The revenue functon was specfed to be of both Leontef s form and the translog form. Although the Leontef s form s more popular n prevous studes, we had proposed the translog form n ths study. Accordngly, both forms were tred. However, based on prelmnary results the Leontef s form was found to be superor to the translog form. Perhaps, ths may be due to the restrcton that revenue shares n the latter should need to add up to one. For ths reason, the Leontef s revenue functon s chosen n ths study. Mathematcal detals underlyng the Leontef s revenue functon, dervaton of output supply functons to be estmated and formulae nvolved n calculatng varous elastctes are gven n Appendx 4. For part of 1991 nformaton on bgeye tuna was mssng n the HDAR longlne data. It s suspected that bgeye may have been ncluded wth one of the other tuna speces. Because of ths, the 1991 data are excluded n revenue functon analyss.
4 2.4 Estmaton Procedures As mentoned above, dfferent speces harvested by longlners have been aggregated to sx speces or speces groups. The problem s that all vessels do not land all speces n every trp. Ths resulted n mssng nformaton on outputs and prces for a szeable proporton of the observatons. For example, of the 6,666 matched trp-level observatons 2,293 or 34% had complete output and prce nformaton. For estmatng the output supply functons, the data need to have complete nformaton on outputs and prces of all speces ncluded n the model. For ths reason, the observatons wth no nformaton on any of the output or prce varables were excluded. In vew of dfferences n harvestng technologes and output composton among dfferent trp types, revenue functon analyses have been carred out separately for swordfsh, mxed and tuna trps. Ths wll provde nformaton on how dfferent trp types respond to changes n prces and fshng effort. For comparson purpose, analyss has also been conducted on all trp types combned. The estmaton procedure s drven by two assumptons. Frst, we assume that the fsherman s decson to harvest a gven quantty of a partcular fsh speces s nfluenced by the level of current prce he/she gets for that speces. Second, we assume that fshermen base ther decsons on some knd of expected prces rather than on the current prces. Accordngly three dfferent types of revenue functons are specfed. The frst one relates the current trp-level outputs to current trp-level prces for each vessel. In the second model, current trp-level outputs are expressed as functons of prces obtaned n the mmedate precedng trp. In the thrd model, the trp-level data are aggregated to the quarterly level and total quarterly outputs are related to ther average prces n that quarter. The second and thrd models are the models based on expected prces. These models would also avod the possble smultanety problem n usng current outputs and current prces. The systems of output supply functons obtaned from the generalzed Leontef s revenue functons for dfferent trp types and for dfferent assumptons for output and prce relatonshps are estmated usng Zellner s seemngly unrelated regresson estmaton (SURE) technque. Under the SURE framework the cross-prce coeffcents are symmetrc. Besdes the parameter estmates for the systems output supply functons, assumptons of nonjontness-n-nputs and nput-output separablty are also tested. Fnally, own-prce, cross-prce and effort elastctes of output supply functons for sx speces are computed. These results are summarzed next. 2.5 Results The three models estmated (trp-level current prce, trp-level lagged prce and quarterly) were qute smlar n terms of sgn, magntude and sgnfcance level of estmates for own-prce, cross-prce, and effort elastctes. Furthermore, the models were also generally n agreements wth respect to tests of hypotheses for nonjontness n nputs and nput-output separablty. However, n terms of behavoral assumptons, the results obtaned from the second model (.e. fshermen respondng to the prces receved n the prevous trp rather than what they would receve n the current trp) seem more plausble. Hence, the results summarzed here pertan to the model wth current trp-level outputs wth lagged prces. The varables (outputs, prces and effort) nvolved n estmatng the lagged model are summarzed n Appendx Table 5. 2.5.1 Parameter Estmates
5 The parameter estmates for the supply functons are presented n Appendx Table 6. Ownprce and cross-prce elastctes presented n Secton 2.5.3 are estmated based on the estmated coeffcents of supply equatons. A number of estmated coeffcents n most supply equatons were sgnfcant at the 0.05 level. Of partcular nterest are the effort and squared effort coeffcents. Effort coeffcents are postve n most of the supply equatons and also sgnfcant n supply equatons of major speces targeted under each trp type. The coeffcents for squared effort are mostly negatve, ndcatng that effort s characterzed by dmnshng margnal productvty. However, the magntude of the squared effort coeffcents mples that changes n margnal productvty would be neglgble. 2.5.2 Tests of hypotheses The results of the tests of hypotheses for nonjontness n nputs and nput-output separablty are presented n Appendx Table 7. The results rejected both null hypotheses for all trp types as well as all trps combned. The rejecton of nonjontness n producton suggests that there are techncal nteractons among speces targeted by longlners, and sngle speces management would affect the explotaton of other speces. Smlarly, the rejecton of separablty between nputs and outputs suggests that management of the bomass or the aggregate stock of speces s also napproprate. 2.5.3 Elastctes estmates The estmated own-prce, cross-prce, effort elastctes of output supply functons, along wth the standard errors are presented n Appendx Table 8. Elastctes and ther standard errors are calculated at the observed mean values of varables nvolved. The theory suggests that own-prce supply elastctes be postve, but cross-prce elastctes can be postve or negatve. As shown n Appendx Table 8, a number of own-prce elastctes for each trp type are negatve. However, except for other pelagc speces for tuna trps, none of these negatve own-prce elastctes are sgnfcant. More nterestngly, none of the postve own-prce supply elastctes s sgnfcant. Own-prce elastctes obtaned from the other two models (current trp and quarterly) were also not sgnfcant. Thus, these results suggest that output supply decsons of longlners are ndependent of prces they get. As shown n Appendx Table 8, a number of cross-prce elastctes are sgnfcant at the 0.05 level. The cross-prce elastctes between the major targeted speces are negatve and sgnfcant n a number of cases. For example, for the swordfsh trps swordfsh cross-prce elastctes wth respect to yellowfn tuna and bgeye tuna are both negatve and sgnfcant, ndcatng substtuton n producton. However, yellowfn and bgeye cross-prce elastctes wth respect to swordfsh are hgher than swordfsh cross-prce elastctes wth respect to yellowfn and tuna, ndcatng an ncrease n swordfsh prce would reduce the outputs of yellowfn and bgeye more than decrease of swordfsh output due to ncrease n yellowfn and bgeye prces. In mxed trps, bgeye tuna and yellowfn tuna are found be sgnfcant substtutes. However, the effect of a change n bgeye prce on yellowfn output s much larger compared to the effect of yellowfn on bgeye. For tuna trps, bgeye s found to be a sgnfcant substtute for both swordfsh and yellowfn tuna. The bgeye s relatonshp wth yellowfn tuna n tuna trps s smlar to that n the mxed trps, whle ths relatonshp wth swordfsh s opposte of that found n the swordfsh trps. Among the remander of the speces, cross-prce elastctes between albacore and other pelagcs are always postve, ndcatng ther complementary relatonshps n producton. On the
6 other hand, bgeye tuna and marln cross-prce elastctes wth respect to other pelagc speces are always negatve. Therefore, the results suggest that although the outputs are not responsve to ther own-prces, there exst economc and technologcal relatonshps among varous speces harvested by longlners. Thus nether the sngle speces management nor treatng all speces as one aggregate stock wll be approprate n managng the longlne fshery. 2.6 Dscusson The objectve of examnng the effect of area closure on nterrelatonshps among speces could not be completed owng to lmted sample sze. Because of havng to exclude the observatons wth mssng nformaton, there were not enough observatons to estmate the model separately for before and after the area closure. 3. Work Plan for the Next Fscal Year (July 1, 2001 to June 30, 2002) 1. Identfy and estmate varables (such as fshng effort and the number of speces) and estmate the revenue functon model for troll and handlne fsheres; 2. Test for nput-output separablty and nonjontness-n-nputs of the harvestng technology of Hawa-based troll and handlne pelagc fsheres; 3. Estmate own and cross-prce supply and effort elastctes for troll and handlne fsheres; 4. Estmate the mult-speces economes of scope, speces-specfc economes of scale, mult-speces economes of scale, speces-specfc margnal costs and cost elastctes; 5. Incorporate the estmated relatons nto the exstng mathematcal programmng allocaton model as a demonstraton on how the estmated relatons can be used; and 6. Report wrte-up, ncludng preparaton of one or two artcles for journals and nternatonal conferences. 4. Lst of papers publshed n refereed journals durng FY 2001. None 5. Other papers, techncal reports, meetng presentatons, etc. None 6. Names of students graduatng wth MS or Ph.D degrees durng FY 2001. Include ttle of thess or dssertaton. None 7. For mult-year projects, provde budget for the next year on a separate page.
7 Appendx Table 1. Speces Composton of Pounds Sold by Hawa s Longlne Fshery, 1991-1998 1991 1992 1993 1994 1995 1996 1997 1998 Swordfsh trps (n = 1,225) Total pounds sold ( 000) 4,659 5,220 6,503 3,854 1,337 889 1,140 1,428 Speces shares (%) Yellowfn tuna 3.9 2.1 3.6 3.4 5.2 4.2 9.2 3.3 Albacore tuna 2.1 2.9 3.0 5.2 2.8 5.0 3.2 2.2 Bgeye tuna 3.7 6.0 9.5 6.3 8.4 6.2 9.5 8.8 Broadbll swordfsh 80.1 83.1 78.2 78.6 78.3 78.9 68.9 81.4 Marln 3.3 3.3 3.8 4.4 2.9 3.6 4.4 2.8 Other pelagcs 6.9 2.5 1.9 2.1 2.3 2.0 4.8 1.5 Mxed trps (n = 2,205) Total pounds sold ('000) 5,386 5,578 4,891 1,574 2,202 1,965 2,288 2,987 Speces composton (%) Yellowfn tuna 14.3 7.2 7.1 20.9 17.7 13.8 16.5 7.3 Albacore tuna 3.1 4.3 3.5 1.9 2.7 5.8 3.1 3.5 Bgeye tuna 10.2 17.7 24.0 28.6 21.4 16.9 11.4 21.1 Broadbll swordfsh 40.3 55.1 56.3 31.5 42.4 48.0 56.0 55.1 Marln 14.6 8.5 6.9 12.3 10.1 10.0 8.6 8.3 Other pelagcs 17.5 7.3 2.3 4.8 5.7 5.5 4.4 4.7 Tuna trps (n = 3,236) Total pounds sold ('000) 2,735 2,976 4,513 4,308 4,912 5,026 8,163 9,221 Speces composton (%) Yellowfn tuna 8.3 5.6 12.4 11.0 12.4 9.8 11.4 9.4 Albacore tuna 8.2 6.2 8.7 9.9 18.5 22.7 27.8 17.8 Bgeye tuna 16.7 43.2 40.8 48.0 33.7 34.9 34.8 45.7 Broadbll swordfsh 4.1 3.3 2.4 0.8 1.8 2.1 1.4 2.5 Marln 20.4 22.5 18.1 12.4 16.3 13.7 8.9 8.1 Other pelagcs 42.4 19.2 17.8 17.8 17.3 16.8 15.7 16.4 All trps (n = 6,666) Total pounds sold ('000) 12,780 13,774 15,907 9,735 8,451 7,880 11,591 13,636 Speces composton (%) Yellowfn tuna 9.2 4.9 7.2 9.6 12.7 10.2 12.2 8.3 Albacore tuna 3.8 4.2 4.8 6.8 11.9 16.5 20.5 13.1 Bgeye tuna 9.2 18.8 22.8 28.4 26.5 27.1 27.7 36.5 Broadbll swordfsh 47.0 54.5 49.9 36.6 24.5 22.2 18.8 22.3 Marln 11.7 9.5 8.8 9.2 12.5 11.7 8.4 7.6 Other pelagcs 19.0 8.0 6.5 9.5 11.9 12.3 12.4 12.3 n denotes the number of matched observatons from NMFS logbook and HDAR sales data.
8 Table 2. Revenue Shares of Hawa s Longlne Fshery by Speces, 1991-1998 1991 1992 1993 1994 1995 1996 1997 1998 Swordfsh trps (n = 1,225) Total revenue ($1,000) 12,788 15,123 19,329 12,686 4,416 2,798 3,089 3,223 Speces shares (%) Yellowfn tuna 4.3 2.1 3.3 3.3 5.1 5.0 8.6 3.9 Albacore tuna 0.9 0.6 0.8 1.0 0.6 1.3 1.3 1.1 Bgeye tuna 11.0 9.5 12.0 10.0 9.0 12.0 11.4 10.5 Broadbll swordfsh 81.6 85.6 81.2 81.7 82.4 76.4 74.9 81.3 Marln 1.1 1.2 1.3 1.9 1.2 1.5 1.7 1.8 Other pelagcs 1.0 1.0 1.4 2.1 1.6 3.9 2.2 1.4 Mxed trps (n = 2,205) Total revenue ($1,000) 14,430 15,064 13,844 4,976 6,051 5,975 6,364 6,666 Speces shares (%) Yellowfn tuna 14.0 8.6 7.1 19.7 16.4 14.3 14.9 8.5 Albacore tuna 1.4 1.7 1.3 0.6 1.1 2.5 1.4 2.0 Bgeye tuna 29.5 24.9 30.3 37.5 26.4 22.1 15.4 26.8 Broadbll swordfsh 46.8 57.4 57.5 34.3 49.3 52.8 62.6 55.1 Marln 4.7 4.0 2.5 5.5 3.2 4.0 2.9 4.4 Other pelagcs 3.5 3.5 1.3 2.3 3.6 4.3 2.8 3.2 Tuna trps (n = 3,236) Total revenue ($1,000) 6,377 7,267 10,461 11,483 10,064 11,389 16,924 19,714 Speces shares (%) Yellowfn tuna 10.4 6.9 14.4 10.8 16.6 12.4 14.9 10.6 Albacore tuna 5.2 3.9 5.3 5.1 9.3 13.3 16.3 10.6 Bgeye tuna 58.1 62.4 59.4 67.1 54.9 54.3 52.5 62.6 Broadbll swordfsh 5.4 3.8 3.2 1.2 3.1 3.5 2.0 2.6 Marln 9.7 12.1 8.5 7.2 6.9 7.3 5.0 4.5 Other pelagcs 11.3 10.9 9.2 8.6 9.2 9.3 9.3 9.1 All trps (n = 6,666) Total revenue ($1,000) 33,594 37,454 43,634 29,145 20,531 20,163 26,377 29,603 Speces shares (%) Yellowfn tuna 9.7 5.6 7.2 9.0 14.1 11.9 14.2 9.4 Albacore tuna 1.9 1.7 2.0 2.5 5.0 8.4 10.9 7.6 Bgeye tuna 27.9 26.0 29.2 37.2 36.6 38.9 38.7 48.9 Broadbll swordfsh 52.2 58.4 55.0 41.9 33.8 28.2 25.2 23.0 Marln 4.3 4.4 3.4 4.6 4.6 5.5 4.1 4.2 Other pelagcs 4.0 3.9 3.2 4.7 6.0 7.0 6.9 6.9 n denotes the number of matched observatons from NMFS logbook and HDAR sales data.
9 Appendx Table 3. Summary Statstcs of Ex-vessel Prces for Longlne Fshery by Speces, 1991-1998 Swordfsh Trps Mxed Trps Tuna Trps All Trps Speces Year Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Yellowfn tuna 1991 3.43 2.16 3.02 1.11 3.08 1.10 3.11 1.37 1992 2.87 1.31 3.33 1.41 2.79 1.15 3.07 1.33 1993 2.74 1.04 2.90 1.33 2.81 0.99 2.82 1.12 1994 3.27 1.50 3.06 1.07 3.16 1.49 3.16 1.40 1995 3.33 1.31 2.79 1.31 2.74 0.96 2.81 1.11 1996 3.23 1.45 3.26 1.34 3.00 1.18 3.09 1.25 1997 2.81 1.13 3.04 1.38 2.87 0.89 2.90 1.02 1998 2.63 1.14 2.74 1.08 2.72 1.06 2.72 1.07 Albacore 1991 1.43 0.81 1.41 0.57 1.65 0.63 1.49 0.65 1992 0.89 0.57 1.32 0.70 1.67 0.64 1.32 0.71 1993 0.92 0.58 1.15 0.68 1.55 0.53 1.26 0.65 1994 1.00 0.67 1.18 0.61 1.58 0.76 1.36 0.75 1995 0.93 0.48 1.11 0.76 1.34 0.66 1.25 0.69 1996 0.89 0.42 1.24 0.44 1.47 0.55 1.37 0.54 1997 1.14 1.09 1.14 0.49 1.44 0.56 1.37 0.61 1998 1.12 0.55 1.19 0.49 1.33 0.51 1.29 0.51 Bgeye tuna 1991 3.48 2.16 3.90 1.95 3.70 1.63 3.77 1.88 1992 4.54 2.77 4.04 2.35 3.58 1.27 3.99 2.17 1993 3.76 2.35 4.06 2.44 3.45 1.12 3.73 1.99 1994 4.84 3.09 4.27 2.38 3.99 1.49 4.28 2.24 1995 4.11 2.21 3.19 1.57 3.38 1.60 3.41 1.68 1996 5.06 3.06 4.37 2.18 3.58 1.29 3.92 1.83 1997 3.45 1.54 3.77 2.18 3.13 0.95 3.28 1.36 1998 2.92 1.24 3.02 1.34 3.08 1.00 3.06 1.11 Swordfsh 1991 2.82 0.86 2.98 1.10 2.23 1.24 2.81 1.11 1992 3.19 0.88 2.88 0.88 2.37 1.37 2.87 1.02 1993 3.21 0.66 2.96 0.80 2.48 1.27 2.96 0.91 1994 3.47 0.93 3.50 1.00 3.05 1.63 3.37 1.19 1995 3.42 0.77 3.13 0.79 3.10 1.39 3.17 1.07 1996 3.24 0.86 3.21 0.98 3.24 1.53 3.23 1.23 1997 2.88 1.03 3.10 0.95 3.11 1.51 3.07 1.23 1998 2.22 0.77 2.31 0.93 2.17 1.27 2.23 1.10 Marln 1991 1.27 0.72 1.06 0.62 1.19 0.59 1.13 0.64 1992 1.40 0.82 1.57 0.76 1.47 0.65 1.50 0.74 1993 1.22 0.54 1.31 0.64 1.21 0.51 1.25 0.56 1994 1.78 0.86 1.53 0.60 1.67 0.57 1.67 0.67 1995 1.68 2.05 1.15 1.00 1.01 0.56 1.12 0.96 1996 1.59 0.62 1.32 0.59 1.29 0.61 1.32 0.60 1997 1.75 1.22 1.51 0.94 1.36 0.56 1.41 0.71 1998 1.83 0.81 1.55 0.77 1.43 0.64 1.48 0.69 Other pelagcs 1991 0.65 0.63 0.94 0.80 0.90 0.62 0.87 0.72 1992 1.33 0.90 1.51 1.02 1.41 0.49 1.44 0.85 1993 1.77 2.39 1.50 1.70 1.29 0.51 1.47 1.55 1994 2.32 3.38 1.87 2.53 1.41 0.53 1.74 2.14 1995 2.38 3.37 2.02 2.82 1.19 0.50 1.51 1.85 1996 6.88 7.59 2.84 3.36 1.39 0.63 2.22 3.18 1997 2.02 3.05 3.09 4.68 1.32 0.47 1.70 2.31 1998 2.19 2.16 2.15 2.31 1.35 0.66 1.58 1.38
10 Appendx 4. Econometrc Model The generalzed Leontef's revenue functon s gven as: R (Z, P) = j β j (P P ) j 1 2 Z + β P Z 2 where R s total revenue accrued from all fsh speces harvested; Z s the amount of composte effort (nput) used, and P s a vector of prces. Accordng to Hotellng s Lemma, dfferentatng the revenue functon wth respect to prces yelds a system of output supply functons as: R (P, Z) = Q P = j β j 2 ( P / P ) 1 2 Z + β Z + β Z j,, j = A, T, B, Y, M and O The symmetry condton requres that β j = β j for j. Separablty between nputs and outputs nvolves the restrcton that β = 0 and nonjontness-n-nputs can be examned by testng the restrcton that β j = 0 j. The estmated supply equatons form the bass for computng own-prce supply elastctes for each speces and cross-prce elastctes among the pars of speces. Accordngly, own-prce elastcty of the th fsh speces (ε ) could be estmated as follows: ε Q = P P Q = 1 2Q j β j ( P / P ) 1 2 Z j Smlarly the cross-prce elastcty of the th speces wth respect to the jth speces (ε j ) could be computed as: Q P j 1 ε ( P / P ) 1 2 j = = βj j Z P Q 2Q j Fnally, effort elastcty (.e. supply response to a change n the composte effort) for the th speces (ε z ) could be computed as follows: Q Z Z Q = j β j ( P / P ) j 1 2 + β Z + 2β Z Q Note that elastctes and ther standard errors are evaluated at the observed mean values of varables nvolved.
11 Appendx Table 5. Summary Statstcs of Varables Involved n Estmatng Trp-Level Output Supply Functons wth One-perod Lagged Prces, 1992-1998 Swordfsh Trps (n = 216) Mxed Trps (n = 551) Tuna Trps (n = 352) All Trps (n = 1,119) Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Outputs (Pounds/trp) Yellowfn tuna 967 1,231 1,192 1,885 1,195 1,582 1,149 1,683 Albacore tuna 747 1,256 471 793 2,097 3,646 1,035 2,305 Bgeye tuna 1,380 1,734 2,211 2,821 3,385 3,945 2,420 3,147 Swordfsh 10,581 9,852 5,408 7,268 582 2,077 4,888 7,632 Marln 899 1,245 1,041 1,402 1,506 1,585 1,160 1,453 Other pelagcs 539 726 624 930 1,939 1,819 1,021 1,398 Prces ($/lb) Yellowfn tuna 3.04 1.13 3.20 1.23 2.93 1.04 3.09 1.16 Albacore tuna 0.99 0.62 1.29 0.67 1.43 0.58 1.28 0.65 Bgeye tuna 4.07 2.08 3.86 1.95 3.45 1.30 3.77 1.81 Swordfsh 3.24 0.92 3.03 0.95 2.91 1.36 3.04 1.10 Marln 1.64 1.24 1.51 0.78 1.35 0.60 1.49 0.84 Other pelagcs 2.03 2.54 1.70 1.53 1.52 1.11 1.71 1.68 Effort Net tonnage 74.6 30.3 66.9 23.6 61.7 23.5 66.8 25.4 Trp length (days) 18.6 7.8 11.6 4.9 12.3 3.7 13.2 5.9 n denotes the number of observatons wth complete nformaton.
Appendx Table 6. Parameter Estmates for Longlne Trp-Level Output Supply Functons (wth lagged prces) Intercept Yellowfn Albacore Bgeye Swordfsh Marln Others Effort Effort 2 Swordfsh trps Yellowfn 1,373.20** 0.188-0.059-0.303 0.220* 0.223** -0.495 8.60E-5 (240.07) (0.116) (0.149) (0.181) (0.130) (0.070) (0.366) (7.30E-5) Albacore 363.62-0.138 0.128-0.149 0.059 0.188-6.00E-5 (241.47) (0.087) (0.111) (0.101) (0.055) (0.360) (7.40E-5) Bgeye 799.33** -0.635** -0.045-0.102* 1.292** -3.00E-5 (325.70) (0.280) (0.118) (0.062) (0.490) (9.90E-5) Swordfsh 6,342.03** 0.070-0.156 4.934** -4.90E-4 (1,883.07) (0.142) (0.086) (2.389) (5.73E-4) Marln 950.92** -0.024-0.246 4.67E-7 (244.49) (0.059) (0.359) (7.50E-5) Others 577.56** -0.067 5.40E-5 (141.37) (0.189) (4.30E-5) Mxed trps Yellowfn 945.70** -0.122-0.476* 0.448-0.119 0.247** 0.538-1.60E-4 (215.80) (0.093) (0.250) (0.315) (0.178) (0.111) (0.535) (1.43E-4) Albacore 375.91** 0.136** -0.029-0.119 0.070 0.180-1.80E-6 (91.33) (0.067) (0.081) (0.100) (0.071) (0.210) (6.10E-5) Bgeye 1,184.79** 0.276-0.232* -0.260** 2.290** -4.90E-4** (320.33) (0.383) (0.132) (0.086) (0.715) (2.12E-4) Swordfsh 2,412.74** 0.748** 0.030 3.030* -5.50E-4 (808.41) (0.167) (0.105) (1.593) (5.35E-4) Marln 608.01** -0.006 0.351-2.40E-4** (158.35) (0.096) (0.333) (1.06E-4) Others 384.66** 0.346-9.00E-5 (105.47) (0.218) (7.00E-5) Tuna trps Yellowfn 982.81** 0.211-1.538 0.363 0.338 0.258 1.453* -4.10E-4 (301.99) (0.458) (0.436) (0.253) (0.277) (0.266) (0.890) (3.86E-4) Albacore 400.47-0.427 0.362 0.333 0.006 3.598* -1.93E-3** (693.47) (0.590) (0.373) (0.380) (0.355) (1.894) (8.86E-4) Bgeye 1,389.05* -0.688* 0.105-0.168 6.728** -1.77E-3* (745.81) (0.358) (0.281) (0.330) (1.950) (9.51E-4) Swordfsh -229.64 0.025 0.706** 1.048-4.60E-4 (370.34) (0.192) (0.208) (0.921) (4.72E-4) Marln 1,176.57** -0.090-0.659 8.80E-5 (302.49) (0.232) (0.832) (3.88E-4) Others 1,452.24** -0.590 1.63E-4 (339.85) (0.886) (4.35E-4) All trps Yellowfn 1,144.72** -0.186-0.290* 0.062 0.167 0.193** 0.212-6.00E-5 (129.55) (0.129) (0.152) (0.159) (0.108) (0.083) (0.278) (6.10E-5) Albacore 698.17** -0.054 0.161-0.069 0.207** 0.597* -1.90E-4** (177.30) (0.109) (0.128) (0.103) (0.077) (0.324) (8.40E-5) Bgeye 1,677.82** -0.679** -0.057-0.160** 2.430** -3.70E-4** (241.06) (0.266) (0.092) (0.076) (0.473) (1.12E-4) Swordfsh 1,739.99** 0.198* -0.078 3.854** 4.60E-5 (551.82) (0.100) (0.092) (0.959) (2.55E-4) Marln 870.61** -0.050 0.198-1.60E-4** (110.87) (0.061) (0.221) (5.20E-5) Others 782.21** 0.387** -1.20E-4** (105.96) (0.188) (5.00E-5) ** = Statstcally sgnfcant at the 0.05 level. * = Statstcally sgnfcant at the 0.10 level. 12
13 Appendx Table 7. Tests Of Hypotheses of Nonjontness n Inputs and Separablty Between Inputs and Outputs (trp level wth lagged prces) Test Statstc F Value Degrees of Freedom Crtcal Value (α = 0.05) Decson Swordfsh trps Nonjontness n nputs (β j = 0 j) 2.49 15;1,263 1.83 Reject null Input-output separablty (β = 0) 2.83 6;1,263 2.17 Reject null Mxed trps Nonjontness n nputs (β j = 0 j) 3.12 15;3,273 1.83 Reject null Input-output separablty (β = 0) 2.96 6;3,273 2.17 Reject null Tuna trps Nonjontness n nputs (β j = 0 j) 2.63 15;2,079 1.83 Reject null Input-output separablty (β = 0) 3.79 6;2,079 2.17 Reject null All trps Nonjontness n nputs (β j = 0 j) 3.10 15; 6,681 1.83 Reject null Input-output separablty (β = 0) 8.44 6; 6,681 2.17 Reject null
Appendx Table 8. Prce and Effort Elastctes for Trp Level Analyss (wth lagged prces) Wth respect to the prce of Yellowfn Albacore Bgeye Swordfsh Marln Others Effort elastcty Swordfsh trps Yellowfn -0.053 0.082-0.052-0.239* 0.124* 0.139** -0.172 (0.473) (0.051) (0.132) (0.143) (0.073) (0.044) (0.174) Albacore 0.327-0.174-0.278 0.230-0.190 0.085 0.190* (0.203) (0.152) (0.176) (0.199) (0.129) (0.079) (0.100) Bgeye -0.027-0.036 0.422-0.304** -0.015-0.039* 0.422** (0.069) (0.023) (0.591) (0.134) (0.040) (0.023) (0.208) Swordfsh -0.021* 0.005-0.050** 0.071 0.003-0.009** 2.469** (0.012) (0.004) (0.022) (0.693) (0.007) (0.005) (0.598) Marln 0.247* -0.095-0.058 0.080-0.152-0.022-0.059 (0.146) (0.065) (0.153) (0.165) (0.367) (0.054) (0.128) Others 0.374** 0.057-0.198-0.271-0.030** 0.067** 0.036 (0.131) (0.059) (0.134) (0.167) (1.00E-4) (0.010) (0.054) Mxed trps Yellowfn 0.022-0.026-0.178* 0.148-0.028 0.061** 0.218 (0.783) (0.020) (0.093) (0.104) (0.042) (0.027) (0.307) Albacore -0.166 0.044 0.202** -0.038-0.111 0.069 0.126* (0.126) (0.213) (0.099) (0.106) (0.093) (0.070) (0.077) Bgeye -0.079* 0.014** 0.078 0.045-0.027* -0.032** 1.064** (0.042) (0.007) (0.829) (0.062) (0.015) (0.010) (0.332) Swordfsh 0.034-0.001 0.023-0.098 0.040** 0.002 3.450** (0.024) (0.004) (0.032) (0.977) (0.009) (0.006) (0.639) Marln -0.067-0.043-0.144* 0.412** -0.155-0.003 0.356** (0.101) (0.036) (0.082) (0.092) (0.509) (0.039) (0.160) Others 0.220** 0.040-0.255** 0.026-0.004-0.028 0.253** (0.105) (0.043) (0.089) (0.097) (0.304) (0.066) (0.093) Tuna trps Yellowfn 0.242 0.048-0.544** 0.118 0.075 0.060 0.064 (1.153) (0.096) (0.143) (0.076) (0.057) (0.058) (0.381) Albacore 0.056-0.090-0.123 0.096 0.060 0.001 1.084* (0.122) (1.602) (0.170) (0.099) (0.069) (0.068) (0.589) Bgeye -0.163** -0.032 0.272-0.073* 0.008-0.013 1.605** (0.046) (0.044) (1.304) (0.038) (0.020) (0.025) (0.689) Swordfsh 0.244 0.169-0.500* -0.265 0.011 0.341** 0.731** (0.170) (0.174) (0.261) (0.906) (0.088) (0.101) (0.238) Marln 0.129 0.088 0.043 0.009-0.245-0.025 0.427 (0.105) (0.101) (0.116) (0.073) (0.928) (0.064) (0.352) Others 0.072 0.001-0.051 0.196** -0.017-0.201** 0.721* (0.078) (0.073) (0.106) (0.061) (0.973) (0.049) (0.366) All trps Yellowfn 0.048-0.048-0.130* 0.025 0.047 0.058** -0.020 (0.457) (0.033) (0.068) (0.064) (0.030) (0.025) (0.172) Albacore -0.130-0.014-0.042 0.112-0.033 0.107** 0.280* (0.090) (0.385) (0.084) (0.088) (0.050) (0.040) (0.150) Bgeye -0.050* -0.006 0.201-0.117* -0.007-0.021** 0.699** (0.026) (0.012) (0.568) (0.046) (0.011) (0.010) (0.238) Swordfsh 0.006 0.010-0.072** 0.048 0.013* -0.006 3.429** (0.015) (0.008) (0.028) (0.629) (0.007) (0.007) (0.402) Marln 0.097-0.026-0.036 0.113* -0.127-0.021 0.212** (0.063) (0.038) (0.058) (0.057) (0.329) (0.026) (0.092) Others 0.118** 0.081** -0.109** -0.047-0.021-0.023 0.219** (0.054) (0.032) (0.055) (0.060) (0.269) (0.030) (0.076) ** = Statstcally sgnfcant at the 0.05 level. * = Statstcally sgnfcant at the 0.10 level. 14