UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Two- and three-dimensional velocity structure of the southwestern Canadian Cordillera from seismic refraction… Zelt, Barry Curtis 1994

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

Item Metadata

Download

Media
831-ubc_1994-954183.pdf [ 3.55MB ]
Metadata
JSON: 831-1.0075296.json
JSON-LD: 831-1.0075296-ld.json
RDF/XML (Pretty): 831-1.0075296-rdf.xml
RDF/JSON: 831-1.0075296-rdf.json
Turtle: 831-1.0075296-turtle.txt
N-Triples: 831-1.0075296-rdf-ntriples.txt
Original Record: 831-1.0075296-source.json
Full Text
831-1.0075296-fulltext.txt
Citation
831-1.0075296.ris

Full Text

IMPLICATIONS OF TENURE FOR FOREST LAND VALUE AND MANAGEMENTIN BRITISH COLUMBIAbyDAOWET ZHANGB.Sc. South-central Forestry University, 1984M.Sc. Beijing Forestry University, 1989A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDYTHE FACULTY OF FORESTRYDepartment of Forest Resource ManagementWe accept this thesis as conformingt the equired standTHE UNIVERSiTY OF BRITISH OLUMBIAJUNE 1994© DAOWET ZHANG, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(SignatureDepartment of______________________The University of British ColumbiaVancouver, CanadaDate k (, L.DE-6 (2/88)11AbstractThe forest tenure system in British Columbia is a fundamental element of provincial forestpolicy with important economic implications. The effect of tenure on land value and forestmanagement has generated a great deal of speculation, but the lack of empirical information hashindered informed debate.This thesis contributes to this need for empirical information. Forest property rights aredescribed in terms of bundles of characteristics. Analyzing differences in these characteristics locateseach form of tenure in the spectrum from complete property rights to the complete absence ofproperty rights. A model for predicting land value and the intensiveness of forest management isdeveloped.An event study of forest policy changes, and a hedonic study, are used to analyze the valueof thinly traded forest tenures and that of frequently traded tenures, respectively. The results suggestthat the particular property characteristics embodied in a specific form of tenure are important factorsin determining the value of forest lands.This thesis also analyzes empirically the effect of forest tenure on silvicultural investment andthe quality of forest practices. The results show that variations in silvicultural investment and forestpractice are strongly related to the specific characteristics of forest tenure.Variations in land value, silvicultural investment and productivity are attributed to differencesin tenure characteristics: a strong and complete form of tenure leads to high land value, highinvestment and high outputs. Policy implications of this thesis are discussed, and research needs areidentified.111Table of ContentsPageAbstract iiTable of Contents iiiList of Tables ivList of Figures vAcknowledgement viI. Introduction: The Research Project III. The Context: Forest Tenure Arrangements in British Columbia 14Ill. A Model of Predicting Land Value and Forest Management 24IV. An Event Study of the Effect of Forest Tenure on Land Value 40V. The Effect of Forest Tenure on Land Value: A Hedonic Study 63VI. The Effect of Forest Tenure on Silvicultural Investment 79VII. The Effect of Forest Tenure on the Quality of Forest Practice 95VIII. Summary and Conclusions 116Bibliography 125ivList of TablesPageTable 1. Distribution of AAC and Harvests among Forms of Tenures in 1991-1992 15Table 2. Characteristics of Major Forest Tenures in British Columbia 22Table 3. Harvesting Volume Distribution of Forest Products Firms in 1987 50Table 4. Parameter Estimates for Capital Asset Pricing Models 53Table 5. Performance of Stock around the Announcement Date of Forest Policy Change 55Table 6. Parameter Estimates Using Multiple Regression Method 59Table 7. Variable Definitions, Sources and Statistics of Forest Land Transaction Variables 66Table 8. Some Statistics on Unmanaged and Managed Forest Lands, Timber Licenses 68Table 9. Empirical Results of Log-Log Equation for Forest Land Values 74Table 10. Defmitions and Statistics of Silvicultural Investment Variables 83Table 11. Some Statistics on Private Forest Lands, Timber Licenses, Tree Farm Licensesand Forest Licenses 85Table 12. Empirical Results of Log-Linear Equation on Silvicultural Investment 89Table 13. Definitions and Statistics for Dependent Forest Practice Variables 101Table 14. Some Statistics on Private Forest Lands, Timber Licenses, Tree Farm Licensesand Forest Licenses: Dependent Forest Practice Variables 101Table 15. Empirical Results for Logistic Equation: Appearance of Not SatisfactorilyRestocked Lands 103Table 16. Estimates of Percentage of Not Satisfactorily Restocked Lands 107Table 17. Empirical Results for Logistic Equation: Planting Occurrence 109Table 18. Empirical Results for Linear-Linear Equation: Months to Planting 112VList of FiguresPageFigure 1. Stock-Market Performance around the Announcement Date of Forest Policy Change 57Figure 2. Silvicultural Investment among Forest Tenures 91Figure 3. Probability of Not Satisfactorily Restocked Land Occurrence 105Figure 4. Probability of Planting Occurrence 111viAcknowledgementFirst and foremost, I would like to thank my research supervisor, Dr. Peter H. Pearse, whoencouraged me to do this project and has supported my endeavour in all stages, and who has givenme the freedom to explore and timely guidance in focusing my work.I also would like to express my appreciation to the members of my advisory committeefor their collective and individual assistance: to Dr. Clark S. Binkley for his provoking thoughts,motivation and for his supervision when Dr. Pearse was on sabbatical; to Dr. Anthony D. Scott andDr. Douglas Allen for their tremendous grasp of the theory of property rights, for their willingnessto share their knowledge and insights with me; to Dr. Russell S. Uhier for his thoughtfulness ineconometric work, generous support and valuable suggestions.Special thanks are due to Drs. ilan Vertinsky, David Haley, G. C. Van Kooten, DianaWhistler in the University of British Columbia and Dr. Peter N. Duinker in Lakehead University fortheir time and encouragement.I am grateful to the University of British Columbia for granting me full graduatefellowship for three years, and to the Forest Economics and Policy Analysis Research Unit forproviding fmancial assistance in the summer of 1993, a place to work and other research support.Data on land values were provided by British Columbia Assessment Authority, forestcompanies and individuals. Data on silvicultural investment and forest practice were provided byBritish Columbia’s Ministry of Forests. Special thanks are due to Mr. Mike Lane of the BritishColumbia Assessment Authority, and Messrs. Henry Benskin, Mike Blackstock, Bruce Bell, SteveHunt, Rob Gretchen and Ms. Meiching Tsoi of the Ministry of Forests for their much appreciatedhelp.Many people participated the discussion of this project or contributed some data. Theyare: Messrs Bill Rosenberg, Robert Sitter of International Forest Products Limited, D. A. Lang ofPope & Talbot Ltd., Rod Beaumont of Weldwood of Canada Limited, Hans A. Svendson of RiversideForest Products Limited, Karsten Holmsen of Holmsen Forestry Limited, Maurice Ayers and StuartMcPherson of Sterling Wood Group Inc., Dean Nielson of Niho Land & Cattle Company Ltd., AlanD. Fry of British Columbia’s Ministry of Environment, Lands and Parks. Thank you all!Finally, I would like acknowledge my appreciation for my family, friends and graduatestudent colleagues for their moral support.1I. Introduction: The Research ProjectThis thesis investigates the economic implications of property rights in forests. This firstchapter outlines the current status of literature on this subject, the rationale and plan for this study.The chapter first discusses the economic nature of property in general and property rightsin timber specifically. Section 2 summarizes current literature on the characteristics and effects oftenure, the theoretical interpretation of the effects, and methodologies used in empirical studies.Section 3 explains the rationale of this study and Section 4 describes the plan of this thesis.1. BackgroundIn most parts of the world, governments influence management of natural resources suchas forests, minerals and fisheries for a variety of purposes: to compensate for the failure of markets,to account for the full social costs and benefits of resource use, and to achieve certain distributiveobjectives. The degree of government involvement ranges from the traditional socialist state wheregovernment owns, manages and utilises the resources, to the market state where govermnentintervention is limited to regulating private resource owners. Canada lies within this range,characterized by public ownership of resources that are utilized by private enterprises.The link between publicly owned resources and the privately owned enterprises is thetenure system. In Canada, and in developing countries such as China, Malaysia and Kenya whichhave sought to use market forces to develop their economies, tenure systems have had profoundimpacts on the economic efficiency and social impacts of natural resource sectors. As a result of the2practical importance of tenure arrangements, the economics of property rights and contracts hasemerged over the last three decades, as a significant, specialized field of study.Forest tenure refers to rights over forest land. In Canada, where the majority of forestlands are publicly owned, most forms of forest tenure are contractual arrangements between thegovernment and those who use forest resources. The forest tenure system is the instrument forallocating public timber to private enterprises. It has played a major role in forest policy since earlyEuropean settlement. The result of a century of evolution is a complicated array of rights over forestland and timber. The characteristics of different forms of tenure affect the management and thedevelopment of resources, the division of responsibilities for these activities, the efficiency of thedependent industries and the distribution of the economic benefits from resources used (Pearse1 990b).2. Literature Review2.1 Characteristics of TenureIn law, property is often defmed as a bundle of rights (Pearse 1990a). In order to analyzethe effect of tenure as a whole on resource management, tenure can be described in terms of itsintegral parts or “bundles” of rights (Scott 1990; Pearse 1990a). Depending on the author, betweensix and eight such dimensions of property can be identified.Comprehensiveness — the range of benefits from an asset which a property holdercan claim.Exclusiveness — the extent to which a property holder can prevent others from3freely sharing in the benefits of the asset.Duration — the time over which the rights can be exercised.Flexibility — the extent to which the specifications of a particular property can bemodified during its duration.Quality of title or security — the extent to which a person’s ownership of aproperty right is secure, socially acknowledged and enforceable. In Canada,security has no statutory or common law basis but depends on the socio-politicalenvironment in which a particular set of property rights is granted (Scott 1990).Transferability — the rights of property holder to transfer the entitlement to anotherparty through sale, donation, bequest or the similar mechanisms.Benefits conferred to the tenure holder — how much of the economic returngenerated by an asset can be captured by the holder? Property rights may be limitedby the imposition of taxes, fees and restrictions which require the property to bemanaged or maintained in certain ways. Even freehold title does not convey theright to all benefits from property because rules and regulations restrict the use ofassets and taxes expropriate part of the gains.Although all seven characteristics bestow benefits on the owner of a right, they can beseparated into two classes according to the nature of the benefits conferred (Scott 1991). Increasesin the first four characteristics (comprehensiveness, exclusiveness, duration and flexibility) yieldbenefits to tenure holders whether or not they are in a position to adjust the amounts of labour andcapital they employ. These dimensions are valuable because they facilitate better use of inputs, bypreventing overlap of property possession with successors or neighbours. For example, anenhancement of duration allows more efficient allocation overtime, while an increase in exclusivenessenables the holder to avoid common property problems with neighbours in the same space.Increases in the second three characteristics (quality of title, transferability and the benefits4conferred) yield benefits to the owners chiefly by assisting them to change their factor/land ratioprofitably. They are valuable because they simplify the process of attaining the best combination oflabour and capital (or allow the transfer of property to those with better access to suitable labour andcapital). Thus, an increase in transferability of land promotes its “highest and best use”. Moreover,superior quality of title facilitates access to credit; the value of such benefits has been demonstratedfor rural Thailand where farmers with secure title were able to increase their capital/land ratio andthe productivity of lands (Feder et al. 1988).2.2 Effects of Forest Tenures and Their CharacteristicsForest tenures in Canada are the primary means of reconciling the interests of private userswith those of public landlords (Pearse 1990b). They also have a major influence on how resourcesare allocated, managed and developed, and on the distribution of the economic benefits fromresources used. Commercial timber contracts and small-scale tenure arrangements determine theincentives for depletion of, and investment in, forest resources (Hyde, Mendelson and Sedjo 1991).Poorly-defined property rights in tropical countries distort incentives for efficient land use. Alongwith policy spill-over from other sectors, poorly designed and enforced tenures are widely recognizedas a major factor in wasteful deforestation.Among the characteristics of property, comprehensiveness, security, duration andtransferability are widely regarded as the most significant factors affecting Canada’s resource tenurearrangements (Pearse 1988, 1993; Schwindt 1992). Hyde and Newman (1991), using examples indeveloping countries, suggest that security, comprehensiveness, and duration are key factors in longterm stewardship of lands and forests. Their conclusions are summarized in the following paragraphs.5First, the absence of security encourages over-exploitation and disinvestment. Securityencourages resource users to invest in the long-term management of the resources, includingconservation and resource protecting investments. Second, without property rights extending to thefull range of potentially valuable resources, local people (often squatters) have no incentive to protectthe long-term productivity of the lands, or to produce non-marketable values from the forests. Rather,the incentive is to extract what is commercial in the short-run and to move on. Third, timberconcession arrangements are often too short to encourage silvicultural activities beyond the firstharvest.Feder et al. (1988) emphasize the importance of transferability. Without it, property holdersforgo not only access to credit, which may be used to support long-term conservation investments,but also the right to sell, which limits the potential returns on their investments by making the holdingperiod equal to the physical maturity of the investment. Without the opportunity to make land salesand transfers, there is no incentive to leave property in good condition should the holders choose tomove.Why do tenures and their characteristics have these effects? The economic theory of propertyrights postulates that property rights steer economic behaviour within a society assuming that propertyrights are exogenous factors of any given finn (Libecap 1989). Jensen and Meckling (1979) arguethat a firm’s position on its production function is constrained by the structure of property rights, justas the production function is constrained by the state of technology. This concept is related to thatof Coase (1960), who argues that property rights are factors of production.More specifically, the structure of property rights affects the range of choice available to a6firm, and the costs of using alternative inputs. By influencing the relative costs of inputs (capital,labour and land), property rights affect the quantity of input used, which in turn determines theoutputs or the location on the relevant production functions. Furthermore, by affecting the economicincome of the firm, property rights determine its incentives to invest. For example, restrictions onthe length and transferability of a tenure are likely to dull the tenure holder’s incentive for long-terminvestment. With reduced investment, production falls and output decreases, and the resultingreduction in income lowers the value of the tenure itself.If one treats the tenure system endogenously at society level rather than specific firms,however, the above conclusion about more inputs and more outputs is less certain. In this case, thetenure system is subject to change, but transaction costs of the change might overshadow the benefits.Then the forms of tenure currently used may be the best ones from society’s point of view in spiteof poor economic performance of individual firms.To make this point clear, it is necessary to step back and look at the role of transaction costsin the evolution of property rights. In his seminal article, Coase (1960) articulates what has cometo be known as the Coase theorem: that when transaction costs are zero, the gains from trade aremaximized independently of any initial distribution of property rights. Transaction costs, as statedby Allen (1991), are the costs of establishing and maintaining property rights. Thus regardless ofhow the government assigns property rights and liabilities among interest groups, if the costs ofcreating and maintaining property rights for various users of forest resources are zero, then the forestresources will always fmd the same uses. All parties will calculate costs and benefits and reach thesame decisions on resource uses. In other words, zero transaction costs mean that the initial7distribution of rights is unimportant, because the eventual outcome will always be the same’.Transaction costs are always positive in the real world since property rights to assets cannotbe perfectly delineated due to incomplete knowledge and the heterogeneity of the assets2 (Barzel1989). According to the Coase theorem, if transaction costs are too high, they might block anyreallocation, and the initial allocation determines the efficiency of resource use and distribution ofincome. For example, rearrangement of common property can increase productivity or resource rents(Cheung 1970). However, the social benefit of reorganizing property rights cannot in itself justifythe changing of property rights; the costs of organizing and enforcing such rights (and the distributionof net gains) must be considered as well. As Coase (1960) noted, the existence of common poollosses does not necessarily mean that it is in society’s best interest to take actions to more completelydefmed property rights: “[But] the reason why some activities are not the subject of contracts isexactly the same reason why some contracts are commonly unsatisfactory — it would cost too muchto put the matter right.” Therefore, transaction costs as well as the net gains (and their distribution)of property right adjustment determine the evolution of property rights over time (Demsetz 1967;Libecap 1989).The above arguments need qualification. Even if changing the property rights does move theproduction possibilities curve outwards, such changes might affect the distribution of wealthadversely. Therefore, from the viewpoint of positive economics, it is impossible to evaluate the1 Several recent articles show that initial distribution of rights does matter. For example,Kahneman, Knetsch and Thaler (1990) argue that measures of willingness to accept greatly exceedmeasures of willingness to pay. They label the increased value of a good to an individual when thegood becomes part of the individual’s endowment the “endowment effect.”2 These conditions require that every exchange contains some mechanism for measurement andenforcement. See Barzel (1989) and Cheung (1974) for discussion about the heterogeneous (or multidimensional) nature of assets (gasoline and house renting, respectively).8impact of changes in property rights on social welfare (Furubotn 1985). Eggertsson (1990) alsoobserves that it is difficult to resolve the distributional conflicts inherent in major changes inownership arrangements. Considering the complexity of distributional effect, most authors discussthe effect of tenure from an utilitarian point of view (the more, the better) and implicitly employ thePareto potential compensation principle, i.e., if the gainer from an action can potentially compensatethe loser, the action should be approved.2.3 Methodology Used in Empirical StudiesThere have been several studies on the effects of different forest tenures and theircharacteristics on economic efficiency and resource allocation. Luckert (1988) and Luckert and Haley(1989, 1990) use an interview method to analyze the effect of tenure characteristics on the distributionof economic rent, the allocation of forest resources, and investment in silviculture in BritishColumbia. Leffler and Rucker (1991) use the transaction costs method to analyze the effect oftimber-harvesting contracts in Southern United States. Giffis (1990) analyzes descriptively forestconcession management and revenue policies in tropical countries. Vincent (1990) illustrates howthe inefficiency of tropical timber royalty systems affects the feasibility of tropical forest management.There are apparently no other empirical studies done in the forestry sector.More studies have been conducted on agricultural tenure arrangements. Feder et a!. (1988)measure the benefits of providing secure tenure to farmers in agricultural communities in ruralThailand and conclude that the benefits, in terms of additional long-term productivity, outweigh theinitial administration costs. Mighot-Adholla et al. (1990) apply the model provided by Feder et a!.(1988) to examine agricultural tenures in sub-Saharan Africa. They find that indigenous African9(Kenya, Ghana and Rwanda) institutions adapt as needed and that less secure tenure does not retardagricultural development in these countries. A study by Anderson and Lueck (1992) observes thatland tenures on American Indian reservations have significant impacts on agricultural productivity.The relationships between the methodologies used in these studies can be summarized asfollows. First, the interview method (Luckert 1988; Luckert and Haley 1989, 1990) reveals thepreference of tenure holders; the observation method (Leffler and Rucker 1991; Anderson and Lueck1992) shows their behavioral choices. Analyzing actual behaviour is always preferable given thewell-known problems associated with the use of the interviews to predict actual behaviour (Cummingset at. 1986). Second, the tenure-characteristics method attributes the effect of a specific form oftenure to its characteristics (Feder et a!. 1988; Luckert 1988; Luckert and Haley 1989, 1990) and thetransaction costs method attributes the effect of tenure to transaction costs (Leffler and Rucker 1991;Anderson and Lueck 1992). They are related or sometimes even similar methods. To understand thisrelationship, it is necessary to address the relationship between tenure characteristics and transactioncosts.Tenure characteristics and transaction costs are related concepts. Changes in somecharacteristics of a tenure (comprehensiveness, exclusiveness, duration, flexibility and benefitconferred) may increase or decrease the transaction costs of the tenure itself. However, theenhancement of these characteristics yields benefits to the tenure holder, and therefore may helpreduce the transaction costs of factor inputs (e.g., easy access to credit and labour). Enhancement ofthe others (security and transferability) will not only reduce the transaction costs of factor inputs, butalso the transaction costs of the tenure itself, by definition. Thus the effect of tenure could beattributed to either tenure characteristics or transaction costs or both, depending on the particular10circumstance.The conceptual model tested by Feder et al. (1988) illustrates this point. The model looksat the indirect relationships between “security ownership of land”, measured by the holding of a legaltitle, and land productivity. Farmers with greater security of title have a higher probability ofrecouping the benefits from land improvements and are therefore more inclined and able to makemedium or long-run land improvements. Moreover, because greater security implies a greaterlikelihood of repayment, lenders are more willing to offer credit (or in other words, the transactioncosts of capital are reduced), leading to easier financing of improvements and inputs. Therefore theirconclusions could be attributed to either the characteristics of the tenure (security) or transaction costsor both.3. The Rationale for This ResearchAlthough there have been recent studies on the effect of forest tenure, there is a paucity ofresearch on the implications of the forest tenure system, and a still greater scarcity of empiricalguidance in this important forest policy issue. The consequence is serious, especially for BritishColumbia, where more than 95 percent of forest lands are publicly owned, with the rights to harvesttimber and to manage the forests granted to private industry through various tenures.The importance of tenure in the British Columbia forest industry is widely recognized, andit is the impetus behind many forest policy changes. For example, in 1987 the government of BritishColumbia attempted to change most Forest Licenses, a volume-based tenure covering over 50 percentof timber harvests, into the area-based Tree Farm License (Ministry of Forests and Lands 1987; Haley11and Leitch 1992). The move was intended to give companies “more tenure security” and to “createa positive environment for investment in forest industry.” A similar recommendation was made bythe British Columbia Forest Resource Commission (1991), based on the similar assumption that theinsecurity of tenure and lack of equity in future timber crops act as major disincentives for privatesector to invest in silviculture on Crown lands held under various tenure arrangements. However, thequestion remains, all else equal, how much more will a Tree Farm Licence holder invest than a ForestLicense holder? There is no substantial empirical evidence to back up the assumption that the tenuresystem is an impediment to investment. Not surprisingly, these recent efforts to reform the tenuresystem have failed to win public support and have not been implemented.This project has emerged from the importance of and apparent need for empirical studies onthe forest tenure systems. Pulling further relevant threads of economic theory and testing thosetheories by economically sound empirical studies are the major tasks of this project. Specifically, theobjective is to address the following questions:• How do the characteristics of forest property rights affect the value of the property?• What is the role of tenures in detennining silvicultural investment?How do tenures affect outputs and silvicultural performance?By contributing to a better understanding of forest tenures, this study not only has importantimplications for British Columbia forest policy, but it also has relevance to forestry problems in otherjurisdictions. In particular, the conclusions of this thesis may be applied to countries who use orattempt to use the private sector to manage public forests. In addition, the theory and methodologydeveloped in this thesis can be adopted to address many kinds of property rights and leasing12problems.2.4 Thesis PlanThe main body of the thesis is organized in three parts, each of which addresses one of thethree questions.Part one includes chapters II and ifi. Chapter II reviews the forest tenures in BritishColumbia in terms of their characteristics. Chapter III establishes a theoretical model for predictingthe value of forest property rights, silvicultural investment and output, based on forest tenurecharacteristics. The development of the model generates the general hypothesis of this thesis.Part two consists of chapters IV, V, VI and VII, all of which deal with empirical studies.Chapters IV and V analyze the effect of forest tenure on property (land) value. Since most Crownforest lands under various tenures in the province are rarely traded, their market value can only beapproximated, based on the stock market’s implicit valuation of harvesting rights. The study reportedin Chapter IV is based on the premise that given the relative efficiency of stock markets, anyreduction in harvesting rights should be reflected in the stock prices of affected finns.Chapter V employs a hedonic study method to examine the more frequently traded foresttenures, concentrating on private forest lands and Timber Licenses. Chapters VI and VII address theeffect of forest tenure on silvicultural investment and the quality of forest practice, respectively.All chapters in this part follow a general outline: introduction, methodology, data, empirical13results, and conclusions and discussions. Each chapter itself is a more-or-less self-contained study.Part three consists of Chapter VIII. It summarizes the study and its contributions to thegeneral knowledge of property rights. It also addresses the policy implications of the findings. Thereport concludes by identifying areas for further research.14II. The Context: Forest Tenure Arrangements in British ColumbiaThis thesis focuses on the effects of forest tenures. Based on a description of existing privaterights to forest resources on Crown lands and private forest lands in British Columbia, this chapterdescribes the relative importance of each specific tenure, the rights and obligations conveyed by eachform of tenure and the key characteristics that distinguish one tenure from another. By doing so, thischapter will indicate which types of tenure are the subject of this study, locate specific interests inthe range from complete property to the absence of any private rights, and help to single out thetenure characteristics to which the fmdings of this thesis should be attributed, and to whichimprovement of future policy could be usefully directed.Table 1 presents the distribution of actual harvests and the committed annual allowable cut(AAC) among forest tenures in B.C. Four types of tenures — private forest lands, Tree FarmLicenses, Timber Licenses and Forest Licenses — accounted for, respectively, 8.2, 16.4, 6.3, 53.5percent, together 84.5 percent of actual harvests in 1991-1992. They also made up some 80 percentof the committed AAC in the same year. These tenures are the subject of this study. Other tenures,such as Woodlot Licenses and Timber Sale Licenses (Major) are not considered here because theyhold an insignificant portion of annual timber harvests and because there is too little information onthem3 to make useful analysis possible. Minor Timber Sale Licenses (issued under Small BusinessForest Enterprise Programs) are excluded for two reasons. First, they are not transferable andtherefore no market price can be observed. Second, the government assumes the managementresponsibilities after harvesting (Gillespie 1991, p.5), and the silvicultural activities are recordedHowever, Timber Sale Licenses (Major) are included in the event study on land value inChapter IV since they are included in the forest policy changes. See Chapter IV for discussion aboutchanges of forest policy in B.C. in September, 1987.15Table 1. Distribution of AAC and Harvests among Forms of Tenures in 1991-1992Form of Tenure AAC(103m) % Harvest (103m) %Private LandsaWithin Tree Farm License 419 0.56Outside Tree Farm License 5,752 7.68Sub-total 6,172 8.24Crown LandsTimber License1’ 4,695 6.27Tree Farm License° 17,322 23.61 12,312 16.43Forest License 41,562 56.70 40,109 3.54Woodilot License 482 0.64 407 0.54Minor Timber Sale License (SBFEPd) 9,216 12.56 8,465 11.30Timber Sale License (outside SBFEP) 3,110 4.24 1,083 1.45Others 1,657 2.26 1,505 2.01Sub-total 68,575 91.53Federal Lands and Indian Reserves 173 0.23Total 73,405 100 74,920 100Source: British Columbia Ministry of Forestry, 1991-1992 Annual Report.a All Crown Granted lands.b Both within and outside Tree Farm License.AAC figures include both Schedule “A” and “B” lands; harvest figures include “B” lands only.d Small Business Forest Enterprise Programs.16differently from other tenures. Therefore this form of tenure cannot be directly compared to theothers.1. Private Forest LandsThe origins of the private forest lands in B.C. can be traced back for more than 100 years4.Before 1906, the Crown had granted extensive areas of forest lands to private users as fee-simplelands, now often called Crown granted or private forest lands. Private lands are the most completeform of right over forest lands conveyed to private parties. The right is comprehensive, includingboth the land and the timber. It is exclusive, freely transferable, flexible, secure and perpetual.Furthermore, private land holders reap all of the economic benefits after paying for property tax andbear all of the management and development costs.Few regulations apply to private forest lands5. As a general rule, private forest owners mayharvest their timber and manage their lands as they wish. They may classify their lands as eithermanaged or unmanaged forest lands. The holders of managed forest lands must make a work planand a commitment to practice sustainable forestry, in return for a preferential property tax treatment.The owners of unmanaged forest lands enjoy fewer restrictions. Most industrial private forest landsin B .C. are private forest lands within Tree Farm Licenses (see below), which are usually classifiedas managed forest lands as well6. Unlike the owners of other private forest lands, the owners ofSee (Pearse 1992) for a succinct summary of the evolution of forest tenure in B.C.Log export restrictions apply to some private forest lands and to all of the other tenuresdiscussed below. These may have a large, negative impact on land and timber values, and thereforeon the absolute level of investment but not the relative level across tenures.6 R.B. Townshend. Personnel Communication. 1993.17these managed forest lands within Tree Farm Licenses have to report harvesting and silviculturalactivities to the B.C. Ministry of Forests, although they do not need cutting permits to harvest theirforests.2. Tree Farm LicensesThe Tree Farm License is a relatively long-term, large-scale tenure, serving large industrialenterprises which are often required, as a condition of the license, to operate a timber processingfacility in B .C. Five characteristics of Tree Farm Licenses are important for this study. First, theholders have rights to the timber on the land only. Their rights are therefore less comprehensive thanthose of private forest owners. Second, Tree Farm Licenses have a limited term of 25 years, withprovisions for “evergreen” replacement7.Third, Tree Farm Licenses may include private forest landsand Timber Licenses (called “Schedule A” Lands), in combination with Crown lands (“Schedule B”Lands). Fourth and most important, the forests and lands under Tree Farm Licenses have to bemanaged under approved management plans and the holders of these tenures are obligated to obtaincutting permits from the Ministry of Forests. Lastly, Tree Farm License holders pay stumpage at anappraised rate for timber harvested, and land rental for standing timber on “B” lands, and they mustcarry out silviculture and road building (until 1987 some of the costs were reimbursed by thegovernment).In contrast to private forest lands and Timber Licenses, transfers of Tree Farm License requirethe consent of the Minister of Forests. Furthermore, the transactions should include part or all of the“Evergreen” replacement means that, after 10 years of the license term have expired, the holdermay call for a new 25-year license to replace the original one and the government is obligated to offera replacement license with only minor modifications of the terms and conditions.18appurtenant manufacturing facilities (Haley and Leitch 1992; Schwindt 1992). When a Tree FarmLicense is transferred, five percent of the allowable annual cut (AAC) attributable to “Schedule B”lands is retracted by the Crown. In short, the transaction costs for Tree Farm Licenses aresignificantly higher than for private lands.Holders of Tree Farm License are required to make 50 percent of the harvests from “ScheduleB” lands available for harvesting by independent contractors. Tree Farm Licenses are furtherrestricted by cut control, which dictate that licensees must harvest within 50 percent of AAC annuallyand within 10 percent over a 5-year period. Up to 5 percent of the AAC of a Tree Farm License maybe taken without compensation during the 25-year term8. These restrictions, added to legislatedchanges9 to the Tree Farm License have weakened the security of this tenure (Luckert 1991).3. Timber LicensesTimber Licenses came into existence with the conversion of the Old Temporary Tenures’°since 1978. Tn this form of tenure, the Crown owns the land and timber and the licensees are givena non-renewable right to harvest the mature timber within a specified period. The right is notcomprehensive since it excludes the land. It also has finite duration. Nevertheless, it is exclusive,8 There is some uncertainty over the maximum amount which may be deducted withoutcompensation. According to some interpretations Section 53 of the Forest Act allows for 5 percentdeletion for highways, pipelines and similar rights of way and an additional 5 percent for otherpurpose (Schwindt 1992, p.76).The two most recent legislation changes to Tree Farm Licenses occurred in 1978 and 1987. SeePearse (1976) for the proposal that led to the former change and Chapter IV of this thesis for thelatter.10 See Pearse (1992) for discussion of the historical origin of Old Temporary Tenures.19and transferable under the condition that the holders pay five percent of the higher of the marketvalue of the timber standing in the area of the license and the value that the licensee has declared tothe Crown. Since 1987, the holders of Timber Licenses pay an annual rental and at their choice,either a fixed royalty or a variable stumpage on timber harvested. If the royalty is chosen, the holdersmust bear the costs of all works, notably reforestation and road-building required under the ForestAct and other related regulations. As Pearse (1992) notes, “with few exceptions, Timber Licenseholders chose the royalty option”. Therefore, it is safe to say that the Timber License holders areresponsible for the reforestation costs.Timber Licenses exist both within and outside Tree Farm Licenses, accounting for less than5 million cubic metre, or, 6.3 percent of the total 199 1-1992 billed harvests. Roughly half of thatvolume came from areas within Tree Farm Licenses. All Timber Licenses contained in Tree FarmLicenses are subject to the Tree Farm License agreement and its management and working plans, andonce harvested, they are rescheduled as “B” lands. Those Timber Licenses outside Tree FarmLicenses are subject to an operating plan, which must be submitted to the Chief Forester for approvaland they simply revert to the Crown once harvested. All harvesting operations on these TimberLicenses must be carried out under cutting permits that conform to approved operating plans preparedby professional foresters. Because of these regulations Timber Licenses are less flexible and lesssecure than private forest lands, and probably Tree Farm Licenses. Moreover, because TimberLicenses outside Tree Farm Licenses will revert to the Crown once harvested, the rights associatedwith these Timber Licenses are weaker to their holders than that with Tree Farm Licenses in term ofequity in future crops.4. Forest Licenses20Forest License is the most important type of tenure in term of harvests in British Columbia(Table 1). It is a volume-based license” by which the licensees has a “quota” or right to cut aspecified volume of timber per year within a broad administrative area. The specific location ofoperations is designated from time to time. As in the case of the Tree Farm License, Forest Licenseholders have rights to the timber only; thus their rights are less comprehensive than those of privateforest land holders. The licenses are issued for 15 years and most of them are renewable orreplaceable on an “evergreen” basis. They are transferable, subject to Ministerial consent and a take-back of 5 percent of AAC. All harvests under Forest Licenses must be conducted under the termsof cutting permits from the Ministry of Forests, and in return for the cutting rights, the licensee mustsubmit successive five-year management and working plans which include a description of operationsto be conducted and silviculture treatment to be undertaken. The licensees pay stumpage and rentalto the Crown and practise silviculture at their own expense to ensure successful regeneration withina specified period. A cut control similar to Tree Farm Licenses applies to Forest Licenses; thelicensee must harvest within 50 percent of the AAC each year and to within 10 percent over a 5 year“Traditionally area-based tenures refer to the rights to timber in a given geographical area, andvolume-based tenures refer to the rights to certain volume of timber irrespective of location. The area-based tenures in B.C. (Tree Farm License, Timber License and Woodlot License) are consistent withthe above definition: they all have a specified geographic area; harvest and silvicultural activities areconfined within the boundary of the geographic area.In contrast, the volume-based tenures have slightly changed their meanings in B.C. There isno specific geographical area for volume-based tenures (Forest License and Major Timber SaleLicense). However, there are broad areas (Timber Supply Areas) that define the operational boundaryfor these licenses. Therefore licensees’ actual harvest areas (and subsequent silvicultural activities)are changed over time although licensees usually have enough timber to log for 3-5 years once theyset up a log camp. Because the licensees are responsible for their actual harvest area until they have,after 15-20 years, completed their silvicultural responsibilities, the volume-based tenure also fits thetraditional definition of area-based tenures: which have well-defined locations in this period.Since volume-based tenures do not guaranteed that their holders may come back to the samearea where they logged and made investment in silviculture a long time ago, they are considered asless attractive than area-based tenures in terms of silvicultural investment. This hypothesis is verifiedby empirical studies in this thesis.21period.The obligation of licensees to provide independent contractors with the opportunity to cut aportion of the authorised harvest now applies to Forest Licenses as well (Pearse 1992). In addition,up to 5 percent of AAC can be withdrawn from Forest Licenses without compensation during the 15-year term. All Forest Licenses bear a lesser degree of security than Tree Farm Licenses since ForestLicense holders perceive that they are unlikely to return to the same area in which they invest insilviculture, and thus can not recoup the benefits.5. SummaryTable 2 summarizes the rights conferred by the more important forest tenures. Collectivelythese tenures occupy a wide range within the spectrum of “property”. They are similar inexclusiveness’2,but vary considerably in the characteristics of comprehensiveness, duration, security,transferability and benefit conferred. These are the key characteristics, and the rest of this studyinvestigates their effects on land value and forest management.12 Forest License is exclusive in terms of harvesting rights, but not exclusive over time.22Table 2. Characteristics of Major Forest Tenures in British ColumbiaPrivate Tree Farm Timber ForestLands License License LicenseComprehensivenessLand and Timber Yes No No NoTimber only No Yes Yes YesExclusiveness Yes Yes Yes YesDurationTerm Perpetuity 25 years Until timber 15 yearsremovedReplacement Perpetuity Evergreen No EvergreenSecurityDeletion Conditions No Yes No YesArea or Volume Based Area Area Area VolumeGeneral Security Yes Less secure Yes Less secureTransferabilityFree Transferable Yes Permitted with Yes Permitted withconsent consentTakeback when Transfer No Yes Yes Yes23Table 2. continuedPrivate Tree Farm Timber ForestLands License License LicenseBenefit ConferredProperty Tax Yes No No NoCrown Charge: Stumpage No On “B” lands No YesCrown Charge: Royalty No On some Timber Yes NoLicensesCrown Charge: Rental No Yes Yes YesObligation: Reforestation On managed Yes Yes Yesforest landsRestriction: Cut Control No Yes No YesRestriction: Log Export Some Yes Yes YesRestriction: Contractor No Yes No YesClause24Ill. A Model of Predicting Land Value and Forest ManagementThis chapter examines the expected effect of forest tenure on land value and forestmanagement by applying a model incorporating a simplified version of forest tenure in BritishColumbia and the traditional theory of capital. Results from the model suggest hypotheses about thedirection of the effect of an increase in the characteristics of tenure on land value, silviculturalinvestment and outputs.The chapter is divided into three sections. The first states assumptions about the land andland market, capital market, tenure, tenure holder and production. The second develops the modelitself’3 and determines the expected effect of tenure. The thesis hypothesis and methodology areaddressed in Section 3.1. AssumptionsCapital theory postulates that the value of a tract of forest land (or any other capital asset),is the present value of the future net revenues that the asset is expected to produce. The present valuecan be viewed as the demand price of the assets: the maximum price that a buyer would be willingto pay for the rights to the asset’s expected income. It can also be viewed as the supply price of theasset: the minimum amount that a seller would be willing to accept to relinquish the rights to the13 The model developed in this chapter is consistent with other models. It is similar to the modelof Feder et al. (1988), but has gone one step further by incorporating more tenure characteristics. Ithas implicitly used the concept of transaction costs since the key factor considered in the model, theprobability of eviction, affects the transaction costs in terms of maintaining the tenure. Therefore itis by and large consistent with the transaction cost method in Leffler and Rucker (1991) as well.However, the model in this chapter is deterministic and those in Feder et al. (1988) and Leffler andRucker (1991) are stochastic.25income. In this context, the present value of the expected net revenues is the asset’s marketequilibrium price (Washburn 1990). Based on these concepts, a model of forest land value,investment, outputs and tenure characteristics can be developed for British Columbia under thefollowing simple assumptions, necessary to focus on the main issues.1.1 Land and Land Marketa. Forest lands are suitable only for growing timber and are uniform in quality, accessibilityand species composition; but differ in tenure arrangements. (This assumption is relaxedin empirical study by using suitable indices to control the heterogeneity of land quality,location and species composition.)b. All lands are divisible and transferable without transaction costs.c. Tenure arrangements cannot be changed by tenure holders alone.1.2 Capital Marketa. All tenure holders can obtain the necessary amount of capital from the capital market atfixed, real interest, r.1.3 Tenurea. All forest tenures are exclusive and flexible. The latter refers to the specification that a26tenure can be modified during its duration.b. Timber is the only benefit conferred to a tenure holder, and all tenure holders pay auniform tax, say, a certain percentage of net income. (This assumption is relaxed in theempirical study.) Thus, all tenures have the same conditions of comprehensiveness andbenefits conferred.Under these assumptions, only security and duration are left. The most importantcharacteristic is security, which is affected in part by duration (Pearse 1990a; Schwindt 1992) andtransferability. Renewability of tenure is also a factor of security where the tenure is renewable, butfor simplicity it is not considered here.c. Free-hold tenure is secure; others are not. Under free-hold tenure, the probability oflosing the property and/or of losing a proportion of the property are zero. There is a nonzero risk of eviction and/or expropriation of a proportion of the property under othertenures. Security is positively related to duration. In other words, the longer the duration,the smaller the likelihood of losing tenure.1.4 Tenure Holdera. A tenure holder begins with purchasing or negotiating some amount of forest lands, onwhich there is mature timber standing, under a particular form of tenure. She or heharvests the mature timber in t years, and decides the amount of investment in two kindsof forest activity (discussed below). The tenure holder maximizes net terminal wealth27after harvesting the mature timber’4.b. The tenure holder is risk averse and has a risk premium (s) in risky activities such asinvesting in tenured forest lands. The risk premium goes up when the risk of evictionincreases. The discount rate for the tenure holder is r-i-s.1.5 Productiona. The yield of the mature timber is not affected by any human inputs, but by time alone.Therefore, it does not directly enter the maximization of terminal net worth for the tenureholder.However, the presence of the mature timber represents an opportunity for the tenureholder to capture part of the economic rents generated from it, which may be used directlyto fmance the investments in forest lands, andlor to use it as a collateral to borrow moneyin capital market. The risk of eviction affects the likelihood of the tenure holder tocapture the rents and to use it as collateral, which affects the risk premium (s) of thetenure holder to invest in forest lands. In other words, higher risk of eviction (higherprobability of losing the mature timber and the lands) leads to higher financial costs ofinvestment. Therefore, by affecting the risk premium, the chance of losing the maturetimber affects indirectly the terminal net worth of the tenure holder.b. After harvesting the mature timber, the holder can invest in two types of activity withAn alternative is to assume a tenure holder will maximize present net worth when acquiringa tenure. This does not change the results, but alters the equations slightly.28regard to forest lands. The first is Silviculture, K=B-i-I, where B is the costs of basicsilviculture’5 (seed collecting, site preparation, regeneration and brushing), which ismandatory and uniformly carried out for all tenures; I is costs of intensive silviculture(tending, thinning, pruning and fertilization), and is dependent on the perception of theholder of the possibility of acquiring the value of second growth forests, which, in turn,is affected by the characteristics of tenure. The second is land improvement andinfrastructure, M (road building, soil conservation practice and so on), which increasesthe value of bare land and depends on the tenure holder’s perception of the likelihood ofcollecting land rents, and the amounts of these.The cost function for these activities is linear with respect to K and M.(1) c=f(k,m); ci>O,cm>O;c=O,cmm=Owhere c is per-hectare cost; k is per-hectare silvicultural expenditure and m is per-hectareland improvement expenditure.c. The production function of second growth (and the subsequent) forests (Y) exhibitsconstant returns to scale in land, silviculture and land improvement. Productivity is alsoaffected by rotation age (T).(2) y=y(k,m,T)15 In B .C. Forest Act, basic silviculture refers to harvesting methods and silvicultural operations,including seed collecting, site preparation, artificial and natural regeneration, brushing, spacing, standtending and other operations that are required for the purpose of establishing a free growing crop ofa commercially valuable tree species. In practice, spacing and tending are often seen as intensivesilviculture. Here all lands are assumed to be within the intensive margin of timber production (Pearse1990a); i.e., it is economically worthwhile to invest in silviculture and land improvement if tenurewere secure.29where y1 = ai > 0, y < 0, y1 > 0 for i * j, i = k, m, T and lowcase letters denote per-hectarevariables.d. The rotation age (T) is identical for all forests under different tenure systems. Allharvesting involves clear-cutting.e. Investments in land improvement and infrastructure (which only occur right afterharvesting the mature timber) increase the net terminal value of bare lands (Q) (after therotation of second growth forests), but have decreasing marginal returns:(3) q = q(m) q’= dciJdm> 0; q” <0f. The products of forest lands under different tenure arrangements are homogeneous andthe prices of output are set to unity.2. Development of Model: The Expected Effects of TenureThe per-hectare net terminal wealth of the holder at the time when the mature timber isharvested can be expressed as:(4) V = y(k, m, T) exp(-rT-sT) + q(m) exp(-rT-sT) - c(k,m)where the first term is the value of the second growth; the second term is the net value of bare landsafter harvesting the second growth (both terms are discounted to the date when the mature timber isharvested); the third term is cost after harvesting the mature timber. The tenure holder chooses k andm to maximize the terminal wealth.30First-order conditions for a maximum require that:(5) y exp(-rT-sT) - c= 0(6) Ym exp(-rT-sT) - q’ exp(-rT-sT) - Cm = 0The Hessian Matrix of the system (5) and (6) is given by:(7) (H11 H12H=IH H21 22whereH11 = y exp(-rT-sT) <0(8) H22 = (y, + q”) exp(-rT-sT) <0H12 = H21= Ymk exp(-rT-sT) > 0The second-order conditions for a maximum require that the determinant of H be positive:(9) H11 H22 - H122 = (y mm + q” y - y)By the concavity of the per-hectare production function, y mm > y2. The other term isclearly positive, so the determinant is positive.The effect of tenure insecurity is demonstrated by assuming a small increase in the likelihoodof eviction, which leads to a small increase in risk-premium, s. The effect of an increase in thelikelihood of eviction is shown by differentiating equations (5) and (6) with respect to s, which yieldsthe following comparative static results:(10) dk/ds = T Yk /y <031(11) dm/ds = T (y + q’)I (Ymm + q”) <0Proposition 1. The value of the tenure is related positively to the security and duration of theproperty rights.This proposition is supported by equation (4). When the risk of eviction goes up, the riskpremium increases and the value of equation (4) decreases.Proposition 2. Both silvicultural and land improvement expenditures are related negativelyto tenure insecurity and are related positively to tenure duration.This point is given by equations (10) and (11).Proposition 3. The productivity of forest lands is related negatively to tenure insecurity andrelated positively to tenure duration.This point can be seen by comparing equations (2), (10) and (11). As Yk >0, y, >0 and dk/ds<0, dm/ds <0, output, y, is related negatively to riskiness of tenure, but positively to duration oftenure.A Word about TransferabilitySince forest tenures are assumed to be transferable, the above model does not deal explicitlywith the effect of transferability. Forest tenures in B.C., as I have shown in Chapter II, have different32degrees of transferability: there is no restriction for private lands and Timber Licenses, but somerestrictions apply to Tree Farm Licenses and Forest Licenses.From an economic viewpoint, the importance of transferability is that it affects efficiency ofresource allocation. Restrictions to marketability of an asset may prevent it from being transferredto those who can use the asset most productively and apply it to its best use. Thus, it is expected thatthe value of forest tenure is related positively to the degree of transferability.A high degree of transferability also leads to more investment. Since those who are able tobid tenures away from others are always the more efficient producers in an economy, they would beexpected to use optimal inputs at a given output level. High transferability means a high possibilityto recover the potential returns on investment at any time. On the contrary, if transferability istruncated, holders can not liquidate the tenures as they wish and will face high risk to recover thepotential returns on investment. Therefore, their incentives to invest are affected’6. Low investmentresults in a low output level. The above points can be summarised as:Proposition 4. The value of forest tenure is related positively to the degree of itstransferability. When forest tenure is highly transferable, the tenure holder has an incentive to investmore and hence gets a high level of output.3. Hypothesis and Methodology16 When the transfer of tenures is subject to official consent or approval, holders may recoupsome benefits of their investments from those to whom they sell, but the benefits will be less thanfull market value of the property because of the costs and uncertainty associated with obtaining therequired authorization.33The above discussion leads into the hypothesis of the thesis:The characteristics of forest property rights affect the value of forest land, owners’ investmentbehaviour, the productivity of lands, and thus the economic efficiency of forest landmanagement.Given this general hypothesis, the following part of this section discusses the linkage betweentheory and observation, and designs some empirical analysis frameworks to be tested.The empirical analysis of this study tests how institutions (tenures) affect inputs and outputs,rather than testing the traditional relationship between inputs and outputs by assuming that institutionsare invariant. The methodology which will be used quantifies the relationship between dependentvariables, observed economic indicators which reflect forest land value and management andindependent variables, tenure systems, and other factors. Cross-section regression analysis wifi beapplied.3.1 Econometric Model to Test the Influence of the Form of Property Rights on the Value of ForestPropertyThere are two ways to test empirically the effect of tenure on the value of property rights.The first, more direct method, is to analyze the market transaction value of forest lands underdifferent tenures. It can be hypothesized that the expected future net revenue of a tract of forest landis affected by (1) its forest cover characteristics such as species, timber volume and size of trees; (2)34its natural attributes such as size, soil quality, topography which determine the natural productivityof land; (3) its location and distance to markets; and (4) the nature of property rights over it.However, the regulatory environment, such as the sustained yield policy, can modify theproperty rights. In particular, the allowable cut effect’7 can affect the value of forest land (Pearse1965). Under this regulatory environment, potential buyers can afford to pay a higher price for a tractof forest land if they can add it to a sustained yield forest because this will enable them to increasethe regulated harvests on their other lands. This implies that private lands added to Tree FarmLicense as “Schedule A” lands are worth more than private lands alone because of their impact onAAC from “Schedule B” lands. This can be expressed as a hedonic equation’8:(12) P = fCC,, L, Cf, T,, ACE)where P = per hectare market value of a tract of forest landC, = natural attributes of land (site index, size, etc.)L = locationCf = forest characteristics (volume, species composition, etc.)T, = tenure typeACE = allowable cut effect‘ The allowable cut effect was first described and formally analyzed by Schweitzer et at. (1972)although it was realized at least as early as 1965 by Pearse (1965). The use of allowable cut effectin allocating investment in timber production was rigorously analyzed by Binlcley (1980).It should be noted that taxation can affect land value directly. It can be treated in two ways.One is to treat it as an independent variable and explicitly put it in the right-hand-side of equation(12). As taxation also represents a characteristic of tenure (economic benefit conferred to the holder),it can also be explicitly ignored (and its effect will be caught in the tenure variable). The secondmethod is used in this paper for simplicity.35Since specification of the functional form of a hedonic equation is usually arbitrary, the BoxCox transformation technique or the maximum likelihood method can be used to find the functionalformulation. The regression results will estimate the contributions to forest land value of thecharacteristics of the land, forest, and tenure type.It should be noted that the concept of land value used in this study is not the same as thatused by forest land appraisers in B.C. The latter is in fact a “bare land value”, i.e., the value of theland after the mature timber is removed. It is conventional for the appraisers to estimate the valueof a forest property by assessing the value of the merchantable forest first and then, according to thetimber growing capacity of the land, assessing its “residual land value” or so-called “land value”. Incontrast, the land value used in this thesis corresponds to the economic definition of an asset’s value,i.e., its discounted future net income of land. For example, if a land tenure has only a duration of20 years, its holder may only appreciate the income derived from the land in this 20-year period.This is likely to be less than the value under a tenure with a 100 year term. In this simple case, thedifference in land value can be attributed to one of the characteristics of tenure, duration.The application of this land valuation method in B.C. is conspicuously limited. Althoughprivate lands are heavily traded in B.C., Crown lands under tenure arrangements other than TimberLicenses are not. When transactions of Crown lands do take place, the prices are usually blurred bymill and chip transaction arrangements. In other words, the dependent variable, market price (P), forforest tenures other than private forest lands and Timber Licenses is not observable. Therefore thismethod is only applied to private lands and Timber Licenses in Chapter V.The second, indirect method, tries to find price information from stock markets. It is widely36accepted that the stock market is relatively efficient (commonly referred to weak-form efficiency inthe context of financial economics) (Malkiel 1990; Fama 1991). Whenever there is a firm-specificevent or regulation change that affects a firm, it will reflect on the stock price of the firm or that ofa series of firms. Under this assumption, an “event study” is carried out to examine the impact ofa forest policy change in British Columbia, involving cancellation of a percentage of the allowableannual cut attached to timber companies under certain forms of license, in an attempt to quantify thevalue of these cutting rights, or of the lands withdrawn. The results are reported in Chapter IV.In summary, a hedonic study will be applied to frequently traded tenures and an event studywill be used to examine thinly traded tenures. The hedonic study will reveal the total value of forestlands, while the event study only examines the marginal value of forest lands since the AAC takebackwas only a fraction of the total AAC committed to companies. Moreover, the hedonic study is alsomore intuitive than the event study.3.2 Econometric Model to Test the Effect of Forest Property Rights on Silvicuittiral InvestmentThe effect of forest tenure on silvicultural investment can be tested by using an equationsimilar to equation (12):(13) I = f(C1, L, Cf. C, T)where I = per hectare silvicultural investment on a tract of forest landC =producer’s characteristicsIt is necessary to clarify the role of property in silvicultural investment decisions. Property37rights provide a framework of incentives and constraints within which their holders operate and makedecisions. All else being equal, the characteristics of tenure determine the future returns tenureholders receive, if any, and how much they will invest. Consider two types of tenures. One is secureand lasts in perpetuity; another has a 25-year duration, renewable on an “evergreen” basis, and itssecurity is in doubt. The holders of the former tenure are likely to invest more than their counterpartssince they have more certain rights to future crops. In this simple case, the difference in silviculturalinvestment, if any, could be attributed to a characteristic of the tenure, security.Another example is area-based tenure versus volume-based tenure. Since the main purposeof volume-based tenures is to allow the government to maintain flexibility in forest resource use, theycannot be an effective instrument for attracting private investment. The holders of volume-basedtenures are unlikely to return to the same areas in which they have invested, and therefore cannotexpect to recapture the benefits of investments in the subsequent crops. They would invest less thanthe holders of area-based tenure who could recapture some of the benefits of their silviculturalactivities. Again, all else being equal, the difference in silviculture investment in these two types oftenure can be attributed to whether the tenure is area-based or volume-based.Three other influences on silvicultural investment behaviour must be noted. One is mandatorysilvicultural requirements. All major forest tenure holders in B.C. are required to ensure the successof reforestation activities (trees must achieve free to grow status) within a certain period of time.However, such a requirement must be flexible enough to allow field foresters to cope with varioussite classes and other factors. Furthermore, regulation does not specify how much tenure holdersshould invest and what kind of activity they should do. As long as they are able to meet therequirement and avoid penalties, they can use the least cost method if they do not expect to capture38the future return. Otherwise, if they have the incentive to invest, they can adopt whatever methodthat yields them the greatest benefit relative to cost, which often call for investing more. Therefore,if the characteristics of tenure are significantly different, their implications for investment behaviourcan be revealed.Another influence is the allowable cut effect. Silvicultural activities undertaken beyondcontractual requirements may result, through the allowable cut effect, in immediate increases in thevolume of mature timber that licensees may harvest (Luckert and Haley 1991). However, theallowable cut effect applies to all area-based tenures. Therefore, the dummy variables used for area-based tenures (see Chapter VI) capture the allowable cut effect as well.Finally, the lack of an open timber market in B.C. is another factor which could affectlicensees’ decisions. It has been pointed out that the constraints on the timber market (e.g., partialabsence of competitive stumpage, log and chip markets; log export restrictions) lead forest tenureholders to transfer some economic rents from timber growing and harvesting activities to forestproducts manufacturing (Luckert and Haley 1991). This issue, too, is controversial. Even if it is true,it happens in all kinds of tenure and thus can be seen as a residual factor that need not be explicitlyconsidered in this study.The empirical study on the effect of tenure on silvicultural investment will be addressed inChapter VI.3.3 Econometric Model to Test the Effect of Tenure on Outputs and Forest Practice39Given the revealed amounts of silvicultural investment, the next step looks at variables thatdescribe forest growth (forest coverage, absence of not-satisfactorily restocked land, survival rate ofplantations, etc.). In the same site, the productivity of forest lands under tenure forms that provideincentives for more investment can be expected to be higher than that on other forest lands.Therefore it can be hypothesized that these growth related factors (Y) are a function of silviculturalinvestment (I), time after standing timber is harvested (Ta), and all the factors in equation (13):(14) Y1 = f(I, C1, L, Cf. C, T, T)A challenge for using this model in forestry is to find good output indicators. An answer tothis challenge and the empirical results, are reported in Chapter VII.40IV. An Event Study of the Effect of Forest Tenure on Land Value1. IntroductionOn September 15, 1987, a provincial-wide “new forest policy” was announced in BritishColumbia. The new forest policy: (1) shifted the responsibility for silviculture from the governmentto private companies; (2) transferred immediately five percent of the AAC from all replaceablelicenses (Tree Farm License, Forest License, Timber Sale License, and Timber Sale HarvestingLicense), with another five percent to be taken upon renewal of the licenses (to provide more scopefor the Small Business Forest Enterprise Program); and (3) increased stumpage and other forestcharges from $580 million to an estimated $680 million annually.This new forest policy was one of the most significant forest policy changes since the 1976Pearse Royal Commission and the subsequent Forest Act in 1978. The media intensively coveredit and reactions were mixed. Small forest operators generally welcomed the change. Companies thatwere dependent on licenses to Crown forests were distressed, even “very dismayed” (Financial Post1987). According to newspapers, the stock prices of publicly traded B.C. forest firms fell after theannouncement (Financial Post 1987; Vancouver Sun 1987).The fall of B .C. forest products firms’s stock price may have been related to the objectivesof the new policy. The new policy intended to: (1) rebuild, through small business, the jobs that wereeliminated by big forest products finns during the 1981-1985 recession; (2) generate a fair return tothe provincial government and cut government spending; and (3) replace the 15 percent countervailingduty imposed by the US. In pursuing the first objective, timber companies lost 5 percent of their41previously “secured” rights to Crown timbers. The second and third objectives increased the cost tothe forest industry. This policy was regarded as an unfavourable development, especially for the bigfirms.The opposition to this policy change cited weakened investor confidence and industrycompetitiveness. If investors lose their confidence in B.C. ‘s forestry industry, it becomes difficultfor the industry to attract capital. Without enough capital investment, the industry could lose itscompetitiveness and reduce employment still further. Since the forest industry is the predominantsector in the province, any negative changes in the industry would have a major impact on theprovincial economy. How significantly was investor confidence affected by the new forest policy?The answer lies in the performance of the shares in forest products firms, around the policyannouncement date.This chapter investigates the attitude of investors towards the new forest policy by determiningif the policy change is indeed a negative event from the perspective of shareholders, i.e., if thechanges in stock prices were abnormal, after accounting for risk and market-wide effects. Moreimportantly, this chapter catches the unique opportunity of the AAC takeback to study the value ofthe forest tenures, which, as discussed in Chapter ifi, are rarely traded. The results will provide anindication of the relative perceived values of forest lands held by private companies through foresttenures.Because of the imposition of the 15 percent export tax on Canadian softwood lumber exportsto the United States on December 30, 1986 and the declared intention of the B.C. government toreplace this tariff by increasing stumpage, it is plausible that the market would have already reacted42somewhat to the increase in stumpage before the arrival of the new policy. In fact, the formerPremier announced that the government was undertaking a review of its stumpage system inSeptember, 1986 (Vancouver Sun 1986), roughly one year prior to the new policy announcement.Moreover, the increase in stumpage of 100 million is very small relative to market capitalization offirms in this study. Therefore the results of this chapter cannot be interpreted to be the effect of thestumpage bike (Financial Post 1987).The next section reviews the methodology of event study, followed by a description ofhypotheses, data and results in Section 3. Section 4 interprets the results and discusses policyimplication.2. MethodologyEvent-study methodology has been used by financial economists to determine the impact ofa specific financial decision on shareholder returns (e.g., Desai and Stover 1985; Zinkhan 1988), orto analyze the impact of a variety of regulations and regulation changes on the expected profits offirms (e.g., Schwert 1981; Binder 1985a, 1985b; Boardman, Vertinsky and Whistler 1992). Althoughthe method can be applied to a single firm, event studies typically focus on a group of firms thatexperience similar events.There are two distinctive cases when several firms are considered. In the first, each firmexperiences a series of events, but the event periods are specific to the firm and bear no relation tothe events of any other firms in the sample except that the events are of the same kind (mergers,takeovers, etc.). The second case is one of identical events across a sample of firms, e.g., regulatory43changes or industry-wide shocks. To simplify the following discussion, these two cases will beidentified as non-regulatory and regulatory events.The event study methodology, along with the underlying assumptions, varies in each case.However, the methodologies can be classified into two basic approaches. The older, more common,method is called residual analysis. It has mainly been applied to non-regulatory events, but itsapplication to regulatory events has grown. The second method is a multiple regression analysis,known as “covariance method” (Boardman, Vertinsky and Whistler 1992). It is a “one-step” methodin which the intercept and risk parameters of the market model, and the shift in the mean return dueto an event, are estimated simultaneously by multiple regression analysis. It has been usedexclusively in regulatory events. The basic ideas of these two methods and their advantages anddisadvantages are reviewed in the following paragraphs.Suppose that there is only one event. At a single-firm level, the residual analysis partitionsthe data into two periods, uses the “non-event” period data to establish a “benchmark” for what isexpected to happen in the “event period” in the absence of event. The remaining data (with event)are then used to estimate “abnormal” and “cumulative abnormal” returns by comparing what actuallyhappened to the benchmark scenario. To set a benchmark, the capital asset pricing model (CAPM)developed by Sharpe (1964) and Lintner (1965) is often used to control for risk and market-wideeffects. According to CAPM, a security’s expected return is a positive function of that proportionof total risk that investors cannot divert (commonly known as systematic risk). The market modelshown below is generally used to estimate the parameters of the capital asset pricing model:(15) R. = c, + 3, + Etwhere44Rd = the rate of return for stock i on day t;= the rate of return on the market portfolio on day t;a1, f3 = regression parameters.= a random disturbance term, assumed to be normally distributed as N(O,1), independentof the explanatory variable Rmt.After the regression parameters are estimated by using non-event period (0, T0) data, dailyabnormal returns for security i can be calculated for each day of the event period’9:(16) AR, = = Rd - t>T0The cumulative abnormal returns, CARs can be constructed as either(17) CARd= E A1?r>TOor(18) CARI = H(1+AR11 ) —lt>ToThe hypothesis test statistic for CAR,=O is a familiar t statistic when the non-event period islarge (so that CARI has a normal distribution). The variance of CARd is assumed, with somemethods of adjustment20,to be the same as that of the non-event period.‘ The AR and CAR used in this chapter are all random variables, not true variables.20 See Collins and Dent (1985), Dann and James (1982), Desai and Stover (1985), Theil (1971),Boardman, Vertinsky and Whistler (1992) for different adjustment methods in residual analysis andportfolio approach.45Applying residual analysis to a set of (say, N) firms involves aggregation of CARft andcomputation of the relevant variance. After estimating all parameters of the N firms during the non-event period and computing the CAR1, the mean CARL of the N firms is calculated as weightedaverage of CAR. Computation of the variance of CARS is based on either a weighted averagevariance of each firm, or the residual variance of the weighted portfolio of the N firms in the non-event period (Collins and Dent 1984; Desai and Stover 1985). In each case, the weight can be equalor unequal, depending on the assumption made (Dann and James 1982).Note that the estimation implicitly assumes that there is no contemporaneous cross-correlationamong equations (firms), which is probably not a valid assumption for industry-wide regulatoryevents. Also, when equal weight is used to get the variance of CARE, it is assumed that all firms haveequal residual variance. This is also probably inappropriate in most empirical settings.A variant of residual analysis is the portfolio approach, which uses an appropriate weightmechanism to form a portfolio (Thompson 1985), and estimates the portfolio analogously to a singlefirm. It is a simple and effective way to conduct event study. The portfolio will have parametersequal to the weighted average of individual security parameters. Thompson (1985) points out thathypothesis testing can be carried out with the portfolio residual variance estimate; moreover, thevariance estimate is consistent and enables asymptotically (as T0 goes to infinity) valid inferences tobe drawn about the true population parameters, even if the portfolio weighting scheme is based onan invalid covariance matrix assumption. Thus, the use of a simple average in forming a portfoliois growing in popularity. This approach avoids the computational burden of searching for appropriateweight, such as the inverse of an individual firm’s residual variance.46Since residual analysis and its portfolio approach variant have the appealing characteristicsof simplicity and usefulness, they have been used increasingly to study regulatory events (e.g., Dannand James 1982; Boardman, Vertinsky and Whistler 1992). However, their use could have seriousconsequences if the underlying assumptions of equal variances across firms and no contemporaneouscorrelations among equations are invalid (Collins and Dent 1984; Binder 1985a, b). Theseshortcomings of residual analysis can be overcome in multiple regression analysis.A multiple regression analysis begins by parameterizing the abnormal returns ‘y in theindividual return equations:(19)using the dummy variable equals one during the event periods and zero otherwise. }1k is arandom error which is independent of R and has a normal distribution of N(O, 1). Note the identityof in equation (15), and p and + Pil in equation (19) during the no event period and the eventperiod, respectively. In equation (19), y1 measures the average abnormal return for firm i during theevent period.When the explanatory variables in the return-generating process are the same for each of theN firms the multiple equations below can be estimated jointly as a seemingly unrelated regressionmodel (Zeiiner 1962; Theil 1971):R1 = a1 + r1 Rmt + yjD1 + Pit(20) R2 = 2 + Rmt + y2D + P2tRNt = a + N rnt + “(lDNt + PNt47This approach allows that individual abnormal returns and residual variances differ acrossfirms. It also incorporates the cases where the contemporaneous covariances of the disturbancesacross equations E(p, lit) are non-zero, whereas the non-contemporaneous covariances hit-k) allequal zero. Note that the multiple regression analysis gains no efficiency in estimating coefficientsand the residual variances, producing estimates which are identical to those obtained from OLSestimation of the individual equation (Theil 1971, Chapter 7). The advantage of this approach is inhypothesis testing since heteroscedasticity across equations and contemporaneous dependence of thedisturbances are explicitly incorporated into the hypothesis tests. This technique avoids the statisticalproblems encountered in the application of residual analysis in regulatory events.Three hypotheses are of primary interest in event studies, especially studies of regulatorychange. The first (H1) is that the sum of the abnormal returns during the event period across the Nequations equals zero (i.e., Yy1 = 0). The second (H2) is that all abnormal returns during the eventperiod equal zero (i.e., y = 0, for all i). The third (H3) is that some abnormal returns during the eventperiod equal zero (i.e., ‘ = 0, for some i). Tests of H2 and H3 will be more powerful than tests of H1if an event affects the sample firms but the abnormal returns differ in sign. These joint hypothesistests are of special importance in regulatory events since there are good reasons to believe thatregulation benefits some firms and hurts others.Although tests of H1 and H3 in the residual analysis and multiple regression analysis are wellspecified, using the portfolio approach to test this hypothesis is preferred because it is both correctlyspecified and economically and computationally simpler (Thompson 1985). Tests of H2 and H3 cannot only be done by using residual analysis, but also by using portfolio approach if one candistinguish the firms who will benefit from those who wifi be hurt. The following analysis employs48both the portfolio approach and multiple regression analysis.3. Hypotheses, Data and ResultsAs discussed earlier, there are two main components of the new forest policy left to address.They have different impacts on big and small firms. Small firms which can bid for the SmallBusiness Forest Enterprise Programs have access to more timber under the new policy, but do notbear the responsibility for silviculture. Clearly they gain from the policy change. The big firmswhich cannot bid for the Small Business Forest Enterprise Programs lose five percent of their AACunder forest tenures. At the same time, they must shoulder the silvicultural costs. These companiesseem to lose as a result of the new policy. Therefore it is expected that the new policy should havea negative impact on the share prices of big forest products firms. Therefore the main null hypothesisof this chapter is that the aggregate of the abnormal returns to big firms during the event periodequals zero. A rejection of this hypothesis would mean that shareholders of the big firms in theindustry suffer from the policy change. As none of the small firms are listed on the stock market,this study does not draw any conclusions about them.The second null hypothesis is that the abnormal return for every big firm equals zero. Theeffect of regulation may not be even among firms since the value of the stock is a function of boththe content of the regulation and the circumstances of each finn. While the forest policy may harmfirms heavily dependent on Crown timber obtained under the affected tenures, it will benefit thosewho own rights to close substitutes, such as private lands and Timber Licenses. Firms that havediversified themselves by operating in other provinces would be in a better position than those whohave all of their forest operations in British Columbia. In addition, because B.C. is such a big player49in the Canadian forest industry, the forest policy may also positively affect firms that do not operatein B.C., but face the same major market (the US market). Thus, a group of non-B.C. firms is alsoincluded in this study.Initially the sample consisted of fifteen firms that had operations in B.C. and six non-B.C.firms, which were on the TSE I Western Data Base of the Toronto Stock Exchange. Four of the B.C.firms were dropped because they were involved in mergers within a year prior to the event date. Theeleven firms used in this study are: Canfor Corporation, Canadian Pacific Forest Products Limited,Crestbrook Forest Industries Limited, Doman Industries Limited, International Forest ProductsLimited, Macmillan Bloedel Limited, Slocan Forest Products Limited, Scott Paper Limited, Weidwoodof Canada Limited, Westar Timber Limited and West Fraser Timber Co. Limited. The non-B.C.group consists of Abitibi-Price Inc., Cascades Inc., Consolidated Bathurst Inc., Domtar Inc., DonohueInc., and Tembec Inc..The hypotheses are first tested using the portfolio approach. Four portfolios are formed inlight of the characteristics of each firm and the possible impacts of the policy on them. In additionto the portfolios of B.C. firms and non-B.C. firms, there are two other portfolios which are furtherderived from the portfolio of B.C. firms:Less-diversified B.C. Companies. These companies own small quantities of timber land, holdsmall Timber Licenses and operate mainly in B.C. They should be hit hardest by the forest policychanges. In this study seven companies are used. They are Crestbrook Forest Industries, DomanIndustries, International Forest Products, Slocan, Weldwood, Westar Timber and West Fraser Timber.Table 3 shows these firms on average only harvested 2.2 percent of their timber from their private50Table 3. Harvesting Volume Distribution of Forest Products Firms in 1987Firms Private & Timber Tree Farm License Other TenuresaLicense (%) (%) (%)Less-diversified B.C. 2.2 12.5 85.8Crestbrook forest Industries 1.8 8.4 89.9Doman Industries 0.0 0.0 100.0International Forest Products 7.0 0.0 93.0Slocan 0.0 12.7 87.3Weldwood 4.3 19.4 76.3Westar Timber 2.1 46.9 51.0West Fraser Timber 0.0 0.0 100Diversified B.C. 31.3 58.3 10.4Canfor Corporationb 29.2 60.5 10.3Canadian Pacific Forest Products 38.2 31.2 30.6Macmillan Bloedel 51.3 46.2 2.5Scott Papere 2.4 97.6 0.0B.C. firms 12.4 29.2 58.5Sources: Timber Harvesting Branch, Ministry of Forests, British Columbia.a Forest Licenses, Timber Sale Harvesting Licenses, Major Timber Sale Licenses and others.b Figure of 1992 estimated by the company.Because Scott Paper has mainly operated outside of B.C., it is classified as a diversified B .C. firm(although it neither owns much private industrial forest lands, nor holds lots of Crown timber throughTimber License in B.C.).51lands and Timber Licenses in 198721; the rest came from Tree Farm Licenses and Forest Licenses,for which they have to lose 5 percent of AAC under the new policy. Henceforth this group isreferred to as “less-diversified B.C.”Diversified B.C. Companies. These companies are the large integrated forest productscompanies that own most of the private industrial forest lands and Timber Licenses in B.C., and havediversified themselves outside of the province (or have actually come from other parts of Canada).They are subject to conflicting effects. Reduced access to Crown timber will be harmful, butincreased prices for their own timber and the better position that results from operating outside of theprovince will be beneficial. Four companies (Canfor, Canadian Pacific Forest Products, MacmillanBloedel and Scott Paper) are used in this study. These firms on average harvested 31.3 percenttimber from their private lands and Timber Licenses in 1987 (Table 3). Henceforth this group isreferred to as “diversified B.C.”The market model during the non-event period is then estimated separately for these fourportfolios. The daily return to each portfolio is calculated as the equally weighted daily return to eachsecurity in the portfolio. The TSE 300 price return index is used as the market index22:21 Ideally, I would want to examine the sources of timber for the mills of these firms. Forestproducts firms in B.C. purchase some timber through the open market (e.g., Vancouver log market)and through contracts. Because the relevant data cannot be found, the distribution of timber harvestedunder different tenures is used as a substitute for the source of timber supply. The implicit assumptionis that all of the timber harvested by a firm will go to the finn’s mills, and that all of the firms havethe same proportion of timber that comes from the open market and contracts. In reality thisassumption is not strictly valid. Thus the classification of the less-diversified B.C. and diversified B.C.groups is loose in terms of source of timber supply.22 Another form of capital asset pricing model is:-Rft = a + [R - R]+where Rf is a risk-free rate (say, a one-month T-bill rate). This model is often used in monthly andannual studies, where Rft changes. When Rft (and f3) is constant, this model will reduce to equation(21). This model has not been used here since the day-to-day Rft is not available and Rft (and )52(21)p={a,b,c,n};t={1...T0}Where p = a, B.C. forest products firms;= b, less-diversified B.C. forest products firms;= c, diversified B.C. forest products firms.= n, non-B.C. forest products firms.Equation (21) is estimated by OLS using 147 observations, beginning on February 2, 1987and ending on September 1, 1987, two weeks prior to the announcement date. Observations priorto 1987 are excluded to eliminate the impact of the 15 percent countervailing duty which wasthreatened by the US in the middle of 1986, and subsequently imposed on December 30 of 1986.Observations in January of 1987 are also dropped to avoid the possible “January effect”. The resultsare reported in Table 423 24A number of diagnostic tests are performed to assess the suitability of these equations. Thefits, as measured by theR2-adjusted, are quite high. The hypothesis for normality among residualscannot be rejected for all of the four portfolios. Furthermore, the tests for heteroscedasticity (B-P-Jtests and ARCH tests) reveal that the models are appropriate. However, Durbin-Watson statisticshardly changes during the 147-day estimation period used in this study.23 Equation (21) is a “seemingly unrelated regressions” model. The correlation ofcontemporaneous residuals across equations is expected. However, since each equation has the sameexplanatory variables OLS provides efficient parameter estimates.24(21) has also been estimated by including one period lead and lag of the market returnindex to control the thin trading (Scholes and Williams 1977). Since all of the results are similar tothese reported here and the coefficients of the lead and lag of market return index are not significantat the 10 percent level, the lead and the lag are dropped.53Table 4. Parameter Estimates for Capital Asset Pricing ModelsPortfolioExplanatoryVariables B.C. Firms Less-diversified B.C. Diversified B.C. Non-B.C.Rmt 1.2251** 1.2806** 1.1720** 0.9094**(9.7596) (8.1736) (7.5045) (8.0942)Constant 0.0008 0.0005 0.0009 -0.0010(1.0160) (0.4672) (0.8781) (-1.2331)R2-adjusted 0.3923 0.3107 0.2748 0.3065D.W 1.8299 1.9527 1.9748 1.6523Observation 147 147 147 147t statistics in parentheses.** Significant at the 5 percent level.indicate that the null hypotheses of no serial correlation of estimated residuals can be rejected at the5 percent level for three of the four portfolios25.The cumulative abnormal returns for the four portfolios are then computed using an equationanalogous to equation (18), where the subscript i is replaced by p for portfolio for a 23-day interval,starting at September 2 (day 8)26. For computational purposes it is useful to note that:25 To control the autocorrelation of the residual, autocorrelation regressions are run for equation(28). The follow-up prediction uses the estimated serial correlation coefficient RhO. However, theprediction results are not significantly different from those reported in this chapter. These results arenot surprising because most estimated RHOs are not significantly different from zero at the 10 level.There is little agreement in the literature regarding when the event “window” should start andfor how long it should last. Therefore a trial-and-error method is often used to choose the startingdate. Desai and Stover (1985) start the window at -20 (20 days before event); Dann and James (1982)54(22) CAR = CAR1 + AR + AR1, CAR1Assuming no serial correlation and that the null hypothesis of no effect is true, the varianceof CAR can be estimated as (Boardman, Vertinsky and Whistler 1992):(23) Var(CARj) = Var(CAR1+ Var () + Var(s1)Var(CAR1)where Var(s) is the estimated residual variance of the no-event period. The CARs and test statisticsare reported in Table 5; the daily CARs are presented in Figure 1.The CARs for the 23-day event interval for all B.C. firms are negative but not statisticallysignificant at the 10% level. The negative sign of CARs prior to the event date indicates thatinformation might have been leaked to investors. Also, the sign of CARs following the event datemeans that new information was absorbed by investors as negative. However, the CAR reboundsback to its pre-event level at October 5 (day 14).The CAR curves for less-diversified B.C. and diversified B.C. have similar shape and theCARs are all not significantly different from zero at the 10% level. However, the CARs are quitedifferent in sign. The less-diversified B.C. group has negative CARs during the 23-day event interval,but the CARs for diversified B.C. have positive signs in 10 days of the 23-day event interval. TheCARs for non-B.C. firms follow the same pattern as those of diversified B.C. group. These findingsindicate that the new forest policy had a minor impact on diversified B.C. and non-B.C. forestproducts firms. In short, the results show: (1) investors perceive the new forest policy as astart at -10. Zinkhan (1988) and Boardman, Vertinsky and Whistler (1992) use -5 as the starting day.Equation (29) is estimated by using day -10, -8, -6, -4, -2, -1 and 0 as the starting date. The resultsdoes not significantly differ with those reported here.55Table5.PerformanceofStockaroundtheAnnouncement Dateof ForestPolicyChangeB.C.FirmsLess-diversifiedB.C.DiversifiedB.C.Non-B.C.dayPercentPercentt-StatisticsPercentPercentt-statisticsPercentPercentt-statisticsPercentPercentt-statisticsofARof CARof CARof ARofCARof CARofARofCARofARofCARofCARofCAR-8-1.944-1.944-1.601-1.943-1.943-1.522-1.005-1.005-0.790-0.082-0.082-0.090-70.107-1.839-1.272-0.089-2.030-1.1251.001-0.014-0.008-0.115-0.198-0.153-60.328-1.517-0.8571.630-0.433-0.1960.7980.7840.3560.9210.7220.455-5-0.380-1.891-0.925-0.952-1.381-0.5410.5791.3670.537-0.6340.0830.045-4-0.514-2.396-1.048-0.556-1.929-0.6760.1011.4700.517-0.357-0.274-0.134-3-0.038-2.433-0.9710.231-1.703-0.545-0.7330.7260.233-1.563-1.832-0.818-20.406-2.037-0.7530.941-0.778-0.2300.3931.1220.3332.0800.2090.086-1-0.617-2.641-0.913-0.850-1.621-0.4490.5141.6410.4600.9471.1570.4470-0.290-2.923-0.9530.283-1.342-0.351-0.4981.1350.297-0.1551.0000.3641-0.947-3.843-1.188-0.723-2.056-0.509-0.5010.6280.156-0.5400.4550.1572-0.529-4.352-1.283-0.521-2.566-0.606-0.885-0.263-0.062-0.1920.2620.0863-0.042-4.392-1.240-0.170-2.731-0.6170.226-0.038-0.009-0.409-0.148-0.047UiVI56Table5.continuedB.C.FirmsLess-diversifiedB.C.DiversffiedB.C.Non-B.C.dayPercentPercentt-StatisticsPercentPercentt-statisticsPercentPercentt-statisticsPercentPercentt-statisticsofARofCARof CARofARofCARofCARofARofCARofCARofARofCARofCAR4-0.099-4.487-1.2170.141-2.594-0.563-0.187-0.225-0.0490.1560.0080.0025-0.127-4.608-1.204-0.416-2.999-0.6280.191-0.035-0.007-0.964-0.957-0.2796-0.204-4.803-1.2120.174-2.830-0.572-0.791-0.825-0.1670.300-0.659-0.18670.002-4.800-1.1730.402-2.440-0.4780.098-0.728-0.143-0.474-1.130-0.3098-0.273-5.061-1.200-0.743-3.164-0.601-0.268-0.994-0.189-0.298-1.425-0.3789-0.037-5.096-1.174-0.951-4.085-0.7540.272-0.724-0.1340.636-0.798-0.20610-0.250-5.333-1.1960.296-3.801-0.683-0.823-1.541-0.2780.404-0.397-0.10011-0.725-6.020-1.3 16-0.503-4.285-0.750-0.163-1.702-0.2990.078-0.320-0.078121.374-4.729-1.0092.396-1.991-0.3400.695-1.019-0.1771.0760.7530.179131.914-2.905-0.6051.131-0.883-0.1472.4041.3610.2280.9911.7510.408140.387-2.529-0.5160.8910.0000.0000.6171.9870.3251.7813.5640.81257Figure 1. Stock-Market Performance around the Announcement Date of Forest Policy Change4.:—/,__:-1B.C. Firms Less-diversified B.C. Date___ Diversified B.C. — Non-B.C.seemingly negative, but not statistically significant event for B.C. forest products firms as a whole;(2) the less-diversffied B.C. group is impacted slightly, but the diversified B.C. group and non-B.C.firms have no significant gains or losses (Figure 1).The above analysis is conducted by using multiple regression analysis as well. A system of11 equations for B.C. firms were run. The Breusch-Pagan lagarange multiplier test statistic for thediagonal covariance matrix for these equations is 239.80. Based on this statistic, the null hypothesisof no contemporaneous correlation across equations is rejected at the 5 percent level. Thus, themultiple regression method gains efficiency in hypothesis testing by accounting the contemporaneous58dependence across equations. The results, however, are the same as those of the portfolio approach.Seven of the eleven dummy variables in the 11 equations are negative, and the remaining four arepositive (Table 6). But none of them is significantly different from zero at the 10 percent level. Notsurprisingly, the null hypotheses that total abnormal return equals zero, and that all of the abnormalreturns equals zero cannot be rejected at the 5 percent level (the Wald X2 test statistics for them havethe value of 0.27 and 7.92 with 1 and 10 degrees of freedom, respectively).4. Conclusions and DiscussionThe 1987 changes in forest policy in British Columbia evidently were regarded as a negative,but not statistically significant event for B.C. forest products firms as a whole. The new policyresulted in minor losses for less-diversified B.C. firms. Diversified B.C. firms and other Canadianforest products firms which operate outside of B .C. have not experienced significant gains or losses.These conclusions seem to contradict media reports27,but they can be supported by carefulreasoning. First, the 5 percent reduction in AAC may not affect the equilibrium of B.C. timbermarkets very much because B.C. bans log export. The timber that is taken away from big forest27 The Vancouver Sun (1987) reports:“[I]nvestor’s reaction was negative to the B.C. government’s plans to reduce the annualharvests of most of the province’s forest companies and to increase the price on timber itleaves them.[Tihe Toronto Stock Exchange’s pulp and forest products sub-index, dominated by B.C.-based firms, was down 52 points — or about one percent—in early trading. The indexrecovered somewhat by mid-session and was off 39.09 points to 5,359.69.”The Financial Post (1987) reports:[Wie are dismayed. I mean, very dismayed,’ says Michael Apsey, president of theCouncil of Forest Industry of B.C., the industry associate. The dismay was shared by thestock market. Shares of major B.C. companies — all of which are currently reporting recordearnings — declined on the news of the new policy.”59Table 6. Parameter Estimates Using Multiple Regression MethodFirms Rmt Constant Dummya R2Less-diversified B.C.Crestbrook forest Industries 1.5490** 0.0016 -0.0079 0.1299(6.5453) (0.9998) (-1.3635)Doman Industries 1.7127**-0.0007 0.0162 0.1067(4.2288) (-0.2318) (1.6185)International Forest Products 0.9l86** 0.0011 -0.0102 0.0429(2.3183) (0.4089) (-1.0455)Slocan 2.0491** 0.0014 -0.0057 0.2073(6.3451) (0.8137) (-0.9685)Weldwood 0.3544 0.0004 -0.0053 0.02 14(1.5091) (0.2620) (-0.9106)Westar Timber 1.3543**-0.0010 -0.0001 0.0895(3.9339) (-0.3973) (-0.0118)West Fraser Timber 1.l794** 0.0006 -0.0023 0.1374(4.9464) (0.3814) (-0.3944)Diversified B.C.Canfor Corp 1.1031** 0.0006 0.0029 0.1299(4.8870) (0.3586) (0.5250)Canadian Pacific Forest Products l.4901** 0.0020 -0.0017 0.1537(5.3 136) (0.9979) (-0.2521)Macmillan Bloedel 1.4495** 0.0014 -0.0057 0.1978(6.0410) (0.8137) (-0.9685)Scott Paper 0.5795**-0.0001 0.0037 0.0675(3.3689) (-0.0677) (0.8820)8 Daily abnormal return.t statistics in parentheses.** Significant at the 5 percent level.60products companies will still end up in the B.C. timber markets, and those who lost their AAC canrecover at least a portion of it. Indeed, Gillespie (1991) reports that the big firms which cannot bidfor timber under Small Business Forest Enterprise Program circumvent the restriction by “surrogatebidding”, a practice where a large firm, in return for logs, allegedly provides financial backing, to asmall firm who is eligible to bid on the sale.Second, timber in B.C. may be fully priced with respect to all kinds of tenure. That is, thetimber companies pay the same price for the same kind of timber, regardless of how they get it —through tenure, private lands or the open market. This point is supported by Uhier (1991), SterlingWood Group Inc. et at. (1986) and Heaps (1988). Uhler (1991) finds that during the years includedin his study (1969 to 1984), timber pricing under forest tenures was not below competitive levels inat least three out of six forest regions in B .C. These forest regions— Vancouver, Rupert and Nelson— together account for 50 to 60 percent of the timber harvested in the province. Sterling WoodGroup Inc. et at. (1986) show that once differences in logging cost are taken into account, for standswith similar grade and species mixes the prices paid by competitive bidders under the Small BusinessForest Enterprise Programs and the prices paid by tenure holders are not significantly different. Thisis another way of saying that timber is fully priced under all different tenures. If the timber is fullypriced, then a reduction in AAC would have no effect on firm’s earnings, and therefore no effect onstock prices.Third, the net present value of the AAC reduction is small. The gross residual value of logstraded in the Vancouver log market, before stumpage or royalty, is $16.95 per cubic meter for allspecies and all grades in 1987 -1989 (CWC Canadian Western Capital Ltd. 1991). Tn the sameperiod, the average stumpage price in the Vancouver forest region is $9.30 per cubic metre.61Therefore, one can argue that the maximum net profit for the timber harvesting industry in theVancouver forest region is about $7.65 per cubic metre. This amount is also, in short-run, themaximum excess price a company would be willing to pay and the maximum excess stumpage pricethat a forest land owner can get. Given this figure, and assuming that the regional differences in thisprofit margin are negligible, a five percent AAC reduction to the seven less-diversified B.C. firms,would reduce average net earnings by $1.03 mfflion per year. This amount is not likely to have asignfficant effect on these companies, which recorded average total sales of over $400 millionannually in the 1987 to 1989 period (Price Waterhouse, various years).Lastly, as mentioned above, the market may have already reacted to the stumpage increase.Taking all of these considerations into accounts, the collection of new requirement, and especiallymandatory silviculture at the licensees’ expense does not seem to have been sufficient to detractsignificantly from investors’ valuation of B.C. forest product companies.One implication of these finding is that the marginal value of renewable forest tenures is smallor close to zero. This point is related to the argument that timber being fully priced in B.C. Whilethe value of forest tenures as a whole cannot be denied, marginal increases or decreases of tenuredforest lands do not appear to have much value to investors. Similarly, Woodbridge and Mackenzie(1992) report that “secure access to fibre” seems to be more important than “secure tenure”. In otherwords, as long as companies can get access to fibre, they are not concerned about whether the timbercomes from tenured forest lands or other lands.While the short-nm effects of the forest policy change do not seem to have harmed B.C.forest companies’s investors, the long-nm effects are less clear. On one hand, the policy change has62forced the American counterparts to drop (at least temporarily) their charge that B.C. timber issubsidized. On the other hand, the companies must pay more. The latter may not be a seriousproblem in times of economic prosperity, but it may prove to be a heavy burden in times of recession.The results of this chapter should be interpreted with caution. This study only measures themarginal effects of forest policy change, concentrating on the AAC reduction and the shift ofsilvicultural responsibility. Although the marginal AAC reduction has small effect, the effect of largeAAC takeback could be very significant. This study provides no evidence on this issue. Thereforeone cannot infer, based on the fmdings of this chapter, that another 5 or 10 percent AAC reductionwill have little impact on investment in the B.C. forest industry. Furthermore, the scenario of the1987 AAC reduction is different from that of other recent reductions in AAC. In the latter case, thetimber has became totally unavailable for the industry. Any further AAC cuts or imposition of coststo the industry may erode investors’ confidence, which in turn could have a big impact on theindustry which drives B.C.’s economy. Policy-makers do have some room to regulate forest resourceusers without affecting investment activities in the industry, but it is necessary to understand thedynamic relationship between investment and regulations and to evaluate the effects of regulationbefore major changes in regulation are imposed.63V. The Effect of Forest Tenure on Land Value: A Hedonic Study1. IntroductionPrivate forest lands and Timber Licenses are frequently traded in B.C. Enough observationsare available to estimate a hedonic pricing model that captures the effect of forest tenure on landvalue. This chapter measures and interprets the significance of these forms of tenure for the valueof the land. As a by-product, this study also examines the determinants of forest land value in BritishColumbia, and the role of the allowable cut effect in the decisions of large and integrated firms topurchase forest lands.The literature on determinants of land value can be found in real estate (e.g., Vrooman 1978;Coulson and Robins 1987) and agriculture (e.g., Hushak and Sedr 1979; Palmquist and Danielson1989). With the exception of Armstrong (1975) and Washburn (1990), few studies of this kind arefound in forestry. Furthermore, the literature apparently contains no research on the effect ofinstitutional instruments on land value in any of these sectors.The next section describes the data for estimating equation (19). Section 3 discusses theempirical results, and conclusions based on these fmdings are presented in Section 4.2. Data28To demonstrate the effect of forest tenures on land value, cross-sectional data have been28 The data used in this chapter are available from the author.64collected on transactions of private forest lands and Timber Licenses in the period from 1987 to 1992.This six-year period is long enough to cover a whole business cycle, thereby controlling for changein macroeconomic conditions. The starting year, 1987, has been chosen because major changes inB.C. private forest legislation took place in January of that year. The study area includes the B.C.Coast (Vancouver forest region) and the Southern Interior (Kamloops and Nelson forest region),which together account for some 54 percent of the timber harvested in 199 1-1992 and at least halfof the productive forest lands in the province. Managed and unmanaged forest lands, defmed inChapter II, are treated as two separate tenures29.The data used in this study came from various sources. Information on private forest landshas been mainly provided by British Columbia Assessment Authority (BCAA), which assesses allprivate forest lands in the province and records the transactions that involve them. There were 1084transactions in private forest land province-wide during the study period. The use of thesetransactions is limited to 247 by excluding: (1) all properties that apparently do not have a value (i.e.,the values are attached to something else or have yet to be determined); (2) all properties outside thetwo study regions.Since the data from BCAA do not include information on the forest inventory, speciescomposition and potential products of each property, a mail-out survey was conducted to determine29 If the lands are outside Tree Farm Licenses, B.C. legislation allows re-classification of thesetwo categories from one to another provided that the holders pay certain tax differences. For example,if holders of managed forest lands find that their lands would be more valuable if classified asunmanaged, they can have them reclassified upon payment of the accumulated tax savings they haveenjoyed by having classified the lands as managed forest lands since 1987. Therefore, the differencebetween managed forest lands outside of Tree Farm Licenses and unmanaged forest lands is limitedto taxation and commitment for sustainable forestry practice. However, most managed forest landsused in this study are within Tree Farm Licenses. Therefore unmanaged and managed forest landsare treated as two different types of tenures.65these forest characteristics. Among the 247 properties surveyed, responses from some 115 propertieswere collected, but only 82 of these (45 from the managed forest lands and 37 from the unmanagedforest lands) are useful. The rest are excluded from the study because the owner either could not,or would not reveal all the information needed for this study. Most of the managed forest lands areincorporated in Tree Farm Licenses, but none of these 82 properties have improvements such aslogging facilities and buildings. This circumstance simplifies the data analysis, and helps avoid errorsdue to the difficulty of estimating the value of the improvements.Data on Timber Licenses were provided by the Ministry of Forests, Timber License holdersand independent appraisers. There were some 170 Timber Licenses traded between 1987 and 1992in the Coast and Southern Interior, but full sets of information have been obtained only for 24because of the confidentiality requirements of some Timber License holders. The appraised valuesof these Timber Licenses were used as market prices. This usage is appropriate since the B.C. ForestAct (Section 50.4) specifically requires that an independent current appraisal be conducted when aTimber License changes hands. Some of these Timber Licenses could be incorporated into Tree FarmLicenses as Schedule “B” lands.Table 7 describes the variables used in this study. Price per hectare is the dependent variable.While the mean price per hectare for all observations is $3115.80. There is a significant differenceamong tenures (Table 8). However, any conclusions regarding the effect of tenure on land value canbe firmly drawn only after a full analysis of the determinants of land value, since these other factorsdiffer among tenures as well. To facilitate presentation, the twenty variables are categorized into fivegroups: tenure, forest cover, natural attributes of the lands, location and others, and producer’scharacteristics.66Table 7. Variable Definitions, Sources and Statistics of Forest Land Transaction VariablesVariable Mean Standard Definition SourcesReal price of land per hectare asof December, 1992 Cs)0.3491 0.4789 Dummy: managed forest land (1if managed forest land, 0 otherwise)0.4245 0.4966 Dummy: (1 if unmanaged forestland, 0 otherwise)Estimated timber inventory (m3)Percent of Douglas firPercent of hemlock and balsamPercent of cedarPercent of products as peelerand polePercent of products as sawlogTract size (hectare)Dummy: location (1 if Interior,0 otherwise)ACCESS 1 0.5283 0.5016 Dummy: distance from a mill (1if between 32-64 1cm, 0 otherwise)ACCESS2 0.2547 0.4378 Dummy: distance from a mill (1Value3115.80Deviation4980.20PRICEPFLmPFLuVOLUMED_FIRHBCEDARPRODUCT1PRODUCT2SIZECOAST205.0528.7233.6816.8010.09188.1929.2424.8219.4520.13BC Assessment Authority(BCAA), Ministry ofForestsBCAA, Tenure holdersBCAA, Tenure holdersBCAA, AppraisersAppraisers, Tenure holdersAppraisers, Tenure holdersAppraisers, Tenure holdersAppraisers, Tenure holdersAppraisers, Tenure holdersBCAA, AppraisersBCAA, AppraisersBCAA, AppraisersBCAA, Appraisers55.85111.230 .424530.08254.290.4966if greater than 64 km, 0 otherwise)67Dummy: average slope (1 if lessthan 40 degrees, 0 otherwise)0.415 1 0.495 1 Dummy: average slope (1 if between40-60 degrees, 0 otherwise)28.04 27.26 Percent of good soil quality45.42 31.87 Percent of medium soil quality12.26 20.28 Percent of poor soil quality45.16 21.25 Number of month from transactiondate to January, 19870.9323 0.0684 Consumer Price Index(December, 1992=1)0.0862 0.0192 Risk-free interest rate (3-monthCanadian Treasury Bill rate)0.4245 0.4966 Dummy: purchaser (1 if largeforest firm; 0 otherwise)Tenure. The three types of tenures considered in this study are converted into two dummyvariables (PFLu and PFLm) for analysis. PFLu takes the value of unity if the property is anunmanaged forest land, and the value of zero otherwise. Similarly, PFLm takes the value of unityif the property is managed, and the value of zero otherwise. The Timber License is treated as thebase type; therefore, the dummy variables for managed and unmanaged forest lands are of primaryTable 7. continuedVariable MeanValue0.4340StandardDeviation0.4980Definition SourcesTOPOG1TOPOG2SOIL_GSOIL_MSOIL_PDATECPITNTPRODUCERBCAA, AppraisersBCAA, AppraisersBCAA, AppraisersBCAA, AppraisersBCAA, AppraisersBCAAStatistics CanadaStatistics CanadaMinistry of Forests68Table 8. Some Statistics on Unmanaged and Managed Forest Lands, Timber LicensesVariable—Managed Forest Lands Timber LicensesMean StandardUnmanaged Forest LandsMean Standard Mean StandardValue Deviation Value Deviation Value DeviationPRICE 3768.00 4398.30 3406.30 6193.30 1565.30 2493.00VOLUME 127.75 111.43 199.58 202.77 334.46 190.92D_FIR 40.46 34.04 28.97 25.13 10.17 17.34HE 16.87 20.52 39.49 25.18 48.71 13.58CEDAR 13.92 22.94 12.28 15.17 29.75 15.32PRODUCT1 17.57 29.33 4.11 8.41 9.75 14.41PRODUCT2 43.11 34.98 59.56 29.75 67.37 14.25SIZE 85.53 204.60 68.55 125.79 230.87 424.08COAST 0.3243 0.4746 0.4444 0.5025 0.5417 0.5090ACCESS 1 0.5676 0.5023 0.5333 0.5045 0.4583 0.5090ACCESS2 0.2162 0.4173 0.3778 0.4903 0.0833 0.2823TOPOG1 0.5135 0.5067 0.5227 0.5052 0.1250 0.3378TOPOG2 0.4054 0.4977 0.3864 0.4901 0.5000 0.5 108SOIL_G 34.18 31.74 20.38 19.28 32.92 30.07SOIL_M 44.57 35.70 51.76 29.65 34.83 27.63SOIL_P 11.30 23.75 12.02 16.92 14.21 21.04DATE 43.97 18.69 51.20 19.37 35.67 15.67PRODUCER 0.1081 0.3143 0.7556 0.4346 0.2916 0.4643Observation 37 45 2469interest in this study. The variables for both managed and unmanaged forest lands are expected tohave significant positive signs since the characteristics of private lands favour the owners more thanthose of Timber Licenses favour the license holders. The coefficients for PFLu and PFLm areexpected to be equal.Forest Cover. The average volume of timber per hectare on each property is included as avariable (VOLUME), and it is expected to have a positive sign. Four species (Douglas fir, hemlockand balsam, cedar) are singled out in this study in order to measure the effect of species compositionon land price. These species account for more than 90 percent of the timber transacted through theVancouver log market (Zhang and Binkley 1994) and 92 percent of the timber harvested on the Coastin 1991-1992 (Ministry of Forests 1993). Although these species are not so dominant in the interior,they still account for 39 percent of the timber harvested in 1991-1992 (Ministry of Forests 1993).To simplify the analysis, hemlock and balsam are treated as a single species because both have almostexactly the same quality, price, and end-use for the same grade, and in fact, the Ministry of Foreststreats them as a single species in stumpage appraisal. This would reduce the analysis to three speciesvariables — D_FIR, HB and CEDAR — which measures the percent of Douglas fir, hemlock andbalsam, and cedar, respectively, on each property. The coefficients of these variables indicate theeffect of these species on the land price per hectare, compared with the remaining species (a mixtureof spruce, pine and hardwood species). The signs of these variables are expected to be related totimber prices by species.Three forest products — peeler (and pole), sawlog, and pulpwood— are considered in thisstudy. Two potential product variables (PRODUCT1 and PRODUCT2) measure, respectively, thepercent of timber inventory that is for the purpose of producing peeler (and pole), and the percent of70timber volume that is for producing sawlogs. The coefficients of these variables indicate the effectof the average tree size or potential timber products on the land value, compared to pulpwood. It isexpected that both variables have positive signs.Natural Attributes. The size of each property in hectares is included as a variable (SIZE).The price per hectare should vary inversely with the size of the tract because the market for largetracts of private forest lands is thinner. In other words, few buyers are willing to pay the costs ofsubdividing the lands. Furthermore, harvest costs vary inversely with the per hectare inventory,although perhaps weakly. Total timber volumes being equal, the value per hectare is inverse to thesize of the parcel.The distance to market is measured as the approximate distance from the property to theclosest mill or log dump. Three categories of distance are used30. Two dummy variables(ACCESS 1 and ACCESS2) are assigned the value of unity if the property is less than 32 kilometresaway and between 32-64 kilometres away from the closest mill or log dump, respectively, and zerootherwise. The coefficients for these variables reveal the effect of these distances on land value,compared to the property that is greater than 64 kilometres away from the closest mill or log dump.Positive signs for these variables are expected, with ACCESS 1 greater than ACCESS2. Similarly,three categories of topography are incorporated in the property assessment. Two dummy variables(TOPOG1 and TOPOG2) are assigned the value of one if the tract is flat (with an average slope ofless than 40 degrees) and steep (with an average slope of between 40-65 degrees), respectively, andzero otherwise. The coefficients of these variables reveal the effect of flat and steep topography on30 Distance to the market, and topographical variables can be treated as continuous variables.However, since data from BCAA record each of them in discrete form, this study follows thispractice.71the per hectare value of a property, compared with very steep topography (with an average slope ofgreater than 65 degrees).Four categories of soil quality (good, medium, poor, inoperable and non-productive) measurethe natural productivity of the land. Three variables are used to take into account the effect of soilquality on the price of land. SOIL_Cl is a variable that measures the percentage of good soil qualityincluded in a property; SOIL._M is a variable that measures the percent of medium soil quality andSOIL.Y is a variable that measures the percent of the poor soil quality. The coefficients of thesevariables indicate the effect of good, medium and poor soil quality lands on land price compared withthe effect of inoperable and non-productive lands. All of the soil quality variables should havepositive signs, with SOIL_Cl greater than SO1LM, and SOIL_M greater than SOIL_P.Location and Others. Although three forest regions are included in this study, Kamloops andNelson are combined as a single region (Southern Interior). A dummy variable (COAST) is assignedto a value of unity for each property in the Coast; all other properties are assigned zero. COAST isa location variable, and is expected to have a positive sign since the Coast is closer to populationcentres and to markets for forest parcels.A date variable (OATh) is included to capture the time trend of land price. The monthlyCanadian Consumer Producer Index is included as a variable (CPI) to test the effectiveness of forestland as a price hedge during inflation. A variable of risk-free interest rate (INT), which takes thevalue of the 3-month canadian treasury bill rate, is added to catch the financing cost of purchasingforest lands.72Producer’s Characteristics. Finally, a dummy variable (PRODUCER) accounts for the effectof each producer’s characteristics. It takes the value of unity if the purchaser of a property is a largeintegrated forest product firm and zero otherwise. The criterion used here to distinguish large forestproduct firms from others is the holding of committed cutting rights in the province. The top 20companies31,which collectively hold more than 74 percent of the committed annual allowable cutare designated as large firms. A significant positive sign indicates that these companies are willingto pay a higher price to hold more forest lands and timber. The explanation for this result could bethe allowable cut effect (Pearse 1965) or economies of scale.3. Empirical ResultsThe functional form of the hedonic equation is selected empirically by applying the Box-Coxtechniques to the most common functional forms (linear-linear, linear-log, log-linear, and log-log).The log-log form has proven to be preferable32. The regression results are given in Table 9.31 These companies are: Macmillan Bloedel, Fletcher Challenge, Canfor, West Fraser/Enso,Weldwood, Doman, Slocan, Westar, Canadian Pacific, Weyerhaeuser, Tolko Industries, Lakeland,Crestbrook, Repap, Ainsworth, Louisiana Pacific, Carrier Lumber, Pope and Talbot, and Lignum.32 Two methods are used here to choose the function forms. Both lead to the same conclusion.The first is maximum likelihood method. Spitzer (1982) and Judge et a!. (1988) show thatmaximizing the Box-Cox likelihood function is equivalent to minimizing the residual sum of squaresfor the regression where the dependent variable is divided by its geometric mean prior totransformation. This method is used in Palmquist and Danielson (1989) and Washburn (1990). Idivided each dependent variable by its geometric mean and estimated the four functions. The residualsum of squares is 3333 for linear-linear; 216 for linear log; 2400 for log-linear and 181 for the log-log model. Thus the log-log function form which has the smallest residual sum of squares ispreferable.The second method is comparison of R2. Goldberger (1968) promotes this method. Since thefour R2’s of functions which have different dependent variables are not directly comparable,comparable measures have to be proceeded. A log-linear equation exemplifies this method. First,compute the ‘‘s, the calculated values from log-linear function; take their anti-logs, ‘‘ = antilogThese are obviously estimates of the absolutes rather than logarithmic values. Second, computethe R2 between Y1 and Y.”. This is comparable to R2’s of linear-linear and linear-log functions, which73Column 1 of Table 9 is the result of a regression without the variable of PRODUCER. Mostof the results for the explanatory variable are reasonable. Out of twenty (20) parameters estimated,sixteen (16) of them have the expected signs. Of those that have counter-intuitive signs, none isstatistically significant from zero. Among the parameters with the expected sign, nine (9) aresignificant at a 90 percent confidence level or better. The following part of this section describessome parameters in detail.Tenure. The coefficients for unmanaged and managed forest lands indicate that tenure is asignificant factor in determining the land price. These parameters are significantly different from zeroat the 90 percent confidence level. The regression results imply that the value of Timber License,expressed respectively as a percentage of the average value of managed and unmanaged private forestlands, is 22.6 percent and 34.1 percent. This means that, while all observations have a mean valueof $3115.80 per hectare, Timber Licenses have the value of only $704.04-$1063.84 per hectare.One may argue that these results are surprising, given the fact that everything else beingequal, the difference in land price between private forest lands and Timber Licenses is the residualvalue after harvesting the mature timber. However, on closer examination, the results may bereasonable. First, private forest land owners own other things such as minerals and can use the landfor non-timber purposes. In particular, unmanaged forest lands may be more valuable because theycan be used in agricultural, recreational and other uses. Managed forest lands may be more valuableare the R2 between‘and Y1. The same logic applies to log-log function. I computed the estimatedR2’s for log-linear and log-log functions as 0.2437 and 0.5463. Comparing them with the R2’s for log-linear and linear-log functions (0.2614 and 0.4806), it is evident that log-log function is the best.Therefore the log-log function has been chosen. Notice also the constant price elasticityproperty of the log-log function.74Table 9. Empirical Results of Log-Log Equation for Forest Land Values(1) (2)VariableCoefficient T-ratio Coefficient T-ratioTenurePFLu 1.4874 2.7550** 1.6447 2.9926**PFLm 1.0746 1.8599* 0.9111 1.5190Forest CoverVOLUME 0.0011 1.9900** 0.0009 1.8122**D_FIR -0.0641 -0.9410 -0.0478 -0.6947HB -0.0066 -0.0814 -0.0140 -0.1732CEDAR -0.0459 -0.6686 -0.0491-0.7191PRODUCT1 0.0579 0.3243 0.0040 0.0225PRODUCT2 -0.1181 -0.1235 -0.0244-0.0956Natural AttributeSIZE -0.0905 -0.6121-0.0557 -0.3732ACCESS1 0.7915 1.8082* 0.8996 2.0250**ACCESS2 -0.3656 -0.7411 -0.3 175 -0.6450TOPOG1 0.0625 0.1155 0.1595 0.2938TOPOG2 -0.0960 -0.1995 -0. 1083 -0.2260SOIL_C 0.0543 0.7577 0.0298 0.4060SOIL_M 0.1999 2.4334** 0. 1955 2.3898**SOIL_P 0.1511 2.9194** 0.1583 2.4448**Location and OthersCOAST 0.4065 1.0075 0.1992 0.463975Table 9. continued(1) (2)VariableCoefficient T-ratio Coefficient T-ratioDATE -1.2009 2.1528* -1.3602 2.4017**CPI 12.8220 1.8044* 13.2590 1.8732*INT 0.8263 0.7517 1.1991 1.0932INTERCEPT 14.0100 2.9194** 14.9740 3.1017**PRODUCER 0.6727 1 .3580R2 0.4408 0.4528R2 -adjusted 0.3093 0.3 161Observation 106 106since the owners enjoy the benefit of allowable cut effect (see below). Second, most Timber Licensesstudied here will not expire before 2015. This means that some returns from harvesting the maturetimber will not be captured immediately.The test for identity of the coefficients for PFLu and PFLm, the dummy variables forunmanaged forest lands and managed forest lands, indicates that they are not identical at the 20percent level (with T-ratio of 0.9034 and 84 degree of freedom). However, when the PRODUCERis added into the equation (column 2 of Table 9), their identity is rejected at 20 percent interval (withT-ratio of 1.43 17). This may be because PFLu is picking some effects that were not included in themodel, such as potential for development and non-timber uses.Forest Cover. The value of forest land is strongly related to its forest inventory. The76elasticity of the price per hectare of forest land with respect to average per-hectare volume is 0.0011.In other words, for a one percent increase (decrease) in average volume per hectare, the value offorest land per hectare increases (decreases) 0.0011 percent. Translating this information into realterms, a 2.05 cubic meter increase in timber volume (one percent of the mean timber volume) wouldresult in a land value increase of $3.43 (0.0011 percent of the mean land value per hectare)33. Therelationship between forest land value and species reveals that land value is not affected by speciescomposition. No significant difference among the potential products is found, but the productvariables do have the expected signs.Natural Attributes. Soil quality is found to be significantly related to the land value. Thesignificant positive coefficients of medium and poor soil quality indicate that the value of forest landtends to increase as the percent of medium and poor soil quality on the property increases. Thevariable for good soil quality has a positive sign, but it is not significantly different from zero. Thecoefficients for the land size variable are not significantly different from zero. But the variable signconfirms the studies in agricultural, forestry and real estate lands (e.g., Armstrong 1975; Vrooman1978; Palmquist and Danielson 1989; Washburn 1990) in which the unit land value is found to benegatively related to tract (parcel) size.The variable that measures the 0-32 km category of distance to mill is significantly differentfrom zero, but the 32-64 category km is not significantly different from the greater than 64 kmcategory. No significant relationship is found between land value and topographical variables.This result is surprising given that the average stumpage was 7-10 dollars per cubic meterduring the study period. One possible explanation is that some immature trees were reported by someowners.77Location and Other Variables. The coefficient for COAST indicates that the land value ispositively related to location, but the per hectare value of forest lands on the Coast is not significantlydifferent from that of the Southern interior once all other factors have been accounted for.The coefficient for the transaction date indicates that the land value is negatively related totransaction date, and therefore the forest land price is declining over the study period. This trend isperhaps related to the 1990-1992 recession. CPI is significantly different from zero, but INT is not.These results mean that forest land is a good price hedge during inflation, and that the movement ofinterest rate does not affect the value of forest land very much. The latter may be due to too littlemovement over the sample period.Producer’s Characteristics. The results in column 2 of Table 9 reflect the addition of thevariable of PRODUCER. This addition has little statistical relation to any of the variables other thanPFLm. The t-ratio for PFLm is reduced and the variable becomes an insignificant variable at the 10percent level with this addition. This outcome indicates that there is some collinearity between PFLmand PRODUCER. In fact, 34 out of the 45 managed forest lands are held by the large firms. Theseresults imply that, in combination with the tenure of managed forest lands, the allowable cut effecthas affected forest land value in B.C. (Pearse 1965). Large forest products firms do consider theallowable cut effect when they purchasing forest lands, especially when the lands can be added toTree Farm Licenses.4. ConclusionsThe estimates of this chapter show that the per hectare value of the Timber License is only78about 23-34 percent of that of private forest lands and that unmanaged forest lands are slightly morevaluable than managed forest lands. This finding complements to the results of the preceding chapter,which asserted that AAC based on lands held under Tree Farm Licenses and Forest Licenses havelow marginal values. More importantly, these findings could be applied to the values of lands underother tenures. For example, the characteristics of Forest Licences are not as complete as those ofTimber Licenses. If so, lands under Forest Licenses must have a lower value than private forest landsas well.The high value of private forest lands comes as no surprise, since in almost every aspect,private forest land is the most favourable form of tenure for the holder. Furthermore, the owners ofunmanaged forest lands, by avoiding the commitment of sustainable timber production practice, canput a high value on their lands for recreation, fisheries, hunting and other uses. The owners ofmanaged forest lands, on the other hand, benefit from the allowable cut effect if they hold their landin conjunction with Tree Farm License “Schedule B” lands.79VI. The Effect of Forest Tenure on Silvicultural Investment1. IntroductionInvestments in silviculture increase future wood supplies. The role that tenure plays inproviding incentives to invest in silviculture, has been the subject of many theoretical studies (e.g.,Pearse 1976, 1985, 1993). Previous empirical studies in this area are limited to the interview method(Luckert 1988; Luckert and Haley 1990).Observed differences in silvicultural expenditure patterns among tenure types may be theresult of several factors other than tenure arrangements themselves, including site quality, speciescomposition, size, location and general economic trends. Although recognizing these other factors,Luckert (1988) and Luckert and Haley (1990) deal explicitly only with tenure type and soil quality.Furthermore, these studies are limited to two types of tenures — private lands and Tree FarmLicenses— which are important, but do not represent the majority of forest tenures in BritishColumbia. There have been no other empirical studies dealing with the relationship betweensilvicultural investment and tenures in B.C. and elsewhere in Canada, although there are a few similarstudies in other sectors, notably in agriculture (e.g., Feder et al. 1988).This chapter presents and discusses an empirical study examining the effect of forest tenureon silvicultural investment. An econometric model, as expressed in equation (13), is used to controlall factors that affect silvicultural investment. The results have profound policy implications.The chapter is organized as follows. The next section describes data used to estimate the80silvicultural investment model. Section 3 discusses the empirical results, and Section 4 presentsconclusions and policy implications.2. DataThis study examines four major tenure types: private forest lands within Tree Farm Licenses,Crown lands within Tree Farm Licenses34,Timber Licenses35,and Forest Licences. Data on thesetenures have been collected in the period from September 15, 1987 to May 15, 1993. The startingdate is chosen because it marks the beginning of period of the recent changes in B.C. forestlegislation whereby all major tenure holders were given the responsibilities and expense ofsilvicultural activities36. The date with the most recent information available is chosen as the enddate. This five and half year period covers an entire business cycle (the boom years of 1988-1989and the recession year of 199 1-1992), and thus controls for changes in macroeconomic conditions.The study examines silvicultural investment in the four tenure types in three regions in B.C.As in Chapter V, the regions studied include the Coast (Vancouver forest region) and the Southerninterior (Kamloops and Nelson forest region). The observation unit of this study is a cutbiock, whichIn other words, Tree Farm Licenses excluding private forest lands and Timber Licenses.Includes Timber Licenses both within and outside Tree Farm Licenses. As indicated in Table1, roughly half of the timber harvest under Timber Licenses in the province come from TimberLicenses within Tree Farm Licenses. The distribution of Timber Licenses in the sample studied inthis thesis is unknown.36 Before that, three mechanisms had been used in silviculture: direct government silviculturalinvestment, licensees’ voluntary investment and government reimbursement of approved silviculturalinvestment undertaken by licensees. Under that scenario, the comparison of actual silviculturalinvestment among tenures is very difficult since it is hard to distinguish these different types ofinvestment. This is probably one reason that Luckert (1988) and Luckert and Haley (1990) used theinterview method rather than actual observed data.81can be seen as a homogeneous unit in terms of the natural land attributes, the location, harvestingactivity and species composition. The Ministry of Forests designates each cutbiock as a singleobligation and its licensee is required to report all pre-harvest silviculture prescription (PHSP)information. In addition, harvesting and silvicultural activities for each cutbiock (or obligation) mustbe reported every four months or on a yearly basis.Silvicultural investment. There are 28 types of silviculttiral activities, or treatments)37,designated by the Ministry of Forests. The Ministry requires licensees to report their silviculturalactivities and their subsequent expenditures38 in each Forest District39 every year. If a licenseeholds a few cutbiocks under more than one type of tenure in a district, and has initiated a silviculturalactivity on lands under all tenures within a year, the licensee is allowed to report the total costs ofThese treatments are classified into four groups:(1) regeneration — planting, replanting, fill planting, natural regeneration, mechanicalplanting, direct seeding;(2) surveys — regeneration survey, sit preparation survey and free growing survey;(3) site preparation — partial slash burning, whole cutblock burning, partial chemicaltreatment, whole cutblock chemical treatment, guard, partial treatment, mechanical treatment, roadrehabilitation;(4) stand tending — chemical vegetation control, stem clipping, crop covering, fertilization,animal brush control, manual vegetation management, infected tree removal, density control, pruning,seedling protection and spacing.38 All expenditures reported in this thesis are licensees’ alone and do not include expendituresfrom federal-provincial Forest Resource Development Agreement (FRDA). This is because FRDAI (1986 - 1990) was focused on restoring backlogs (not satisfactorily restocked lands prior to 1982),and FRDA 11(1991 - 1994) is focused on stand tending (spacing, pruning and fertilization). Theobservations (cutblocks) included in this thesis are not qualified to FRDA I, and the trees on thesecutblocks are too young to be qualified for spacing and pruning treatments. Furthermore, fertilizationonly happened on 7 of the 2311 observations studied here and cannot have a significant effect.There are 23 forest districts in the study area. They are:(1) Chilliwack, Campbell River, Duncan, Mid Coast, Port Albemi, Port McNeil, QueenCharlotte, and Squamish in the Vancouver forest region;(2) Clearwater, Kamloops, Lillooet, Merritt, Pentiction, Salmon Arm and Vernon in theKaniloops forest region;(3) Arrow, Boundary, Cranbrook, Golden, Kootenay, Invermere, and Revelstoke in the Nelsonforest region.82that silvicultural activity as a single figure, while reporting the treatment area under each cutbiockseparately. Thus the unit cost of silvicultural treatment for a specific cutbiock is unknown. Toprotect the confidentiality of the licensee, the Ministry aggregates both the costs of every silviculturaltreatment type and the size of treatment in each district every year before releasing the informationto the public.Given this situation, the investment figure of a silvicultural activity implemented on eachcutblock in this study is computed by multiplying the area of the activity undertaken by the licenseeand the average unit cost of that same activity in the same district and the same year. The figure isthen compounded or deflated to real dollar values in December of 1992 by using the CanadianConsumer Price Index. Adding up all investment figures of every silvicultural activity undertakengenerates the total real silvicultural investment for each cutblock. Therefore the total investmentfigure of a cutbiock is the accumulation of investment in all silvicultural activities undertaken on thatblock after harvesting completion to May 15, 1993. Here, it is assumed that the silvicultural industryin the study area is competitive and that conditions after logging are sufficiently similar so that alllicensees incur roughly the same costs for the same silvicultural treatment in a given district, and thatthere is no systematic difference in the quality (and therefore the cost) of silvicultural activity acrosstypes of licenses.For the present study, 2311 observations have been obtained, each pertaining to a cutbiock,and revealing all of the information needed in this study. These cutbiocks were logged during theperiod of September 1987, to December, 1988. Table 10 describes the variables used in this study.The real silvicultural investment per hectare is the dependent variable. The mean investment perhectare for all observations is $750.58, and the sample statistics for investment and other variables83Table 10. Defmitions andMeanVariableStatistics of Silvicultural Investment VariablesStandard DefmitionValue DeviationINVESTMENTPFLTLTFLBALSAMCEDARD_FIRHEMLOCKSPRUCESIZESOILGSOIL_MBGC1BGC2750.580.01990.13200.25796.27139.97 145.553413.657322.910429.686 10. 12380.81480.3872622.450.13970.33850.437514.934219.230419.221125.559536.822929.24140.32940.38850.4872Real silviculture investment per hectare as of Dec., 1992 ($)Dummy: private forest land (1 if private forest land, 0otherwiseDummy: Timber License (1 if Timber License, 0 otherwise)Dummy: Tree Farm License (1 if Tree Farm License, 0otherwise)Percent of balsam regeneratedPercent of cedar regeneratedPercent of Douglas fir regeneratedPercent of hemlock regeneratedPercent of spruce regeneratedSize of cutbiock (hectare)Dummy: soil quality (1 if good, 0 otherwise)Dummy: soil quality (1 if medium, 0 otherwise)Dummy: biogeoclimatic zone (1 if Coast Douglas Fir orCoast Western Hemlock or Montane Hemlock, 0 otherwise)0.2756 0.4469 Dummy: biogeoclimatic zone (1 if Engelman Spruce-subalpine Fir or Montane Spruce or Sub-boreal Spruce, 0otherwise)0.0884 0.2836 Dummy: biogeoclimatic zone (1 if Interior Douglas Fir, 0otherwise)BGC384Table 10. continuedVariable Mean Standard DefinitionValue DeviationCOAST 0.3981 0.4896 Dummy: location (1 if coast, 0 otherwise)DATE 56.09 11.27 Number of month from January, 1987 to last silviculturecomplete dateTNT 0.0849 0.0243 Risk-free interest rate (3-month Canadian Treasury Bill rate)PRODUCER 0.6582 0.4755 Dummy: Producer Characteristics (1 if integrated firm, 0otherwise)among the tenures is presented in Table 11.Tenure type. The four types of tenures are converted to three dummy variables (PFL, TFLand TL) for analysis. PFL takes the value of unity if the property is private forest land within TreeFarm Licenses, and zero otherwise. Siniilar1y, TFL takes the value of unity if the property is TreeFarm License ?Schedule B’ lands, and zero otherwise; TL takes the value of unity if the property isunder a Timber License,and zero otherwise. The Forest License is treated as a base type, andtherefore the dummy variables for private forest lands, Tree Farm License and Timber Licenses areof primary interest in this study.The variable for private forest lands is expected to have a significant positive sign because,as explained earlier, the characteristics of private lands favour the owner more than those of ForestLicenses. The signs of TFL and TL should also be positive, for two reasons. First, the characteristicsof both Tree Farm Licenses and Timber Licenses are either less diluted than, or at least equivalentto those of Forest Licenses. Furthermore, Tree Farm Licenses are area-based tenures which may85Table 11. Some Statistics on Private Forest Lands, Timber Licenses, Tree Farm Licenses and ForestLicensesPrivate Lands Timber License Tree Farm License Forest LicenseVariableMean Standard Mean Standard Mean Standard Mean StandardValue Deviation Value Deviation Value Deviation Value DeviationINVESTMENT 839.44 471.26 645.38 572.39 800.69 811.62 745.95 607.22BALSAM 8.5000 12.9178 11.4066 19.1222 6.6544 15.5626 4.8805 13.1600CEDAR 18.8478 27.0259 22.8525 25.2338 11.6795 19.9085 6.0455 15.2022D_FIR 26.0652 36.7897 9.9868 24.6404 5.4647 18.6050 3.9090 16.5475hEMLOCK 21.0435 21.0259 37.9442 32.9603 17.0772 27.4843 6.4831 18.0449SPRUCE 4.0435 8.0497 0.1085 27.3075 20.3758 35.3749 27.3519 38.9378SIZE 31.4652 31.7023 30.6571 27.1463 28.2926 30.6571 27.1463 27.5759SOIL_G 0.1087 0.3147 0.1013 0.3065 0.1864 0.3897 0.1031 0.3042SOIL_M 0.8696 0.3405 0.8072 0.3952 0.7671 0.4231 0.8344 0.3718BGC1 0.9783 0. 1474 0.8758 0.3303 0.4193 0.4939 0.2455 0.4305BGC2 0.0000 0.0000 0.0228 0.1499 0.2662 0.4424 0.3457 0.4758BGC3 0.000 0.0000 0.0000 0.0000 0.0549 0.2280 0.1256 0.3316COAST 0.9783 0.1774 0.8758 0.3303 0.4193 0.4938 0.2636 0.4408DATE 54.52 12.06 57.93 10.04 53.41 12.92 56.97 10.66PRODUCER 0.9130 0.2849 0.8399 0.3673 0.7520 0.4322 0.5701 0.4952Observation 46 306 601 1377provide an incentive for their holders to invest more than in the case of a volume-based tenure suchas a Forest License. On the other hand, this study has included a mixture of Timber Licenses, bothwithin and outsides Tree Farm Licenses. Therefore, the sign of TL is expected to be positive86although Timber Licenses outside Tree Farm Licenses have no residual values to the licensees.Species. Regeneration costs differ among species (Smith 1986). Five species (balsam, cedar,Douglas fir, hemlock and spruce) are singled out in this study to measure the effect of speciescomposition on silvicultural investment. These species account for more than 70 percent of thetimber harvested in the province and the second growth crops should roughly be composed of thesame proportion. So there are five species variables, BALSAM, CEDAR, D_FTR, HEMLOCK andSPRUCE, which measure the percentage of balsam, cedar, Douglas fir and spruce regenerated on acutbiock, respectively. The coefficients of these variables measure the effect of these species onsilvicultural investment per hectare, compared with the remaining species (a mixture of pine, cypressand hardwood species). Since the regeneration costs for these species are not available, the signs ofthese variables cannot be predicted.Natural Attributes. The size of each cutbiock in hectares is included as a variable (SIZE).Again, the sign of this variable is not obvious because the effect of scale on forest regeneration isunknown. Three categories of soil quality (good, medium and poor) measure the natural productivityof the lands. Two dummy variables (SO1L_G and SOIL_M) are used to account for the effect of soilquality on silvicultural investment. SOIL_G is a variable taking the value of unity if the site classof a cutbiock is good, and zero otherwise. SOIL_M is a variable taking the value of unity if the siteclass of a cutbiock is medium, and zero otherwise. The coefficients of these variables indicate theeffect of good and medium soil quality lands on silvicultural investment compared with the effect ofpoor soil quality lands. Both variables should have positive signs.There are eight biogeoclimatic zones encountered in this study. To simplify the analysis,87these zones have been classified into four groups, in consultation with forest ecologists40,and threevariables are assigned (BGC1, BGC2 and BGC3). BGC1 takes the value of unity when thebiogeodimatic zone is either Coastal Douglas Fir, Coastal Western Hemlock, or Montane Hemlock,and zero otherwise. BGC2 takes the value of unity when the biogeoclimatic zone is either EngelinanSpruce-Subalpine Fir, Montane Spruce, or Sub-boreal Spruce, and zero otherwise. Similarly, BGC3takes the value of unity when the biogeoclimatic zone is Interior Douglas Fir, and zero otherwise.The coefficients of these variables measure the effect of these biogeoclimatic types on silviculturalinvestment, compared with the biogeoclimatic type of Interior Cedar Hemlock41. The signs for thesevariables are not obviously predictable.Producer’s Characteristics and other Variables. Similar to Chapter V, a dummy variable—PRODUCER— is included to take into account the effect of the producer’s characteristics. It takesthe value of unity if the purchaser of a property is a large integrated forest products firm and zerootherwise. The top 20 companies42,which collectively hold more than 74 percent of the committedannual allowable cut, are designated as large firms. A significant positive sign will indicate that thesecompanies invest more in silviculture than other firms.° Professors Philip J. Burton and Gordon Weetman. Personnel Communication. 1993.41 Originally the Interior Douglas Fir is treated as the base biogeodimatic zone. It is recognizedlater that, since there are no private lands in this base zone, such design may cause statistical error.Therefore, the base biogeodimatic zone is changed to the Interior Cedar Hemlock. This change affectsthe results of all biogeoclimatic variables. However, it does not have any noticeable effect on themagnitude of the tenure variables.42 See note 31 in Chapter V. Fourteen of these 20 firms happen to be included in this study. Theyare: Ainsworth, Canadian Pacific, Canfor, Crestbrook, Doman, Fletcher Challenge, Interfor, MacmillanBloedel, Pope & Tolbot, Slocan, Tolko Industries, Weidwood, Westar and Weyerhaeuser. All of themhave area-based tenures and therefore the variable PRODUCER could capture some allowable cuteffect as well.88A date variable (DATE) is included to capture the time trend of silvicultural investment. Arisk-free interest rate variable (TNT), which takes the value of the 3-month canadian treasury bill rate,is added to catch the financing cost of silviculture investment.Location. Originally a dummy variable (COAST) is assigned a value of unity for eachcutbiock in the Vancouver forest region and zero otherwise to capture the effect of location onsilvicultural investment. This variable is dropped later due to its significant correlation with BGC1.The correlation coefficient between them is 0.9724. This can even be seen in Table 11 by comparingthese two variables.3. Empirical ResultsThe functional form of equation (20) is selected empirically by applying the Box-Coxtechniques to the most common functional forms (linear-linear, linear-log, log-linear, and log-log).The log-linear form is preferable43. The regression results are given in Table 12.Most of the results for the explanatory variable accord with prior expectations. Out ofeighteen (18) parameters estimated, sixteen (16) of them are significant at the 90 percent confidencelevel or better. The following portion of this section describes some parameters in detail. Again, theeighteen variables are categorized into four groups: tenure, species composition, natural attributes,producer’s characteristic and others.As discussed in note 32 of Chapter V, once the dependent variables have been transformedappropriately, the function form with the smallest residual sum of squares has been chosen. Theresidual sum of squares 3852 for linear-linear; 3926 for linear log; 2754 for log-linear and 2860 forlog-log model. Thus the logarithmic transformation is consequential with respect to the dependentvariable, but inconsequential to the independent variables.89Table 12. Empirical Results of Log-Linear Equation on Silvicultural InvestmentVariable Coefficient T-ratioTenurePFL 0.5931 3•433**TFL 0.2425 4.187**TL 0.1419 1.749*Species CompositionBALSAM -0.0121 6.476**CEDAR 0.0046 2.608**D FIR 0.0060 3f9**HEMLOCK -0.0096 6.150**SPRUCE 0.0053 7.0432**Natural AttributeSOIL_G 0.2646 2.283**SOIL..M 0.3975 4.078**BGC1-0.4462 3.846**BGC2-0.2819 4.148**BGC3-0.9736 10.453**SIZE-0.0005-0.598Location and OthersPRODUCER 0.2986 5.678**DATE 0.0188 3.675**INT 0.1022 4.386**90R2 -adjustedObservation** Significant at 95 percent level.* Significant at 90 percent level.Tenure. The coefficients for PFL, TFL and TL indicate that tenure is a significant factor indetermining the level of silvicultural investment. In fact, these parameters are significantly differentfrom zero even at the 99 percent confidence level for PFL and TFL, and at the 92 percent level forTL. The regression results imply that per hectare silvicultural investment under Forest Licenses,expressed as a percentage of the investment in private forest lands within Tree Farm Licenses, TreeFarm License Schedule “B” lands and Timber Licenses, is 55.26, 78.47 and 86.77 percent,respectively. In other words, per hectare silvicultural investment in private forest lands, Tree FarmLicense Schedule “B” lands and Timber Licenses is 180.96, 127.44 and 115.25 percent, respectively,of the investment in Forest Licenses. Therefore, ceteris pan bus, at the mean silvicultural investmentamount of $745.95 per hectare under Forest Licenses, lands under private ownership, Tree FarmLicenses and Timber Licenses receive $1349.87, $950.64 and $859.71 investment per hectare (in thereal dollar values of December, 1992). Figure 2 presents these predicted silvicultural investmentusing the regression results, at the sample means for all the data, and shows the impact of tenuredifference alone.Table 12. continuedVariableINTERCEPTCoefficient T-ratio3.9780 8.223**R2 0.20110. 1951231191IaC,,Species Composition. Silvicultural investment is strongly related to species. The significantlypositive relationship between silvicultural investment and cedar, Douglas fir and spruce reveals thatlicensees invest significantly more in regenerating these species than in a mixture of “other species”(other than balsam, cedar, Douglas fir, hemlock and spruce). This implies that these are morevaluable species. In contrast, regenerating balsam and hemlock costs less than these “other species”.Natural Attributes. Soil quality is found to be significantly related to silvicultural investment.The positive coefficients of good and medium soil quality indicate that silvicultural investment tendsto increase in good and medium soil quality lands. This finding is justifiable, given that good andFigure 2. Silvicultural Investment among Forest Tenures—140&”120&-iOoo”800600400200fI-...-....-..-i-.,--.-..’.--1 I. -. - -Private Tree Farm License Timber License Forest License92medium soil qualities yield higher returns than poor soil quality. The three variables that measurebiogeoclimatic types show that silvicultural investment varies in different biogeoclimatic zone. Thecoefficient for cutbiock size variable is negative, but not significantly different from zero.Producer’s Characteristics and Other Variables. The coefficient of PRODUCER is positive,and significantly different from zero at the 95 percent level. This result could be interpreted withreference to the allowable cut effect, and economies of scale in the forest industry. Large andintegrated firms that hold most of the area-based tenures can benefit from the allowable cut effect.They may also be efficient enough in generating revenue that they can invest more in silviculture.In addition, they may want to invest more in silviculture to ensure their current and future timbersupply, and therefore protect their investment in large forest products manufacturing facilities.A significant positive relationship is found between silvicultural investment and silviculturalending date. Therefore, silvicultural investment is increasing over the study period. This trend isperhaps related to the 1992-1993 timber price surge, and the increasing emphasis on silviculturalinvestment in recent years. The coefficient of INT is positive, and significantly different from zero.This is counter-intuitive, because the investment in silviculture is expected to decrease when realinterest rate rises, not increase as indicated here.4. Conclusions and DiscussionBasic economics postulates that an investment activity must be justified by its return in thefuture. Silvicultural investment is no exception. Thus, tenure holders’ investment decisions will bejudged in terms of the future benefits that they expect the investment will bring to them.93The forest tenures in B.C., differ from one another in qualities such as comprehensiveness,duration, security and transferability, thereby defining the future benefits from silvicultural investmentthat accrue to the holders. Therefore, the tenure characteristics should affect silvicultural investmentitself. In particular, the findings of this study show that the per hectare silvicultural investment ofForest Licenses is only about 55 percent of that on private forest lands, 78 percent of that on TreeFarm License schedule “B” lands and 87 percent of that on Timber Licenses, respectively. Thesefindings are broader than those of Luckert (1988) and Luckert and Haley (1990), but consistent withtheir conclusions about private forest lands and Tree Farm Licenses.The high silvicultural investment on private forest lands comes as no surprise, since in almostevery aspect, private forest lands, as a tenure, favour the owners. The strong positive effect of TreeFarm Licenses and Timber Licenses on per hectare silvicultural investment verifies that thecharacteristics of these two tenures are favourable to their holders, and that area-based tenures arepreferable to volume-based tenures.These results should not be interpreted as asserting that lands under one tenure are managed“better” from a social point of view than lands under another tenure. Indeed, as Luckert and Haley(1990) observe, “the public and private objectives for management of forest differ and it is impossibleto conclude that the intensity of forest management under a tenure (such as Forest License) is sub-optimal.” However, it is possible to judge the findings of this study— the adequacy of silviculturalinvestment under four major types of tenures in B.C.— in the context of public objectives.A recent study by Vertinsky, Wehrung and Brumelle (1990) shows that the forest industry,the provincial government, and the public all identify the stability of economic wood supply as their94highest silvicultural priority. This goal calls for more silvicultural investment. However, as Pearseobserved, “the enthusiasm of Canadians for public ownership (of forest lands) is not matched by anenthusiasm for big bureaucracies to manage them.” Thus, attracting investment from the privatesector seems to be the only choice. This analysis indicates that tenure is one of the most importantfactors in influencing the private tenure holders’ investment decisions. Therefore, the rewards ofpursuing the objective of restructuring the current tenure system will be significant.95VII. The Effect of Forest Tenure on the Quality of Forestry Practice1. IntroductionThis chapter investigates whether or not the current tenure system encourages private industryto practise good silviculture on public lands in B.C. No literature has been found, to my knowledge,that deals directly with the relationship between forest tenure and silvicultural performance. Thereare, however, a few similar studies in other sectors, notably in agriculture (e.g., Feder et al. 1988;Anderson and Lueck 1992), which address the relationship between land tenure and productivity(outputs).This chapter is a further development of Chapter VI. Most data used in this chapter have thesame source, definition and simple statistics as in Chapter VI. More importantly, the results of thesetwo chapters are inter-related: forest tenures affect silvicultural investment, the amount of silviculturalinvestment determines the choice of silvicultural treatments and affects outputs. In addition, foresttenures may directly affect the forest practice and outputs as well.The chapter is organized as follows. The next section starts with measuring outputs andsilvicultural performance in forestry; followed by a brief discussion of the incentives and constraintsassociated with forest tenures that have determined tenure holders’ silvicultural performance. Section3 and Section 4 respectively describe models and data. Section 5 discusses empirical results, andconclusions and discussion are presented in Section 6.2. Measuring Output and Silviculture Performance in Forestry96Relating land tenure and silvicultural performance calls for a measurement indicator. Theindicator often used in agriculture literature is output or land productivity. It is difficult to have anoutput indicator in forestry for two reasons. First, there is no market for young stands under tenuresother than private lands in B.C. Second, the development of other quantitative indicators is a lengthyprocess. For example, even indicators such as the achievement of free-to-grow status can only beavailable after regeneration activities have been undertaken for more than ten years in B .C.Therefore, only short-term readily available output measurements such as the existence of not-satisfactorily restocked landsM and the percentage of not satisfactorily restocked lands45 (ascomparing to gross harvested area) are used in this study.The alternative is to measure the silvicultural activities performed by tenure holders byassuming that different silvicultural inputs lead to different outputs. Silviculture, however, hasmultiple dimensions. It includes the broad categories of site preparation, reforestation and standtending, which aggregately consist of 28 officially recorded silvicultural activities in B.C.Measuring each and every silvicultural activity is neither efficient nor necessary. For one thing, these‘ Not satisfactorily restocked lands refers to lands without stocking to a prescribed standard. Thestandard varies from site to site, but a typical nile calls for at least 750 trees per hectare of acceptablespecies to be established within three years of denudation on the Coast and 5 years in the interior(Pearse, Lang and Todd 1986). It is a general category that the Ministry of Forests uses to describecurrent forest land status. The appearance of not satisfactorily restocked lands is a gRd outputindicator since it reveals the results of human efforts and the nature.The percentage of not satisfactorily restocked lands reveals the relative proportion of notsatisfactorily restocked lands to the cutbiock size. Using it as an indicator does not diminish the roleof the existence of not satisfactorily restocked lands, since there are only about 10 percent of allcutblocks studied in this thesis have this category and the average proportion for these cut blocks isless than 0.1 or 10 percent. With the remaining 90 percent of cutblocks having the value of zero, thedependent variable itself bias toward zero and may overshadow the effect of any independent variable.Therefore, using this indicator only may not be appropriate to catch the whole picture of the effectof tenure and other variables.See note 37 in Chapter VI.97activities are, at least partly, reflected in the fmdings on silviculture investment in Chapter VI.Furthermore, there is some debate about the impacts of certain activities (e.g., broadcast burning andchemical treatment). Third, some treatments such as spacing, pruning and fertilization, are not widelyused because there has been too little time to implement these treatments (e.g., spacing and pruning)since the latest legislation change in 1987. Lastly, certain treatments (such as fertilization)48 arerarely used. Therefore, this study has chosen the reforestation method, i.e., planting or naturalregeneration, and the regeneration period (from the end of harvesting to the completion of planting)as two silvicultural performance indicators49.What is the role of tenure in silvicultural performance? As in the case of silviculturalinvestments, tenure provides a framework of incentives and constraints within which its holdersoperate and make decisions. Everything else being equal, the characteristics of the tenure determinewhether tenure holders receive the future returns of silviculture, the quantity of these returns, theamount of their investment and the type of activities they choose. In turn, investment and silviculturalactivity detennine outputs such as not satisfactorily restocked lands.The examples discussed in Chapter III make the point clear. The first example looks at twotypes of tenures. One is secure and lasts is forever; another has a 25-year duration, renewable on an“evergreen” basis, but its security is in doubt. Since they have a stake in future crops, the holders‘ See Chapter IV for a discussion about the changes in B.C. legislation in September, 1987.48 Fertilization only occurs in seven out of the 2311 cutblocks studied in this chapter.The legitimacy of choosing reforestation method as an performance indicator is thatreforestation is the primary silvicultural costs and the difference between the costs of planting andthe costs of natural regeneration is huge. It is hoped that revealing the choice of tenure holders onregeneration method will throw some light on their attitudes in costly activity. This is also the purposefor choosing regeneration period, since licensees who have different attitudes in costly activity mayhave different length of time to carry the activity.98of the former tenure may invest more and perform better silviculture, thus minimizing notsatisfactorily restocked lands, than the holders of the latter tenure. In this simple case, the differencein silviculture investment, performance, and the existence of not satisfactorily restocked lands can beattributed to a characteristic of tenure, security.The second example looks at area-based tenure versus volume-based tenures. The holders ofarea-based tenures may recapture some benefits of their silvicultural activities and therefore canquickly regenerate denuded lands. On the other hand, the holders of volume-based tenure, whoseonly incentive is to minimize their silvicultural costs, may delay regeneration as long as they are ableto meet regulatory requirements. Again, everything else being equal, the difference in silviculturalperformance in these two types of tenure can be attributed to whether the tenure is area based orvolume based. The same logic applies to all of the indicators used in this chapter, namely, theseindicators being the appearance of not satisfactorily restocked lands, the regeneration method, and theregeneration period.3. ModelsThe models used in this chapter are variations of equation (21) discussed in Chapter III. Theycan be written in the form of a structural, economic model, based on cutbiock (tract) observations:(24) Occurrence of Not Satisfactorily Restocked Lands = f(I, C1, L, Cf, C, T1, T)(25) Percentage of Not Satisfactorily Restocked Lands = f(I, C1, L, Cf, C, T, T,)(26) Occurrence of Planting = f(C1, L, Cf, C, T, T)(27) Regeneration Period of Planting = f(I, C1, L, Cf, Ci,, T, T,)99Equation (24) states that the existence of not satisfactorily restocked lands is a function ofsilvicultural investment, time, and all of the variables in equation (13). Equations (25) and (27) havethe same independent variables as equation (24), but the dependent variables are, respectively, thepercentage of not satisfactorily restocked lands, and the regeneration period of planting. Equation(26) shows the determinants of planting occurrence. It has the same variables as in equation (24)with the exception of the investment variable. This difference prevents the duplication of relatedvariables on both sides of the equation and thus avoids man-made correlation between planting andinvestment50.Equations (24) and (26) are equations with qualitative dependent variables. The logisticregression is used to estimate them5’ 52• The Box-Cox transformation technique or the maximum-likelihood method is used to find the functional formulation for equations (25) and (27). Theregression results of these equations will reveal the contributions of tenure and other factors tosilvicultural performance.Notice, however, that equation (13) can be nested in equations (24), (25) and (27). As aresult, a strong relationship between silvicultural investment and the tenure variables in equation (13),50 See Section 2 of Chapter V for discussion of computing the investment figure for each cutbiockstudied.51 The linear logistic function has the formlog(p/1-p) = a + ‘Xwhere p is the probability of an event occurring, cx is the intercept parameter, and is the vector ofslope parameters. Sop= exp(ôc + j3’X)/(l + exp(à + frX))where a and 13 are estimates of the intercept and slope parameters.52 Equations (24) and (26) were also estimated by using the linear probit function. As would beexpected, the results are similar to those reported below.100as shown in last chapter, could detract from the influence of the tenure variables in these equations53.Therefore the following analysis is done both with and without the investment variable.4. DataAll independent variables used in this chapter have the same origins, definitions and simplestatistics as those of Chapter VI. The definitions and statistics of each dependent variable arepresented in Tables 13 and 14. This section only discusses the expected signs of independentvariables.Tenure. It is expected that the variable for private forest lands (PFL) will be significantlynegative in equations (24) and (25), and significantly positive in equation (26) and (27), since thecharacteristics of private lands favour the owners more so than those of Forest Licenses. The signsof TFL and TL should be the same as that of PFL for reasons similar to these given in Chapter VI;the characteristics of both Tree Farm Licenses and Timber Licenses are either stronger than, or atleast equivalent to those of Forest Licenses. Furthermore, Tree Farm Licenses and Timber Licensesare area-based tenures which may provide their holders with stronger incentives to improve theirperformance than volume-based tenures such as Forest Licenses.Investment. The investment variable assigned is expected to be significantly negative forequations (24) and (25), but positive for equation (27).One may argue that estimation of equations (24), (25) and (27) involves two-stage least squareregressions. In the first stage, equation (13) is estimated and the predicted investment, I, is obtained.By using the predicted investment instead of the actual investment in the regression of equations (24),(25) and (27), the second stage reveals if tenure variables have additional effects on outputs. Thisapproach is also tried. The results are indifferent from those reported in this chapter.101Table 13. Definitions and Statistics for Dependent Forest Practice VariablesVariable Mean Standard DefinitionValue DeviationNSR 0.0979 0.2972 Dummy: not satisfactorily restocked lands (1 if notsatisfactorily restocked occurs, 0 otherwise)% NSR 6.4952 30.4803 Percent of not satisfactorily restocked lands (NSR/Size)PLANTING 0.7514 0.4323 Dummy: planting (1 if planting, 0 otherwise)PERIOD 34.93 12.44 Regeneration period (number of month after harvesting tocompletion of planting)Table 14. Some Statistics on Private Forest Lands, Timber Licenses, Tree Farm Licenses and ForestLicenses: Dependent Forest Practice VariablesPrivate Lands Timber Licenses Tree Farm Licenses Forest LicensesVariableMean Standard Mean Standard Mean Standard Mean StandardValue Deviation Value Deviation Value Deviation Value DeviationNSR 0.0222 0.1491 0.0915 0.28879 0.0915 0.28858 0.1046 0.3061% NSR 0.2728 1.8305 4.6134 17.6466 5.2042 20.5029 7.6854 36.2847PLANTING 0.8889 0.3 178 0.7222 0.4486 0.7604 0.4272 0.7443 0.4363PERIOD 32.95 13.46 37.83 11.29 31.11 14.05 36.10 11.47Species. The signs of the species variables cannot be predicted.Natural Attributes. The sign of the size variable is expected to be positive in equation (24)since not satisfactorily restocked lands are more likely to occur in big cutbiocks. However, it should102be negative in equation (25) since holding the size of not satisfactorily restocked land and everythingequal, the percentage of not satisfactorily restocked lands should be negatively related to the cutbiocksize. Its sign in equation (26) and (27) should also be positive since natural regeneration is unlikelyto be successful in large cutbiocks. Both soil quality variables should have negative signs inequations (24) and (25), but positive signs in equations (26) and (27). The signs for thebiogeoclimatic type variables are less obvious.Others. The signs of PRODUCER and DATE are not obvious.5. Empirical Results5.1 The Occurrence of Not Satisfactorily Restocked LandsTenure. Column (1) and column (2) of Table 15 present the logistic regression results ofequation (24) with and without silviculture investment as an independent variable, respectively.Private lands contribute negatively to the probability of not satisfactorily restocked land appearancein both cases. The coefficients of TFL and TL are negative but not significantly different from zeroat the 20 percent level. This result suggests that private forest ownership has a significant negativeeffect on the probability of occurrence of not satisfactorily restocked land.Figure 3 demonstrates the effect of forest tenure on the probability of not satisfactorilyrestocked land occurrence. This figure is based on the regression results in column (2) of Table 15,and on the assumption that the soil quality is medium, the producer is a large integrated firm, thebiogeocimatic zone variable is BGC1, and all continuous variables take their mean values. Also,103Table 15. Empirical Results for Logistic Equation: Appearance of Not Satisfactorily Restocked Land(1) (2)VariableCoefficient Wald X2 ratio Coefficient Wald X2 ratioTenure and InvestmentPFL -1.0521 1.8956*-1.0829 2.0181*TFL -0.0808 0.1822 -0.0932 0.2244TL -0.2226 0.7541 -0.2549 0.9244INVESTMENT -0.0447 9.8042**Species CompositionBALSAM 0.0077 21.6105** 0.0077 21.5149**CEDAR -0.0028 1.4450 -0.0032 1.9556D_FLR 0.0049 3.6248** 0.0042 2.7703*HEMLOCK 0.0095 12.6650** 0.0089 11.4460SPRUCE 0.0031 3.4468** 0.0030 3.1590**Natural attributeSOIL_G -1.0749 9.6090**-1.0550 9.4129**SOIL_M -0.7681 9.4084**-0.7842 9.8731**BGC1-0.7917 6.0467** -0.6304 3.9526**BGC2-0.1456 0.3703 -0.1545 0.4296BGC3 0.9926 14.3300** 1.0257 15.4987**SIZE 0.1247 3.2807** 0.1282 2.9709**OthersPRODUCER 0.1334 0.5935 0.0933 0.2950DATE 0.2339 0.4581 0.1396 0.1662104Table 15. continuedObservation** Significant at the 10 percent level.* Significant at the 20 percent level.the logistic equations were transformed to compute the probability of not satisfactorily restocked landoccurrence.Investment. The coefficient of the silvicultural investment variable is negative andsignificantly different from zero at the 10 percent level. This finding is consistent with commonknowledge: increased investment in silviculture is expected to reduce the appearance and amount ofnot satisfactorily restocked lands. When this variable is dropped from the equation, the effect of thetenure variables (and most other variables) increases (column 2 of Table 15).Species Composition. The coefficients for the species variables show that, compared to themixture of “other species”, regenerating balsam, Douglas fir, hemlock and spruce contribute positivelyto the probability of having not satisfactorily restocked lands, but the effect of cedar is not differentfrom that of “other species”.(1) (2)VariableCoefficient Wald X2 ratio Coefficient Wald X2 ratioINTERCEPT -0.9354 0.4027 -0.9445 0.4150-2 log L 84.136 75.472Score 91.401 79.8172311 23116—’Natural Attributes. Not surprisingly, good and medium soil quality contribute negatively, andcutbiock size contributes positively, to the probability of having not satisfactorily restocked lands.The coefficients for the biogeoclimatic zone variables suggest that the probability of not satisfactorilyrestocked land occurrence varies in different biogeoclimatic zones.Others. The coefficient of PRODUCER indicates that the probability of having notsatisfactorily restocked lands does not vary among producers, once all factors are accounted for. Thecoefficient of the DATE variable reveals that the time trend of the probability of not satisfactorilyrestocked land appearance is static during the study period.Figure 3. Probability of Not Satisfactorily Restocked Land Occurrence10581.4..j ;21’0 :z:zzi ;.;zz:iPrivate Tree Farm License Timber License Forest license1065.2 The Percentage of Not Satisfactorily Restocked LandsThe above analysis does not account for the relationship between the size of not satisfactorilyrestocked land and cutbiock size. To strengthen the results of Table 15, equation (25) is estimatedby using linear-log function form in which the continuous independent variables take the form ofnatural logarithm. The results (Table 16) support those reported in Table As expected, tenure,silviculture investment, size, site class, biogeoclimatic zone, and others all contribute significantly tothe percentage of not satisfactorily restocked lands. All variables except CEDAR, SIZE andPRODUCER have the same signs as in Table 15. While the coefficients of CEDAR andPRODUCER are not significant in both cases, change in the sign of the SIZE is reasonable—everything else being equal, the percentage of not satisfactorily restocked lands should get smalleras the cutbiock size gets bigger.5.3 The Occurrence of PlantingTenure. Table 17 presents the logistic regression results of planting occurrence as expressedin equation (26). PFL is positively related to the logistic probability of planting occurrence at the 10percent level. The coefficients of TL and TFL are insignificant at the 10 percent level. Therefore,private lands contribute negatively to the probability of planting occurrence. The relationship betweentenure and the probability of planting occurrence is presented in Figure 4, where the probabilities ofplanting occurrence are predicted using the results of Table 17, and assuming that the soil qualityThe low R2 indicates the lack of fit in the model estimated. This is primarily due to the factthat 90 percent of the cutbiocks have zero percent not satisfactorily restocked lands, and therefore theindependent variables are zero; and the remaining 10 percent cutbiocks have independent variablesclose to zero. This limit of this regression is overcome by including the preceding regression in thischapter, namely, the regression on the occurrence of not satisfactorily restocked lands.107Table 16. Estimates of Percentage of Not Satisfactorily Restocked Lands(1) (2)VariableCoefficient T-ratio Coefficient T-ratioTenure and InvestmentPFL -0.0682 1.438* -0.0729 1.535*TFL -0.0152 -0.969 -0.0161 -1.022TL -0.0366 1.683* -0.0383 1.760*INVESTMENT -0.0054 2.795**Species CompositionBALSAM 0.0004 2.552** 0.0004 2.626**CEDAR 0.0001 0.126 0.0000 0.046DYIR 0.0004 1.733* 0.0003 1.495HEMLOCK 0.0004 1.904* 0.0004 1.835*SPRUCE 0.0002 1.185 0.0002 1.108Natural AttributeSO1L_G -0.1066 ..3375**-0.1062 3360**SOIL_M -0.1033 3.879**-0.1050 3.940**BOC1 -0.0454 1.784*-0.0342 -1.359BGC2 -0.0079 0.423** -0.0074 0.393**BGC3 0.6318 2.502** 0.0693 2.720**SIZE -0.0161 2.589**-0.0168 2.703**OthersPRODUCER -0.0036 -0.250 -0.0067 -0.462DATE 0.0146 0.513 0.0087 0.303108Table 16. continued(1) (2)VariableCoefficient T-ratio Coefficient T-ratioINTERCEPT 0.3080 2.470** 0.2929 2.348**R2 0.0251 0.0218R2 -adjusted 0.0179 0.0159Observation 2311 2311** Significant at the 10 percent level.* Significant at the 20 percent level.is medium, the producer is a large integrated firm, the biogeocimatic zone variable is BGC1, and allcontinuous variables take their mean values.Species Composition. The coefficients for the species variable show that, as comparing to themixture of “other species”, balsam and hemlock contribute negatively, and Douglas fir and sprucecontribute positively to the probability of planting at the 10 percent level. These results imply thatartificial regeneration is more often used for Douglas fir and spruce than for “other species”, andnatural regeneration is more often used for balsam and hemlock than “other species”. These resultsare expected since Douglas fir and spruce are the most valuable species.Natural Attributes. Cutblock size and good soil quality affect positively the probability ofplanting occurrence. The positive effect of medium soil quality is insignificant at the 10 percentlevel. The coefficients for the biogeoclimatic zone variables suggest the probability of planting variesin different biogeoclimatic zone.109Table 17. Empirical Results for Logistic Equation: Planting OccurrenceVariable Coefficient Wald X2 ratioTenurePFL 0.4436 3.1199**TFL -0.0080 0.0118TL 0.0588 0.3503Species CompositionBALSAM-0.0036 29.0322**CEDAR 0.0011 1.3061D_FIR 0.0027 6.5860**HEMLOCK -0.0032 lO.6326**SPRUCE 0.0018 6.7670**Natural AttributeSOIL_G 0.2841 3.5951**SOIL_M 0.0641 0.278 1BGC1 -0.1077 0.7935BGC2-0.1806 3.9426**BGC3-0.9063 65.1497**SIZE 0.7161 6.3646**OthersPRODUCER-0.0118 0.0297DATE-0.3557 6.7462**INTERCEPT 1.9524 10.6878**-2logL 180.144110Table 17. continuedScoreObservation** Significant at the 10 percent level.* Significant at the 20 percent level.Others. The coefficient of PRODUCER shows that the large firms and the small firms areindifferent in terms of choosing regeneration methods. The coefficient of the DATE variable revealsthat the time trend in planting is declining during the study period.5.4 The Regeneration Period of PlantingEquation (27) is estimated by using 1732 cutbiock observations, where tenure holders havechosen to plant rather than naturally regenerate. The functional form of the equation has beenselected empirically by applying the Box-Cox technique to the most common functional forms (linearlinear, linear-log, log-linear, and log-log). The linear-linear form is found to be preferable55. Theregression results are presented in Table 18.Tenure. The coefficients for the tenure variables indicate that tenure is a significant factorin determining the period of regeneration, irrespective of the inclusion of investment variable. Thenegative coefficients of PFL and TFL are significantly different from zero at the 5 percent level.However, the negative coefficient of Th is insignificant. The regression results imply that, at meanThe residual sum of squares is 93 for linear-linear; 153 for linear log; 104 for log-linear and144 for the log-log model.184.34123119Of8O7I6Oi5o3O2O10’Figure 4. Probability of Planting OccurrencePrivate Tree Farm License Timber License Forest Licenseillregeneration period of some 35 months (Table 13), private lands are regenerated in 32 months andTree Farm License schedule “B” lands are regenerated in 34 months.Investment. The coefficient of INVESTMENT is positive and significantly different from zeroat the 10 percent level. This finding comes as no surprise: tenure holders prefer to defer costlyactivities such as planting.Species Composition. The regeneration period is strongly related to species. Thesignificantly positive coefficients for balsam, Douglas fir and hemlock reveal that it takes longer for1..,...’InE:1:1112Table 18. Empirical Results for Linear-Linear Equation: Months to Planting(1) (2)VariableCoefficient T-ratio Coefficient T-ratioTenure and InvestmentPFL -3.3393 3.251** -3.2571 3.l72**TFL -1.3550 3•935** -1.3328 3.859**TL -0.5403 -1.117 -0.5402 -1.120INVESTMENT 0.0004 1.787*Species CompositionBALSAM 0.0449 4.047** 0.0434 3.915**CEDAR 0.0158 1.526 0.0164 1.587D....FIR 0.0189 1.927* 0.0203 2.070**HEMLOCK 0.0471 5.096** 0.0468 5.068**SPRUCE 0.0041 0.902 0.0054 1.194Natural AttributeSOIL_G 0.0510 0.074 0.1000 0.145SOIL_M -0.0604 -1.042 -0.5641 -0.974BGC1 0.8005 1.158 0.6835 0.993BGC2 -0.6801 1.673*-0.7659 1.896*BGC3 1.5695 2.806** 1.4191 2.565**SIZE 0.0139 2.909** 0.0137 2.878**OthersPRODUCER 0.0136 0.043 0.0517 0.165DATE 0.9052 70.577** 0.9059 70.641**113Table 18. continuedR2 -adjustedObservation** Significant at the 5 percent level.* Significant at the 10 percent level.these species to be regenerated than “other species”. The regeneration period for cedar and spruceis similar to “other species”.Natural Attributes. Soil quality is found to be an insignificant factor in determining theregeneration period. This may reflect current provincial regulations, which require that all cutblocksbe regenerated regardless of soil quality. The regeneration period varies in different biogeoclimaticzone. The coefficient for the cutbiock size variable is positive and significantly different from zeroat 10 percent level. This means that the regeneration period is positive related to cutblock size.Others. The regeneration period is not different among producers. The coefficient for DATEindicates that the length of regeneration period is positively related to the ending date of silviculturalcompletion, and therefore, the regeneration period is getting longer over the study period.(1) (2)VariableCoefficient T-ratio Coefficient T-ratioINTERCEPT -16.3190 17.274**-16.0846 l7.185**R2 0.7351 0.73470.7331 0.73291732 17326. Conclusions and Discussion114The estimates of this chapter show that:(1) not satisfactorily restocked lands are less likely to occur on private lands than onCrown lands under Tree Farm Licenses, Timber Licenses and Forest Licenses;(2) where not satisfactorily restocked lands occur on private lands and TimberLicenses, the proportion of these lands to the total cutblock size is smaller thanthe proportion for Tree Farm License Schedule “B” lands and Forest Licenses;(3) artificial regeneration is more frequently used by private land owners than thelicensees of Tree Farm License Schedule “B” lands, Timber License and ForestLicense; and(4) if planting is chosen as the regeneration method, the regeneration period forprivate lands is shorter than the periods for Tree Farm License Schedule “B”lands, Timber Licenses and Forest Licenses; and Tree Farm License Schedule “B”lands have a shorter regeneration period than lands under Timber Licenses andForest Licenses.The findings confirm that the characteristics of tenure affects the tenure holder’s silviculturalperformance. When a tenure is comprehensive, secure, transferable and long-term, such as the caseof private lands, their holders shows a stronger interest in the future crop. Through investing moreand choosing the most effective methods, they tend to eliminate not satisfactorily restocked lands.They apparently use costly and thus maybe more effective method (planting) to regenerate denudedforest lands. Furthermore, they regenerate cutblocks more quickly and avoid opportunity costs ofprolonging the regeneration period.On the contrary, if a tenure is not comprehensive, is insecure, is short-term and has limits on115transferability such as Forest License, their holders may not benefit much from their silviculturalactivities. Thus, the only incentive left to them is to meet the requirements of regulations withminimum cost. In fact, they can become so concerned about minimizing their costs that their actionscan lead to an increase (at least temporarily) in not satisfactorily restocked lands. They are likely toavoid using more expensive methods, such as planting, to regenerate the lands. And, likewise, in aneffort to avoid costs, they may postpone regeneration activities.One can argue that all of the not satisfactorily restocked lands will ultimately be eliminatedby the licensees at their own expense. Moreover, licensees must meet the requirement of free to growstatus within a certain period of time. However, the occurrence of not satisfactorily restocked landsgives rise to opportunity costs, because the benefits of future crops are postponed. Delays inregenerating denuded forest lands represent such a loss to the public as well.It is important to note that one should not infer from this that planting is better than naturalregeneration. Given that planting costs more than natural regeneration, the fact that holders of sometenures consistently and deliberately avoid planting, merely implies that the characteristics of thesetenures do not provide enough incentives for them to invest.The fmdings of this and previous chapters reveal that property institutions do make adifference in land value and forest management. These results and the policy implications aresummarised and discussed in the concluding chapter.116VIII. Summary and Conclusions1. Summary and ConclusionsThis study has investigated the role of tenure in determining forest land value and forestmanagement. It attempts to describe these relationships through a detailed empirical study of actualtenure holder behaviour, an approach which has not been applied to the area of forest tenure. Theprimary results of the analysis can be summarised in the following paragraphs.First, forest tenures in B.C. fall along the entire spectrum of “property”. They are similar inexclusiveness, but vary considerably in other characteristics, namely comprehensiveness, duration,security, transferability and benefits conferred. The development of the model, based on thetraditional theory of capital and a simplified version of forest tenures, predicts that forest tenure hasa significant impact on land value and forest management.Second, the event study of the forest policy change in 1987 shows that the marginal value ofcutting rights under some forest tenures (Tree Farm License, Forest License, Major Timber SaleLicense and Timber Sale Harvesting License) is small. Explanations for this fmding include factorssuch as existing log export restrictions, the earlier market anticipation, the insignificance of the netloss due to annual allowable cut reductions compared to the total assets of medium and largecompanies, and the plausible scenario that timber is fully priced for all tenures. However, noevidence is provided here on the total value of these tenures or the cutting rights under them.Third, the estimation of forest land value reveals that private lands are significantly more117valuable than Timber Licenses. This is because private forest lands, as a tenure, include the residualvalue of the bare land, and provide for more flexible use of the land and current stand of timber.Moreover, the owners of unmanaged forest lands are able to put a high value on their lands forrecreational, agricultural and other non-timber uses, and owners of managed forest lands can enjoythe benefits of the allowable cut effect. Likewise, since all of the characteristics of Timber Licensesare less tnmcated than those of Forest Licenses, one would expect that private lands are morevaluable than Forest Licenses, but the data are not available to demonstrate this.Fourth, private lands, Tree Farm Licenses and Timber Licenses receive more investment thanForest Licenses. It is evident that there is a direct relationship between forest tenure and silviculturalinvestment. Silvicultural investment tends to be high when a tenure is comprehensive, long-term,secure and when all of the benefits of future crops are conferred to its holders. The opposite is truewhen a tenure is short-term, insecure and few benefits are left for its holders.Therefore, greater investment in silviculture on private lands can be attributed to thecomprehensiveness, security and perpetuity of private ownership, which allow the owners to gain allof the benefits of future crops after property taxes. Tree Farm Licenses have the advantage of beinglong-term and area-based, which could give their holders a sense of security in silviculturalinvestment. The result for Timber Licenses is less easy to explain. However, it can be explained,at least in part, by the fact that about half of Timber Licenses are within Tree Farm Licenses and willeventually become part of Tree Farm Licenses.Fifth, the nature of forest tenures affects outputs. Specifically, not-satisfactorily-restockedlands are less likely to occur on private lands than on Forest Licenses, and even if they appear on118private lands, their proportions to the total harvest area are smaller than those of Forest Licenses.Little difference among Tree Farm Licenses, Timber Licenses and Forest Licenses is found in outputs.Similarly, forest tenure affects the performance of its holder. Costly silvicultural treatmentssuch as artificial reforestation, are more often used by private forest owners than the holders of ForestLicenses. If planting is chosen as the regeneration method, the regeneration period for private forestland is shorter than those for Tree Farm Licenses, Timber Licenses and Forest Licenses; likewise,Tree Farm Licenses have a shorter regeneration period than Timber Licenses and Forest Licenses.The results can only be explained by the fact that private land owners (and to a lesser extent, TreeFarm License holders) have the necessary security and benefits conferred in their property. Again,the incentives and constraints of tenure affect the performance of tenure holders.In summary, this study suggests four key conclusions.(1) The value of private forest lands is higher than that of Crown lands under other tenures.(2) Silvicultural investment behaviour is related to the characteristics of the property rights.The stronger and more complete a tenure, the more silvicultural investment.(3) The outputs of land depends on tenure characteristics. When the characteristics of atenure are less attenuated, the outputs will be greater.(4) The behaviour of tenure holder can be predicted from their tenure characteristics. Whena tenure is stronger and more complete, its holder will invest more and have better119silvicultural performance.2. Consistency of Results with Theory of Property: Contribution of This StudyThe theory of property suggests that property is a bundle of rights through which the holdercan enjoy the benefits of an asset; each right has its characteristics, and differences in thecharacteristics of property rights have profound economic consequences. The more “complete” thecharacteristics, the more complete the property right, and thus the more valuable the property itself.Property rights also govern the incentives for holders to invest and perform, which in turn determineeconomic outcome. Any restrictions on full private rights are likely to blunt a property holder’sincentives to invest, and the ultimate result will be lower output than there would be in the absenceof restrictions.The findings of this thesis are consistent with the theory. Forest tenures which are less“complete” than private ownership have lower values; the holders have weak incentives to invest morethan the minimum requirements of regulations, and, as a result, the output of these lands is lower thanthe output of private lands. It is evident that the structure of property rights affects both the relativeprices of inputs and the allocative decisions of the holders, and these alterations can be interpretedas adjustments of a given production function. Property rights also dictates which production functionand technology the holder chooses.3. Policy ImplicationsBefore discussing the policy implications of this study, it is important to note the complexity120of the issues surrounding forest tenure. First, the various tenure characteristics are not independent.That is, changing one characteristic modifies the others as well. The interdependency of tenurecharacteristics is best exemplified by a look at the dimension of economic benefit conferred, whichis virtually affected by all other characteristics. Another example is belief in security, which can beaffected by duration, transferability, and area- or volume-based tenure arrangements. Therefore,designing an optimal tenure is a difficult task.The second, more serious, problem is to find a social utility function and then find the bestpublic-private intersect, both of which influence the optimal specification of characteristics. Theproblem associated with finding a social utility function is well documented in economic textbooks.To choose the best public-private intersect could be exemplified by flexibility, referring to the extentto which the specifications of a tenure can be changed. Governments, as landlords, have beenattempting to maintain the flexibility of tenures in order to cope with rising public demands on forestresources. In fact, it is one of the main motivations behind the of creation of volume-based tenures.However, as I have shown earlier, volume-based tenures have limitations in attracting privatesilvicultural investment.A related issue is the transaction costs. The lack of private silvicultural investment in ForestLicenses does not necessarily mean that it is in society’s best interests to take actions towards a morecompletely defined property rights. The transaction costs must be considered and weighted againstthe benefits of changing any tenure.Despite these complexities inherent in tenure policy, decision-makers must choose a courseof action regarding the structure of tenure policy. The findings of this study could provide some121insights for future policy directions.The most important implication of this study is that, in order to attract more investment in theforest industry, and to generate stronger incentives for forest users to perform good silviculture, morecomplete forms of forest tenure should be designed.Tenure should be comprehensive, but this does not mean that forest companies need to havethe rights to all resource values on forest lands. As long as exclusivity prevails for each resourceinterest, the must advantageous balance among users can be achieved as long as every interest isrepresented by someone who can bargain with others when conflicts arise (Pearse 1988, 1993).The balance between duration (and renewability) and flexibility is also important. A longerterm tenure is warranted if it does not jeopardize the necessary flexibility of the government. To takefull advantage of market mechanisms to ensure that forest resources are used efficiently, forest tenuresmust be transferable.Second, area-based tenure is more attractive to private investment than volume-based tenure,and improves tenure holders’ silvicultural performance. This implication is drawn from thecomparison of silvicultural investment and the quality of forest practice between Tree Farm Licenseand Forest License. Everything else being equal, area-based tenure gives holders more security andan incentive to manage. One way to change this may be to modify the current Tree Farm Licensesand to redesign some, if not all, Forest Licenses along the lines of Tree Farm Licenses. In fact,International Forest Products Limited has recently proposed a “Supply Block Management Unit” to122convert chart areas of a Forest License into, in effect, area-based tenure56.Third, lack of equity in future crops prevents benefits and costs from being intemalised bytenure holders and leaves them with little incentives to invest and to manage. This lack of equitycould be achieved by replacing the current stumpage system with some revenue device that does notvary according to the timber produced, such as initial, lump-sum charge for the forest tenure or afixed land rent.Hitherto, no comment has been made about the characteristic of property referred earlier —the economic benefit conferred to the holder. Everything else being equal, the more economicbenefits left to the holders, the more they wifi invest in silviculture. In order to change this importantcharacteristic one may immediately pay attention to the stumpage system. Luckert and Haley (1989)argue that, although the current stumpage system in B.C. makes some indirect allowance for basicsilviculture, it aims at extracting all the rents for the public owner, and therefore leaves no rewardto those who produce more forest growth. It thus fails to attract much investment in intensivesilviculture (Luckert and Haley 1989). A cropsharing system, which leaves a share of the rents tothe licensees, would mitigate this problem. Therefore, replacing the current stumpage system withcropsharing system could provide some economic incentive for improved management.Likewise, if tenure holders are fully compensated when their rights are terminated, they willregard their holdings as more secure. The issues surrounding compensation are discussed extensivelyin Schwindt (1992), who argues that when termination occurs, the forest property rights should notbe compensated, but the costs of tenure holders should be reimbursed. The results of this studyClark S. Binkley. 1994. Personnel Coniniunication.123support the principle of compensation.One way to advance these objectives involves selling the rights to timber production on forestlands, but retaining public ownership of the land itself. These arrangements fall just short of privateownership of land, but strengthen the economic benefits conferred to the holders, and provide themwith the necessary economic incentives. This approach has been initiated by the New Zealandgovernment (New Zealand Forestry Corporation 1989), and certainly suggests an alternative to thepublic forest tenure system in British Columbia.The characteristics of property rights affects land value, silvicultural investment, and outputs.Simply knowing that the tenures such as Forest Licenses attract less silvicultural investment thanother types of tenure does not necessarily imply that these tenures should be abolished or changed.However, this thesis has demonstrated that the potential benefits of pursuing this objective, toparticularly to the B.C. govermnent which owns more than 95 percent of forest lands, are huge.4. Further StudyThis study has drawn attention to the need for research in several related issues. One is thetransaction costs associated with different types of tenures. Forest tenure change can only achievebeneficial results if the net gains outweigh the transaction costs. A second is to examine the politicalfeasibility or public acceptance of tenure change: an investigation that relates more closer to politicalscience than to economics. Without enough political wifi or public support, any tenure change isunlikely to be successful. A third is the environmental issues associated with forest tenures. Thegeneral perception is that property rights for non-timber interests are weak. This thesis does not124explicitly deal with this issue. Further study is warranted in designing the property rights for theseuses.125BibliographyAllen, Douglas W. 1991. What Are Transaction Costs? Research in Law and Economics 14:1-18.Anderson, T.L. and D. Lueck. 1992. Land Tenure and Agricultural Productivity on IndianReservations. Journal of Law and Economics. 35 (October):427-454.Armstrong, F.H. 1975. Valuation of Vermont Forests: 1968-1974. Department of Forestry,University of Vermont. USA.Barzel, Yoram. 1989. Economic Analysis of Property Rights. Cambridge University Press. NewYork, USA. l22p.Binder, J.J. 1985a. Measuring the Effects of Regulation with Stock Price Data. Rand Journal ofEconomics 16:167-183.Binder, JJ. 1985b. On the Use of the Multivariate Regression Model in Event Study. Journal ofFinancial Accounting Research 23:370-383.Binkley, Clark S. 1980. Economic Analysis of the Allowable Cut Effect. Forest Science 26(4):633-642.Boardman, A., I. Vertinsky and D. Whistler. 1992. Modelling the Impacts ofRegulatory Changeson Stock Prices: Protecting the Spotted Owl. Faculty of Commerce, University of British Columbia,B.C., Canada.British Columbia Forest Resource Commission. 1991. The Future of Our Forests. Chairman: A.L.Peel. 700-747 Fort Street, Victoria. B.C. V8V 1X4. 1 17p+appendices.Cheung, Steven N.S. 1970. The Structure of Contract and the Theory of a Non-exclusive Resource.Journal of Law and Economics 13 (April):49-70.Cheung, Steven N.S. 1970. The Theory of Price Control. Journal ofLaw and Economics 13 (1):53-72.Coase, Ronald H. The Problem of Social Cost. Journal of Law and Economics 3 (1):1-44126Collins, D. and W. Dent. 1984. A comparison of Alternative Testing Methodologies Used in CapitalMarket Research. Journal of Financial Accounting Research 22:48-81.Coulson, N.E. and R.P. Robins. 1987. Testing the Functional Form of Statistical AppraisalEquations. The Appriasal Journal. January.Cummings, R.G., D.S. Brookshire and W.D. Schuize. 1986. Valuing Environmental Goods: AnAssessment of the Contingent Valuation Method. Totowa, NJ: Rowman and Allanheld.CWC Canadian Western Capital Ltd. 1991. Managing British Columbia’s Forest Resources:Structural Alternatives, Financial Considerations and Valuation Estimates. Pp: 1-64 in Backgroundpaper for BC Forest Resource Commission (volume 4).Dann, L.Y. and C.M. James. 1982. An Analysis of Impact of Deposit Rate Ceilings on the MarketValues of Thrift Institutions. Journal of Finance 37:1259-1275.Demsetz, Harold. 1967. Toward a Theory of Property Rights. America Economic Review 57(2):347-359.Desai, A.S. and R.D. Stover. 1985. Bank Holding Company Acquisitions, Stockholder Returns, andRegulatory Uncertainty. Journal of Financial Research 8:145-156.Eggertsson, Thrainn. 1990. Economic Behaviour and Institutions. Cambridge University Press.New York, USA. 385p.Fama, Eugene F. 1991. Efficient Capital Market II. Journal of Finance 46:1575-1617.Feder, Gershon, T. Onchan, Y. Chalamwong and C. Hongladarom. 1988. Land Policy, and FarmProductivity in Thailand. World Bank Research Publication. Baltimore, MD: Johns HopkinsUniversity Press. l65p.Financial Post. 1987. B.C. hikes for tax bite. 21 Sept.Furubotn, Birik G. 1985. The Gains from Privatization: A General Equilibrium Perspective.Working Paper. University of Texas at Arlington. Department of Economics.127Gillespie, D.W. 1991. Small Business Forest Enterprise Program I Woodlot Program. Report toB.C. Minister of Forests. Webber and Company, Kamloops. B.C. V2C 3K8. 20p.Gillis, Malcolm. 1992. Forest Concession Management and Revenue Policies. Pp: 139-75 inNarendra P. Sharma (ed.) Managing the World’s Forests: Looking for Balance between Conservationand Development. Kendall/Hunt Publishing Company. P.O.Box 539, Dubuque, Iowa. USA.Goldberger, Auther S. 1968. Topics in Regression Analysis. The Macmillan Company. New York,USA. l43p.Haley, David and J. Leitch. 1992. The Future of Our Forests- Report of the British ColumbiaForest Resources Commission: A Critique. Canadian Public Policy XVffl (1):47-56.Heaps, T. 1988. An Econometric Analysis ofLog Production in Coastal British Columbia. WorkingPaper 108. Forest Economics and Policy Analysis Research Unit. University of B.C., Vancouver,Canada.Hyde, William F. and David H. Newman. 1991. Forest Economics in Brief-with SummaryObservations for Policy Analysis. Report to the World Bank. 1 17p+appendix.Hyde, William F., R. Mendelson and R. Sedjo. 1991. The Applied Economics of TropicalDeforestation. Unpublished discussion paper.Jensen, Michael C. and Meckling, William H. 1979. Rights and Production Functions: AnApplication to Labour-Managed Firm and Codetermination. Journal of Business 52 (4):469-506.Judge, J.G., R.C. Hill, W.E. Griffiths, H. Lutkepohl and Tsoung-Chao Lee. 1988. Introduction tothe Theory and Practise of Econometrics (second edition). John Wiley & Sons.Kahneman, Daniel, Jack L. Knetsch and Richard H. Thaler. 1990. Experimental Tests of theEndowment Effect and the Coase Theorem. Journal of Political Economy 98 (6):1325-1348.Leffler, Keith B. and R.R. Rucker. 1991. Transaction Costs and the Efficient Organization ofProduction: A Study of Timber-Harvesting Contracts. Journal ofPolitical Economics 99:1060-1087.Libecap, Gary D. 1989. Contracting for Property Rights. Cambridge University Press. New York,USA. l32p.128Lintner, J. 1965. The Valuation of Risk Assets and the Selection of Risky Investments in StockPortfolios and Capital Budgets. Review of Economics and Statistics 47:13-37.Luckert, Martin K. 1988. The Effect of Some British Columbia Forest Tenures on the Distributionof Economic Rents, the Allocation of Resources, and Investment in Silviculture. Unpublished Ph.Dthesis. University of British Columbia. 238p.Luckert, Martin K. 1991. The Perceived Security of Institutional Investment Environments of SomeBritish Columbia Forest Tenures. Canadian Journal of Forest Research 21:318-325.Luckert, M.K. and David Haley. 1989. Funding Mechanismsfor Silviculture on Crown Land: Status,Problems and Recommendations. Working Paper 131-A. Forest Economics and Policy AnalysisResearch Unit. University of British Columbia. 59p.Luckert, M.K. and David Haley. 1990. The Implications of Various Silvicultural FundingArrangement for Privately Managed Public Forest Land in Canada. New Forest 4:1-12.Luckert, M.K. and David Haley. 1991. Canadian Forest Tenures and the Silvicultural InvestmentBehaviour ofRational Firms. Working Paper 166. Forest Economics and Policy Analysis ResearchUnit, University of British Columbia. l7p.Malkiel, B.G. 1990. A Random Walk down Wall Street. W. W. Norton & Company, New York,USA. 440p.Mighot-Adholla, S., P. Hazell, B. Barel, and F. Place. 1990. Land Tenure Reform and AgriculturalDevelopment in Sub-Saharan Africa. Washington, DC: World Bank, ARDD, Unpublished discussionpaper.Ministry of Forests and Lands. 1987. News Release. Victoria, B.C. September, 15.Ministry of Forests. 1993. Annual Report of 1991-1992. Victoria, B.C.New Zealand Forestry Corporation. 1989. Sale ofState Owned Forests in New Zealand: Prospectus.Wellington: New Zealand Forestry Corporation Limited.Palmquist, R.B. and L.E. Danielson. 1989. A Hedonic Study of the effects of Erosion Control and129Drainage on Farmland Values. American Journal of Agricultural Economics 7 1:55-62.Pearse, P.H. 1965. Distortions in the Market for Forest Land. Forestry Chronicle. 41 (4):406-418.Pearse, P.H. 1976. Timber Rights and Forest Policy in British Columbia. Report of RoyalConiniission on Forest Resources of British Columbia. Queen’s Printer. Victoria (2 vols).Pearse, P.H. 1985. Obstacles to Silviculture in Canada. Forestry Chronicle April:91-96.Pearse, P.H. 1988. Property Rights and the Development of Natural Resource Policies in Canada.Canadian Public Policy XIV (3):307-320.Pearse, P.H. 1990a. Forest Tenure Policy in Canada: The Inte,face ofPrivate and Public Interests.Paper prepared for the IUFRO XIXth Word Congress, Montreal, Canada.Pearse, P.H. 1990b. Introduction to Forestry Economics. University of British Columbia Press.204p. Vancouver, B.C.Pearse, P.H. 1992. Evolution of the Forest Tenure System in British Columbia. Faculty of Forestry,University of British Columbia. Vancouver, B.C.Pearse, P.H., A.J. Lang and K.L. Todd. 1986. Reforestation Needs in British Columbia: Clarifyingthe Confusion. Information Report 85-13. Forest Economic and Policy Analysis Unit, University ofBritish Columbia. Vancouver, B.C.Pearse, P.H. 1993. Forest Tenure, Management Incentives and the Search for SustainableDevelopment Policies, Pp: 77-96 in W.L. Adamowicz, W. White and W.E. Phillips (eds.) Forestryand the Environment: Economic Perspective. CAB International. United Kingdom.Pejovich, Svetozar. 1984. Origins and Consequences of Alternative Property Right. Pp: 163-175in Adrian B. Gamache (ed.) Selling the Federal Forests. The University of Washington, College ofForest Resources. Seattle, WA.Price Waterhouse. various years. Forest Product Industry Survey.Scholes, M. and J. Wiffiams. 1977. Estimating Betas from Nonsynchronous Data. Journal of130Financial Economics 5:309-327.Schwert, G.W. 1981. Using Financial Data to Measure Effects of Regulation. Journal ofLaw andEconomics 24:121-158.Schweitzer, D.L., Robert W.S. and Schailau, C.H. 1972. Allowable Cut Effect: Some Physical andEconomic Implications. Journal of Forestry 70 (7):415-418.Schwindt, Richard. 1992. Report of the Commission of Inquiry into Compensation for the Takingof Resource Interests. Queen’s Printer for British Columbia. Victoria. 150p+appendices.Scott, Anthony. 1990. The Market for Characteristics of Property Rights. Unpublished discussionpaper. Department of Economics University of British Columbia. Vancouver, Canada V6T 1Z3.Scott, Anthony. 1991. Individual Right to Use The Forest. Department of Economics, Universityof British Columbia, Vancouver, Canada V6T 1Z3. 209p.Sharpe, W.F. 1964. Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk.Journal of Finance 19:425-442.Smith, D.M. 1986. The Practice of Silviculture. John and Wiley and Sons. USA.Spitzer, J.J. 1982. A primer on Box-Cox Estimation. Review ofEconomics and Statistics 62:307-313.Sterling Wood Group Inc. et al. 1986. Comments on US Department of Commerce October 1986Preliminary Determination. Supplementary Joint Report to Canadian Forest Industries Council,Vancouver, B.C.Statistics Canada. various years. Canadian Economic Observer. Catalog No. 62-001.Theil, H. 1971. Principles of Econometrics. NY: John Wiley and Sons.Thompson, R. 1985. Conditioning the Return-Generating Process on Firm-Specific Events: ADiscussion of Event Study Methods. Journal ofFinancial and Quantitative Analysis 20(2): 151-68.131Uhier, R.S. 1991. Canadian Public Timber Policing and the Great Subsidy Debate. Pp: 73-93 inUhier, R.S. (ed) Canada-United States Trade in Forest Products. UBC Press. Vancouver, B.C.Vancouver Sun. 1986. Premier Says other will Follow if B.C. Hikes its Stumpage Rate. 9September.Vancouver Sun. 1987. Forestry Policy Fuels Stock Sell off. 16 September.Vertinsky, I., D.A. Wehrung and S. Brumelle. 1990. Priorities for Silviculture Investment: Public,Government and Industry Perspectives. Forest Economics and Policy Analysis Research Unit,University of British Columbia.Vincent, Jeffrey R. 1990. Rent Capture and the Feasibility of Tropical Forest Management. LandEconomics 66 (2):213-223.Vrooman, D.H. 1978. An Empirical Analysis of the Determinants of Land Values in AdirondakPark. American Journal of Economics and Sociology 37:165-177.Williamson, Oliver B. 1989. Transaction Cost Economics. Chapter 3 in R. Schmalensee and R.D.Willig (ed.) Handbook of Industrial Organization.Washburn, C.L. 1990. The Determinants of Forests Value in the U.S. South. Unpublished Ph.Dthesis. Yale University.Woodbridge, P. and N.S. Mackenzie. 1992. A Strategic Framework for Growth in BritishColumbia’s Forest Sector: Vision 2000. Report Prepared for Forest Summit Conference 1992,Vancouver, B.C.Zeliner, A. 1962. An Efficient Method of Estimating Seemingly Unrelated Regressions and Testsfor Aggregation Bias. Journal of the American Statistical Association 57:348-368.Zhang, D. and C.S. Binkley. 1994. The Inter-temporal Efficiency of Vancouver Log Market andStorage Returns. Journal of Canadian Forest Research 24:550-557.Zinkhan, F.C. 1988. The Stock Market’s Reaction to Timberland Ownership RestructuringAnnouncements: A Note. Forest Science 34:815-819.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075296/manifest

Comment

Related Items