Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Fleet reduction through licence removal : an application of auction theory Gregory, Mary 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2000-503054.pdf [ 5.56MB ]
JSON: 831-1.0089804.json
JSON-LD: 831-1.0089804-ld.json
RDF/XML (Pretty): 831-1.0089804-rdf.xml
RDF/JSON: 831-1.0089804-rdf.json
Turtle: 831-1.0089804-turtle.txt
N-Triples: 831-1.0089804-rdf-ntriples.txt
Original Record: 831-1.0089804-source.json
Full Text

Full Text

F L E E T R E D U C T I O N T H R O U G H L I C E N C E R E M O V A L : A N A P P L I C A T I O N OF A U C T I O N T H E O R Y B y Mary Catherine Gregory B . A . , Memorial University of Newfoundland, 1988 M . A . , Dalhousie University, 1991 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Department of Resource Management and Environmental Studies) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF B R I T / S H C O L U M B I A January, 2000 © Mary Catherine Gregory, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ^UOSHYU. HWft The University of British Columbia Vancouver, Canada DE-6 (2/88) 11 A B S T R A C T Auction theory has been recognised as an appropriate valuation and allocation rule when non-standard items are transacted. This paper presents a study of the Pacific Salmon Licence Retirement Program (1996) to provide an analysis of this type of auction program in practise. The licence retirement program was motivated by a combination of overcapitalisation in the Pacific salmon fleet and resource declines. The first section discusses the excess capacity problem in fisheries management and alternative measures to address it. A description of the Pacific salmon fishery sets the context of the study and an outline of the Pacific Salmon Licence Retirement Program (1996) provides details of the auction used. A review of auction theory literature presents the basis for use of this allocation method. The data analysis of 1,722 separate bids indicates that licence values are difficult to assess but are related to licence and observable vessel attributes. The multi-dimensional scaling analysis finds that the bids submitted to the Program are close to one another in terms of reasonableness. The regression analysis supports the correlated or common value model as an appropriate bid valuation method in the gillnet, seine and troll fleet sectors. Elements of program design in this instance were specifically focused on collecting data to mimic "market" behaviour. The use of multiple bidding rounds produced cost savings to the Program of approximately $6.7 mil l ion and signaled to second round bidders appropriate licence values. The study finds that the initial passive role of the auctioneer permitted bidders to reveal information on bid values, but that revelation of the information obtained in the first bidding round could contribute to program success. Equity concerns, arising from variation in bid valuation might also be addressed through establishment of multiple bidding rounds at the outset of such a program. iii Table of Contents ABSTRACT II LIST O F TABLES IV LIST O F FIGURES V OVERVIEW 1 1.1 I N T R O D U C T I O N 1 1.2 E X C E S S C A P A C I T Y IN T H E F I S H I N G I N D U S T R Y 4 1.3 PACIFIC SALMON FISHERY SYNOPSIS 10 AUCTION THEORY LITERATURE REVIEW 21 PACIFIC SALMON LICENCE RETIREMENT 1996 33 3.1 I N T R O D U C T I O N 3 3 3.2 P R O G R A M D E S C R I P T I O N 3 5 3.3 T H E C H O I C E O F A N A U C T I O N P R O C E S S 3 8 3.4 R U L E S A N D I M P L E M E N T A T I O N O F T H E L I C E N C E R E T I R E M E N T P R O G R A M 4 2 3.5 B I D F O R M U L A T I O N A N D S O U R C E S O F U N C E R T A I N T Y 4 6 3 . 6 D A T A S E T A N D M E T H O D O L O G Y 5 2 PACIFIC SALMON LICENCE VALUATION 58 4.1 I N T R O D U C T I O N . . . 5 8 4.2 M U L T I - D I M E N S I O N A L S C A L I N G : R A N K I N G R E A S O N A B L E B I D S 6 0 4.3 R E G R E S S I O N A N A L Y S I S : R E S U L T S O F C O R R E L A T E D V A L U A T I O N M O D E L 7 2 INFORMATION ISSUES 89 5.1 T H E R O L E O F I N F O R M A T I O N IN T H E S E C O N D A U C T I O N R O U N D 8 9 5.2 S I G N A L I N G 1 1 4 5.3 E Q U I T Y 117 SUMMARY OF LESSONS LEARNED AND CONCLUSION 122 6.1 S U M M A R Y O F L E S S O N S L E A R N E D 122 6.2 C O N C L U S I O N 1 2 4 BIBLIOGRAPHY 129 APPENDIX 1 133 APPENDIX 2 136 iv L I S T O F T A B L E S T A B L E 1: P A C I F I C S A L M O N F L E E T S I Z E C H A N G E S 1 9 9 5 - 9 7 17 T A B L E 2: D I S T R I B U T I O N O F B I D R A N K I N G S F O R S U B M I T T E D 6 8 A N D A C C E P T E D B I D S T A B L E 3: D I F F E R E N C E I N S A M P L E P R O P O R T I O N S O F 6 8 A C C E P T E D B I D S B Y Q U A R T I L E T A B L E 4 : S U M M A R Y O F R E G R E S S I O N R E S U L T S B Y G E A R 81 S E C T O R F O R B O T H B I D D I N G R O U N D S T A B L E 5: G I L L N E T B I D S : D E S C R I P T I V E S T A T I S T I C S 9 8 T A B L E 6: T R O L L B I D S : D E S C R I P T I V E S T A T I S T I C S 101 T A B L E 7: S E I N E B I D S : D E S C R I P T I V E S T A T I S T I C S 1 0 4 T A B L E 8: G I L L N E T A C C E P T E D B I D S : D E S C R I P T I V E S T A T I S T I C S 1 0 7 T A B L E 9: T R O L L A C C E P T E D B I D S : D E S C R I P T I V E S T A T I S T I C S 1 1 0 T A B L E 10: S E I N E A C C E P T E D B I D S : D E S C R I P T I V E S T A T I S T I C S 113 T A B L E 11: D I F F E R E N C E S I N P R O P O R T I O N S O F A C C E P T E D B I D S B Y 114 A U C T I O N R O U N D T A B L E 12: A V E R A G E P A Y M E N T B Y G E A R T Y P E 1 1 9 V L I S T O F F I G U R E S F I G U R E 1: S E I N E M D S U N R O T A T E D 6 4 F I G U R E 2: S E I N E M D S R O T A T I O N 6 4 F I G U R E 3: G I L L N E T D A T A S E T 6 M D S U N R O T A T E D 6 5 F I G U R E 4: G I L L N E T D A T A S E T 6 M D S R O T A T I O N 6 5 F I G U R E 5: T R O L L D A T A S E T 5 U N R O T A T E D 6 6 F I G U R E 6: T R O L L D A T A S E T 5 M D S R O T A T I O N 6 6 F I G U R E 7: G I L L N E T M D S P E R C E N T A G E R A N K I N G S F R E Q U E N C Y 6 9 D I S T R I B U T I O N F O R S U B M I T T E D A N D A C C E P T E D B I D S F I G U R E 8: S E I N E M D S P E R C E N T A G E R A N K I N G S F R E Q U E N C Y 7 0 D I S T R I B U T I O N F O R S U B M I T T E D A N D A C C E P T E D B I D S F I G U R E 9: T R O L L M D S P E R C E N T A G E R A N K I N G S F R E Q U E N C Y 71 D I S T R I B U T I O N F O R S U B M I T T E D A N D A C C E P T E D B I D S F I G U R E 10: G I L L N E T V E S S E L L E N G T H V S . B I D A M O U N T P E R 8 2 F O O T F I G U R E 11: S E I N E V E S S E L L E N G T H V S . B I D A M O U N T P E R F O O T 8 3 F I G U R E 12: T R O L L V E S S E L L E N G T H V S . B I D A M O U N T P E R F O O T 8 4 F I G U R E 13: G I L L N E T V E S S E L L E N G T H V S . B I D A M O U N T F O R 9 2 A L L R E S U B M I T T E D B I D S F I G U R E 14: T R O L L V E S S E L L E N G T H V S . B I D A M O U N T F O R A L L 9 3 R E S U B M I T T E D B I D S F I G U R E 15: S E I N E V E S S E L L E N G T H V S . B I D A M O U N T F O R A L L 9 4 R E S U B M I T T E D B I D S F I G U R E 16: G I L L N E T V E S S E L L E N G T H V S . B I D A M O U N T F O R 9 6 A L L S U B M I T T E D B I D S I N R O U N D 1 F I G U R E 17: G I L L N E T V E S S E L L E N G T H V S . B I D A M O U N T F O R 9 7 A L L S U B M I T T E D B I D S I N R O U N D 2 F I G U R E 18: T R O L L V E S S E L L E N G T H V S . B I D A M O U N T F O R A L L 9 9 S U B M I T T E D B I D S I N R O U N D 1 F I G U R E 19: T R O L L V E S S E L L E N G T H V S . B I D A M O U N T F O R A L L 1 0 0 S U B M I T T E D B I D S I N R O U N D 2 F I G U R E 2 0 : S E I N E V E S S E L L E N G T H V S . B I D A M O U N T F O R A L L 102 S U B M I T T E D B I D S I N R O U N D 1 F I G U R E 2 1 : S E I N E V E S S E L L E N G T H V S . B I D A M O U N T F O R A L L 103 S U B M I T T E D B I D S I N R O U N D 2 F I G U R E 2 2 : G I L L N E T V E S S E L L E N G T H V S . B I D A M O U N T F O R 105 A C C E P T E D B I D S I N R O U N D 1 F I G U R E 2 3 : G I L L N E T V E S S E L L E N G T H V S . B I D A M O U N T F O R 1 0 6 A C C E P T E D B I D S I N R O U N D 2 L I S T O F F I G U R E S F I G U R E 2 4 : T R O L L V E S S E L L E N G T H V S . B I D A M O U N T F O R A C C E P T E D B I D S I N R O U N D 1 F I G U R E 2 5 : T R O L L V E S S E L L E N G T H V S . B I D A M O U N T F O R A C C E P T E D B I D S I N R O U N D 2 F I G U R E 2 6 : S E I N E V E S S E L L E N G T H V S . B I D A M O U N T F O R A C C E P T E D B I D S I N R O U N D 1 F I G U R E 2 7 : S E I N E V E S S E L L E N G T H V S . B I D A M O U N T F O R A C C E P T E D B I D S IN R O U N D 2 1 CHAPTER I Overview 1.1 Introduction The fishing industry in Canada has been characterized by excess harvesting and poor economic performance for participants for more than two decades despite the widespread use of limited entry programs and input controls. Open access to a common pool resource was initially argued to be the problem and restrictions on inputs, including the number of fishers, was the proposed solution. Restricting entry did not alleviate the situation in many cases. Destructive competition continued over the common pool resources leaving the industry and its participants non-viable without government support. New proposals were developed to deal with this second problem - intense competition among fishery participants to secure a share of the common pool resource. Competitive fishing practices also gave rise to overinvestment in the industry adding excess capacity to the existing problems of overexploitation and rent dissipation. Methods to deal with the incentives to improve harvesting technology and continually invest in bigger, more powerful boats and more efficient fishing gear became necessary to prevent resource depletion as well as to ensure the economic survival of the industry. Among the proposed solutions, the most influential is the use of Individual Quota (IQ) management systems for fisheries. This method specifies an amount of harvest available to each participant at the outset of the fishing season. IQ management has been 2 adopted in many fisheries around the world including Canada, Australia and New Zealand. However, this solution is not always readily able to be implemented for all fisheries facing excess capacity problems. The Pacific salmon fishery in 1996 is arguably a poor candidate for conversion to an IQ management program. The absence of a total allowable catch set prior to the fishery opening and lack of a catch monitoring program as well as the need for in-season management due to the difficulty in estimating run timing and size mean that the introduction of an IQ program would be problematic. The serious problems facing the Pacific salmon fishery indicate clearly that changes to the fisheries management system were needed to address overexploitation, conservation problems, excess capacity and rent dissipation. One element proposed to solve the problem involved the permanent removal of limited entry vessel licences from the fishery through a licence retirement program. The program operated as a "reverse" auction where licence holders submitted sealed bids indicating the amount of payment they required from the Government of Canada (which funded the program) in order to relinquish their salmon fishing privileges. This program wi l l be the focus of this study using a data set provided by the Department of Fisheries and Oceans to the author, who is employed by the Department and assisted in the program's design and delivery. Traditional auction theory suggests that, in the presence of asymmetric information, the passivity of the seller is appropriate behaviour. In the reverse auction, the buyer (the government) is passive, revealing little or no information about its position while the sellers possess information about their valuation of their fishing privileges. The thesis w i l l address the question of whether the auction method assists in the revelation of information that can be mutually-accepted to mimic market behaviour. 3 The retirement program conducted two rounds of bidding within a three-month period, with the second round closely following the first. A total of 1,722 separate bids were received in both rounds (relating to 1,274 licences) and 797 licences were removed with approximately an equal share taken in each auction round (James, 1996). The thesis w i l l consider the bids from both rounds and attempt to determine the effects of the multi-round process in defining a quasi-market and on bidding behaviour. The reverse auction method was used to allocate public funds for the purpose of resource conservation. The thesis w i l l address this policy issue in an attempt to determine i f the auction method is an appropriate allocation rule for the circumstance and might have applications for other fisheries facing overcapacity and in other resource management settings. A n overview of the program wi l l be presented to highlight lessons learned from this exercise with a view to the use of auction processes in similar situations. The paper begins with a discussion of the excess capacity problem in fisheries management and alternative measures to address problems arising in this situation. The next section provides a background description of the Pacific salmon fishery to set the context of the study. The third section provides a detailed description of the Pacific Salmon Licence Retirement Program. The following section is a literature review of auction theory. Subsequently, the data set used in the analysis is discussed in detail, followed by statistical analyses of the data. A presentation of interesting theoretical and policy implications is contained in the next section. Finally, conclusions are summarized. 4 1.2 Excess Capacity in the Fishing Industry The problems of overfishing and rent dissipation ultimately arise when access to fishing opportunities is not restricted. This theory was developed by H . Scott Gordon in his influential 1954 article. If access to a resource is open with participants who compete aggressively for a share of the resource, the result w i l l be overexploitation and a dissipation of available resource rents (Gordon, 1954). This occurs because no participant has incentives to conserve the resource (if left unexploited by any participant, another w i l l surely take his place due to open access) and the free entry w i l l occur until available profits disappear. Gordon's article recommended limited entry as a solution to the overexploitation evident in common pool fisheries where open access had been the norm. He theorized that restricting the number of participants would encourage those remaining in the fishery to recognize the destructive nature of their competition and to better use the resource. Limit ing entry moves the common pool resource on the spectrum away from open access and prevents continual growth of participation in times of resource abundance or when other employment opportunities are dismal. A t this time, fisheries managers followed biological models proposing maximum sustainable yield ( M S Y ) as the appropriate basis on which to target stock exploitation (Clark, 1985). This model assumed an underlying logistic growth model and proposed that a fishery resource could sustain such a maximum harvest level over time. The model eventually lost favour as its weaknesses became evident, including disregard for the uncertainty inherent in attempting to manage a complex biological resource with many inter-stock and inter-species relationships (Botsford et al, 1997; Healey, 1993)). The simplistic model was replaced by an 5 optimal sustainable yield (OSY) model that attempted to incorporate a wider array of complex and often conflicting objectives (Botsford et al, 1997). Economic, social and biological considerations were all incorporated into the O S Y framework. These models also failed to meet their expectations in practice possibly because incorporation of these competing objectives leads to a clear trade-off between conservation goals and socio-economic effects, such as employment levels and community stability (Botsford et al, 1997). Political pressure is most often strongest on the side of sustaining existing levels of fishing activity and continuing to support its dependent communities (Safina, 1998). Gordon's original theory found favour with other academics whose articles contributed to the literature. A n influential school began to write and formulate policy prescriptions for the "tragedy of the commons" as it applied to fishery resources (e.g. Clark, Gordon, Scott). The limited entry proposal remained a focus of discussions and its benefits were believed to include indirectly resource conservation. These studies also reviewed the possibility of the fishery being controlled by a single owner (Scott, 1955). The monopolist was hypothesized to restrict supply - which in the renewable resource context implies conservation objectives would be met - thereby restricting entry and extracting resource rent from participants. This model would ensure that overexploitation did not result and that rent dissipation would not occur, although the results are dependent on the discount rate assumed. A simple example involved the control of fishing on a lake by a sole owner who would licence access to the lake and control harvest. In most tidal water fishing nations, the control rests with a level of government which acts as a sole owner with public interest objectives. This early work focused on one input control mechanism - entry - to eliminate the common pool problem of open access. In practice, it became evident that human nature is 6 innovative. Limits placed on one input, the number of licences or participation, gave rise to a second class of the common pool problem. While participants were fixed in number, they still competed destructively for the available resource due to the common pool nature of their limited access. This is referred to as the "race for fish" type of harvest which gave way to technological advances and substitution of other inputs for the restricted input (Crutchfield, 1979; Scott, 1983). More gear, more efficient gear, more crew members to tend it, larger vessels, more powerful vessels and a vast array of fish finding equipment were added to the fishers' investments resulting in overcapitalization in the industry (Safina, 1998). This increased investment was supported in part by government funding programs and subsidies to fishers available in Canada (Swenerton, 1993). The limited entry fishery still suffered from overexploitation and rent dissipation and added excessive capital investment to the list of difficulties facing the industry. A new solution to this second common pool problem has been developed. It involves a further limitation of fishery participation and definition of a new type of access that moves further away from common pool toward "private" property. The use of individual (transferable) quotas - IQ or ITQ programs - has been proposed to address the excess capacity, capitalization and overexploitation found in many fishing industries worldwide (Crowley and Palsson, 1990). This solution has been adopted in a number of jurisdictions, including Australia and New Zealand and some Canadian fisheries such as Pacific halibut, G u l f crab and Southwest Nova Scotia groundfish (Falloon, 1993; Geen et al, 1993; Crowley and Palsson, 1990). The central idea behind the introduction of IQ programs is that by providing participants with a specified individual percentage of the available, harvestable resource at 7 the outset of the fishing season, fishers are better able to plan their fishing activity. This type of management system is designed to remove the incentives for destructive competition and the race for fish that develops (Muse and Schelle, 1989). The knowledge of an IQ at the start of the fishing season permits more orderly harvesting techniques, permits fishers to plan their investment according to their known harvest opportunities and provides incentives for resource protection (Geen et al, 1993). A further development has been proposed in the transferability of quotas which allows participants to trade quotas, in theory, from the least efficient to the most efficient thereby improving industry viability (Geen et al, 1993). These IQ programs work well in certain settings. The following is a list of conditions which provide good indications about the potential success of IQ programs: small number of participants; stable resource; total allowable catch ( T A C ) or harvest level set before the fishing season; little in-season management; catch monitoring program; and localized fishing (Falloon, 1993; Geen et al, 1993). Some of these criteria are more significant than others in determining the appropriateness of IQ programs and their potential successful implementation. In particular, in-season management, the lack of reasonably certain harvest level targets prior to fishery opening and catch monitoring are very important. The number of participants might be an obstacle but i f fishers are organized and agree to the IQ program then a large group might not impede the success of such a program. The direct advantages of IQ programs include: enhanced security of access to the resource and stability for industry participants; increased monitoring of fishing activity and better enforcement; improved industry viability and more orderly harvesting with the elimination of the need to race for fish. The indirect advantages are equally attractive such as the increased focus on conservation through improved catch monitoring and the incentive for 8 participants to protect their "investment" for the future (Falloon, 1993; Geen et al, 1993; Muse and Schelle, 1989). The IQ solution to excess capacity has proven its success in many fishery management problems in Canada and throughout the world. The Scotia-Fundy mobile gear groundfish fleet in Atlantic Canada adopted an ITQ program and saw the size of its fleet reduced by half in a few years (Crowley and Palsson, 1990). Fishery managers in Australia and New Zealand have been particularly aggressive in implementing IQ and ITQ programs in these jurisdictions. These foreign jurisdictions did not necessarily have similar experience with traditional fishery management methods, such as limited entry programs, prior to the introduction of IQ programs (Muse and Schelle, 1989). The implementation of these IQ programs was not without controversy. For example, New Zealand's fishery management reforms took place in the context of larger government-wide reforms of state-funded agencies and took a full year longer to implement than has been initially anticipated (Muse and Schelle, 1989). In addition, the orange roughy fishery which was put on ITQ management faced severe biological devastation following the introduction of the program (Batstone and Sharp, 1999). This example illustrates that catch monitoring is essential to the success of IQ programs and this type of management might not be applicable to all situations. Orange roughy is a long-lived species that takes years to reach maturity and was especially vulnerable to excessive harvesting. The implementation of IQ or ITQ programs has involved difficulties for many jurisdictions and taken years to achieve (Falloon, 1993; Muse and Schelle, 1989). One important obstacle encountered is the lack of support among fishery participants and the necessary distributional questions that surround such changes in fisheries management. 9 When fisheries experience downturns or biological collapse, the need to act quickly often causes jurisdictions to look to more immediate interventions than the implementation of management changes that could take years. Licence and vessel retirement has been used increasingly to address such situations. Holland, Gudmundsson and Gates (1999) present a detailed survey of fishing vessel and licence retirement programs undertaken in numerous countries - Canada, the United States, Denmark and Norway - for a variety of fisheries. The authors find little to recommend such programs from an economic standpoint, preferring the use of fisheries management changes to address problems in the industry, specifically IQ management. The paper suggests that many licence and vessel retirement schemes begin with competing objectives that cannot hope to be met simultaneously. This study presents a detailed analysis of one licence retirement program. Evaluation of this tool w i l l contribute to the debate concerning such measures and offer some insight into why such programs persist, despite criticism. 10 1.3 Pacific Salmon Fishery Synopsis The previous discussion outlines the reasoning behind proposed fishery management programs such as limited entry and IQ systems to address chronic fishery problems including overexploitation and excess capacity. The Pacific salmon fishery has been characterized by these factors and rent dissipation for at least 15 years, as described in the 1982 Commission of Inquiry report (Pearse, 1982), and probably for more than 30 years (Healey, 1993; Swenerton, 1993). The Pacific salmon resource suffers from overexploitation due to the large, highly mobile, technologically advanced commercial fleet that harvests the resource and due to increasing pressures from the recreational and aboriginal sectors of the industry as well (Healey, 1993; Pacific Policy Roundtable Report, 1995). Initial attempts to introduce limited entry in the late 1960s proved controversial (Pearse, 1982) and ultimately had little success in reversing the rent dissipation and overharvesting evident in the fishery. The second class of common pool problems, that of substitution of inputs and excessive investment, occurred following the introduction of limited entry programs. Fishery participants used government funding programs to capitalize their fishing enterprises through the purchase or upgrade of vessels, gear and technologically advanced equipment. This combination resulted in the Pacific commercial salmon fleet - described as one of the most powerful and mobile small-boat fishing fleets in the world (Pearse, 1982; Healey, 1993; Pacific Policy Roundtable Report, 1995). 11 The Pacific salmon fishery includes three sub-component user groups: commercial fishers, recreational fishers and aboriginal fishers. Each user group seeks to maintain (at a minimum) or enhance their access to the resource and secure a greater share of the returns from it (Pacific Policy Roundtable Report, 1995). The commercial sector includes three gear types - gillnet, seine and troll. The recreational sector includes sport fishing lodges, which are commercial businesses, as well as individual fishers from British Columbia and other parts of Canada and the world. The aboriginal sector has a unique place in the Pacific salmon fishery as the First Nations who possess constitutionally protected rights to harvest traditional species of fish. Fleet change began with the limited entry program of the late 1960s, called the Davis Plan after then federal Minister of Fisheries Jack Davis (Swenerton, 1993). This Plan consisted of four phases: the introduction of limited entry vessel licensing (applied to those who could demonstrate salmon resource attachment) was designed to create a permanent "cap" on fleet size or maximum number of salmon fishing vessels; a fleet reduction would then be achieved through licence retirement; improved vessel standards and product quality would be encouraged; and economic regulations introduced to improve fishing effort (Swenerton, 1993). These changes did not meet the approval of many in the industry and their ultimate implementation resulted in changes to the original proposals that diluted the desired effects (Swenerton, 1993). The reasons behind licensing vessels and not fishers remain unclear; this method was not applied on the Atlantic coast when limited entry was introduced in the 1970s (Levellton, 1981). The Davis Plan did achieve some success and ensured that the salmon fleet would not continue to grow. The licence buy-out removed 361 licences at a cost of $6 mil l ion and the licensing provisions designated some licences as " B " 12 (less industry attachment demonstrated) licences which were phased out over time (Swenerton, 1993). In 1982, an independent Commission of Inquiry on Pacific Fisheries Policy was appointed by the Governor-General. The resultant report of the Commissioner, Peter Pearse, responded to the broad-ranging terms of reference with an encompassing review of all aspects of the Pacific fisheries: policy, allocation, enforcement, sportfishing, consultation, habitat, the Fisheries Act, and Aboriginal fisheries (Pearse, 1982). The report's bold statement that "Canada's Pacific fisheries were at a crisis point" (Pearse, 1982, p.vii) caught attention. Concerning the salmon resources, the report made buoyant predictions about the future potential for harvest expansion. O f the five salmon species harvested in the Pacific fishery, only chinook received a bleak outlook for resource decline. A l l others were predicted to be stable or increasing in trend for the future. The report also indicated that the full benefits of resource potential could be realized through enhancement and fish production efforts. Habitat destruction in rivers leading to spawning grounds and on the grounds itself was cited by Pearse as a factor that would have a dramatic effect on salmon resource health. Rapid urbanization, logging practices and pollution play a role in fish habitat degradation. Later, the 1994 Fraser River report would elaborate on this problem as a contributing factor to diminished biodiversity of the salmon species and stocks in British Columbia rivers (Fraser River Sockeye Public Review Board, 1995). This effect decreases the ability of the resource to withstand negative impacts from fishing effort, environmental and ocean conditions and ultimately means the resource survival could be at risk. 13 The Pearse report highlighted other factors affecting resource abundance including fishing pressure and natural predation mortality. However, overfishing is cited as the major problem facing the industry and excess capacity - too many participants and too many vessels with too much gear - was believed to be the root of this problem. Limited entry did not achieve its objectives because investment in better, more efficient boats and gear ensued (Healey, 1993). Solutions to the excess capacity problems that Pearse proposed included changes to the licensing system, a fleet reduction program and changes to fisheries management which would all take place over a ten-year timeframe. The licensing changes included single-gear and area licensing for salmon and strict vessel replacement rules to attempt to control the continual expansion of the fleet (Pearse, 1982). The Commission's report made a point of highlighting the weaknesses of the Pacific commercial fishery and predicted poor future performance in the absence of change. A 50 per cent reduction in the fleet was suggested to improve the economic returns to the remaining fleet and address the bankruptcy facing many participants. (Some 30 per cent of the government-guaranteed loans to fishery participants were in default at the time.) The report discussed the advantages of individual quotas in fisheries management and suggested further study of their application to all Pacific fisheries, including salmon. The Commission's recommendations had little support with the industry and were never acted upon. In 1984, recommendations of the De Bane report reiterated many of the Pearse Commission. Area and single-gear licensing, fishery quotas, catch royalties, development of the sportfishery and a 45 per cent fleet reduction were to be delivered through a new Pacific 14 Fisheries Restructuring Act (Swenerton, 1993). The 1984 federal election intervened and these proposals were never implemented. Following the change in government, the focus of the government's role in the fishing industry shifted as the free trade agenda was pursued. The Pacific commercial salmon fishery enjoyed healthy returns from the resource throughout the 1980s, government assistance programs were eliminated and changes to the fleet did not occur (Swenerton, 1993). The early 1990s were a particularly controversial period for the Pacific salmon fishery. In 1992, a government inquiry was launched into "missing fish" on the Fraser River (Pearse and Larkin, 1992). This situation focused on increasing Aboriginal involvement in the fishery assisted through the Aboriginal Fisheries Strategy (AFS) proposed by the federal government. Aboriginal access to the fishery resource is guaranteed under the Constitution for food, social and ceremonial purposes and their access to fishery resources for commercial trade has been increased in recent years (Gislason, 1998). The recent attempts to focus on aboriginal commercial fishing opportunities is related in part to the expectation that the treaty settlement process currently under way wi l l result in increased aboriginal involvement in the Pacific fishery. In 1994, the Fraser River Sockeye Public Review Board investigated the conditions of this fishery and made a number of recommendations to the Minister of Fisheries and Oceans (Fraser River Sockeye Public Review Board, 1995). A l l these recommendations were accepted, though action on them was not as immediate as many expected. One recommendation was the establishment of a Pacific Salmon Roundtable on Policy to be convened to discuss options to manage the three sectors of the fishery. 15 Meanwhile, the salmon resources in the Pacific presently show a stable to declining trend into the future (May, 1996). Salmon resources are cyclic in nature but resource trends and salmon production in the 1970s and 1980s appeared to be increasing. The situation appears to be undergoing a significant change and the prognosis is for less resources to be available for harvest in the future (Botsford et al, 1997; Gislason, 1998). This is partly due to a perceived shift in oceanic conditions that affect the ocean survival of the species as well as ocean predation on salmon fry. These conditions led to a serious decline in the salmon fishery in 1995 that saw landings and value drop to less than half their 1991-94 averages (Gislason, 1998). Certain species are at more serious risk than others, such as chinook and coho. These latter species are the staple of the recreational fishery sector and the troll sector in the commercial fishery (Healey, 1993; Gislason, 1998). The market conditions for salmon harvested in the Canadian Pacific play an important role in the industry. The future for Canadian wi ld salmon in the international marketplace reinforces the declining resource trends (Pacific Policy Roundtable Report, 1995). Whereas British Columbia wi ld salmon previously represented a large portion of the global supply of salmon, it now must compete with farmed product from many nations, especially Norway and Chile (McKinsey, 1998). These aquaculture-based competitors produce salmon at a much lower cost than the B C commercial fleet harvests wi ld salmon. The result of this increase in supply and lower-cost competition is lower salmon prices in the marketplace - a trend which is predicted to continue in future (Gislason, 1998). The Pacific Policy Roundtable Report in 1995 was the result of a multi-stakeholder exercise that examined the serious problems facing the industry and proposed action to address them. The industry had reached a crisis point with poor returns in 1995 and 1996 16 predicted to be a low year in the cycle. Representatives from the commercial, recreational and aboriginal sectors all agreed that the status quo could not continue (Pacific Policy Roundtable Report, 1995). The Pacific Salmon Revitalization Strategy was announced in March, 1996, as the government's response to the Roundtable report. The proposed strategy was a six-point plan: conservation through risk-averse management; new licensing measures; licence retirement program; resolution of the complex intersectoral allocation difficulties; new institutional mechanisms; and transition measures. The new licensing measures were single-gear licensing, area licensing and licence stacking. Single-gear licensing changes the previous licence privilege which permitted seine licence holders to fish using seine, gillnet or troll gear and permitted gillnet and troll licence holders to fish with either gear. The change required that each licence holder designate one of the permitted gear types for the licence that would permanently establish fleet sector sizes for each of the gear sectors - seine, gillnet and troll. Area licensing requires each licence holder to designate one area where the operator wi l l be permitted to fish. Seine licence holders had a choice of two areas while gillnet and troll licence holders could opt for one of three areas. Licence stacking was introduced to permit licence holders to expand their fishing opportunities by transferring additional area licences to existing vessels. Licence stacking would then imply that fewer vessels would be used to fish with the existing number of licences. These new licensing measures were designed to reduce the pressure on the salmon resource through further input restrictions. Coupled with the licence retirement program, the licensing changes were anticipated to result in a reduction in the size of the fleet of up to 50 per cent over four or five years. 17 The licence retirement program was described in government press announcements as a "kickstart" to the fleet restructuring program which would remove licences before the 1996 season opening and do so permanently (Press Release: March 29, 1998). The focus of this study w i l l be the licence removal program to permanently retire salmon licences from the fleet. A complete description of the program is provided in later chapter. Following the retirement of salmon licences permanently from the fleet in 1996 and licence stacking arrangements, the number of active salmon vessels declined to 2,881 at the beginning of 1998, see Table 1. A vote on the future use of licence stacking provisions was held in 1998 among licence holders and a majority chose to continue with this measure (Gislason, 1998). Table 1 Pacific Salmon Fleet Size Changes 1995-1997 1995 1996 1996-1997 December 1997 Initial Fleet Licences Retired Licences Stacked Licences Fleet Active Vessels Gillnet 2,543 446 371 2,081 1,739 1,703 Seine 536 48 119 488 369 365 Troll 1,288 303 87 992 876 813 Total 4,367 797 577 3,561 2,984 2,881 Source: This table is modified from Gislason, 1998, p. 3-3. Notes: This table does not include 35 communal F licences This table does include 254 N N F C licences Concern regarding the sustainability and conservation of chinook and coho heightened in advance of the 1998 fishing season. Changing ocean climate conditions were affecting survival rates for these species and to subject them to traditional fishing pressure was believed to place some of the stocks at undue risk for survival, especially coho in the Upper Skeena and Thomson River systems (Press Release: Junel9,1998). Chinook and coho 18 are targeted by the troll sector in the commercial fishery and by the recreational sector (Gislason, 1998). In addition, coho are ubiquitous and it is very difficult for the traditional harvest methods used in the Pacific salmon commercial fishery to avoid catching coho, even i f other species are targeted. Again in the summer of 1998, major changes were announced for the Pacific salmon fishery. These changes focussed once more on the need to reduce pressure on the salmon resource through severe fishing restrictions that closed much of the coast to commercial, recreational and Aboriginal harvests. In response to these closures, the federal government announced measures to respond to the situation including a substantial salmon licence retirement program, valued at approximately $200 mil l ion over three years. Another significant component of the response was a program of habitat restoration and enhancement, allocated $100 mil l ion over five years. Along with commercial fleet reduction and improved habitat management, all sectors of the salmon fishery - commercial, recreational and Aboriginal - were encouraged to adopt selective fishing methods that would allow fishing on strong stocks but would minimize impacts on those stocks at conservation risk. Other response measures included a proposal for early retirement for fishers, modeled on similar programs offered in Atlantic Canada, assistance for recreational operators and adjustment for affected fishers and communities (Press Release: June 19, 1998). The continuation of difficulties in the Pacific salmon fishery following closely on the heels of significant management changes in 1996 indicates the depth and seriousness of the crisis. The management changes announced in 1996 were in response to an industry downturn involving price declines, slipping market share, overcapitalization and a declining 19 trend in available catch. The 1998 situation was driven by dramatic conservation concerns and the need to protect weak salmon stocks that the traditional mixed stock fishery methods would threaten. The 1998 changes focused on the way the Pacific salmon fishery is prosecuted, in contrast to the 1996 changes which involved fishery management primarily. The response to both situations highlighted the need to reduce the size of the fleet and the number of participants who depend on the salmon resource for their livelihood. In the follow-up to the conservation concerns of 1998, a series of papers were released to stimulate discourse on key issues: A N e w Direction for Canada's Pacific Salmon Fisheries, Allocation Framework for Pacific Salmon 1999-2005 and Selective Fishing in Canada's Pacific Fisheries (Press Release: June 18, 1999). A licence retirement program for Pacific salmon licence holders began in 1998 and the initial two rounds completed by the fall o f 1999 have resulted in the removal of 746 salmon licences at a cost $107 mill ion. The size of the salmon fleet remaining in the fall of 1999 is 2,557 licences (Press Release: June 18, 1999). The 1999 salmon fishing season was managed using conservation measures similar to those enacted in 1998, including a closure of sockeye fishing on the Fraser River due to low run size estimates. A n important agreement was reached with the United States in advance of the 1999 fishing season related to the implementation of the Pacific Salmon Treaty (PST) that had been the cause of conflict in previous years. In 1992, the original arrangements of the 1985 treaty expired and agreement on coastwide fishing plans could not be established in any of the following six years. The 1999 arrangements modernize the treaty and emphasize resource conservation and sharing both the costs and benefits of salmon production. The new framework uses abundance-based management to replace catch ceilings which w i l l respond 20 to conservation concerns. Pacific Salmon Treaty Endowment Funds w i l l provide $140 mil l ion to invest in the North and South for habitat, stock enhancement, science and salmon management measures in both countries (Press Release and Backgrounder: June 14, 1999) The 1990s w i l l likely be viewed as a pivotal period in the history of the Pacific salmon fishery. This decade has seen the commercial fishery decline from its peak performance in terms of landed value to face a series of crises relating to overcapitalization, price decreases, conservation concerns and management changes. 21 CHAPTER II Auction Theory Literature Review The economic theory of auction methods in goods allocation largely concerns a review of alternative auction mechanisms observed in the real world. Auctions are typically seen in markets which are thin - that is, in instances where there are few participants and usually only single items to be traded. Standard auction formats include the English or open (outcry) auction and the Dutch auction. Auction theory also covers procurement examples involving the distribution of contracts, often on the part of government, to competing bidders through the sealed bid process. The market thinness associated with auctions is a crucial feature of the model. Using the example of selling a painting for instance, the market has only one seller (who has a monopoly position) and few potential buyers of the piece of art. The item to be sold is generally considered unique or at least rare and indivisible. The scarcity of the item contributes to the reasoning that an effective method to determine its value is to offer it for sale and seek opposing bids from interested potential buyers. The seller knows the value of the item to herself but not necessarily to other bidders and is wil l ing to sell as long as the winning bid is above her "reservation" price, a minimum valuation below which she w i l l not engage in the trade. Mi lgrom (1989) outlines the reasons that the auction method might be chosen as an allocation rule. Standard auctions, he argues, are designed to benefit the bid-taker, usually the seller, to the maximum potential. In addition, other possible objectives outlined by the 22 author are to minimize transactions costs of trade, to prevent collusion or corrupt trades and to achieve efficiency. The latter is not usually associated with auctions since the prevailing view tends to be that auction results are sub-optimal. Mi lgrom indicates that auctions are commonly chosen when prices are highly variable or unstable. Wang (1993) presents a comparison of an auction method with a posted-price allocation rule, where a price is announced for the item and anyone wishing to trade at that price may do so. Wang's model studies the effect of price variability on the choice of allocation process in the context of the administrative costs of each option. There are three common auction methods cited in the literature: English, Dutch, and sealed bid auctions. The English auction is also referred to as the ascending bid or open auction and involves the auctioneer (an independent agent) starting a series of bids with his initial price (which may be the seller's reservation price) and continuing to accept bids until no bidder is wil l ing to increase. The bidder with the last and highest value obtains the item and pays an amount approximately equal to the second to last bid as the price (the winning bid must only just exceed its next rival). The model assumes that people act rationally and do not gain enjoyment from participation in the bidding process itself. In practice, many auctions include fees for entry and the auction house which are not modeled explicitly. The Dutch auction is similar in that it involves open participants calling out bids, but in the opposite order to the English auction. Dutch auctions involve descending bids from an opening high bid and the winner is the person who bids first and must pay a price equal to his bid. Dutch auctions often count down on a clock in practice, and the clock is stopped by a bidder when her valuation is reached. This method permits only the winning bid to be 23 observed, in contrast to the open and sealed bid auctions which provide information on bid valuations and their variation among participants. The final auction method often studied is the sealed bid auction which involves participants submitting written bids to a central agent who awards the item to the bidder who bids the highest value. There are two types of sealed bid auctions differing in the ultimate price paid. The first-price sealed bid auction calls for the winner to pay a price equal to her bid, while the second price calls for the winner to pay a price equal to the second highest bid. These different auction methods all have some similar characteristics outlined in the literature. A l l involve a seller who sets a "reservation" price and optimal strategies for bidders in each case involve setting a "reservation value" as the bid value. This reservation property is similar to the reservation wage set in the search theoretical unemployment models (McCal l , 1976). This reservation property means that trade may not always occur if, for example, all bids are less than the seller's reservation price. This theory is based on the following assumptions: bidders are risk neutral; the bid distribution from which bidders sample randomly is common and known; and the bid distribution has nondegenerate and finite variance and the distribution does not change over time. The final assumption regarding changing bid distribution over time can be modeled and results in a dynamic model that still exhibits the reservation property. The types of auctions outlined in the literature generally involve situations where a single, indivisible item is sold or where a single contract for service is awarded. Some authors do examine auctions which involve multiple identical or similar items for sale such as wine or horse auctions (Ashenfelter, 1989; Schotter, 1976). Auction theory is commonly studied in natural resource economics as an allocation mechanism for access rights, such as 24 exploration rights for oi l fields, petroleum or forestry leases to extract resources and in the fishery as an efficient method to allocate fishing privileges. In each of these natural resource applications it is assumed that a central monopolist, usually the government, controls the resource and there are a number of interested participants who wi l l bid to secure access to the resource. Timber and petroleum leases are commonly allocated using auction mechanisms and an example of a fishery quota auction exists for the Washington state geoduck fishery (Huppert and Brubaker, 1999). Huppert and Brubaker suggest the use of auctions to allocate individual quotas in fisheries using this management tool. A prevalent phenomenon recently has been the allocation of pollution quotas by government to industry (Tietenberg, 1992). These quotas have many of the same properties as fishing licences, inspired by a similar theoretical framework, in that they represent annualized privileges that may be traded among industry participants. Such quotas also have been capitalized in the value of the firm because without the quota, the firm cannot operate effectively. One problem with these pollution quotas that has been recognized in practice is that the system may not encourage firms to adopt new, more environmentally sensitive production strategies. Instead, the firm has incentive to acquire sufficient quota for its future production needs, perhaps due to institutional impediments (Putting Markets to Work, 1997). One might envision that at some future time, these quotas wi l l have to be reduced or eliminated and payment offered to the firms who w i l l relinquish quotas (Pembina Institute for Appropriate Development, 1996, p.45). A n auction to sell art would involve a number of participants who have different valuations of the piece which may be based, in part, on the item's resale value as an 25 investment. Presumably, the bidders' values would also vary according to taste or other individual reasons (such as complementing an existing collection) that are subjective or personal. A n auction to award access to natural resources, such as timber rights, would likely be based on information that is relatively objective. The information could (and likely would) vary considerably among auction participants depending on the amount of information publicly available, the cost of obtaining more or better information and the ability to predict future conditions, such as market prices. Ashenfelter (1989) describes the information collected through the auction process as an externality - one that the auctioneer cannot exclude from others. The price information available to observers at an auction, including unsuccessful bidders, is useful to others than the actual traders. Due to the thin nature of markets which use auction methods, the information permits others to estimate values of items similar to those sold at auction without incurring the inconvenience or cost of selling at auction. In addition, Ashenfelter argues that the public auction mechanism can serve to protect uninformed sellers and can allow trades to take place outside the auction system. The price information is certainly an important result of the auction process. The issue of secrecy of reservation price is examined in the literature. Theoretical models assume that the ascending auction begins with bids at the reservation price, but Ashenfelter asserts that, in practice, this is not the case. Instead, there is empirical evidence that reservation prices are hidden from potential bidders. Theory suggests that this is not useful, in the presence of risk neutral bidders, because truth-telling is dominant strategy. This implies that the value of the item to the bidders would not be altered by knowing the reservation price and there is no advantage to the seller in concealing it. One possible explanation for secrecy is 26 the avoidance of bidding "rings" who might collude to alter the bidding strategy in order to reduce the selling price. If truth-telling is the dominant strategy, as in the English, Dutch and sealed bid auctions, bidding rings would not survive because incentives exist for individuals to abandon the ring. One of the most intriguing results of auction theory is due to Vickrey (1961) and relates to the question of which auction method affords the highest revenue or benefit to the seller. Vickrey believed that the open auction and sealed high bid auction offer the seller the same expected revenue, under certain assumptions. This revenue equivalence theorem holds i f bidders are risk neutral, know their own valuations but not others', each bidder has same beliefs about competitors and the set of beliefs is common knowledge. In each case, the individual with the highest valuation secures the item (which means that the result is Pareto optimal and efficient) and the social value of the trade is also identical under each auction method. The reasons for employing one auction mechanism over another must then relate to other factors. These might include the cost associated with gathering all potential bidders in one location to conduct the open auction, either due to the large number of bidders and observers or travel cost. Many authors argue that sealed bid auctions protect from bidding rings more so than open outcry auctions can. Auction theory concerns asymmetrical information, usually favouring the seller, who adopts a passive role once the rules of the auction are determined. Ashenfelter and Riley (1981) and Schotter (1976) examine the passivity of the seller in an effort to determine how she might fully benefit from her monopoly position. Sellers who choose the auction mechanism seek to provide little information to potential buyers and refuse to participate 27 actively in price negotiation. Instead, the seller relies on the auctioneer to implement the rules of the auction process and permits the potential buyers to compete vigourously for the item, hoping to extract the most revenue possible from the trade (Schotter, 1976). A s such, the rules of an auction set prior to its conduct are important and should be clear and easily understood. Ashenfelter (1989) and Riley (1981) describe the common features of auctions that demonstrate revenue equivalence and, in fact, are identical in all outcomes - winning bidder, price paid, revenue to seller and optimal bidding strategies. The features are: each bidder bids an amount equal to his reserve price; the highest bidder wins; the auction rule is anonymous; and there exists a common equilibrium bidding strategy. They find also that the seller's reservation value is independent of the number of bidders and is strictly greater than the seller's personal valuation of the item. The model employed assumes that bidder values are independent - that one bidder's valuation conveys no information about any other's valuation. This assumption seems unlikely to hold in the case of auctions for natural resource rights or leases and w i l l be discussed further in the next section. Schotter (1976) describes a set of four conditions which would be viewed as desirable in an allocation rule that could be considered "optimal" and rates the auction method against these conditions. The Pareto criterion is the first which the author argues is not met by auctions. This may be true in some cases, but in fact later authors have shown that, under the assumptions outlined above the individual who values an item most highly wins it through the auction method - satisfying the Pareto criterion. The mistaken view that this is untrue may result from witnessing additional post-auction trades characteristic of , for example, the 28 sports drafts. However, these are non-standard auctions and do not adhere to the set of assumptions generally used. Schotter's second criterion is path independence or neutrality which requires that only the preferences of participants in the auction determine the outcome, the order of proceedings is irrelevant. In auctions involving multiple, like goods, this condition is not met because the sequence of items put up for auction alters the outcomes. Anonymity is argued to be a desirable characteristic of allocation rules and one which auction methods guarantee. In fact, the anonymity of auction processes is the reason they are used most often to award government contracts. Finally, low transactions and administrative costs is a desirable condition and one that most auctions, in particular the sealed bid, meet. Schotter asserts that the final two conditions are those which auctions meet most readily and for which they are chosen as an allocation rule. He also argues that this mechanism is important empirically and has tended to receive little attention. This has changed somewhat in the interim, especially in view of the ability to model auctions in the game-theoretic framework. Standard economic demand models assume generally that prices for identical items do not vary. Exceptions occur, for example, in instances where quantity discounts are offered (i.e. i f you buy three, you save). Ashenfelter (1988) offers empirical evidence from experience with actual wine auctions that price variability is common when the auction method is used. Usually, prices decline for lots of wine as additional lots are sold. The author suggests that the practice may be due to quantity constraints. Some wine auctions permit the first winning bidder to purchase all available wine lots of similar variety for the same price as the initial winning bid. If bidders know in advance that this offer may result in 29 all the wine selling to one buyer, efforts to secure an amount of the wine require aggressive bidding at the outset. Ashenfelter also explores the possibility that bidders have risk averse preferences which leads to similar results. The study of auction theory under the assumption of risk aversion among bidders has been examined by several researchers. Standard auction theory assumes that bidders are risk neutral which is important in determining the revenue equivalence result mentioned above. Riley and Samuelson (1981) and Riley (1989) demonstrate that the revenue equivalence result does not hold when bidders are risk averse. From the seller's perspective, this implies that choice of auction method might influence the profit to be realized from the sale of the item. The authors show that the sealed high bid auction best exploits buyers' risk aversion to the benefit o f the seller through greater expected revenue, compared to both open auctions and the sealed bid second-price auction. The effect of assuming that bid values are correlated is also examined by Riley (1989). The standard auctions assume independent valuations among bidders, though generally bids are modeled as drawn randomly from a common, known distribution. B i d valuations are related if, for example, bidder j has a low(high) valuation, it is more likely that bidder k also has a low(high) valuation. Riley's result is that correlated valuations mean the seller can profit from using an English auction more than the other types. The reasoning behind the finding is that, i f bidders realize that bids are correlated, the entire bid distribution shifts left (right) and bidders would adjust their bids accordingly. If not the bidders would be subject to the Winner's curse (Milgrom, 1989) - meaning that a naive bidder would bid as i f the distribution had not shifted left and therefore would bid too high. While the bidder would 30 win the item, he would have paid in excess of the true valuation of the good. In a sealed bid auction, this effect benefits the bidders but in an open auction, it benefits the seller. (Essentially, the Winners Curse occurs because the bidder (incorrectly) bases her bid on the expected outcome rather than the conditional expected outcome, conditional on bidding the most. The result is an overestimate of the value of the item.) Correlated valuations is the situation believed to exist in auctions for natural resource leases, such as for oi l or forestry extraction, that was discussed earlier. In these cases, it seems reasonable that various bidders seek information to quantify the value of the resource base and would likely have valuations that are closely related. Usually, this situation is modeled with the valuation having a common component and individual-specific component which may differentiate the degree or quality of information gathered (Chakravorti et al, 1995). This model can be represented by the following equation: v; = v + ei where v;= individual i 's bid; v = value known commonly to all ; and ei = individual i 's specific valuation from private information. In the case of the common value or correlated values model, the revenue equivalence result outlined earlier is no longer valid. Instead, the English auction is the appropriate process i f revenue generation is the primary goal. In addition, when asymmetric information is also assumed in the common value model, auctions may not be efficient and theory predictions are ambiguous relating to many other outcomes of the auction process. A s such, little analysis of these non-standard auctions is available in the literature. 31 Finally, the literature discusses the case of auctions where many similar items are offered for sale. These auctions can proceed sequentially through the items; however, Vickrey (1976) notes that in an open auction format, sequential auctioning wi l l have sub-optimal results. Bidders may wait, anticipating lower prices i f a maximum of one item is desired by any bidder, or may face monetary constraints i f more than one item is to be purchased. Schotter (1976) also comments on auctions involving many similar items which might interact, such as horse auctions, noting that the sequence of goods offered can have important implications for the results of the auction. This implies that the path independence property described earlier is violated. Sequential auctioning can result in a variety of outcomes on the number of items traded, the prices paid and revenue to the sellers depending on the order of items offered. The seller is then faced with the additional problem of setting the sequence as well as the auction rules and reservation prices. Vickrey (1976) argues that a simultaneous auction would be preferred to a sequential auction when many like goods are to be traded. The appropriate mechanism in such a case would be the sealed bid auction which could mimic these conditions. To relate to the previous discussion of the Pacific Salmon Licence Retirement Program, the sealed bid method was used because the auction was designed to simultaneously compensate many licence holders for relinquishment of their fishing privileges. Using the language of the auction literature, the program being studied conducted a simultaneous, sealed-bid, multi-round "reverse" auction. Little literature on the conduct of such auctions is available. One recent exception is the United States auction of 12 national licences to use the radio spectrum for personal communications services (Crampton, 1995). This auction was conducted as a 32 simultaneous, open, multi-round process which was logistically feasible, given the small number of participants. The next section w i l l outline how the data set available relating to the Pacific salmon licence removal program w i l l be used to address the issues discussed in the literature. This study w i l l focus on empirical evidence from the program and test elements of auction theory predictions outlined above against actual outcomes seen in the data. 33 CHAPTER III Pacific Salmon Licence Retirement 1996 3.1 Introduction Input controls relating to vessel size, such as length and hold capacity, and fishing gear are attempts to reduce overfishing. These efforts have proved largely futile in controlling capacity and have created instead incentives to substitute input factors and advanced technology. The removal of licence privileges is arguably a permanent and definitive method of reducing one aspect of capacity - the number of licensed participants. Once a licence is canceled or removed, fishing activity associated with that licence privilege terminates. In terms of conservation objectives, removal of licence privileges represents a straightforward instrument for permanent reductions in participation. The main issues surrounding removal of licence privileges are the appropriate method and any accompanying payment (May, 1996; Holland et al, 1999). Fishing licences in Canada are a privilege extended to the licencee by the Minister of Fisheries and Oceans, on behalf of the Government of Canada. In the legislation, licences can be revoked for specified reasons, including violation of regulations and conservation concerns (Commercial Fisheries Licensing Policy, 1996). In practice, however, licences have never been revoked for conservation reasons without payment (May, 1996). Canada has recent historical experience with licence removal mechanisms that have compensated the former licence holders. In the Pacific, salmon licences were "bought back" in 1972, lobster licences on the East coast were removed in 1978, salmon licences in 34 Newfoundland beginning in 1979 and continuing at intervals until 1998, and groundfish licences on the East Coast were retired over the 1993-96 period and again beginning in 1998 period in response to the Northern cod moratorium (Holland et al, 1999;). One of the largest programs of licence removal completed in Canada to date has been the Pacific Salmon Licence Retirement Program (PSLBP) announced on March 29, 1996. This program was an element of the Pacific Salmon Revitalization Strategy and was designed to remove 20% of the commercial salmon fleet (or 800 licences) for a cost of up to $80 mill ion. This program wi l l be the focus of study. Traditional auctions involve many participants who bid to secure a single item (a painting or contract) or a series of like items (livestock, forestry rights). Auction methods are used in circumstances where the item has intrinsic value but its value is unknown or believed to be subjective, varying among the bidders. Auctions are used frequently as an allocation rule in natural resource sectors. Economic theory proposes the use of auctions to allocate resource privileges, such as fishing, forestry and petroleum drilling. The reverse auction method differs from the traditional auction in that there are many, heterogeneous items (in our case, fishing privileges) which are to be "bought" by a single participant, the resource manager (the Government of Canada). Licence holders submitted sealed bids representing the value they would have to be paid in order to surrender their licence privileges. The "reverse" component refers to the nature of the auctioneer's role as buyer rather than his traditional role as seller. (This should not be confused with the ascending (English) or descending (Dutch) bidding processes discussed in the literature review.) The bids were then reviewed by an independent committee and acceptable bids were recommended to the Minister of Fisheries and Oceans. 35 The Department then transacted those bids which adhered to Treasury Board guidelines. This reverse auction method had been used in the East coast groundfish licence removal program. This method differs slightly from the traditional auction used as an allocation rule. 3.2 Program Description The licence buy-back program was designed as an auction to permit licence holders to evaluate the amount of compensation required to permanently give up their fishing privileges. Fishers with multiple licence privileges could still participate in the fishery by relinquishing only one of their licences. The program did not purchase fishing vessels or gear. The program's goal was to remove 20% of eligible licences, approximately 800 licences, contributing to the overall objective of a 50% reduction in licences through the Pacific Salmon fishery restructuring plan. The program was open to all commercial licence holders who held a salmon licence in 1995 and had paid their current (1996) licence fee. These criteria exclude licences held communally such as " N " licences and those held by the Northern Native Fishing Corporation ( N N F C ) . Information packages outlining the changes announced under the Pacific Salmon Revitalization Strategy including details about the licence retirement program and an application form for the program were sent to each licence holder in the week following the public announcement by the Minister (sample attached as Appendix 1). The policy guidelines outlined for the program were summarized by four goals. Value for money was the primary, meaning that licence values to be paid should be reasonable and not unduly inflated resulting in windfalls to program participants. Fairness 36 was another stipulated policy aim to ensure that all eligible licence holders be given the same opportunity to participate. The timing of the program had the objective of removing as many licences as possible prior to the 1996 fishing season. The final policy guideline related to the program's scope to reduce the number of licences in each gear sector and not specifically target one gear type for fleet reduction (James, 1996). The bid assessment process involved the review of all sealed bids by an independent Industry Advisory Committee ("the Committee") which would then recommend to D F O which bids should be transacted and associated licences removed. The application form indicated that the bids would be "assessed primarily on the basis of lowest cost per licensed foot". Other considerations specified included the maintenance of the proportion of reduced fee licences in the fleet (Aboriginal licences); the balance of fleet reduction among the three gear sectors (gillnet, seine and troll); the Committee might take into account the age, length and condition of the vessel licensed; and vessels licensed only for salmon fishing would be given preference over those with multi-species licences (James, 1996). The licence retirement program accepted applications until May 24, 1996, and sealed bids were accompanied by a Relinquishment, Release and Contribution Agreement. This form meant that the bids were binding on the part of the applicant unless withdrawn prior to acceptance of the bid by D F O . The relinquishment process stipulated that all salmon-related licence documents for the vessel must be returned to D F O upon notification of the acceptance of the bid. In the case of vessels licensed to fish only salmon, the relinquishment of the salmon licence would require that all licence and registration documents associated with the vessel be returned to D F O (James, 1996). 37 Initially, the licence retirement program was scheduled to end on June 30, 1996. However, a second round of bidding was announced after the initial round resulted in the removal of 400 licences at a cost of $42 mill ion. The government had committed up to $80 mil l ion to the program and announced that a second round would be conducted to attempt to meet the original target of removing 20 per cent of the eligible fleet, or approximately 800 licences. The second round of bidding immediately followed the first, beginning on June 19, 1996, and closing on July 3, 1996. The results of both rounds of bidding were the removal of 797 licences at a cost of $79 mil l ion (James, 1996). Payments to licence holders varied by licence gear type but were comparable in each auction round within gear sectors. The first round of bidding involved 1,111 applications which were reviewed by the Committee against the criteria outlined in the application form. The Committee members made every effort to balance the fleet reduction among the gear sectors but were constrained by the lack of participation by seine licence holders. The second round of the auction attracted 611 applications, some of which were resubmissions from unsuccessful first-round applicants with revised bids. The total number of licences removed by the program was 798 -447 gillnet, 48 seine and 303 troll (Gislason, 1996). In aggregate, the program was successful in meeting its goal of removing 20 per cent of the eligible fleet. A total of 4,113 licence holders could have applied to the program and 19 per cent were transacted for retirement. (Slightly more were recommended by the Committee but were withdrawn by the licence holders prior to completing the relinquishment agreement.) The program was much less successful in its attempts to balance fleet reduction across the gear sectors. Clearly, the seine fleet was not reduced by 20 per cent. This is attributable to the few applications for the program received from seine licence holders. The 38 program resulted in the removal of 85 reduced-fee salmon licences, leaving the level of Aboriginal participation in the fleet virtually unchanged at 15 per cent. O f the 798 licences removed from the fleet, more than 600 were from single- salmon licensed vessels. (James, 1996) 3.3 The Choice of an Auction Process One important aspect of the program of licence removal was the choice of an auction as the allocation mechanism. The literature review in the previous section outlined the reasons that auction methods are usually chosen including, a lack of a traditional market mechanism for the items to be traded and few participants who would take interest in the trading. These circumstances certainly applied to the commercial salmon licence removal program under consideration. In addition, little knowledge was available relating to private trades in licensed vessels because there was no central registry of transactions. Contributing to the lack of information about the potential value of salmon licences was the restructuring plan announced by the federal government which would have an impact on the value of salmon licences. In fact, the program to remove licences with compensation of up to $80 mil l ion for participants in itself would have an effect on licence values. Estimates of the volume of trade in salmon licensed vessels annually in the years prior to the program indicated that fewer than 5% of licences change hands (DFO Commercial Licensing Statistics, 1996). While data concerning licensed vessel trades in the private market was advertised in trade publications revealing the sellers' valuation, no verifiable record of the value of such 39 transfers was collected or available. Using insurance estimates to value licensed vessels was possible but was believed to overstate the vessel value and understate the licence value because the licence is not considered property and cannot be insured. This lack of information meant that estimates of the value of a licence could be conjectured only. Further complicating matters was the fact that, prior to the announcement of the program, licence transfers included the associated vessel (Holland et al, 1999). In fact, estimates based on an earlier Aboriginal licence and vessel retirement program indicate that the overall value of a licensed vessel is comprised of 26 per cent vessel value and 74 per cent licence value (Holland et al, 1999). This seems to be confirmed by evidence that, without a fishing licence attached to it, the value of the vessel in re-sale is diminished considerably (Safina, 1998). Essentially, this meant that a new "market" was being developed for the licence alone. A n alternative to an auction process is a posted price mechanism. Such a system would involve the announcement of a price or value for salmon licences and those licence holders who wished to relinquish their privileges for this level of payment could do so. A comparison of auction and posted price allocation mechanisms is presented in Wang (1993) using an individual private values model. Wang asserts that in the absence of auction cost, auctions are always an optimal allocation rule. In general, auctions are used when the demand for the item is uncertain or unknown. Mi lgrom (1989) indicates that non-standard goods are preferably allocated using auctions as are goods for which market prices are unstable. Wang accepts this intuition that dispersion of the item value might influence the choice of allocation mechanism , where greater dispersion would favour an auction method, and formulates a model to investigate. Wang highlights that the assumption of a private values model limits the applicability of analysis. The program under study does not meet the 40 requirement of a private values model; however, an examination of the data revealed through the auction mechanism might offer some insights. A t the outset, the choice of auction was made over the posted price mechanism. One problem with a posted price system is that it requires knowledge regarding the appropriate value of a licence prior to the program announcement. Given the lack of available information about salmon licence values, it would be difficult to determine the appropriate value in each of the gear sectors. The controversy surrounding the restructuring plan might have been focused on any posted price, had one been chosen. The federal government had experience with both options - auction and posted price -in the context of the licence retirement programs on the East coast following the imposition of the Northern cod moratorium in 1992. The initial government response package, the Northern Cod Adjustment and Recovery Program, included a licence retirement component where payments were based on net landed values up to a maximum of $50,000 per licence holder (Gardiner Pinfold Consulting Economists, 1994). For this payment, a licence holder would have had to agree to leave the fishery permanently (and not return to harvesting or processing) and forego any benefits under other components of the program, such as income support. This program did not have much success in attracting participants and resulted in 376 licence retirements (with over 9,000 eligible) at a cost of $13.4 mill ion, representing mainly licence holders with marginal fishing activity (Gardiner Pinfold Consulting Economists, 1994). When the successor to N C A R P , the Atlantic Groundfish Strategy (TAGS) , was designed, the licence retirement component was altered to follow a reverse auction process (though other design elements were similar) and was allocated $271 mil l ion initially over the 41 five years of T A G S (Holland et al, 1999). The licence retirement component of T A G S suffered from budget shortfalls associated with the overall implementation of T A G S and its funding was reduced to $60 mill ion (Holland et al, 1999). Three rounds of licence retirement were completed and a total of 466 licences were removed (Press Release: August 8, 1996). Despite criticism of the T A G S program, representations from all East coast provinces to the Post -TAGS review process indicated that a follow-up licence retirement program at the demise of T A G S would be received with more interest (Harrigan, 1998). Finally, another method of valuing fishing licence privileges that warrants consideration is the present discounted value calculation of the future stream of benefits and costs from the licence. Valuation using such a method involves important assumptions about the future, including price and cost data and resource projections. This method has been rejected as unreliable in attempts to apply it to other fishery valuation situations, such as the determination of quota value in trade (Harmsma and Davidse, 1993). The reported poor performance of this method and its reliance on assumptions about uncertain future circumstances mean that is has not been used in practice. The auction mechanism was chosen as the method to implement the licence retirement the program. It was not without challenges which wi l l be presented in the section below. The rules, costs, second auction round and review committee wi l l be discussed, followed by descriptions of bid formulation models and the role of uncertainty in the licence removal auction. 4 2 3.4 Rules and Implementation of the Licence Retirement Program Each participant in the auction was required to submit an application form indicating the vessel length of the licensed vessel, its gear type, ownership information and bid amount. The form had to be signed by each of the registered owners and notarized. This process was designed to ensure that legal transactions could be undertaken once the bid had been accepted. A l l bids were binding on the part of the applicant once submitted to Fisheries and Oceans. Bids could be altered or withdrawn prior to acceptance by the review committee and signature for transaction with written notification from the applicant(s). The application form also served to indicate the proportion of the total bid payment to be paid to each registered owner in the cases where more than one owner existed. Other aspects of the Pacific Salmon Revitalization Strategy may have had impacts on the bidding behaviour of program participants. Those who applied to the licence removal program were given a longer time period in which to designate their licence gear type (e.g. gillnet or troll). This factor is reported to have caused some individuals to submit bids which they later withdrew, having had no intention to get out of the salmon fishery, but simply wishing to delay designating the area selection for their licences. The extent to which this might have inflated the initial number of program participants cannot be determined with certainty. Another factor which is believed to have influenced the altering of bids is information about the total number of applications to the program. The overall Revitalization Strategy was controversial and sparked public protest which resulted in the occupation of several Fisheries and Oceans offices. These disruptions meant that processing applications to the 43 licence removal program was interrupted and delayed, leaving the impression among some fishers that few applicants had come forward. When information became available to the public (following a court injunction to permit the resumption of work in Fisheries offices) that hundreds of applications had been received for the licence removal program, individuals who wished to exit the industry had incentives to ensure their bids were competitive. There was no administrative fee to submit, alter or withdraw applications for the program. A l l applicants were required to renew their commercial salmon licences for the 1996 season before the privilege could be relinquished at a cost of between $1100 and $3600 per licence, depending on the gear type. In addition, an indirect application fee that might have been incurred was the cost to notarize the application form (perhaps $50). Because of time pressure and occupations of Fisheries and Oceans offices at various locations, the application forms were submitted by mail, facsimile (with original mailed) or in person at Fisheries and Oceans offices. The bulk of applications were received during the final days of the application period and were entered directly into the computer system designed for the program. A s a result, many incomplete applications were received, some of which could not be considered serious bids (in excess of $1 million). A n administrative fee or penalty might have prevented "joke" applications and either would be feasible given that all applicants were also licence holders. Payments were made to successful bidders following acceptance by the Minister of Fisheries and Oceans of the Industry Advisory Committee's initial recommendations concerning the first bidding round. The transactions were completed upon signature by two designated departmental officials. Note that payments under this type of program are considered as income for purposes of taxation and, in Canada, the disposal of the fishing 44 licence privilege is taxed as a capital gain, i f one is realized. Participants to the Program were advised to investigate the tax implications of their individual circumstances with respect to surrender of their licence. The second round of bidding occurred immediately following the first and after the acceptance and transaction of the successful first round applications. A l l licence holders who were unsuccessful in the first round had the opportunity to bid in the second round as did any eligible licence holder who had not previously submitted an application. The purpose of the second round was to encourage as much participation as possible from the fleet, particularly from the seine fleet sector whose first round participation was lacking with only 50 applicants from a pool of more than 500 licence holders. The minimal participation of the seine sector may be attributed to their preference to use the licence stacking option to rationalize their fleet, rather than the licence retirement program. The second round was administered in exactly the same manner as the first round with exactly the same rules and same application form. Both new participants and unsuccessful first round participants were required to submit a complete and new application form. The Committee reviewed the bids and made recommendations to the Minister which were accepted. Transactions were completed and payments made to successful bidders. Since virtually all the allocated funds for the program were dispersed, the program closed. The Committee was comprised of representatives from the fishing industry including each of the three commercial gear sectors, the processing sector, the aboriginal sector and was chaired by J im Matkin, a lawyer. The United Fishermen and Al l i ed Workers Union was asked to participate but declined for the first round because of their opposition to the overall 45 Revitalization Strategy. The U F A W U was represented during the second stage of the Committee's work. The Committee worked together and initially focused on the program's objectives, participant's aims and bidding behaviour as well as licence trades in the private market to assist in their review of the bids. A l l decisions of the Committee were unanimous and their recommendations regarding acceptable bids were approved by the Minister. The initial bidding round left the Committee with the problem of rejecting too many bids because they were unreasonable. The Committee recommended a second auction round and indicated that the majority of applicants did not seem to have enough knowledge or information to bid reasonably. I w i l l return to this topic in the next section. The Committee set a goal for itself that the bids in the second round would not exceed the successful first round bids on the basis of cost per licensed vessel foot. A s such, they built in a measure of equity to those licence holders who participated in the first round and whose bids formed the basis for the Committee's view of a reasonable bid. The Committee made every effort to ensure that windfall gains did not accrue to any individual. Concern was expressed by the public that second auction round participants would be in an advantageous position. The dissemination of information from the first round was hypothesized to offer unfair gains to participants in the later round. The Committee members recognized this problem but also acknowledged that the reason that the majority of bids were unsuccessful in the initial bidding round was precisely caused by lack of information and knowledge. 46 3.5 Bid Formulation and Sources of Uncertainty The Industry Advisory Committee recognized at the outset that formulating an appropriate bid would be a difficult challenge for many applicants to the licence removal program. Information on the numerous other changes affecting licence holders and therefore licence values was required to properly evaluate the payment required to relinquish a licence. In addition, private trades in licences occurred due to the introduction of licence stacking and each licence holder had to assess her own position in the changing salmon fleet. The perceived need to offer a second chance to unsuccessful bidders was based largely on the belief that many applicants had mistakenly formulated bids that were unreasonable. Information was available to applicants from a variety of sources. Boat brokers were active in the private market for licensed vessel trading and were used by some participants to formulate and submit bids on their behalf. (This information was available on the application form for some bidders who chose a broker to act as his/her agent.) In addition, some licence holders may have attempted to trade in the private market prior to applying to the government program to "test the market". Evidence from the data suggests that many applicants did not seek expert advice. For example, many bids were received from the gillnet and troll sector offering licence relinquishment for $100,000 compensation. This might be explained by individuals performing the following simple calculation. The Program budget was $80 mil l ion and its aim was to remove 20 per cent of the fleet or 800 licences, therefore the average cost per licence should be $100,000. 47 This method is flawed because the program objective to remove licences from all gear sectors meant that seine licences would have to be included whose value is substantially more (averaging $400,000) than small boat (gillnet/troll) licences. Therefore, those who bid on this seemingly naive basis did not acquire sufficient information on which to formulate a bid. Licences from the gillnet and troll fleet ultimately were transacted by the program at an average of $78,000, while the majority of first round bids in these sectors were in excess of $100,000 per licence. Formulating a bid required consideration of many factors and probably included an element of retirement for some licence holders and giving up of a lifestyle and traditional livelihood for others. In addition, the Revitalization Strategy encouraged the private market trade in licences by introducing area-based licensing and permitting the "stacking" of multiple licences on a single vessel. Licence stacking is subject to restrictions on vessel length whereby a vessel of a given length can stack a licence from another vessel whose length is within a 30% range of the initial boat. This restriction might have affected the bidding behaviour of unusually short and long vessels in the fleet whose options to find appropriate stacking partners were few. The applicant would also have to make a decision on the seriousness of the bid and the risk of rejection. For those who truly hoped to leave the fishery, the decision to participate in the licence removal program probably meant that they did not invest the time, effort and money (perhaps $10,000) required to gear up for the summer fishing season. A s such, rejection of their bid would leave them no alternative but to sit out the season. Since a low bid would almost surely be accepted, the bid value is an indicator of the bidder's risk position. In addition, only one bidding round was foreseen at the outset and the total program 48 funds were fixed so that news of high participation rates might have made risk averse bidders nervous. Prior to the spring of 1996 and the implementation of the salmon fishery restructuring strategy, salmon licence values had been quoted in terms of licensed vessel feet (e.g. $1600 per licensed foot). However, little activity in the private market was observed (Adelson Consulting, 1996). The dismal performance of the 1995 fishery left the industry with millions in losses and dire financial circumstances for many individual participants (Gislason, 1996). (This inability of the fleet to withstand an unexpected resource downturn highlighted the need for structural adjustment.) The appropriate method to formulate a bid using this underlying value formulation would then be to determine the appropriate valuation per foot and simply multiply by the vessel length. A s an example, in the troll fleet sector, the appropriate valuation term might be $2000 per foot and the average vessel length 38 feet for a bid of $76,000 to surrender the licence. A n alternative and reasonable bid formulation might be based on an average value of salmon licence privilege in a fleet sector with discount or premium based on specific characteristics. This is a commonly used formulation in natural resource valuation models where items are believed to have a common value component, usually widely known, and an individual value component which requires additional information to evaluate. These models generally hypothesize that the common value component is available to most bidders and that the individual component is based on information asymmetries among participants in the auction. Bidders in the licence removal program might have had access to general information about a licence value for a particular gear sector and then upgrade or diminish their individual 49 bids based on particular characteristics of their vessel licence privilege. These specific attributes might include vessel length, age or hull construction which would have an impact on the licence value in the private sector and the amount of compensation required to voluntarily leave the salmon fishery. The formulation for the common value model bidding process is: vig = vg + eig where: v,g= individual i's valuation of the licence in gear sector g vg = average valuation for licence in the gear sector g eig = individual value component for individual /''s licence in gear sector g. This formulation has the potential to incorporate upgrading bids for newer vessels that might represent greater fishing capacity or discounting bids for older vessels representing less harvesting capacity. Vessels with fibreglass hull construction and newer technology are better able to follow the salmon runs along the coast and are more efficient. A s such the bidder might view these characteristics to be more valuable and worth a premium over the average value in the gear sector. I w i l l use the term overbidding to refer to the situation where an individual bid is above the average in the gear sector (i.e. ej g > 0). The model also accommodates bids discounted because vessels are older or shorter (longer) than average. Older vessels might obviously have downgraded bids due to their inefficiency or lesser harvesting capacity. It is interesting to note that both shorter and longer vessels might be expected to discount bids. A short or long vessel would have few opportunities to sell in the private market or find an appropriate partner to "stack" a licence 50 and remain in the fishery, compared with a licensed vessel of average length. Changes to the licensing regulations permitted licence holders to transfer licences for stacking purposes from or to vessels within 30% of the stacked vessel's length. A s such, unusually short or long vessels would have more difficulty conducting a private market transaction and might view the licence removal program as the best opportunity to leave the industry with compensation. A discount from the average bid would be reasonable to ensure that the bid be acceptable. Such a discount might also be expected to apply to bids involving multi-licensed vessels since the program clearly indicated preference would be given to reasonable bids from single-licensed salmon vessels. The situation of a discount from the average bid (i.e. ej g < 0) w i l l be termed underbidding. The common value model departs from the private market licence trades advertised prior to the auction. This formulation suggests that the government-sponsored auction might establish a separate trading process. These two alternative formulations w i l l be investigated using the data from the auction, relating to transacted bids, to determine which model seems more reasonable in the next chapter. The auction process and formulation of a bid is subject to a variety of uncertainties, especially in the context of a natural resource sector. The greatest source of uncertainty is likely the restructuring strategy which affected all aspects of the salmon fishery - those who would leave the industry and those who would remain. The licensing system changes required each licence holder to choose a single fishing area in which to harvest the resource, limiting access from the previous system. A private market for licence transfer was encouraged by the licence stacking provisions. The risk-averse management program for resource conservation which was introduced in 1995 Fraser Report was re-asserted. Risk-51 averse management would reduce the number of fishery openings, particularly for chinook and coho stocks which are in precarious states and are of particular importance to the troll fleet sector. In addition, the allocation of salmon among the commercial, recreational and aboriginal sectors was to be studied to resolve this controversial sharing arrangement. Uncertainty regarding the future prospects of the health of the commercial salmon fishery was a serious consideration in 1996. Salmon prices were low and competition in the global market from farmed fish was expected to mean continued low prices (Gislason, 1996). Coupled with low landed volumes, low prices contributed to the disastrous 1995 salmon fishing season. The 1996 fishing season was anticipated to be a low year in the salmon cycle which would lead to a second year of poor performance for industry participants. The Pacific Salmon Treaty (PST) with the United States remained unresolved. The Treaty relates to the sharing of the salmon resources that pass through American waters before entering Canadian rivers and streams to spawn. Without resolution, Canadian fishers would continue to lose fishing opportunities to American fishers who have first access to some of the salmon runs. A further source of uncertainty was the conflicting messages associated with the controversial restructuring plan. The federal position outlined a fairly dismal future for the commercial salmon fishery based on conservation concerns, user group conflicts and financial troubles for the participants. The provincial N e w Democratic (NDP) government, in the middle of an election campaign, made the restructuring an election issue and insisted that the dire predictions from the federal government were overstated and the fault of federal mismanagement ("Clark Calls for Provincial Fish Agency", May 10, 1996; "Clark 's Fisheries Takeover Plan Won't Work: Mi f f l in" , May 12, 1996). The N D P , the United Fishermen and Al l i ed Workers Union ( U F A W U ) and some gear sector representatives 52 supported the view that the current difficulties were anomalies, not a long-term trend, and that encouraging the growth of a fishery based on small, independent participants was a better plan for the future than the proposed solution (Annual Report 1996/97 Ministry of Agriculture, Food and Fisheries, 1998). Attempts were made to halt the restructuring by going to court to claim the federal government was overstepping its authority but the attempt failed ( " B . C . to Challenge Salmon Plan", May 22, 1996; "Cuts to Salmon Fleet Approved", May 24, 1996). The overall impact of this controversy is unclear but certainly contributed to uncertainty on the part of licence holders that the restructuring plan might be significantly redesigned or abandoned. A s such, some licence holders might have concentrated more on opposing the plan or awaiting changes to it rather than on making the difficult decisions associated with the announced changes. 3.6 Data Set and Methodology The auction process under consideration is not a simple, single-item case which can be modeled explicitly using developed methodology. The multi-item, multi-round and simultaneous nature of the Pacific salmon licence removal program complicates many of the assumptions used to formulate and solve the theoretical models presented in the literature review. A s such, this study w i l l focus on an examination of the auction data collected and outcomes realized to draw conclusions relating to which aspects of auction theory remain valid in this "imperfect" context. A thorough review of the data w i l l offer insight into the operation of a non-standard auction and its feasibility as an allocation mechanism. In 53 addition, this review permits an overview of this publicly-funded program with a view to assessing how the auction process achieves policy goals. In studying the licence removal process undertaken by the Department of Fisheries and Oceans (DFO), a data set relating to the licence retirement program has been made available. The data set includes an array of information concerning the bids received from salmon licence holders and the successful bids that were transacted by D F O on behalf of the government of Canada. Each applicant to the Pacific Salmon Licence Retirement Program (PSLBP) was required to complete an application form and submit it to D F O . The forms were then entered into a special database designed for the program so that the Industry Advisory Committee (Committee) could review the applications. The data used in this study is derived from this database, respecting the privacy of bidders and successful applicants. The data were sorted in the database by gear sector of licence holder (gillnet, troll and seine). In total, the first round of bidding received 1,111 applications for the program. These included 640 gillnet, 60 seine and 411 troll applications. The second round of bidding attracted 611 applications - 323 gillnet, 27 seine and 261 troll. There were a total of 448 resubmissions to the second round from unsuccessful first round applicants, leaving 163 new applications. Recall that there were two reasons to proceed with a second round: to attract more applications from seine licence holders and to come closer to realize the initial target of a 20 per cent fleet reduction, using the $80 mil l ion allocated. The complete database collected information relating to the vessel, licence holdings, bid amount and the applicant. The vessel information includes: an identifier, vessel length, vessel age, construction material (e.g. fibreglass, wood), builder and a photograph. The 54 information concerning licence holdings relates to the gear type of the salmon licence, the salmon licence status - as full fee or reduced fee (Aboriginally held A-I) and other licence holdings such as shrimp, groundfish. The bid amount is included as submitted by the licence holder in a total dollar amount for the salmon licence offered. The larger information set collected licence holder age and residence on a self-reported basis but this information has not been included in the data set to be used here. Some applicants also chose to include additional information to their bid amount, such as how they calculated the bid and their views on the program which has also been excluded. The data set used in this study includes the following information, a sample printout of which is presented on the next page: vessel identifier (to track bids through the second round but not presented); hull construction material (ranked 0 to 4); vessel age; vessel length (in feet); bid amount ($); licence status (e.g. reduced fee or not) and other licence holdings (e.g. shrimp). Statistical analysis techniques w i l l be applied to the data set to investigate bid formulation and licence valuation. A non-parametric statistical analysis called multi-dimensional scaling (MDS) to rank the reasonableness of the submitted bids. This method is appropriate for the mixture of qualitative and quantitative data available. M D S is an ordination method which is able to provide rank scores on a defined scale, in this case, the reasonableness of the submitted bid. Sample of Data Reb id H u l l Age Length B i d Amount Fee Other 0 1 33 34.51 95000 0 0 0 0 31 34.51 79000 0 0 0 0 27 34.51 90000 0 0 1 0 45 34.51 85000 0 0 0 0 45 34.51 102000 0 0 0 1 26 34.51 84000 0 0 0 1 45 34.51 85000 0 0 0 0 34 34.51 82000 0 0 1 0 36 34.51 89726 0 0 0 0 36 34.51 117368 0 0 0 0 36 34.51 200000 0 0 1 3 4 34.51 89495 0 0 0 3 4 34.51 101095 0 0 0 3 15 34.54 138200 0 0 1 3 16 34.57 87950 0 0 0 3 16 34.57 102250 0 0 2 0 36 34.57 88700 0 0 2 0 43 34.57 100000 0 0 2 0 36 34.57 84000 0 0 0 0 34 34.57 50000 1 0 0 1 56 34.57 85000 0 0 1 3 7 34.57 110000 0 0 0 3 7 34.57 135000 0 0 0 3 1 34.57 60000 0 0 0 0 38 34.64 104400 0 0 2 0 37 34.64 85000 1 0 0 3 1 34.67 209095 0 0 1 3 19 34.67 86000 0 0 0 3 19 34.67 110000 0 0 0 0 27 34.67 72786 0 0 1 3 1 34.67 116000 0 0 0 0 43 34.74 78095 0 1 1 0 40 34.74 126000 0 1 0 0 40 34.74 129000 0 1 0 0 27 34.77 90000 0 0 2 3 26 34.77 79987 1 0 0 0 62 34.80 75000 0 0 0 3 45 34.83 83545 0 0 1 3 17 34.83 85000 0 0 56 Regression techniques w i l l be used to compare and contrast licence values by gear type by first grouping bids into each of the three gear sectors and to investigate the valuation model used in the bid formulation process. The auction theory literature suggests that the common value auction model is appropriate in natural resource settings. This model involves two components in the bid formulation process - a common value known widely among participants and an individual-specific value based on private knowledge or information. This study wi l l use data for accepted bids to determine i f the common value model is applicable to the successful bids. This might suggest that even in such a complex setting, the model theorized to represent a reasonable bidding formula remains valid. The data w i l l be also be analyzed graphically and bids w i l l be correlated to other vessel and licence characteristics to investigate how bids vary and how valuations are determined. Auction theory further suggests that information is an important externality or product of the auction process itself. That is, choosing the auction mechanism usually implies that markets are thin or price variability is common therefore, conducting an auction gathers information previously unavailable and this information itself has value. The licence removal program conducted two auction rounds over a very short time frame and offers an excellent opportunity to examine the information externality. The data set w i l l be separated into each of the two auction rounds by gear sector and wi l l be analyzed to determine i f the bidding behaviour exhibits evidence that information from the first round was available to second round participants. The implications of this information exchange w i l l also be discussed in terms of its effect on the success of the program and equity to participants. A straightforward examination of the bids - both those accepted and rejected - using statistical measures and 57 graphical analysis w i l l provide insight into the extent of information dissemination between the auction rounds. The auction process policy guidelines were concerned primarily with equity and value for money. Concepts of equity w i l l be discussed and the success of the auction method in contributing to equality of opportunity and equitable outcomes wi l l be analysed. The information on vessel and licence characteristics w i l l again be used to analyze the variability of bids based on observable characteristics. The "value for money" policy guideline could be viewed as equity from the perspective of the use of public funds for the program. This issue wi l l also be discussed and the auction process w i l l be reviewed to assess its effectiveness as an allocation rule using public funds. Finally, an overview w i l l be presented of the lessons learned from the Pacific salmon licence retirement program with a practical view to the applicability of such a process to other fisheries and other resource sectors. This program is one of the largest of its kind ever completed in Canada to address the problem of excess capacity in any fishery. Evidence globally suggests that most fishing nations suffer similar difficulties with respect to existing fisheries operating with too much capacity which face resource downturn. The auction process may be applicable to any situation where resource values are unknown, uncertain or highly variable and where at least one side of the market includes few participants. 58 CHAPTER IV Pacific Salmon Licence Valuation 4.1 Introduction The Pacific Salmon Licence Retirement Program conducted a multi-item, multi-round simultaneous sealed-bid auction to retire fishing licences from the Pacific salmon fleet. The data set associated with this program offers an unique opportunity to examine the results of an auction and investigate the valuation of items, in this case, fishing licences, using this mechanism. The data set related to the Program contains a mixture of quantitative and qualitative data. In analysing the data, two statistical techniques are used - multidimensional scaling and regression analysis. The first is a non-parametric technique which results in an ordination of the data, such as the rankings of attributes, along some measured quality (Pitcher, 1999). The second type of analysis is the more familiar linear regression technique, which places some assumptions on the distribution of the data. These two different analytical tools are applied separately to the data for distinct purposes. M D S is applied to the entire data set of submitted bids to the Program in an effort to rank bids along a measure of reasonableness. The regression technique is applied to the data set of bids accepted by the Program's Independent Advisory Committee to ascertain which characteristics contribute to an acceptable bid and which bid valuation model is appropriate for the licence valuation. 59 The M D S technique has been applied to other fishery-related studies (Pitcher, 1999; Pitcher and Power, 1999) in order to rank fisheries according to measures of sustainability. The specific technique used in these studies is called "Rapfish" and a variant of this method has been employed for this analysis (Pitcher and Preikshot, 1999). The attraction of using M D S to analyse the Pacific Salmon Licence Retirement auction data set lies in its ability to rank attributes of both a quantitative and qualitative nature using a non-parametric technique. The set of all submitted bids to the licence retirement contains over 1700 observations on seven characteristics, four of which are qualitative and three quantitative. This volume of data warrants analysis to determine i f a pattern can be discerned in the bidding behaviour related to the reasonableness of the bids. The set of submitted bids to the Program was considered by an Independent Industry Advisory Committee ("the Committee") which reviewed the bids and selected those which were assessed to be reasonable based on the observed characteristics. These acceptable bids were then recommended to the Minister of Fisheries and Oceans for transaction. The M D S analysis offers a statistical interpretation of the "reasonableness" based on the collected data. The ordination technique w i l l rank the bids, using seven attributes, on a scale of reasonable to unreasonable. To structure the analysis, reference points have been fixed using construed extreme "reasonable" and extreme "unreasonable" bids. This chapter presents the M D S and regression analyses. A n explanation of the M D S method wi l l follow along with the results of the procedure. Finally, an interpretation of the ordination w i l l be summarised. The regression analysis is presented in a subsequent section. 4.2 Multi-dimensional Scaling: Ranking Reasonable Bids 6 0 The M D S method is useful when the data set to be analysed contains both quantitative and qualitative variables. The purpose of the analysis is to represent a multi-dimensional information array in fewer dimensions for ease of interpretation and is based on the concept that distances between the data points should correspond to their observed proximities in some manner (Everitt and Dunn, 1991). The technique involves the definition of an objective function which measures the dissimilarities between the observed proximities and fitted distances. A n optimisation algorithm is then used to recover the configuration of points in a specified number of dimensions, by minimising the objective function. In this study, the M D S technique employed results in a two-dimensional output array. The technique is non-parametric and the output ranks the observed data on a scale, in this case, the reasonableness of the bid in rank order. This scale is appropriate for this study because the bids to the licence retirement auction were compared against each other on the basis of reasonableness to determine which would be recommended for acceptance. The criteria used by the committee to determine acceptance was based on the attributes contained in the data set. A n interesting question arises as to whether a statistical analysis of the data, such as M D S , could produce outcomes similar to the Committee's consideration or significantly different from the work of the Committee. Data on each bid to the program include identifiers (such as name, address and personal information which is not recorded here; for analysis purposes, the data are anonymous), vessel age (year built), vessel length (in feet), hull material of the vessel, licence gear type (gillnet, seine or troll, exclusively), licence status (full fee or reduced fee, 61 exclusively), other licence holdings and bid amount. Additionally, the data were coded according to which bidding round the bid was entered and i f the bid was resubmitted following rejection in the first bidding round. Note that 15 bids are removed from the data set of gillnet and troll bids because of missing values on some of the attributes. The seven attributes used in the M D S analysis are: bid status, hull material, vessel age, vessel length, bid amount, licence fee category and the presence of other licences. Three of these characteristics - vessel age, length and bid amount - are quantitative while the other four are qualitative. Vessel age is obtained by subtracting vessel built year from the Program year, 1996. Vessel length is reported in feet and bid amount in nominal Canadian dollars. The qualitative attributes are presented as follows. The status of the bid refers to its submission to the program where 0 = Round 1 submission; 1 = re-submitted bid (or rebid) to Round 2; 2 = second round bid only. Hu l l material refers to the type of construction of the vessel where 0 = wood; 1 = fibreglass-covered wood ; 2 = steel; 3 = fibreglass; 4 = aluminum. The licence fee category reports whether the licence is full fee (=0) or reduced-fee (=1). The presence of other licences captures whether the bidder's vessel has other licence holdings, besides the salmon licence offered for retirement, where 0 = no other licences; 1 = one other licence; 2 = two other licences; and 3 = three or more other licences. The M D S method applied using the Statistical Package for the Social Sciences (SPSS, 1999) for Windows. The M D S procedure restricts the number of cases that can be analysed to approximately 100 because of the difficulty in calculating the ordination (SPSS, 1999). For this study, the seine data set of submitted bids contains fewer than 100 cases (bids) but the gillnet and troll data sets are larger. To overcome this constraint, the gillnet and troll data 62 sets are divided into groupings, in ascending order of vessel length, which the M D S technique can accommodate. In addition, fixed reference points were included in each data set or grouping. The result is 12 data set groupings for the 956 submitted gillnet bids and 9 data set groupings for the 659 submitted troll bids. The M D S technique is applied to the original seine data set of 83 submitted bids and to each of the groupings for the gillnet and troll data sets, supplemented by the four fixed reference points. The Rapfish analysis is designed to allow the resultant rankings associated with each submitted bid to be assessed on a scale of reasonableness. The lower end of the scale represents an unreasonable bid and the upper limit signifies a reasonable bid. The fixed reference points included in each data set grouping permit interpretation of the output scale readily. The extreme unreasonable bid is created by attaching maximal values to the seven attributes in the M D S analysis, while the extreme reasonable bid is constructed using corresponding minimal values for the attributes. (The hull material score used for the extreme unreasonable bid is 0 while the hull material score used for the extreme reasonable bid is 4.) Two other fixed points are construed by assigning half the attributes with maximal scores and half minimal scores for one and the opposite pattern for the other (Pitcher, 1999). The M D S analysis is applied using standardised z-scores thereby ensuring that the different attribute scales do not affect the final ordination (Pitcher and Power, 1999). For ease of presentation, all output rankings from the M D S analysis are "rotated" to situate the reasonable bid fixed reference point on the right-hand extreme of the x-axis. This convention permits comparison of the data set groupings from each of the gillnet and troll data sets. The raw M D S results and "rotated" results are presented graphically for the seine 63 data in Figures 1 and 2 and for one representative of the gillnet (Figures 3 and 4) and troll (Figures 5 and 6) data sets. These examples illustrate the ordination technique results. The rotation of the graph to present the fixed reference point for the reasonable bid at 90° (or the right-hand side of the x-axis) does not alter the M D S results but offers a more direct interpretation of the graphs. Figure 1: Seine MDS Unrotated Figure 2: Seine MDS Rotation • 1.0 0 -7.5 -7.0 -6.5 -8.0 -5.5 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0 Extreme Unreasonable • -0 .5* • • • d> ^ 1.0 1.5 2.0 2.5 3.0 3(5 Extreme Reasonable Axis 1 Figure 3: Gillnet Data Set 6 MDS Unrotated «o 1 -A u * 5 -9 -8.5 * -7.5 -7 -6.5 - i -5.5 -5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 H V I 1.5 2 2.5 3 3.5 •1 -2 • -3 • 1-Axis 1 Figure 4: Gillnet Data Set 6 MDS Rotation 0. -3.5 -9.0 -8.5 -8.0 -7.5 -7.0 -6.5 -6.0 -5.5 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 % - 0 . 5 4 i 1 - 5 20 2S 3.0 3.6 Jo Extreme Unreasonable -1.5 Extreme Reasonable Axis 1 Figure 5: Troll MDS Unrotated Axis 1 Figure 6: Troll Data Set 5 MDS Rotation .0 -8.5 -8.0 -7.5 -7.0 -6.5 -6.0 -5.5 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 Extreme Unreasonable 1.0 1.5 2.0 2.5 3.0 3.5 4|0 Extreme Reasonable a* Axis 1 67 Interpretation of the rotated rankings for the gillnet, seine and troll data sets reveals a clustering of bids together but no other patterns emerge in the results. The stress scores for each analysis (which measure goodness of fit), including all the groupings within the gillnet and troll data sets, are below .25 which is considered acceptable (Pitcher, 1999). The M D S analysis is designed to reveal i f bid rankings exhibit patterns, however, the absence of separate clusters or other interesting arrangements of the rankings suggest that no distinguishing groupings occur. In particular, a pattern corresponding to the reasonableness of the bid is sought. The results suggest that the determination of reasonableness is complex - it is not easy to distinguish reasonable bids from unreasonable ones. The single cluster of bid rankings offers little insight into the determinants of reasonableness, but does indicate the difficulty in classifying the bids on this basis. To simplify the exposition, the Rapfish "rotated" rankings are projected onto a scale of reasonableness from 0 to 100 per cent. The scale is based on the Rapfish results of ranking the fixed reference points for each data set. This method permits comparison of the results from each of the groupings within the gillnet and troll submitted bids data sets. Table 1 presents a summary of the results of the percentage rankings for the data based on gear types, where the percentage rankings have been grouped according to categories. For example, the gillnet M D S results indicate that 105 submitted bids are between 65 and 70 per cent "reasonable". O f those 105 submitted bids, 62 were accepted by the Committee's assessment. Figures 7, 8 and 9 present the percentage rankings graphically. These results mirror the clustering of bids evident in the M D S results. 68 Table 2 Distribution of Bid Rankings for Submitted and Accepted Bids Gillnet Seine Troll Total Rank Bids Accepted % Bids Accepted % Bids Accepted % Bids Accepted % 50 0 0 0 0 0 1 1 100 1 1 100 55 1 0 0 0 0 0 0 0 1 0 0 60 4 4 100 2 1 50 1 1 100 7 6 86 65 16 11 69 10 2 20 25 12 48 51 25 49 70 105 62 59 15 7 47 153 78 51 273 147 54 75 354 175 49 30 16 53 317 138 44 701 329 47 80 350 143 41 23 19 83 135 54 40 508 216 43 85 116 55 47 3 3 100 16 6 38 135 64 47 90 10 6 60 0 0 9 6 67 19 12 63 95 0 0 0 0 0 2 2 100 2 2 100 100 0 0 0 0 0 0 0 0 0 0 0 Total 956 456 48 83 48 58 659 298 45 1698 802 47 Using the sample test statistic for differences in proportions to compare the probability of acceptance in the top quartile (proportion 1) to that of the next quartile (proportion 2) for each of the gear sectors yields interesting results presented in Table 3. The probability of acceptance in the top quartile for the gillnet bids is less than the probability of acceptance for the adjacent quartile. For the seine bids, the probability of acceptance is greater for bids in the top quartile of reasonableness than for bids in the next quartile. For the troll bids, there is no distinguishable difference in the probability of acceptance between bids in the top quartile and those in the adjacent quartile. Table 3 Difference in Sample Proportion of Accepted Bids by Quartile Proportion 1 Proportion 2 Number of Bids 1 Number of Bids 2 Test Statistic Gillnet 0.4286 0.525 476 480 -2.9979* Seine 0.8462 0.4561 26 57 4.0327* Troll 0.4198 0.4617 162 496 0.9359 * Indicates significance at the 2.5% level. 69 Figure 7: Gillnet MDS Percentage Rankings Frequency Distributions for Submitted and Accepted Bids 375 - i — 300 — I 50 55 60 65 70 75 80 85 90 95 100 Reasonableness Percentage • Submitted • Accepted Figure 8: Seine MDS Percentage Rankings Frequency Distributions of Submitted Bids and Accepted Bids • [ i 1 I 0 55 60 65 70 75 80 85 90 95 100 Reasonableness Percentage • Submitted II Accepted 71 Figure 9: Troll MDS Percentage Rankings Frequency Distributions of Submitted and Accepted Bids 350 325 300 275 250 225 on 200 . u °o <u 175 c r S-i 150 125 100 75 50 25 0 — 1 1 L 1 I l~L l~ta 50 55 60 65 70 75 80 85 90 95 100 Reasonableness Percentage @ Submitted • Accepted 72 Note that almost no bids fall below the 50 per cent reasonable mark. This is partially influenced by the choice of fixed reference points in the M D S analysis. That is, the submitted bids are "closer" to the reasonable bid fixed reference point than to its unreasonable counterpart. This does not have a qualitatively important impact on the results which involve an ordination or ranking of the data. That is, for example, one cannot assert that all submitted bids are more than 50 per cent reasonable based on this analysis. Instead, the important result is that all submitted bids are close to one another in terms of reasonableness. Summary of MDS Results The M D S results demonstrate convincingly that the bid rankings are clustered together with no identifiable patterns within each fleet sector. This result offers little insight into a method to determine the reasonableness or acceptability of the bids and may be interpreted as support for the use of a Committee to assess the bids. The bid assessment process is challenging in the presence of bids which are close together in ranking, making it difficult to readily distinguish reasonable bids from unreasonable ones. 4.3 Regression Analysis: Results of Correlated Valuation Model While the M D S analysis demonstrates that the submitted bids are clustered in ranked ordination, the regression technique offers an opportunity to investigate the subset of 73 accepted bids to explore the characteristics that contribute to bid valuation. This section presents analysis of the accepted bid data from the 1996 Pacific Salmon Licence Retirement program using linear regression analysis, by gear sector (gillnet, seine and troll). Alternative models of bid valuation are discussed at the outset. The results of the regression equations are then presented for each gear sector. In addition, graphical presentations w i l l illustrate the variation of bids. Finally, the appropriate bid valuation model for each fleet sector, based on the observed behaviour from the auction data, w i l l be discussed along with their practical implications. Alternative Models Economic theory concerning valuation of natural resources indicates that the auction value of access to resources might follow the correlated or "common valuation" model (Riley, 1989) discussed earlier. This model argues that the resource access privilege can be valued as a linear combination of an average value, based on commonly known information, and values that are associated with individual-specific characteristics that differ across bidders who possess asymmetric information on which to base their valuations. The model can be summarised by the following equation: Vj = v + ej + Uj where: Vj = individual i 's bid valuation v = average or common value ej = valuation based on individual i 's specific characteristics/information 74 ui = error term (independently and identically distributed). Regression analysis techniques can be applied to this model by assuming that errors are identically and independently distributed. This statistical tool w i l l permit an analysis of the available data from the licence retirement auction to determine the applicability of the common valuation model to this circumstance. Recall that an alternative explanation of bidding behaviour was suggested in advance of the licence retirement program. The fixed value model assumes that licences are assessed a valuation based on length of the associated vessel with a constant value applied per licensed vessel foot. This idea was based on the observation that private licence and vessel trades were advertised in trade publications in this manner, quoting dollar value per licensed vessel foot (e.g. The West Coast Fisherman, Pacific Fishing). The resulting bid valuation would take the form: Vj = Clj + Uj where: Vj = individual i 's bid c = constant value per licensed vessel foot li = vessel length (in feet) Uj = error term (independently and identically distributed). The constant value "c" may differ across gear types but would be the same for bidders within each gear sector. Note that the regression analysis of the common value model w i l l incorporate the vessel length in the bid valuation and yield a co-efficient on vessel length (similar in interpretation to the constant "c") but w i l l also include other licence and vessel characteristics. 75 The analysis of data from the licence retirement auction w i l l provide insight into the applicability of these alternative bid valuation formulations to this case. Such information is of practical use because the conduct of a licence retirement program could vary considerably, depending on which bid formulation model is supported. For example, the program under study used the auction method to retire fishing privileges. This involved the collection, validation and collation of considerable amounts of data through the sealed bid process. Analysis of the common value model could provide information on which characteristics contribute significantly to the explanation of bid valuation and which are not important. This information would be useful in future programs (applied to any fleet sector) to restrict data collection to those characteristics or attributes that explain bid valuation significantly, thereby potentially reducing the cost and administrative burden of such programs. In particular, i f the valuation method characterised by the fixed value per licensed vessel foot model is favoured as the appropriate model, then the cost and effort to collect individual vessel information (other than vessel length) would be wasted. Instead, efforts might be better focused on determining the correct value of the constant "c" (value per licensed vessel foot) to avoid the auction method and simplify the licence retirement process. This data set also presents an opportunity to analyse the data from two bidding rounds immediately following one another. This permitted rejected bidders from the initial bidding round to re-submit bids in the subsequent round. Analysis of this data w i l l provide information on the bidding behaviour of these bidders and their effect on the licence retirement program. 76 Data Sets Recall that data were collected on each bid to the program that included identifiers (such as name, address and personal information which is not recorded here; for analysis purposes, the data are anonymous), vessel age (year built), vessel length (in feet), licence gear type (gillnet, seine or troll, exclusively), licence status (full fee or reduced fee, exclusively), other licence holdings and bid amount. Additionally, the data were coded according to which bidding round the bid was entered and i f the bid was resubmitted following rejection in the first bidding round. The data set used for the regression analysis is a subset of that used in the M D S analysis. Only those bids which were recommended for acceptance are included in the regression technique in an effort to explore the factors that affect bid valuation. Restricting the analysis to accepted bids is designed to mimic market conditions hence, only those bids that were transacted are considered. Note that the characteristics bid status and hull material are not included in the regression analysis. For this reason, seven gillnet and troll bids which were missing these variable values are excluded from the M D S analysis but are included in the regression analysis. The common value model regression equation takes on the following form when applied to this data set: Vj = ao + ai agei + a2 length; + feej + a4 otherj+ as round2 + Ui where: Vj = individual i 's bid amount 77 age; = age of individual i 's vessel (associated with licence for retirement) length; = vessel length of individual i 's vessel feej = 0 i f full fee licence = 1 i f reduced fee licence other; = 0 i f no other licence holdings = 1 i f other licences held round2 = 0 i f bid submitted in first round = 1 i f bid submitted in second round U ; = error term ao - intercept term, interpreted as the average or common value of licence aj , &2 , aj, a4 = co-efficients calculated by the regression. Applying this regression equation separately to the data for each gear sector w i l l yield three equations outlining which characteristics contribute to explaining the bid variation; the signs of the co-efficients w i l l indicate the direction of influence (inflating bids or reducing them). 1 The data to be analysed includes the bids recommended by the independent Licence Retirement Advisory Committee (the Committee) which comprised representatives of each gear sector, of the Aboriginal fishery, of the processing sector and an independent chair. The Committee's recommendations were made to the Minister of Fisheries and Oceans; all their recommendations on bid transactions were accepted. The sample sizes for the gear sectors include 459 accepted gillnet bids, 48 accepted seine bids and 302 accepted troll bids in the two bidding rounds. The total for the entire data set is then 809 accepted bids which is slightly less than the total number of accepted bids to the program (817) because the data set has been cleaned. A total of eight bids from the 78 gillnet and troll sectors for which incomplete information was available to complete the regression analysis were removed. The sample sizes for the gillnet and troll sectors are sufficiently large to produce meaningful results and assume that the associated error terms are identically and independently distributed. The sample size for the seine sector is relatively small in comparison (because few seine licences were offered to the retirement program) however, enough observations are present to offer valuable insight into the auction process and results. Results The results of the regression equation applied to the separate gear sector data sets for both bidding rounds together yielded the output in Table 4 which presents the co-efficient values and their associated t-statistics. The R-squared for the gillnet regression is 0.42 indicating that the proposed model formulation explains 42 per cent of the variation in gillnet bid values of accepted bids to the program. The results for the gillnet sector indicate that all co-efficients except that on the vessel age characteristic are significant. This suggests that bid variation in the gillnet sector can be explained using the vessel length, reduced fee status, other licence holdings and bidding round indicator along with the common or intrinsic licence value indicated by the intercept. Another specification of the regression equation was suggested in which the vessel length attribute has a non-linear effect on the bid value (see Appendix 2). 79 The non-significance of vessel age in the regression analysis may be explained by the relatively low cost of vessel replacement or refurbishing in this fleet. The licence value exceeds the vessel value in this fleet, according to insurance brokers (Adelson Consulting, 1996). This is consistent with the idea that resource access or licence value is more closely aligned with potential capacity than existing capacity. In that sense, vessel length and licence transferability would contribute positively to licence retirement value while the potential to remain in the fishery and re-use one's skills would detract from the value of retirement. Interpretation of these results offers insight into bidding behaviour and confirms many of the expected outcomes. The significance of the intercept and its value ($33,013) support the common value model and indicates the intrinsic licence value, that is, the value of a licence when all other attributes are ni l . This value may be representative of the inherent value of the salmon fishery in British Columbia based on its historical importance and profitability. Vessel length is a significant contributor to bid valuation and the estimated value per licensed foot of a gillnet licence, when the average value is also accounted for, is $ 1,316. Vessel length can be interpreted as a proxy for physical capacity and bid value would be expected to increase with vessel length. A reduced fee licence is estimated to be valued less than a full fee licence for retirement by $10,172. This can be attributed to the licence transfer restrictions associated with a reduced fee licence and the lower licence fees paid. A salmon licence belonging to an individual with other licence holdings was valued an estimated $10,363 less than a single-salmon dependent licence. The presence of other licences was expected to deflate the value of licence retirement because vessels and licence holders associated with multi-licensed 80 vessels could remain in the fishery. Unlike programs to retire licences on the East coast (e.g. Northern Cod Adjustment and Recovery Program ( N C A R P ) , the Atlantic Groundfish Adjustment Strategy (TAGS)) , licence holders in this program were permitted to remain in the fishery after retiring only the salmon licence. Furthermore, the program stipulated that preference in the bid assessment would be given to single salmon licence holders, so those with other holdings knew in advance that they would have to bid competitively (i.e. lower) to have their bids chosen. Finally, the second round bid dummy variable co-efficient indicates that an accepted bid to the second round would be $8,269 greater than the first round bid, when all other attributes are also considered. This suggests that second round accepted bids were higher than corresponding first round bids and may indicate that information was exchanged between bidding rounds regarding acceptable bid values. This topic w i l l be explored in more detail in the next chapter. 81 Table 4 Summary of Regression Results by Gear Sector for both Bidding Rounds Gillnet (n = 459) Seine (n = 48) Troll (n = 301) Coeff. t-statistic Coeff. t-statistic Coeff. t-statistic Intercept 33,013* 9.3 292,384* 7.12 7,791* 2.23 Age -3.59 -0.70 -395 -1.14 37.5 1.13 Length 1,316* 12.45 2,695* 3.87 1,680* 18.67 Fee -10,172* -6.52 -65,467* -3.2 -10,006* -4.18 Other -10,363* -6.98 -60,495* -3.16 -6,445* -5.34 Round 2 8,269* 8.92 31965** 1.76 8,603* 8.73 * Indicates significance at the 2.5% level. ** Indicates significance at the 5% level. Analysis of the data for accepted gillnet bids to the program suggests that more attributes than vessel length explain the value of licence retirement. The more simplistic constant value per licensed vessel foot formulation does not account for the intrinsic value of the licence nor the other factors, such as fee status and other licence holdings. To illustrate the alternate model, a graph showing the accepted bids where vessel length is plotted against bid value per foot is presented in Figure 10. If the value per licensed vessel foot model is valid, there should be a constant value across all vessel lengths. This is not the case for the data presented. Instead, the graph demonstrates that bid value per licensed vessel foot decreases with vessel length, again supporting the idea that there is an average value component to the licence retirement value. Longer vessels, with greater potential capacity, have greater overall licence retirement values but lower values per licensed vessel foot. 82 S3 Ofj • 1-H PH _L _L _L o in in o in in o in m o in o (N in in r/3 > a* " 2 OT u OT OT u > -4—» CU C o o o o o o o o o o o o o o o o o o i n o i n o i n o i n m m (N (N i - H ($) JOOJ jsd junoine p i g 83 > • i t CD COO • i - H P H L o o o o 00 o o o in o o o C N CD CD C/5 CD > o a, -a IS > ti) e CO 0) C/3 CO CD > 4) '55 3 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o C N ($) ;ooj jsd lunoure ptg 84 ( N U • 1-H j_ _L _L _L _L _L o in in o in in -xT-in m o in C N o in in CO a* >• o " 2 OT > C a) <u OT OT u > o H CN S* bp o o o o o o o o o o o o o o o o o o o i n o i n o i n o i n o i n c \ c \ f N C \ c \ * \ c \ c \ ^ r - ^ - c n m c N c N ^ H ^ H ($) ;ooj jsd rrrnoure pig 85 The seine data set was analysed using the same regression equation model and the 48 accepted bids over two bidding rounds. The resulting co-efficients and their associated t-statistics appear in Table 4. The R-squared for this regression is 0.41, indicating that 41 per cent of the variation in seine bids can be explained by factors included in this model. In this case like the gillnet case, all co-efficients are significant except vessel age. Again bid variation is explained by observable characteristics of the licences submitted to the retirement program based on a common seine licence value of $292,384, estimated as the intercept. The value per licensed vessel foot contribution is estimated to be $2,695. A reduced fee seine licence is estimated to be valued less than full fee licences by $65,467 and a multi-licensed seine vessel's bid is estimated to be worth $60,495 less than a single-salmon licensed seine vessel. Finally, a second round bid in the seine sector had an increased value of $31,965. The signs of these estimated co-efficients confirm the effects expected and mimic the results of the gillnet regression. The estimated co-efficients are larger in absolute value than those resulting from analysis of the gillnet data because seine licences are the most highly valued in the salmon fleet. A seine licensed vessel catches seven times the landings of a vessel from the small boat fleet, on average (Gislason, 1996). Similar patterns emerge for the seine bids as those observed for the gillnet sector - reduced fee licences are estimated to be of lower value and the presence of other licences deflates the estimated accepted bid value. The significance of the second round bid dummy variable indicates that bidding behaviour may change between the two bidding rounds and this topic w i l l be explored further in the next chapter under signaling. 86 The seine data analysis also suggests that the more simplistic fixed value per licensed vessel foot approach to seine licence retirement valuation is inadequate to explain the bid variation. Other observed characteristics contribute to the valuation of licence retirement bids besides vessel length alone. Again, Figure 11 presents the plot of vessel length against bid value per foot for the accepted seine bids in order to illustrate that the pattern is not fixed but decreasing in vessel length. This is again consistent with the common value model but does not support the fixed value model. Analysis of the troll data set of accepted bids (which includes 301 bids over two bidding rounds) applied to the common value model yielded similar results to those for the gillnet data set. The regression results are presented in Table 4 with t-statistics for the associated co-efficients. The R-squared for this regression is 0.65 indicating that variation in the bid values can be explained by the model's characteristics - vessel length, licence status, the presence of other licences, round 2 bid indicator and common value - except the vessel age attribute whose co-efficient is insignificant. The troll regression results are similar in order of magnitude to the results of the gillnet regression. A Chow test can be used to determine whether both data sets might be represented by same regression equation by comparing the results of a regression of both sets of data together with a single regression for the troll subset. The results of this test indicate a value of 1.49 for the F-statistic. The critical F value (at 1 per cent confidence) is 1 for the degrees of freedom (756, 301) and the test statistic does not support acceptance of the hypothesis that the gillnet and troll data sets can be represented by a single regression model. 87 The alternative model, the constant value per licensed vessel foot, suggests that for all vessel lengths, a similar bid value would be observed once corrected for vessel length. The troll accepted bid data exhibit this trend in constant or fixed bid value per foot to a much greater degree than the gillnet or seine accepted bid data. The graph of troll vessel length plotted against bid value per foot presented as Figure 12 illustrates the cluster effect that the constant value model predicts. Similar graphs for the gillnet and seine accepted bid data (Figures 10 and 11 respectively) show that the bid value per foot declines as vessel length increases which supports the common value model rather than the fixed value model. Summary of Regression Analysis Analysis of the auction data set using regression and graphical techniques indicates that the correlated or common value model applies to each of the gillnet, seine and troll sector bids. This suggests that the constant value per licensed vessel foot formulation is not an appropriate valuation method in this situation. Further, the results suggest that the auction mechanism permits the valuation of licences on the basis of observable characteristics. The support for the common value model is important because the appropriateness of the auction model confirms the ability of this mechanism to reveal valuations in the absence of a well-defined "market" for items to be valued. In addition, the data sets have been tested and indicate that a separate model applies to each of the gear types with different co-efficients. 88 Note that the auction mechanism does not inhibit these different bidding models from emerging within the fleet sub-sectors. Instead, the auction method permits each sector to define its own unique "market" for licence valuation - evidenced by the different co-efficients that emerge for licence valuation attributes across fleet sectors. This flexibility is an important feature which contributes to the usefulness of the auction method as an allocation mechanism, particularly in the absence of a well-defined market. Another important advantage of the auction method occurs when multiple bidding rounds are employed. This technique has the potential to use signals from initial bidding rounds to transmit information to participants regarding the process which w i l l be explored in greater detail in the next chapter. 89 CHAPTER V Information Issues 5.1 The Role of Information in the Second Auction Round The previous section of this paper described the bid formulation process undertaken by applicants to the licence removal program. I have argued that the common value model, familiar in natural resource settings, underlies the valuation of the licence privilege relinquishment studied here. The program being considered conducted two rounds of auction in a short timeframe - over a matter of two months. Attention w i l l now turn to the role of information in the auction process. The literature on auction theory makes reference to the information externality evident in open auctions, where all present can learn about others' valuations of the items to be traded. The licence removal program was conducted as a sealed-bid auction because of logistical difficulties. Despite this constraint, the applications from each auction round and the results w i l l be examined to address the issue of information revelation. The role of information is crucial in formulating a bid, particularly for those who participated in the auction's second round. A s noted earlier in this paper, concern was expressed from government officials and fishery participants that second round participants might have advantages over first round bidders (May, 1996). The Industry Advisory Committee recommended a second round, explaining that the bulk (700 bids) of the first round bids were rejected because they were unreasonable. This was viewed as evidence that applicants needed additional information in formulating a bid or, at a mimmum, guidance in 90 order to revise their bids to a more appropriate level (James, 1996). The Committee recommended transaction of the acceptable first round bids and a second auction round. A n examination of the data w i l l shed some light on this information issue with a view to improvements suggested in hindsight. The discussion of bid formulation indicated that sources of information were available to all . The licence changes, guidelines and application form for the licence removal program were sent to all salmon licence holders. Boat brokerage businesses are located throughout British Columbia and are engaged in licensed vessel trades on an ongoing basis. Some applicants to the program indicated that a broker was authorized to act on their behalf which clearly suggests that advice had been sought prior to submitting a bid. The private market in licence transfer, not necessarily licensed vessel trade, was active due to licence stacking provisions and offered licence holders an option to explore valuation of their licence privilege. How many licence holders took advantage of these information-gathering opportunities cannot be known with certainty. Graph ica l Analysis A graphical analysis of the data w i l l be presented to examine the role of information in the auction process. The first sets of figures relate to bids from the first auction round that were resubmitted in the second round. This data is not restricted to the successful bids but includes all bids for each of the three gear sectors independently. The second sets of graphs present the data for all bids submitted in the first and second rounds, not just those bids that were resubmitted and are separated into the three gear sectors. This w i l l give a general sense 91 of how the potential "market" for licences offered changed between the two auction rounds. The final sets of figures refers to only the successful bids from the first and second auction rounds. These graphs describe the actual market for transactions in each round for each of three gear sectors. The final figures consider the same data presented in the earlier section on data results but is separated into the two auction rounds for this analysis. Conclusions w i l l be drawn from the data and discussed along with the presentation and analysis of each set of figures. Resubmitted Bids Beginning with the resubmitted bids of unsuccessful first round applicants w i l l determine i f the review committee's hope of signaling that bid values were too high was achieved. The data indicate that overall bid values fell for those who re-applied to the second auction round. Not all those who applied at the outset decided to participate in the second round - only 400 re-applications occurred. Clearly, the message was received that very high bids were unreasonable and that bid values should be revised downward. One bid that was received in the second round has been excluded from this analysis because it offered licence relinquishment for $38 mill ion. This data observation resulted in awkward graphing and has been omitted. Figure 13 shows gillnet sector bids with first round bid value plotted on the left-hand-side and second round bid values on the right; vessel length is plotted on the x-axis (due to the volume of bids, multiple vessel length categories appear). In general, bid values fell from •o GO •1 -H U H Qn « ^ H 2 3 J L J I I L o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o ( N O t ^ ^ f N O t ^ i A i M O N ^ M O r ^ i r i f N ' t ^ n f n n n r N l t N f N l f N l r H r H r H r H ($) junoury P!H 93 94 CN C/3 X) 4> OO CD b f j fe CD a • CD CD C/3 C/3 CD > 1/3 4) -»-» a o T 3 C/3 > c 4) 4> C/3 C/3 4) > 4) *5 CN 4) bp 9? CO CO CO a> a> to I O <si •— ($) junoury P!H 95 the first round for all vessel length categories. There are bid values which did not decline -some stayed the same and one increased. The troll sector resubmitted bids appear in Figure 14 and again indicate that those who chose to re-apply to the second round generally reduced bid values. Again first and second round bids are plotted on the y-axes and vessel length along the x-axis. The initial unsuccessful bid appears to be a basis from which the new bid amount is determined. The final figure for the seine sector data on resubmitted bids is Figure 15 and contains only 11 data points. This graph shows that the seine bids resubmitted in the second round did not fall dramatically from their initial position. One explanation for this might be the knowledge that few seine licences had been removed by the program and bidders hoped to take advantage of desire to achieve fleet reduction in this sector. All Bids The next set of graphs relate to the data from each auction round, for each gear sector and include all submitted bids (not just those resubmitted). These figures present vessel length on the x-axis and bid amount on the y-axis. Effort has been made to use the same scale in each pair of graphs for ease of comparison. These graphs could be viewed as global, potential "supply" curves for the licence removal program in each gear sector, along the lines of the supply curves analyzed in the previous section. 96 97 ($) limouiv pie 98 The first pair of figures, Figures 16 and 17, presents the data for the gillnet sector. First round bids show an upward trend as vessel length increases which flattens out considerably for bids in the second round. Note that fewer bids of extremely high value were submitted to the second round than in the first. In addition, fewer bids from shorter vessels (under 25 feet) and from reduced fee licences are evident in the second round. Overall, the bids in the second round exhibit a ceiling effect where bids cluster at bid values between $70,000 and $100,000. While a similar cluster is evident in the first round bids, the effect is more exaggerated in the second round data. Table 5 below presents some descriptive statistics which relate to the bid distributions in each auction round. Table 5 Gillnet Bids: Descriptive Statistics Item Round 1 Round 2 No. of bids 634 321 No. of reduced fee bids 34 18 Mean bid amount ($) 113,381 100,092 (standard deviation) (53,838) (56,506) Mean vessel length (feet) 34.3 34.2 (standard deviation) (4.2) (4.2) The next pair of graphs, Figures 18 and 19, for the troll sector bids received in each round show similar trends to the gillnet sector data. There is a significant reduction in the number of bid outliers in the second round, reflected in a lower standard deviation for these bids. Again, fewer shorter vessels are observed and fewer reduced fee licences in the second round data. The troll sector data suggest a similar ceiling effect in the bid value amount as observed in the gillnet data, for average vessel length, between $70,000 and $100,000. 99 100 (000'$) lunoure pie 101 Table 6 presents descriptive statistics concerning bid distributions for all troll sectors bids to the first and second auction rounds. Table 6 Troll Bids: Descriptive Statistics Item Round 1 Round 2 No. of bids 406 254 No. of reduced fee bids 10 4 Mean bid amount ($) 115,476 101,103 (standard deviation) (64,641) (42,984) Mean vessel length (feet) 39.3 39.6 (standard deviation) (5.4) (4.8) This concentration of second round bids near the average accepted value from the first round suggests strongly that information regarding bid values accepted by the program spread throughout the applicant pool. It also suggests that second round bids were concentrated in the upper limits of accepted values for vessels of average length. One source of this information might be boat brokers. The obvious benefit to bidders to find out the largest payment that might be available would provide incentive to boat brokers to collect as much information as possible and offer it to applicants who seek their advice. Some brokers would have had access to information from the first round because they were acting on behalf of licence holders. In addition, individual bidders could have spread information about the first round results but this seems less likely to have had such a pronounced effect. Figures 20 and 21 for the seine sector data follow with fewer observations and different implications. The first round data indicate an upward trend in the potential 102 o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o ($) junouiv P19 103 ($) junouiv Pie 104 supply curve but the second round data are much flatter. There are significantly fewer second round applications and outlier bids. Note that only two bids from reduced-fee licence holders were received in the second auction round. These bids tended to be lower in general than those from full-fee licence holder applicants. This is because reduced-fee licences cannot trade as freely in the private market. The ceiling effect is much less evident than in the small boat sectors which may be due to the lesser volume of applications in this sector. Table 7 presents descriptive statistics for the bid distributions of seine sector bids in each of the auction rounds. Table 7 Seine Bids: Descriptive Statistics Item Round 1 Round 2 No. of bids 83 20 No. of reduced fee bids 9 2 Mean bid amount ($) 593,949 561,611 (standard deviation) (354,581) (165,076) Mean vessel length (feet) 60.8 62.8 (standard deviation) (11.3) (14.1) Accepted Bids The final pairs of figures show the accepted bids in each gear sector for each auction round separately. Each graph plots vessel length on the x-axis with bid value on the y-axis. Figures 22 and 23 for the gillnet sector indicate that the upward trend in the data as vessel length increases is more pronounced in the first round than the second. Table 8 presents descriptive statistics for the accepted bid distributions of the gillnet bids. 105 (000'$) lunouxe pig 106 (000'$) ;unouxe pig 107 Table 8 Gillnet Accepted Bids: Descriptive Statistics Item Round 1 Round 2 No. of bids 236 223 No. of reduced fee bids 28 16 Mean bid amount ($) 74,032 84,745 (standard deviation) (14,098) (8,251) Mean vessel length (feet) 33 34 (standard deviation) (4.8) (3.9) Fewer licences from shorter vessels were accepted to the program in the second round, confirmed by the higher average vessel length and smaller standard deviation, likely because fewer applications were received in the shorter vessel category. The ceiling effect in the first auction round was "created" by the Committee because it chose which bids to accept thus defining the maximum of the payment amount. This effect in the second round is more a function of the applications which took the reasonable bid information available and tailored bid values accordingly. This effectively re-inforced the committee's views and may have established a "posted price" for licence removal. The establishment of an average or maximum payment amount was not fixed - payments in the second auction round were dispersed - however, the variation diminished in the second auction round compared to the first, reflected in the reduced standard deviation. The pair of Figures 24 and 25 for the troll sector data in each round exhibit an upward trend as vessel length increases. Curiously, this effect is more pronounced in the second round troll data than in the first round. There is a ceiling effect evident in the first round data 108 (000'$) junoure p i g 109 • m CN <D L _L _L _L _L o o i n o o C N O C N DO CD > 13 T3 CD -4-» OH CD O O ca s-<a E> C CD CD c/3 co CD > > 12 *o 1-1 H ON CD E o o o o o o o o o o o o o (000'$) *uriouxe pig 110 and a cluster of bids in the second round data at this value. There were more reasonable bids from longer vessels in the second round, reflected in the longer average vessel length and smaller standard deviation, than in the first which account for outliers in this figure. The data suggest that information from the first round was used by second round applicants in formulating bids and that it may have been used more effectively. Rather than simply following the ceiling effect, troll sector applicants seem to have been more successful in translating the information available into their own bid situation. Table 9 presents descriptive statistics for the bid distributions of accepted troll sector bids. Table 9 Troll Accepted Bids: Descriptive Statistics Item Round 1 Round 2 No. of bids 134 167 No. of reduced fee bids 9 4 Mean bid amount ($) 70,125 82,017 (standard deviation) (14,231) (11,419) Mean vessel length (feet) 38 39 (standard deviation) (6.3) (4.8) The final pair of graphs, Figures 26 and 27, relate to the seine data from accepted applications in each auction round. There is strong evidence that the ceiling effect imposed by the Committee in the first round resulted in information for the second round applicants. A s discussed in reference to the potential supply curve analysis earlier, the seine fleet sector had fewer applicants in the second round and might have perceived an advantage because the Committee announced its desire to remove more seine licences than had been achieved in the first round. This did not translate into windfalls for the bidders who did not take the signals I l l o o © as oo o oo >/-> C--o T3 O T3 o so CU OH CD O % C N 1) bp E o o >/-) m o CS o CN fl <u CU so C/3 > T 3 3 c/3 > C 1) u C/3 1/3 CL) > cu _c CN 9> op E o o o •3-o o o o o © CS o o CS o o o o (000'$) Junoure p i g 112 o o o O N o CN c o to o cu c*-l T 3 CD +-> a, OJ o o CS (-1 cS CN 1S> c 3 E o o > 3 CO > 43 00 S3 cu CD CO co CD > CD '53 o CN CN CD l - l SB E o o o o o o o o o o r<-> o CN o o CN o o o in (000'$) lunoure pig 113 of reasonable bid values into account. The second round data indicate that only reasonable bids, as defined in the first round, were accepted in the second round. Other variables relating to the seine data indicate that many of the second round bids were from multi-licensed vessels which could be expected to downgrade bids rather than charge a premium. Descriptive statistics relating to the bid distributions of accepted seine sector bids are presented in Table 10. Table 10 Seine Accepted Bids: Descriptive Statistics Item Round 1 Round 2 No. of bids 38 11 No. of reduced fee bids 2 0 Mean bid amount ($) 402,220 443,475 (standard deviation) (67,908) (51,960) Mean vessel length (feet) 58 64 (standard deviation) (9.8) (17.37) Statistical Differences The data for the total submitted bids in each auction round and the accepted bids in each round taken together reinforce the theory that information from the first auction round was available to second round bidders. There does not appear to have been any windfall to bidders who resubmit to the second round, relative to new applicants, especially since fewer than half those who re-applied to the program were accepted in the second round. However, the test statistics in Table 5 indicate that for the gillnet and troll sectors there was a 114 significant difference in the proportion of bids accepted in the second round compared to the first auction round, with a higher proportion accepted in the second round. The proportion of bids accepted in the second round for the seine sector was not significantly different at the 5 per cent level than the proportion accepted in the first auction round. Table 11 Differences in Proportions of Accepted Bids by Auction Round Gear Round 1 Round 2 Proportion Proportion Test Type Bids Accepted Bids Accepted 1 2 statistic Gillnet 640 220 323 224 0.3438 0.6935 -11.00* Seine 60 37 27 11 0.6167 0.4074 1.84 Troll 411 139 261 166 0.3382 0.6360 -7.87* * Indicates significance at the 2.5 % level. 5.2 Signaling One of the important effects of the availability of bid information is through signaling. The first round results appear to have influenced bidding behaviour in the second auction round by lowering bid values which resulted in cost savings to the Program. A review of the bids accepted in the second round from rejected first round bidders who re-submitted second round bids reveals the saving realised. A n assessment of the gillnet data set bids that were re-submitted to the program following rejection in the initial bidding round reveals that those re-submitted bids accepted in Round 2 were reduced by a total of $3,667,229. This represents a savings and indicates that the signaling function of the auction method can work to provide information to 115 participants concerning acceptable bidding patterns and to reduce the overall expenditure under licence retirement when multiple bidding rounds occur. A calculation of savings resulting from re-submitted and accepted bids in the seine sector indicates a smaller reduction in cost of $332,000. This may be explained by the low participation rate of seine bidders in the licence retirement program but does exhibit the trend that bidders learn from the results of early bidding rounds in a multiple bidding round process. The calculation of savings realised from rejected bidders whose re-submitted bids were accepted in the subsequent auction round for the troll sector yields $2,562,543. This amount again represents a significant reduction in the cost of licence retirement and indicates support for the use of repeated bidding rounds to provide signals to initially unsuccessful bidders that permit learning about the process. Improvements might be considered from reviewing the data with the benefit o f hindsight. The bids received in the first round that were extremely high might have been avoided with an initial application fee for the program. It is evident from the second round data that when information regarding reasonable bids was available, much fewer outlier bids were received. This might be important for similar programs, particularly i f administrative costs might be high. In addition, the focus on resubmissions of first round bids was not necessarily appropriate since many new applications were received in the second round and many of them were accepted as reasonable. This result indicates that not only did those familiar with the program seek out information from the first round results, but new participants took advantage of the availability of information as well . 116 I have argued that boat brokers played an important role in the dissemination of information from first auction round results. The position of boat brokers as independent from both parties to the transaction, the government and the licence holders, is significant. Licence holders whose applications were accepted in the first round may have had no incentive to assist others who did not take the risks that they had. A s a result, the role of boat brokers in collecting information from the first auction round and providing advice to second round applicants who sought it might have been instrumental. The auctioneer, in this case the government, may not believe it is worthwhile to conduct multi-round bidding or to publicize information regarding early rounds i f multiple bidding rounds occur. Auction theory suggests that revelation of the reserve price is optimal in the auction setting. In this case, the analysis suggests that by adopting an initially passive role to extract information from the pool of bidders, it is then advantageous to take a more active role in information dissemination. The sharing of information can attract new bidders who may have been initially apprehensive, can assist unsuccessful bidders in revising their bids and can ultimately reduce the overall cost of licence removal. 117 5.3 Equity The launch of a second auction round to the Pacific Salmon Licence Retirement Program (1996) was criticized by some (Tangled Lines, 1996) as unfair, especially to those who participated successfully in the first auction round. The argument appears, to relate to the additional information to which second round participants may have had access. It is clear from the discussion in an earlier section that a higher proportion of bids in the gillnet and troll sectors were accepted in the second auction round. This could be caused by a variety of factors such as second round bids may have been more reasonable, the Committee may have been more prone to accepting second round bids or a combination of effects. This section w i l l examine the data and some concepts of equity to discuss this situation. There is no one definition of equity and much discussion of how to measure equity. The concept of equity can be used to describe a process, the outcomes of a process, treatments across individuals or across other factors, such as income. The policy objectives outlined at the program's outset indicated that fairness was important. Fairness was described as permitting all eligible licence holders equal opportunity to participate in the program. This section wi l l expand the fairness concept to include an assessment of actual outcomes of the auction process against the equity concepts. The roles of the information externality and bid formulation process w i l l be examined in relation to the program's equity. The issue of fairness in bid outcomes is especially important because public funds were used for payment on a voluntary basis. A l l bids to the licence removal program were voluntary and formulated by individuals who had private information relating to their 118 licences' values and had opportunities to gauge reasonable market values for their licences prior to submitting a bid to the Program. One concept of equity requires that each individual be provided the same information relating to the Program, its objectives, its rules and equal opportunity to participate. This element of the fairness principle seems to have been met without difficulty since all licence holders were mailed a complete information package on these issues, the overall Pacific Salmon Revitalization Strategy and an application form for the Program (Appendix 1). Another concept of equity relates to the amount licence holders received to surrender their licence privileges. One possible outcome involves each participant receiving the same amount as payment for relinquishing the salmon fishing privilege, regardless of gear type, vessel or licence characteristics. This might be viewed as equitable because all participants would be treated the same in terms of outcome. However, this strategy could also be viewed as inequitable because it would not account for differences across individuals who participate in the Program. B i d values that vary due to observable vessel or licence characteristics, such as vessel length, gear type, or vessel age, are not evidence of inequities i f the concept includes the idea that like people are treated the same. Outcomes that differ on the basis of objective and measurable characteristics should be expected. In fact, the application form for the program indicated clearly that vessel and licence characteristics might be considered by the committee that would review the bids. The fact that many of the differences in bid values evident in the outcomes can be explained and attributed to various licence and vessel attributes (discussed 119 in the earlier chapter of this study on bid formulation) is evidence that the auction contributed to a fair evaluation of the value of licence retirement across individual bidders. The primary reason for concern relating to equity in licence payments through this Program seems to be that the second round of bidding that was not foreseen at the outset. The Pacific Salmon Revitalization Plan Review Panel ( a joint federal-provincial group) commented on the perception that licences transacted in the first bidding round went "cheap" (Tangled Lines, 1996). A n earlier section of this study provides an analysis of the extent to which bid value information from the first round might have been used by second round bidders and did find merit in the widely held view that information from the first bidding round was used by second round bidders. Despite the fact that the average payment to licence holders in each round was roughly comparable when vessel length is taken into account, the actual average payments in each auction round did differ within each gear sector. The following table presents the average payments as recorded in the final report on the program (James, 1996). Table 12 A v e r a g e P a y m e n t b y G e a r T y p e ( $ ) R o u n d 1 R o u n d 2 O v e r a l l G i l l n e t 7 3 , 7 1 9 8 4 , 7 0 2 7 9 , 2 6 0 S e i n e 4 0 5 , 1 1 8 4 4 3 , 4 7 5 4 1 3 , 9 0 8 T r o l l 7 0 , 8 8 1 8 2 , 1 3 6 7 7 , 0 0 7 The difference in average bid value contributes to the perception of inequity. However, the causes of the variation in the average payment are worth examining. In general, the first bidding round considered many more bids from shorter vessels than did the second round which would tend to reduce the average licence value in the first bidding round. 120 In addition, the second bidding round also reviewed bids which tended to be from longer vessels, contributing to a higher average bid value. There were fewer reduced fee licences submitted for retirement in the second auction round which contributes to lower average value in the initial bidding round since reduced fee licences had lower bid values than full fee licences. The bidding position of first round bidders may have been risk averse, meaning that they were anxious to leave the fishery and fearful that the program would be oversubscribed (i.e. run out of money quickly). Lower bids from risk averse bidders could also contribute to a lower average bid value in the first bidding round. Note that all these comments relate to successful bids. The graphical analysis presented in the earlier section on information externality attests that the first round of 1,111 bids exhibits greater bid variability i f rejected bid are considered than does the second bidding round. The lower average value of first round successful bids is to be expected from an auction process. The auction is designed to permit the choice of acceptable bids and rejection of all others. That the average value of accepted bids increased in the second round is a feature of the process because the initial round chooses primarily the lower portion of the distribution of submitted bids. With a well-defined pool of potential bidders, later bidding rounds must involve a truncated distribution where more bids are from the middle and upper ranges of the distribution. In terms of concepts of equity, each successful participant was paid an amount that he or she indicated represented the value of their salmon fishing privileges to be surrendered. Those whose licence values exceeded that which the review committee considered reasonable valuations kept their licences without penalty. In this sense, the program must be viewed as 121 equitable. N o participant was subject to negotiation or coercion relating to their bid valuation. The auction process allowed participants to formulate their bid valuations and reveal the true value of salmon fishing privileges. The fact that some individuals requested and received different payment amounts is not inequitable because each bidder stipulated the payment amount. Bidders may have been upset ex post to find out that they bid below average but that in itself is not evidence of inequity. Recall that no "reservation price" was devised prior to the bidding process nor was one determined by the Committee through its bid assessment. In addition, no third party, verifiable source for licence valuation was available. A s such, any bid assessment must be self-referential using the information collected by the auction process itself. This might seem to present a circular argument; however, in each of the small boat fleet sectors, the numbers of successful bidders exceeds 300. With these large sample sizes, using information calculated from the data is reliable in the statistical sense. Through the bid assessment process, the "market" for licences created by the auction was defined. The second auction round bids were much less dispersed than those of the first round and appear to have concentrated near the upper range of values accepted from bids in the first round. The multiple round auction process could address equity concerns, i f designed and known at the outset, as potential opportunities for information exchange and signaling could be anticipated by bidders. In terms of equity, depending on the concept one adopts, an auction method may not be appropriate. 122 CHAPTER VI Summary of Lessons Learned and Conclusion 6.1 Summary of Lessons Learned Detailed analysis of a program such as the Pacific Salmon Licence Retirement Program (1996) offers a learning opportunity and, in this case, the lessons may be of interest to fishery or other resource sector managers who may see the advantages of using auction methods. Many valuable insights emerge in the administration, design and implementation of this type of auction to remove fishing licences. The auction model described in this study requires detailed information to assess bids which must be available to the program administrator. Without access to such data, or the ability to collect and collate it cost-effectively, the method may not work. The collection and organisation of such data can be technologically and labour intensive, as well as costly. Nominal application fees to submit a bid to the auction may avoid the cost and burden of collating bid data from those who are not serious bidders. This program did not use application fees because it was deemed inappropriate for first-time participants who might be unfamiliar with the auction method and easily deterred from bidding. The most important program design lesson available from this study is the importance of the multiple bidding round process, which ideally should be explained to participants at the outset. Multiple bidding rounds have the potential to convey valuable information between the auctioneer and the bidding population. In addition, this process may address equity concerns, reduce costs and signal reasonable bid values. 123 Caution must be taken when conducting multiple auction rounds. This program addressed the issue of speculation in licence values by requiring that applicants to the program demonstrate that the salmon licence was held in the year prior to the announcement of the licence retirement program and by requiring that bids submitted be binding on the part of the applicant. In a multiple bidding round process, speculation in licence values should be discouraged, possibly by requiring that bids be binding; that is, those whose bids are accepted in one bidding round may not reject the offer and re-submit another (presumably higher bid) in a future bidding round. Another important program design issue is the choice of auction over other allocation methods, such as posted price. This study has outlined the advantages of auctions as a flexible valuation tool that can mimic market behaviour. The auction method requires bid assessment and consideration should be given at the outset as to how to ensure impartiality in this aspect. The use of an independent bid assessment committee is crucial and could be addressed by seeking industry representatives or independent consultants or auditors. This program required that the independent advisors be active industry participants who did not apply to the licence retirement program. Program implementation in an auction setting should be set in advance and bid assessment should be conducted within a short period of the applications being received. Information concerning the program should be made available to participants, including such factors as the projected timeframe for the auction process (i.e. number of bidding rounds expected), overall budget, expected reductions in the fleet and future expectations of resource abundance and fishery management changes. In this case, participants were informed of 124 significant changes in fishery management, resource abundance, and that the program's budget was $80 mil l ion to affect a reduction of at least 20 per cent in the fleet. This strong message set a serious tone in the industry and may have helped to deflate expectations in the bidding population about potential windfall benefits from licence retirement. 6.2 Conclusion Examination of the data set relating to the Pacific Salmon Licence Retirement Program (1996) has revealed important insights concerning the method used to conduct the program. This analysis provides practitioners with information regarding how to approach valuation of a resource privilege in the absence of a well-defined market. Auction theory has been recognized as an appropriate valuation and allocation rule when non-standard items are transacted, such as when item values are unknown or likely to vary considerably in transaction. Such is the case for many natural resource harvesting privileges, such as fishing licences, pollution emission permits and forestry tenures. Economic theory suggests that in allocating these privileges, the auction method w i l l optimize outcomes. This paper has examined the opposite situation in that the auction mechanism has been used to effectively revoke these privileges by making payments to fishing licence holders who, on a voluntary basis, agree to surrender their fishing privileges. The Program involved expenditure of $80 mil l ion in public funds to permanently remove 797 salmon fishing licences from the Pacific salmon commercial fleet. In advance of the program, licence values were not known with certainty nor could reliable estimates be obtained. A s 125 such, the auction method permitted the establishment of a quasi-market using the auction data. The licence retirement program was motivated by a combination of overcapitalization in the Pacific salmon fleet and resource declines that placed the fleet in a precarious financial position. Many countries have recognized the need to reduce pressure on fishery resources and have introduced methods to rationalize fleets. One response to such circumstances that is often suggested is changes to the management system that encourages fleet reduction, such as the introduction of individual transferable quotas. This solution involves complex design and implementation decisions that often take years to resolve and implement. Another option that has been used is the removal of fishing licences and/or vessels from the overcapitalized fleet through government-funded retirement. This paper has presented as study of the Pacific Salmon Licence Retirement Program (1996) to provide an analysis of this type of program, including the auction mechanism used to allocate public funds. The data analysis presented in this study indicates that licence values are difficult to assess but are related to licence and observable vessel attributes. The multi-dimensional scaling analysis found that the bids submitted to the Program are close to one another in terms of reasonableness and that no definite clustering is evident in the bids. The regression analysis supported the correlated or common value model as an appropriate bid valuation method in the gillnet, seine and troll fleet sectors. The regression co-efficients related to the vessel length, fee status, presence of other licences and second round bid indicator suggest that these attributes significantly contribute to explaining the bid variation in each gear sector. The vessel age characteristic was not found to be a significant explanatory variable. 126 In addition, in each gear sector the regression intercept offers an estimate of the inherent value of a salmon licence, regardless of the other characteristics. Program design is an important aspect of any mechanism to achieve fleet reduction, especially one that uses public funds. Elements of program design in this instance were specifically focused on obtaining information about licence values from the bidding population which was previously unknown and unavailable from other sources. The collected data is valuable and able to mimic "market" behaviour. The results indicate that the auction process permitted the Industry Committee who assessed the bids to review the submitted bids and determine their recommendations regarding reasonable licence values. That assessment helped to influence bid values submitted in the second auction round. The exchange of information following the first bidding round results was indirect -possibly occurring through the role of boat brokers or perhaps through information sharing of individual successful bidders - but had significant impacts. The second bidding round attracted bids from unsuccessful first round bidders along with new bidders who appear to have used information from the first bidding round to devise reasonable bids. The second bidding round data are much less dispersed that the initial data. This signaling effect produced cost savings to the Program, calculated from those bids accepted in the second round from unsuccessful first round bidders to be approximately $6.7 mill ion. The multiple round bidding process had important impacts on the auction. Exchange of information based on the results of the first bidding round provided new and existing bidders with a signal to devise reasonable bids. 127 The concept and perception of equity in the delivery of licence retirement is another important program design issue. The Pacific licence retirement program was criticised as inequitable primarily because of the bid variation among accepted bids that is a feature of the auction method. In particular, the increase in average accepted licence values in the second auction round was of concern. This effect is to be expected from a multiple round auction but the inexperience of bidders with the process may have contributed to this anxiety. If the multiple round process were known at the outset, it might be able to alleviate equity concerns by ensuring that bidders better understand the process and anticipate information exchange. Finally, the licence retirement program discussed in this study might have application to other fishery and resource situations. As the fishery sector becomes more familiar with these types of programs, licence retirement using the auction process may gain wider acceptance and use to address fishery problems. While the common practice is for programs to be publicly funded, a fleet sector could fund and administer such a program if agreement on objectives could be attained. Furthermore, programs to date have been primarily in response to extreme situations of crisis but licence retirement could be implemented to respond to more specific or localized fishery management issues. Markets in such cases would exhibit many of the same properties studied in this example, namely lack of information regarding licence values and market thinness. Other resource sectors might also find applications for the use of auction methods to retire privileges. The use of pollution emission permits and quotas has increased in this decade. These permits and quotas have many of the same characteristics of fishery licences and quota and have experienced similar 128 problems. One could imagine a situation where a government or community would like to reduce pollution emissions, requiring relinquishment of these privileges. The Pacific Salmon Licence Retirement Program (1996) is not the only program that has been used to reduce the size of a fishing fleet in Canada or the world. Examination of this program provides an analysis of the methods used and outcomes observed which may be valuable to other fishery managers who face similar situations requiring fleet rationalization in a short timeframe. It may also inspire new applications of this type of program to other resource problems. 129 Bibliography Adelson Consulting Limited, Commercial Fishing Licence Index, 1995. Ashenfelter, Orley, "How Auctions Work for Wine and Art ," Journal of Economic Perspectives, Volume 3, Number 3, pp.23-36. " B . C . to Challenge Salmon Plan", The Globe and Mail, May 22, 1996. Batstone, C.J . and B . M . H . Sharp, "New Zealand's Quota Management System: the First Ten Years", Marine Policy, Volume 23, Number 2, pp. 177-190, 1999. Botsford, L . W . , J .C. Costilla and C H . Peterson, "The Management of Fisheries and Marine Ecosystems", Science, Volume 277, July 1997, pp.509-515. Chakravorti, B . , W . W . Sharkney, Y . Spiegel and S. Wilkie , "Auctioning the Airwaves: The Contest for Broadband P C S Spectrum", Journal of Economics and Management Strategy, Volume 4, Number 2, 1995, pp. 345-373. Charting a New Course: Towards the Fishery of the Future, Final Report of the Task Force on Incomes and Adjustment in the Atlantic Fishery, Ottawa: Supply and Services Canada, 1994. "Clark Calls for Provincial Fish Agency", The Globe and Mail, May 10, 1996. "Clark 's Fisheries Takeover Plan Won't Work: Mi f f l i n" , The Globe and Mail, May 12, 1996. Clark, Col in W. , Bioeconomic Modeling and Fisheries Management, Toronto: John Wiley and Sons, 1985. Clarke, K . R . and R . M . Warwick, Changes in Marine Communities: An Approach to Statistical Analysis and Interpretation, Plymouth: Plymouth Marine Laboratory, 1994. Commercial Fishing Licensing Policy, Department of Fisheries and Oceans, Minister of Supply and Services Canada, June 1990. Crampton, Peter C. "Money Out of Thin A i r : The Nationwide Narrowband P C S Auction", Journal of Economics and Management Strategy, Volume 4, Number 2, 1995, pp. 267-343. Crowley, R . W . and H . Palsson, A Critical Review of Rights-Based Fisheries Management in Canada, Economic and Commercial Analysis Directorate, Department of Fisheries and Oceans, Ottawa, 1990. "Cuts to Salmon Fleet Approved", The Globe and Mail, May 24, 1996. Everitt, Brian S. and Graham Dunn, Applied Multivariate Data Analysis, London: Edward Arnold, 1991. Falloon, Roger "Individual Transferable Quotas: The New Zealand Case", in The Use of Individual Quotas in Fisheries Management, Organization for Economic Cooperation and Development, O E C D , 1993. Fraser River Sockeye Public Review Board (Canada), Public Works and Government Services Canada: 1995. Gardiner Pinfold Consulting Economists Limited, Canning and Pitt Associates, Final 130 Evaluation of the Northern Cod Adjustment and Recovery Program, Ottawa, 1994. Geen, G . , W . Nielander and T.F. Meany, "Australian Experience with Individual Transferable Quota Systems", in The Use of Individual Quotas in Fisheries Management, Organization for Economic Cooperation and Development, O E C D , 1993. Gislason, Gordon et al, Fishing for Answers: Coastal Communities and the BC Salmon Fishery, The A R A Consulting Group: 1996. Gislason, Gordon et al, Fishing for Money: Challenges and Opportunities in the BC Salmon Fishery, GSGislason and Associates: 1998. Gordon, H.S. , "The Economic Theory of a Common pool Resource: The Fishery", Journal of Political Economy, Volume 62, February, 1954. Harmsma, Herman and W i n Davidse, "Effects of Property Rights on the Structure of the Fishing Industry", in The Use of Individual Quotas in Fisheries Management, Organization for Economic Cooperation and Development, O E C D , 1993. Harrigan, Eugene, Post-TAGS Review Report, Communications Branch, Human Resources Development Canada, 1998. Healey, M . C . , "The Management of Pacific Salmon Fisheries in British Columbia" in L .S . Parsons and W . H . Lear [eds] Perspectives in Canadian Marine Fisheries Management, Canadian Bulletin of Fisheries and Aquatic Sciences 226, 1993. Holland, D . , E . Gudmundsson and J.Gates, "Do Fishing Vessel Buyback Programs Work: A Survey of the Evidence", Marine Policy, Volume 23, Number 1, pp.47-69, 1999. Huppert, Daniel D . and Hans W . Brubaker, "Distributing Fishery Rent through Quota Auctions: Lessons from Existing Resource Regimes", Working Paper: School of Marine Affairs, University of Washington. James, Michelle, "Final Report on the 1996 Voluntary Fleet Reduction Licence Retirement Program", Pacific Region, Fisheries and Oceans, 1996. Levellton, C.R. , Toward an Atlantic Coast Commercial Fisheries Licensing System, Minister of Supply and Services Canada, 1981. May, A . W . , Altering Course: A Report to the Minister of Fisheries and Oceans on Intersectoral Allocations of Salmon in British Columbia, Memorial University of Newfoundland: 1996. M c C a l l , John, "Optimal Response to a Known Bidding Distribution," in Bidding and Auctioning for Procurement and Allocation: Proceedings of a Conference at the Center for Applied Economics, New York University. Amihud, Yakov (eds.) N e w York University Press: 1976, pp.35-37. McKinsey, Krista, "Struggles with Salmon", Scientific American, Volume 9, Number 3, Fal l 1998. Milgrom, Paul, "Auctions and Bidding: A Primer," Journal of Economic Perspectives, Volume 3, Number 3, pp.3-22. Milgrom, Paul and Robert J. Weber " A Theory of Auctions and Competitive Bidding," Econometrica, November 1982, Volume 50, pp.1089-1122. Muse, Ben and Kurt Schelle, Individual Fishermen's Quotas: A Preliminary Review of Some Recent Programs, Alaska Commercial Fisheries Entry Commission, February, 131 1989. Pacific Policy Roundtable Report: Report to the Minister of Fisheries and Oceans on the Renewal of the Commercial Pacific Salmon Fishery, Department of Fisheries and Oceans Canada: 1995. Pearse, Peter H . , Turning the Tide: A New Policy for Canada's Pacific Fisheries, Final Report of the Commission on Pacific Fisheries Policy, September, 1982. Pearse, Peter H . and Peter A . Larkin, Managing Salmon in the Fraser: Report to the Minister of Fisheries and Oceans on the Fraser River Salmon Investigation, Supply and Services Canada, November, 1992. Pembina Institute for Appropriate Development with Apogee Research International, A Fresh Look at Economic Instruments, Canadian Council o f Ministers of the Environment, Winnipeg, 1996. Pitcher, Tony, J. , "Rapfish, a Rapid Appraisal Technique for Fisheries, and its Application to the Code of Conduct for Responsible Fisheries", F A O Circular Number 947, Food and Agriculture Organization of the United Nations, Rome, 1999. Pitcher, T.J. and D . B . Preikshot, "Rapfish, a Rapid Appraisal Technique to Evaluate the Sustainability of Fisheries", Fisheries Research, forthcoming. Pitcher, T.J. and M . D . Power, " Fish Figures: Quantifying the Ethical Status of Canadian Fisheries, East and West", in Coward, H . , R.Ommer and T.J. Pitcher (eds.) Just Fish: the Ethics of Canadian Fisheries, Institute of Social and Economic Research Press, St. John's, 1999. Press Release: June 19, 1998, Ministers Announce Canada's Coho Recovery Plan and $400 Million for Pacific Salmon Fishery, Fisheries and Oceans Canada. Press Release: June 18, 1999,1999 Salmon Management Plan: Staying the Course for Salmon Conservation and Canadian Fisheries Adjustment and Restructuring Program, Fisheries and Oceans Canada. Press Release and Backgrounder: June 14, 1999, Canada and US Reach a Comprehensive Agreement Under the Pacific Salmon Treaty, Fisheries and Oceans Canada. Putting Markets to Work: The Design and Use of Marketable Permits and Obligations, Organisation for Economic Cooperation and Development, O E C D , 1997. Riley, John G. , "Expected Revenues from Open and Sealed B i d Auctions," Journal of Economic Perspectives, Volume 3, Number 3, pp.41-50. Riley, John G . and Wi l l i am F. Samuelson, "Optimal Auctions," The American Economic Review, June 1981, Volume 71, pp.381-392. SPSS Inc.,, SPSS for Windows, Release 9, 1999. Safina, Carl, "The World's Imperiled Fish", Scientific American, Volume 9, Number 3, Fal l , 1998,pp.58-63. Schotter, Andrew, "Auctions and Economic Theory," in Bidding and Auctioning for Procurement and Allocation: Proceedings of a Conference at the Center for Applied Economics, New York University. Amihud, Yakov (eds.) New York University Press: 1976, pp.1-12. Scott, Anthony, "The Fishery: The Objectives of Sole Ownership", Journal of Political Economy, Volume 63, A p r i l , 1955. 132 Swenerton, D . M . , A History of Pacific Fisheries Policy, Department of Fisheries and Oceans, Pacific Region, 1993. Tangled Lines: Restructuring in the Pacific Salmon Fishery, A Federal-Provincial Review of the Mif f l in Plan: 1996. Tietenberg, Tom, Environmental and Natural Resource Economics, Harper-Collins; 1992. Vickrey, Wi l l i am, "Auctions, Markets and Optimal Allocation," in Bidding and Auctioning for Procurement and Allocation: Proceedings of a Conference at the Center for Applied Economics, New York University. Amihud, Yakov (eds.) N e w York University Press: 1976, pp. 13-20. Wang, Ruqu, " Auctions Versus Posted Price Selling", American Economic Review, Volume 83, September, 1993, pp.838-851. Wilson, Robert, " A Bidding Model of Perfect Competition," Review of Economic Studies, October 1977, Volume 44, pp. 511-518. 133 Appendix 1 APPLICATION FOR 1996 SALMON F L E E T REDUCTION VOLUNTARY LICENCE RETIREMENT ' IMPORTANT* All applications & relinquishment agreements must be received by D F O by May 24, 1996. PART 1: V E S SEL DATA: CFV MA/ DOT# Year vessel built Rebuilt Builder Hull material: Gear type Tab #_ OAL (overall length) m Date of measurement For the purpose of determining the overall length, the estimated or surveyed length of this vessel currently on record with DFO will be used. Should the vessel owner wish to submit a survey for the purposes of updating the vessel record, it must be done in accordance with DFO Guidelines for vessel measurements as amended February 1996, and submitted with this application. All applicants must submit a recent full side view profile (bow to stern) photograph of the vessel with this application. PART 2: DECLARATION: This application must be signed by all true registered vessel owner(s). Attach a separate sheet if required. Full legal name(s) of declarant(s) are required. If the vessel owner is a company, we require that the following company documents be on file with DFO or submitted with this application: 1, Copy of the Company's last Annual Report filed with the Registrar of Companies (form 18 Company Act) . 2. Written acknowledgment from the company listing all signing authorities and that DFO may accept those signatures until such time as notice is received by DFO from the company advising that the signing authority is canceled amended or otherwise changed. I/We Name: Name: Address: Address: Telephone: Telephone: Date of Birth: Date of Birth: HEREBY SOLEMNLY DECLARE; THAT IA/Ve are the true owner(s) of the MA/ . and that the vessel is free and clear of encumbrances except for the following: Attached are written statements from above lien and encumbrance holder(s) stating that they have been advised of this application. I/We, appoint as our contact person. Address of contact person: As above or Phone # of contact person: As above or Postal code I/We agree that all communication from DFO will be with the contact person. I/We declare that the information given on this application is complete, true and correct. I/We consent to any inquiries, audits and inspections necessary to verify the information given. I/We, as identified in Part 2, hereby submit our bid for a salmon licence eligibility retirement as follows: Bid for retirement $ This bid is binding on me/us and l/we will remain the owner(s) of the salmon licensed vessel unless, prior to acceptance by DFO, DFO receives a withdrawal of the bid by FAX at 604-754-0403. SIGNATURE(S): Signature Date Signature Date Sworn before me at in the Province of British Columbia this day of AD 1996. Commissioner for Taking Affidavits for British Columbia Registry* 134 1996 SALMON F L E E T REDUCTION VOLUNTARY LICENCE RETIREMENT The federal government has made $80 million available for retirement of licences to reduce the capacity of the West Coast commercial salmon fleet. The purpose of the program is to bring about an equitable and immediate reduction in the number of licences in the salmon fleet. The program will expire at the end of June, 1996. All applications & Relinquishment, Release & Contribution Agreement forms must be received by May 24, 1996. All salmon vessel owners holding full fee and reduced fee salmon licences are eligible to apply for retirement of their eligibility under this program. Communally held Aboriginal commercial licence eligibilities and those held by the Northern Native Fishing Corporation are not eligible for retirement For vessel owners with more than one salmon licence eligibility attached to their vessel, only the shorter length licence or both licences on the vessel may be retired. Bids are binding on the applicant unless withdrawn prior to acceptance of the bid by DFO. A signed application and Relinquishment Agreement will constitute the applicant's acceptance of the programs' terms and conditions in the event the bid is accepted. Bids must reflect the payment required by the vessel owner(s) of record to permanently retire the salmon licence eligibility on their vessel. The disposal of the unlicensed vessel is the responsibility of the vessel owner(s). It may be used as a replacing vessel in the commercial fishery. Vessel owner(s) applying for licence eligibility retirement under this program must have applied for and been issued a commercial salmon licence in respect of this vessel in 1995 and an application and fees for the 1996 commercial salmon licence must be submitted by May 24, 1996. Where an application to retire a licence eligibility is accepted, all current year licence documents issued to the vessel named in the application must be returned to D F O upon notification of an accepted bid ie. 1996 Vessel Registration Certificate, licences, tabs and plastic vessel card (pvc). If the vessel remains licensed in another vessel licence category, only the salmon licence, tabs, vessel registration certificate and pvc must be returned. A Fleet Reduction Committee will be set up to review all valid offers and recommend to D F O which licences should be retired. Bids will be assessed primarily on the basis of lowest cost per licensed foot. Other considerations are: to maintain the proportion of reduced fee licences currently in existence in the salmon fleet. the balance of fleet reduction among gear sectors. in assessing licence retirement bids, the Fleet Reduction Committee may take into account the age, length, hull construction material, and overall apparent condition of the vessel (on the basis of recent photographs). No independent assessments of the vessel will be made. bids from vessel owners with other licence eligibilities on the vessel will be considered on a case by case basis with preference given to single salmon licensed vessels. If the salmon licence is accepted for retirement, the vessel will remain licensed in the other categories. All bids will be reviewed by the Fleet Reduction Committee early in June and recommendations made to D F O . Upon signature of the Relinquishment Agreement by the Departments' Regional Director General, a bid will be deemed accepted and payment will be processed and mailed within fourteen working days of this acceptance. Voluntary payments will be made in one installment. Notification of accepted bids will be sent to successful bidders with a signed copy of the Relinquishment Agreement. The Minister and the Department of Fisheries and Oceans are not under any obligation to accept any bids. Please mail the completed application & Relinquishment Agreement and other required documents to: Department of Fisheries and Oceans Fleet Reduction Program 60 Front St Nanaimo B C V9R 5H7 F O R ADDITIONAL INFORMATION P L E A S E C O N T A C T : R A E D A L G A R N O Telephone: 1-800-897-2988 or 754-0412 Facsimile: (604) 754-0403 135 1*1 Fisheries and O c e a n s P e c h e s et O c e a n s RELINQUISHMENT. RELEASE AND CONTRIBUTION AGREEMENT IMPORTANT THIS FORM MUST BE SUBMITTED WITH THE APPLICATION BY MAY 24,1996 SECTION I: To be completed by Applicant^: Effective upon acceptance of my/our bid by DFO, I/We owner(s) of •CFV# Vessel Name which is eligible for a Category A (salmon) and relinquish all rights Enter Licence Tab # claims to the vessel licence eligibility that resulted by having had a salmon fishing licence previously issued. In consideration of the voluntary payment, I/we hereby agree as follows: i) ii) iii) iv) v) vi) Name: Name: that this relinquishment is irrevocable; that in the event of default, the Minister may (i) terminate any obligation of the government to contribute funding; (ii) require repayment to Canada of all or part of any funds paid as part of this Agreement, such an amount will be considered as a debt due and owing to the Crown. Procedures will be implemented to recover the amount from recipient(s); or (iii) pursue any other remedy at law. The following constitute events of default: (a) I/we submit false or misleading information to the Minister in relation to this Agreement; (b) I/we breach or do not fulfill any of the terms or conditions of this Agreement or of the Program; to disclose the involvement in this Agreement of former public servants who fall under the Federal Government guidelines; that the Agreement amount may be disclosed to third parties; I/we release the Government of Canada and its Ministers, officers and employees and the members of the Fleet Reduction Committee from all claims, suits, actions or demands of any nature that I/we have or may have and that are related to or arise from this Program or Agreement; and the voluntary payment should be divided and I/we direct DFO to make payment as follows: Amount: Amount: TOTAL: Signature of Applicant(s) _Date_ Date SECTION II: To be completed by the Government of Canada: Recommended by: Certified pursuant to Section 34 of the Financial Administration Act. Agreed to on behalf of the Government of Canada. Director, Operations Date Regional Director General Date Financial Coding: 5545 1011 1177 0002 C a n a d a 136 Appendix 2 An Alternative Regression Model Specification The following table presents an alternative specification for the regression model employed in the analysis. The alternative specification is motivated by a suggestion that the vessel length characteristic has a non-linear effect on the bid value, rather than the linear specification used in the earlier analysis. Instead of using vessel length then, this model employs the natural log of the observed vessel length as a dependent variable, while maintaining other aspects of the original specification. The results are not easily interpreted according to the common valuation model because the intercept term can no longer be interpreted as the intrinsic value of a licence, in the absence of other characteristics. The co-efficient on the log of vessel length is significant and of greatly increased magnitude in each gear sector. Again, the vessel age co-efficient is not significant and the reduced fee and other licence co-efficients are significant and of similar magnitude to those of the linear specification. The explanatory power of the gillnet and seine regressions is slightly increased (.42 in each case) than that observed in the linear specification. For the troll regression, the R-squared increases significantly to .78. Gillnet (n = 459) Seine (n = 48) Troll (n = 301) Coeff. t-statistic Coeff. t-statistic Coeff. t-statistic Intercept -73,390* -6.87 -241,329** -1.42 -8,5665* -9.62 Age -4.03 -0.81 -411 -1.19 4.79 0.18 Log-length 43,015* 14.01 170,510* 3.97 42,882* 16.88 Fee -10,288* 6.82 -64,164* -3.15 -9,448* -5.02 Other -10,414* -7.28 -59,151* -3.13 -4,817* -5.02 Round 2 8,097* 9.03 34182** 1.91 13867* 16.35 Indicates significance at the 2.5% level. * Indicates significance at the 10% level. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items