Institutional determinants of profitable commercial forestry enterprises among First Nations in Canada

Institutional determinants of profitablecommercial forestry enterprises among FirstNations in Canada

Ronald Trosper, Harry Nelson, George Hoberg, Peggy Smith, and William Nikolakis

Abstract: This paper uses survey information to examine several common assertions about the institutional prerequisitesfor successful profitability when a First Nation enters an economic enterprise either independently or in joint effort withan outside firm. In the winter of 2004–2005, we interviewed managers on both the First Nations and private sides of jointventures and other business alliances in Canada, to determine what affected their recent profitability experience. We gath-ered information on the ages, sizes, and activities of the firms. We also gathered information about the firms’ managementstructures and relationship with the First Nation, and the characteristics of the government of the First Nation. With a sam-ple size of 40 firms that responded, we found that several institutional characteristics affected profit positively: strong sep-aration of management from band governance, participation in management planning, and the use of staggered terms inband council elections. We found that the likelihood of profitability decreased if the band had been in third party manage-ment as well as if there was formal participation of elders or hereditary chiefs in decision making. We offer interpretationsof these results.

Resume : Cet article utilise les donnees d’un sondage realise au cours de l’hiver 2004–2005 afin d’examiner plusieurs as-sertions courantes au sujet des prealables institutionnels requis pour atteindre la profitabilite des entreprises initiees par desPremieres nations agissant seules ou en partenariat avec des acteurs prives. Nous avons interviewe des gestionnaires impli-ques dans de tels partenariats d’affaires au Canada et provenant tant des Premieres nations que du secteur prive. Nous vou-lions determiner ce qui affectait la profitabilite qu’ils avaient connue recemment. Nous avons rassemble de l’informationsur l’age, la taille et les activites des entreprises. Nous avons aussi recueilli des donnees sur la structure de gestion des en-treprises, les relations avec les Premieres nations et les caracteristiques de la gouvernance en vigueur dans les commu-nautes d’accueil. Un echantillon, regroupant les quarante entreprises qui ont repondu, nous permet d’observer queplusieurs caracteristiques institutionnelles, comme la separation claire entre la gestion des projets et la gouvernance descommunautes, la participation dans la planification manageriale et le recours a des mandats echelonnes lors des electionsau conseil de bande, affectent positivement les profits. Nous avons aussi remarque que les chances de profitabilite diminu-ent si les conseils de bandes jouent un role en tant que de tierce partie dans la gestion des entreprises et si les aınes ou leschefs hereditaires participent formellement au processus decisionnel. Nous tentons d’interpreter ces resultats.

[Traduit par la Redaction]

IntroductionThis paper investigates several hypotheses about the

causes of profitability among forest enterprises establishedby First Nations either independently or in joint effort withan outside firm. We gathered data with a telephone surveyof such enterprises in the winter of 2004–2005. We wereparticularly interested in the importance of institutional char-acteristics of the relationship between the enterprises and theband government. We were also interested in other varia-bles, such as the age and size of the enterprises and the rela-tionships between the enterprises and the firms that joinedwith the bands in establishing the joint forestry enterprise.

We found significant relationships between some institu-tional variables and profitability.

After providing some background explaining the impor-tance of the topic and the reasons for our hypotheses, we de-scribe our data collection, our statistical approach and modelspecification, and then report results. The final section of thepaper provides a discussion of the results.

Background

Government officials and community leaders have be-come increasingly concerned about the lower socioeconomic

Received 18 September 2006. Resubmitted 9 March 2007. Accepted 28 August 2007. Published on the NRC Research Press Web site atcjfr.nrc.ca on 1 February 2008.

R. Trosper,1 H. Nelson, and G. Hoberg. Faculty of Forestry, 2045-2424 Main Mall, The University of British Columbia, Vancouver,BC V6T 1Z4, Canada.P. Smith. Faculty of Forestry and the Forest Environment, Lakehead University, Thunder Bay, ON P7B 5E1, Canada.W. Nikolakis. Forestry Economics and Policy Analysis Research Unit, 2045-2424 Main Mall, The University of British ColumbiaVancouver, BC V6T 1Z4, Canada.

1Corresponding author (e-mail: [email protected]).

226

Can. J. For. Res. 38: 226–238 (2008) doi:10.1139/X07-167 # 2008 NRC Canada

status, community well-being, and generally poorer levels ofhealth of aboriginal communities in Canada.2 The CanadianRoyal Commission on Aboriginal Peoples (RCAP 1996), inits final report, determined that the cost to Canadian tax-payers for the low socioeconomic status of aboriginal peoplewould increase from $7.5 billion in 1996 to $11 billion in2016 (Anderson 2002). The RCAP concluded that a concertedeffort by provincial, territorial, and federal governments onaboriginal economic development would regenerate socialfabric, redress economic inequity, and contribute $375 mil-lion to the Canadian economy by 2016 (Anderson 2002). TheRCAP framed the discussion in terms of ‘‘an awareness ofgoals redressing identity, tradition and modernity. . .,’’ inwhich economic development was described as importantto social regeneration and self determination and not solelyabout individual wealth accumulation (CANDO 1994).

Given that about 80% of First Nation communities are lo-cated within the Canadian commercial forest zone (Gysbersand Lee 2003), the federal and provincial governments havejointly encouraged forest-based development for aboriginalcommunities. Most of these economic enterprises have takensome form of cooperative activity organized among FirstNation communities and firms already established in the for-est sector (NAFA-IOG 2000; Hickey and Nelson 2005).

This paper explores the institutional conditions (rules fordecision making) that promote profitability in all such FirstNation enterprises. After a brief literature review, the paperdescribes the data collected and the statistical methods used.Our results are focused on the relationship between a fewinstitutional factors and the profitability of the firms we in-terviewed. We find that several institutional variables relatedto First Nations have a significant impact on the odds that aforestry enterprise will be profitable.

Forest sector firms involving both First Nations and outsidefirms are prevalent in Canada for several reasons, includingprovincial government policy choices, legal obligations toprotect aboriginal and treaty rights, and aboriginal controlof forest lands through modern-day land claims. In someprovinces, a common approach to accommodating FirstNation concerns regarding land use and forestry operationshas been to provide short-term timber harvesting rightsthrough existing tenure systems. The National AboriginalForestry Association (NAFA 2003) estimates that aborigi-nal groups and individuals hold about 7 �106 m3 or 4%of the total crown timber across Canada. NAFA pointsout, however, that aboriginal groups are over-representedin minor tenures such as cutting permits for personal orspecialty use and under-represented in tenures with long-term management responsibility or significant timber vol-umes. This situation is most common in provinces inwhich First Nations have not signed agreements limiting

their aboriginal title to forest lands or where that title is inquestion, such as British Columbia and Newfoundland.Court decisions outlining how provincial governments mustprotect aboriginal and treaty rights have provided compa-nies with an incentive to enter into relationships with FirstNations that promise to bring stability rather than conflictto the business environment. Finally, in some regions of thecountry, modern day land claims have given First Nationscontrol over settlement lands. Cases in point are the Nisga’ain British Columbia, First Nations in the Yukon and North-west Territories, and the Cree of Northern Quebec.

Joint activities between First Nations and firms vary inform and motivation.3 Where First Nations receive timberallocations directly, they may choose to enter into long-termarrangements to sell or market the timber. In other cases,First Nations may have agreements with local firms, whichsubcontract a portion of the harvest to the First Nations oroffer employment contracts. First Nations may also seekpartners through which they can access capital and exper-tise, especially if they are considering more capital-intensiveactivities such as manufacturing. Jointly organized enter-prises can offer an opportunity to increase the size and (or)scope of the activity to overcome issues around limited re-sources, achieve certain economies of scale or scope, or togain access to markets. Other authors have suggested thatFirst Nations may enter into such arrangements to developcapacity (NAFA 2002). Finally, often businesses in the for-est products industry, regardless of whether they are band-owned, establish long-term relationships with First Nations,especially around harvesting agreements and processing fa-cilities in order to succeed.

These arrangements differ from those more commonly ob-served among firms in the business sector in several impor-tant ways. Many of these enterprises (or at least the FirstNation) have multiple goals and objectives, of which economicdevelopment may be only one. The type of economic develop-ment the First Nation community is seeking may differ amongcommunities. Hickey and Nelson (2005), in a recent surveyof forest sector agreements, identified seven separate func-tions or goals for these efforts: conflict avoidance, revenue,employment and job creation, capacity building, resource ac-cess, resource management, and social benefits.

While provincial and federal policies continue to encour-age the use of such joint arrangements as well as promoteforest sector opportunities for First Nations more generally,to date there has been no systematic study of the perform-ance of these enterprises or contributions to their profitabil-ity, although a number of studies have identified what arebelieved to be potential factors. Wilson and Graham (2005)provide an overview of provincial policies as well as de-tailed analyses of each provincial and territorial context and

2 Aboriginal peoples in Canada are defined in the Constitution Act (1982) to include Indians, Inuit, and Metis. ‘‘Indians’’ refers to a broadgroup of aboriginal people, some of who are recognized as ‘‘status’’ Indians by the federal government, which has an administrative andtrust responsibility under the Indian Act. Groups of status Indians were granted small land bases called reserves and governed as ‘‘bands.’’The term band was seen to be paternalistic; with the rise of the self-determination movement, terminology changed from band to FirstNation. In this paper, we focus on bands, using the term band and First Nation interchangeably. In a few cases, businesses are owned andoperated by a group of First Nations in a Tribal Council. For more background on the nuances of the terminology used for aboriginalpeoples in Canada, see Department of Indian Affairs and Northern Development (2002).

3 We avoid using the term joint venture to describe these enterprises, as there is not a consistent pattern. We do note that jointly organizedenterprises can encompass from more formal relationships (the establishment of a new entity organized along corporate lines as a jointventure or a partnership) to informal agreements (outside firms allocating a portion of their harvest to the band-owned enterprise).

Trosper et al. 227

# 2008 NRC Canada

offer a number of conclusions and recommendations for im-proving First Nation forest sector participation. NAFA andthe Institute of Governance (NAFA-IOG 2000), IOG(1998), Graham and Wilson (2004), and Hickey and Nelson(2005) all discuss various factors that are believed to con-tribute to the success of these ventures.

Fundamental to a discussion of ‘‘success’’ in such enter-prises is the recognition that not everyone in First Nationscommunities defines success solely in terms of profit(Natcher and Hickey 2002), although in our sample moreidentified profit as the goal than any other. Some First Na-tions community members also pay attention to employ-ment. Many are concerned that following the developmentmodel of mainstream society means compromising othervalues such as harvesting at sustainable rates and caring forother aspects of the forest besides the production of woodfibre. Several authors (Elias 1995; Myers 1996; Duhaime etal. 2004) have pointed to aboriginal communities’ desire, es-pecially in the north, for the maintenance of a ‘‘mixed’’economy that includes elements of traditional land-basedeconomic pursuits such as hunting, fishing, trapping, andgathering. Wuttunee (1992, 2005) focuses on ‘‘what choiceswill be made to meet community and individual goals’’ inindigenous communities and the more fundamental ques-tions that economic development raises for communities:‘‘Will we want communities where the environment is cher-ished and elders and traditions honoured, or will we try tomaximize our returns on investment? Are these objectivesmutually exclusive? What does it mean to us to participatemeaningfully in the economy? What are the benefits, andwhat are the costs?’’

Factors influencing profitable economicenterprise

Many reports and studies identify what could be potentialfactors for successful economic development. Some of thesestudies refer to the results of the Harvard Project on Ameri-can Indian Economic Development, which identified factorsthey believe to be essential to success for economic devel-opment on American Indian reservations. These results havebeen discussed extensively in regards to aboriginal eco-nomic development in Canada (see, for example, AtlanticCanada Opportunities Agency 2003; Graham and Edwards2003; Office of the Auditor General 2003) and are also citedin studies of forest-based development in aboriginal com-munities (see, for example, NAFA 2002; Whiting 2001;Graham and Wilson 2004).

The originators of the Harvard Project, Stephen Cornelland Joseph Kalt, in a series of papers (Cornell and Kalt1992, 1995, 1997), identify three basic elements that theybelieve are crucial to successful tribal economic develop-ment: sovereignty and (or) independence, institutional mech-anism for governance, and cultural match. They argue that

other factors, such as resource endowment and distance tomarkets, are of lesser importance. Indeed their results pointto the importance of factors involving the development ofthe institutional framework as essential ingredients that pro-vide the basis for success. Institutions are defined not as theestablishment of actual organizations but as the rules (formaland informal) governing the economic activity.

In the Cornell and Kalt framework, sovereignty has twocomponents: de jure and de facto. De jure sovereignty refers tothe legal power tribes have in the US in regards to their resour-ces and jurisdiction in other matters (such as the ability to tax,establish business codes, and regulate business activities ontribal land). De facto sovereignty occurs when communitiesact as if they had de jure power or act in areas in which formallegal rules provide them room for action. In both cases, com-munities have the ability to make decisions affecting them-selves, so that decision makers within the communities bearthe consequences of their actions: while they reap the benefits,they must also face the costs of poor decisions.

Effective institutions means that there are rules and proce-dures to check rent-seeking behaviour (and prevent politicalopportunism) and the capacity and ability to implementthose rules. The Harvard Project results also suggest thatthere are preferred ways for aboriginal communities to or-ganize their political and economic institutions. Cornell andKalt (1992) assert that effective economic institutions haveseveral important characteristics. First, they have disputeresolution procedures that work. Second, they insulate eco-nomic decisions from political interference. Third, theyclearly identify the roles and responsibilities people have,not only within the enterprise but also within the govern-ment. These results can be achieved in several differentways. One of the ways to create a separation between eco-nomic and political decision making is through the establish-ment of boards of directors or bodies separate from politicalbodies. There may also be formal or informal rules prevent-ing elected officials from being involved in day-to-day oper-ations.4 To ensure impartiality, independent bodies canadjudicate disputes or, at the very least, there may be formalprocedures to address grievances. In terms of political insti-tutions, stability of the rules governing the environment inwhich the enterprise is operating also enhances performance.This can be achieved by having longer electoral terms orstaggered terms such that political leadership does not com-pletely turn over every few years. Training is also an impor-tant component, not only for the enterprise itself, but alsofor political leadership. It can help clarify the roles and ob-jectives for managers, employees, and politicians, as well asproviding these parties with the skills necessary to effec-tively carry out their responsibilities. Finally, reporting canalso be a key ingredient. Not only does it provide a way tocontinually assess performance and ensure accountability,but it also ensures that leaders and the community are awareof what is going on (Grant and Taylor 2007).

4 Jorgensen (2000) extended Cornell and Kalt’s research, investigating the importance of alternative governance mechanisms (in addition tomarket mechanisms) in explaining variations in the outcomes of tribal forestry enterprises and the success of tribal housing developments.She found that tribes used a variety of internal and non-market governance mechanisms. These mechanisms ranged from a reliance onnorms to ensuring participants carried out their responsibilities to institutional structures that explicitly incorporated performance mechan-isms. In some cases, tribes may formally create internal checks and balances to ensure satisfactory outcomes; Harris et al. (1995) providessuch an example of how the Hoopa in California use two different bodies within the tribal organization to ensure the appropriate manage-ment of their forests.

228 Can. J. For. Res. Vol. 38, 2008

# 2008 NRC Canada

Cultural match means that not only the goals but also theinstitutions governing the activity (including politicalbodies) are seen as legitimate by aboriginal communitymembers. Testing such legitimacy in this study was difficult,primarily because the cultural basis of legitimacy would varyamong the aboriginal communities in complicated ways. Weasked a number of questions in the survey that we hopedwould give us some insight into such cultural issues.

Similarly, Hindle and Landsdowne (2005) report that in-digenous opinion leaders also recognize the importance ofculture and autonomy, the institutional factor of accountabil-ity, and twin capacity skills of cultural and business knowl-edge and ability. While not stating a requirement forseparation of business from politics, Hindle et al. (2005) pointto the need for stable good leadership. In fact, they discusshow some ventures have boards to add stability in the eventof a change in government, in effect creating a separation ofbusiness from politics. Peredo and Chrisman (2006) discusssocial capital and the importance of communities (networksin Hindle’s work) in which entrepreneurship is embedded, an-other way of stating that institutions matter and that these in-clude specific organizational forms and the use of norms astools to sanction behaviour that threatens business success.

All these elements work at several levels. First, at a strate-gic level, a general agreement on community goals sup-ported by appropriate institutions provides the frameworkwithin which decision makers can develop a vision that leadsto business enterprises that will ‘‘fit’’ and how economic ac-tivity can help achieve those goals. Second, in the day-to-dayoperations of the enterprises themselves, these elements pro-vide an environment that ensures sustained economic activitythrough aligning incentives, curbing rent-seeking activity,and ensuring accountability. Cornell and Kalt (1992) arguethat the project-based approach that underpinned most ofthe traditional development efforts in the 1960s and 1970sin the US was doomed to failure because the approach wasdriven by fund availability, rather than generated by anagreed upon tribal ‘‘vision.’’ Indeed, Grant and Taylor (2007)argue that connection to community objectives is a key in-gredient for economic success.

Differences in the policy settingBecause of notable differences between the Canadian and

US contexts, this study focuses on the effects of institutionson economic outcomes. The Harvard Project focused onAmerican Indian tribes pursuing tribal-based developmentutilizing tribally owned and controlled resources (e.g., tim-ber, mineral resources, fish, and game). Their results arebased on the sovereignty tribes enjoy in terms of those re-sources in the US where tribes have won the ability to di-rectly assert control and management over resources onreservation lands. In Canada, sovereignty is sharply limited,not varying much among First Nations except for the fewthat have negotiated special agreements with the federal andprovincial governments. In addition, the effective power ofprovinces in forest land use limits the role of cultural viewsin structuring forest enterprises by focusing on timber yields,

limiting management options, and defining administrativeareas that have little (if any) connection with traditional areas.

Within Canada, most First Nation communities face threeimportant constraints on their involvement in the forest sec-tor. The first is that most reserve land bases are too small inarea to provide any hope for economic self-sufficiency(White-Harvey 1994). These reserves are under federal own-ership and administered through the Indian Act, which givesthe Minister of Indian Affairs the authority to make all deci-sions on behalf of First Nations. Second, most commercialforest area in Canada is under the control of provincial gov-ernments that have allocated access to the commercial forestland and forest resources to forest companies through long-term evergreen tenures. Even thought the federal govern-ment has constitutional authority for ‘‘Indians and lands re-served for the Indians,’’ it has been reluctant to intervene onbehalf of First Nations, as provinces have constitutional re-sponsibility for natural resources (Ross and Smith 2002).Therefore, many First Nations in Canada cannot gain accessto forest resources unless it is through some formal agree-ment with the provincial government or an arrangementwith an industrial partner that has a provincial licence.Third, the Indian Act constrains bands in the type of govern-ment forms they can choose and reduces the ability of theband to seek outside capital. As a result, First Nations inCanada do not generally enjoy the same degree of freedomor autonomy in managing their resources as do those UStribes with reservations of a significant size (although thishas been changing in recent years with the signing of self-government agreements and new land claim settlements).

Therefore, although sovereignty and the effective inde-pendence it can grant did play a key role in the HarvardProject’s findings for the US, we may not observe the sameeffect in Canada. This could be because First Nations do nothave comparable power (and therefore do not have the op-portunity to assert sufficient control to realize those benefits)or because the underlying economic conditions cannot sup-port the development of the necessary infrastructure (i.e.,too small a reserve means it would not pay to have a timbermanagement staff as it could not fund itself). Remotenessmay also play a role by increasing the cost of supplies andaccessing markets. First Nations in Canada may thereforeface a more constrained set of choices in terms of their ap-proach to economic development and how they structuretheir institutions.5 Indeed, NAFA suggests that an additionallink that should be considered in the context of aboriginaleconomic development in Canada is ‘‘the powerful drivingforce that economic realities can have on processes of gov-ernance. . .’’ (NAFA 2002, p. 11).

Finally, cultural factors in Canada may also influence al-ternative approaches to developing or managing resources.These factors may include not only First Nation views oneconomic development and forest management, and the rela-tionship of these views to traditional values, but also Cana-dian expectations and norms of forest management. WithinCanada, there has been a long tradition of expecting naturalresource development to deliver not only economic benefits

5 Louie speaks to the challenges of these constraints: ‘‘We agree, in theory, that the separation of politics from business is important. Butpractically, the talent pool of people in our communities that can undertake band administration and economic development is so shallowit is hard to separate the two’’ (McBride 2001, p. 13).

Trosper et al. 229

# 2008 NRC Canada

but also social benefits. This may influence the expectationsof aboriginal communities and others in developing forest-based enterprises. These cultural factors may legitimize ob-jectives other than strictly economic ones.

Given the variety of goals for these enterprises, the study ofdeterminants of profitability is justified for two main reasons:positive profit is needed for sustainability and enterprise man-agers rank profit-making equal to provision of employment.If an enterprise is not economically viable, then it is unlikelyto meet any other objectives it may have. ‘‘Social programsmust be supported by business programs’’ according to[George] Ho Lem [of the Lheit-Lit’en Development Corpora-tion]. They have found it difficult to integrate training andskill building with business initiatives. . . ‘‘As in any otherbusiness,’’ Ho Lem says, ‘‘if the business isn’t making moneyyou shouldn’t be doing it’’ (McBride 2001, p. 38). While max-imizing profit may not be an ultimate goal for a community,their enterprises have to be profitable to survive and provideemployment and other benefits. That profit is an instrumentfor other goals is recognized by Anderson (2002) and Hindleet al. (2005). We are not claiming that profit maximizationis the only goal, only that having revenue cover costs allowsbusinesses to survive and provide other benefits.

Enterprise managers recognize the need for profit. Increas-ing revenue or employment, if not both, were expressed asgoals by 60 of the 61 survey respondents in Hickey and Nelson(2005). In our survey, profit and employment were listedequally when we asked ‘‘What are the business objectives?’’We also asked ‘‘What three things are considered importantin evaluating the success of the business?’’ The most frequentanswer for the most important thing was profit and themost frequent answer for the second was employment.

In summary, bands within Canada can control the institu-tional structure of enterprises to a degree. They are less ableto exert sovereignty or impose their cultural values on forestdevelopment. Successful enterprises, by providing a flow ofprofit, can assist in attaining de facto sovereignty because ofan increase in financial resources. Cultural match in thestructure of an enterprise is limited both by provincial poli-cies regarding how they expect business to be conducted andby the need to engage with corporate partners. Because in-stitutional structure can be selected by band governance, dif-ferent decisions were made in the different enterprises in oursurvey and we are able to examine variation in the conse-quences of those decisions.

Research methodologyThe results reported in this paper are from a larger re-

search project comprised of both survey work and case stud-ies. The case studies provide in-depth consideration of theissues, as shown in Boyd (2006). Case studies, however,cannot provide the kind of statistical tests available from asurvey. We followed our survey with a workshop with re-search participants to discuss the survey results. To fully ex-plore what the determinants of profitability might be inCanada, we surveyed firms in the forestry sector that hadFirst Nation participation, with an emphasis on enterprisesinvolving both First Nation and industry partners. We used

a questionnaire to gather data in four different areas: (i)business characteristics and performance, (ii) business type,ownership, and control, (iii) goals and objectives, both forthe business and the community, and (iv) characteristics ofbusiness and community political structure.

Our sample size is comparable to those in other studies.Out of an original 173 potential businesses identifiedthrough various databases and internet searches, we deter-mined that 161 may have still been in business. We wereable to make good contact with 55 of those; the rest did notreturn calls or did not answer their telephones after repeatedattempts. We were successful in convincing 40 enterprises tosign consent forms and participate in our telephone inter-views, which took one half to a full hour. Table 1 gives thebreakdown of the sample group by type of business. Hickeyand Nelson (2005) interviewed 61 respondents, but 16 ofthese reported memoranda of understandings and protocolsthat were not necessarily businesses, meaning that their sam-ple number of businesses was about 45. The scoping paperNAFA-IOG (2000) reported information on 47 entities.

Telephone surveys provide the opportunity to reach manyrespondents; but telephone surveys also have limitations. Asthe following discussion shows, we were not able to obtain agreat deal of detailed information about rates of profit oramounts of invested capital. The responses to the questionsrepresent a cross-sectional ‘‘snapshot’’ of conditions at thetime of the survey. As mentioned above, other publicationsfrom the research project report case studies (Boyd 2006),which can go into more detail.

While our initial interest was the profitability of joint ven-tures and other partnership types, we found that in develop-ing the study, the distinction between band-owned and otherenterprises was not that clear and that there was overlapamong the categories. Even among the respondents whoself-identified as band-owned businesses, several identifieddifferent types of contracts with outside firms. For instance,5 of the 10 band-owned enterprises had tenure or harvestingcontracts, while another 2 had employment contracts. Onlythree indicated no explicit business arrangements with out-side entities. Our results, however, found no significant dif-ferences between these two types.6

Table 1. Number of firms surveyed.

Type of business

Bandowned

Joint venture,partnership, orcontract Total

Original list 105 68 173Not in business 6 6 12Total population 99 62 161No good contact 83 23 106Contact made 16 39 55Declined or too busy 6 9 15Interviewed 10 30 40Percentage interviewed

Of population 10 48Of contacted firms 63 77

6 When added to the equation reported in Table 9, the dummy variable for being a joint venture has a p value of 0.541. See the last row ofTable 10 for evidence that the dummy variable does not affect results when the estimation method uses ordinary least squares.

230 Can. J. For. Res. Vol. 38, 2008

# 2008 NRC Canada

We expected that obtaining information on profitabilitywould be difficult. To avoid asking about profitability im-mediately, which might deter respondents as it requires ayes or no answer, we first introduced the idea by asking re-spondents ‘‘Are you satisfied with the profit record this busi-ness has established to date?’’ We then asked for anexplanation of their satisfaction or dissatisfaction. We nextasked, ‘‘Taking all that into consideration, did the businessmake a profit last year?’’ and ‘‘On average, over the past 3years, has the business been profitable?’’ If we knew froman earlier question that the firm had not been in businessfor 3 years, we asked, ‘‘Over the life of the business, has itbeen profitable?’’ (Two firms were only 2 years old.) Theanswers to these two questions became our dependent varia-ble. We also asked for information on changes in profits andin employment over the last 3 years or over the life of thebusiness. With this approach, we accomplished the goal ofobtaining reliable answers to our question of primary inter-est, the profitability of the business. Out of our 40 busi-nesses, 25 reported being profitable in the last 3 years,while 15 did not (Table 2).

Descriptive characteristics of the sampleWe also inquired about the nature of the business to con-

trol for factors such as the influence of size on profitability.We anticipated that information on total revenue and totalinvested capital would be somewhat difficult to obtain be-cause most businesses are sensitive about information thatmay affect their competitiveness. Therefore, respondentswere asked either to give a dollar figure or to select from alist of six ranges for the previous year. Since not everyonegave a dollar figure, we have converted specific figures into

the ranges and used dummy variables for revenue and capi-tal invested as measures of scale.

Table 3 reports the businesses by revenue category. Fourof the businesses did not answer either of the revenue ques-tions. Because many firms did not give an exact figure, wereclassified all firms into the six categories. The 36 busi-nesses that did respond to the revenue questions provided aspread over revenue categories. Only four firms had revenueabove $10 million.

Table 4 reports long-term invested capital by type of firm.Twenty of the firms reported invested capital that was$100 000 or less. None reported invested capital in the cate-gory between $100 000 and $499 000. Six reported investedcapital above $10 million.

Table 5 gives the numbers of employees by type of busi-ness. Most firms (18 of the 40) had between 11 and 29 em-ployees. Comparing Table 5 to Table 6 (which reports dataon full-time employment) shows the large role of seasonalemployment in determining numbers employed. We notedthat most firms have very few full-time employees.

Table 7 reports the ages of the businesses. The oldest hasbeen in existence for 34 years. The majority, or 68% ofbusinesses, were <10 years old and 24% of the businesseswere <6 years old. Table 8 illustrates the different activitiesthat the businesses were engaged in. The table reveals thatlogging and silviculture were the two predominant types ofbusiness activity; of the 15 firms involved in forest manage-ment planning, all were engaged in logging and 11 were en-gaged in silviculture.

Statistical techniquesWe did not ask for profit amounts; we asked simply

whether or not profits were positive during the previous 3years. A positive answer was coded as 1; other answers

Table 2. Enterprises by type of business that reportedwhether or not they were profitable on average over theperiod 2000–2003.

Profitable last 3 years

No Yes TotalType of business

Joint venture 6 11 17Partnership 3 4 7Contract 3 3 6Band(s) owned 3 7 10

Total 15 25 40

Table 3. Revenue category by type of business.

Type of business

Revenue categoryBandowned

Joint venture,partnership,or contract Total

<$100 000 1 3 4$100 000 – $499 999 2 2 4$500 000 – $999 999 1 3 4$1 million – $1.99 million 2 6 8$2 million – $10 million 3 9 12>$10 million 1 3 4

Total 10 26 36

Table 4. Invested capital by type of business.

Type of business

Long-term capital investmentBandowned

Joint venture,partnership, orcontract Total

£$100 000 5 15 20$101 000 – $499 999$500 000 – $999 999 1 2 3$1 million – $1.99 million 3 6 9$2 million – $10 million 1 1 2>$10 million 0 6 6

Table 5. Number of employees by type of business.

Type of business

Number of employeesBandowned

Joint venture,partnership,or contract Total

0 –10 1 8 911–29 8 10 1830–99 0 8 8100–200 1 3 4

Total 10 29 39

Trosper et al. 231

# 2008 NRC Canada

were coded as 0. When profit is a dependent variable, we mustuse an estimation technique appropriate for a dichotomousdependent variable; ordinary least squares will give biasedcoefficients. This is because the effect of the variables cannotbe expected to be linear throughout the range of variation,which is limited to a small range (Long 1997, pp. 38–40).For a binary dependent variable such as this, the appropri-ate statistical procedure is to use either logit or probit esti-mation. We chose to use logit because of the ease ofinterpretation of the results in terms of the odds of profit-ability. Probit does not provide such easy interpretation.7The logit model provides an estimate of the logarithm ofthe odds that a particular outcome will occur. If p is theprobability that a particular outcome will occur (p lies be-tween 0 and 1), the odds that the outcome will occur arep / (1 – p). Our estimation equation is used to estimate thecoefficients of the following model:

logðoddsÞ ¼ log½p=ð1� pÞ�¼ b0 þ b1X1 þ . . .þ bnXn

where b0 is a constant, bi is the coefficient of variable Xi,and Xi is one of our independent variables.

The independent variables can be continuous or binary inthis model. We have chosen to use binary independent vari-

ables, measuring the presence or absence of a factor that wehypothesize may affect the profitability of a business.

Once a model is estimated, the coefficients can be used tocalculate the odds and the probability of any particular com-bination of independent variables. For example, if a firm haszero values for all independent variables, the odds of profit-ability are determined by the constant term and equal eb0 . Ifthe firm also has one other variable present, then the oddsare eðb0þb1Þ or eb0eb1 . The equation is nonlinear and requiresestimation by the method of maximum likelihood throughnumerical methods. When the likelihood function has beenmaximized, it provides estimates of standard errors and sig-nificance.

The small number of observations presents two problems.First, we cannot have very many independent variables. Sec-ond, the statistical properties of the maximum likelihood esti-mator approach the normal distribution only asymptotically.Long (1997) reports that the characteristics of the estima-tors in the logit equation have not been studied for smallsamples; he can only offer rough rules of thumb. Mosttroubling for our work is that he recommends that samples<100 not be used, based on his experience and consultationwith other researchers. We would have had difficultyachieving this sample size even if we had pursued withgreater effort interviewing all 161 firms still in businesson our original list. Another rule of thumb is that there beat least 10 observations per independent variable, suggest-ing a limit of four variables for this study. We end up us-ing five institutional variables. Limitations on the numberof variables prevented complicated tests involving both in-stitutional variables and characteristics of the business suchas invested capital.

In addition to not being able to justify the recommenda-tion that samples <100 are too small, Long (1997, pp. 53–54) reports contradictory advice from the literature regardingthe interpretation of p values. Some people recommend ac-cepting coefficients with larger than normal p values forsmall samples; Long himself recommends using smallerthan normal p values. Fortunately, maximum likelihood esti-mation in ordinary least squares works well with small sam-ples (Long 1997, p. 53). For the purposes of statisticalinference, but not for the evaluation of the meaning of coef-ficients, we also report our equation estimated with standard

Table 6. Number of full-time employees by type of business.

Type of business

Number of full-timeemployees

Bandowned

Joint venture,partnership, orcontract Total

0–10 2 12 1411–25 4 2 626–100 0 5 5

Total 6 19 25

Table 7. Time since establishment by type of business.

Type of business

Time sinceestablishment (years)

Bandowned

Joint venture,partnership,or contract Total

2 0 2 25 0 2 26 1 4 57 1 6 78 1 2 39 0 6 610 0 1 111 2 1 315 0 1 118 1 0 119 1 0 122 1 0 123 1 1 225 0 1 134 0 1 1

Total 9 28 37

Table 8. Type of business activity by whether or not enterprise isinvolved in forest management planning.

Involved in forest managementplanning activity

No Yes TotalLogging 18 15 33Silviculture 9 11 20Nontimber forest products 1 1 2Supplemental activities 3 7 10Trucking 0 3 3Wood products manufacturing 6 3 9Other activities 1 3 4

Total 25 15 40

Note: Some enterprises were involved in multiple activities.

7 We used Statistical Package for the Social Sciences, Version 13. SPSS Inc., Chicago, Illinois. Available from www.spss.com.

232 Can. J. For. Res. Vol. 38, 2008

# 2008 NRC Canada

linear regression. As is the case in the example Long uses, ourresults show that the statistical inference is similar betweenordinary least squares and logit, although the coefficients vary.

In summary, our reason for selecting the logit model forthis work was that it properly modeled the probability ofprofitability, which was our interest. The coefficients pro-vided ready interpretation in terms of changes in the proba-bility of profit or loss. Because of our small sample, we useordinary least squares to double check the statistical infer-ence from the logit model, while using the unbiased coeffi-cients of the logit model for interpretation.

Specification searchA third problem with the small sample is that dividing the

observations into two groups is not feasible. With sufficientobservations, one can develop a model derived by selectingsignificant coefficients in one sample, and then test themodel on the other sample. Our procedure was to start witha short list of potential variables and adopt the ones thatproved to be significant. Two errors are possible in thisprocess: selecting an irrelevant variable that happens tohave a low p value in this sample and excluding a variablethat happened not to have an effect in this sample.

In this circumstance, interpretation of significance has tobe tempered by consideration of the number of variablesthat were considered for significance (Leamer 1978). If weconsider 10 potential variables with a 10% p value cutoffpoint, then by chance one will show up significant.

We began with three sets of variables to consider: struc-tural characteristics of the business (e.g., size, activity), in-stitutional variables (e.g., separation of business andpolitics), and management skill (prior business experience,term of current manager). From the results reported in thispaper, we originally considered 12 variables that related toinstitutional conditions and 4 relating to activities. Notingmoderate significance of two of the related variables, wecombined them to create a new variable that proved to besignificant; therefore, we actually tested 17 variables. Thus,we would expect to find about one or two significant varia-bles by chance. We found more. In the ordinary leastsquares specification, three variables were significant at bet-ter than 5% and the other two were significant at better than10%. In the logit specification, we found three significantvariables at a level better than 5%, one at 10%, and reporteda fifth at 15%.

Model specificationIn our resulting specification, we use five independent

variables to examine the institutional characteristics that af-fect the probability that an enterprise was profitable on aver-age over the past 3 years.

The first variable, separation, was created to measurewhether or not there was separation between political leader-ship and business decisions. We derived this measure bycombining the responses from two questions, one askingwhether the chief and council were involved in the businesson a day-to-day basis and the other asking whether the busi-ness was directly band- or tribal council-owned or whether itwas owned by a development corporation. To qualify asseparate it had to meet both tests, that is, not have the chiefand council involved in day-to-day operations and have a

separate board of directors. Eight out of 40 of our respond-ents reported separation.

To measure the development of political rules designed topromote stability, we used staggered terms. This reflects aconscious choice by First Nations to establish staggeredterms of office to ensure that not all elected officials comeup for election at the same time. To do this, a First Nationmust depart from the Indian Act and use a custom election.Five of our respondents reported staggered terms.

Our third measure, third party management, was a meas-ure of the ability of the First Nation to set up effective inter-nal mechanisms for financial and economic governance. Itasks whether or not the local community has ever beenunder third party management. This happens when the localelected leadership is unable to manage the band financesand an independent third party is appointed by the Ministerof Indian Affairs to manage them until the financial situa-tion is considered stable (INAC 2004). Seven of our re-spondents reported having had third party management.

To measure the scope of the enterprise we included a var-iable, management planning, which indicated whether or notthe enterprise was involved in forest management planning.Fifteen of our respondents were involved in forest manage-ment planning.

Our final variable, formal role for elders and hereditarychiefs, indicates whether or not there was a formal role forelders and hereditary chiefs in the government’s decision-making processes. We treated this variable as a decision toexplicitly incorporate cultural values through organizationaldesign. This variable results from two of the questions onthe questionnaire. Twenty-four of our respondents reportedthat either elders or hereditary chiefs had a formal role ingovernance.

Research resultsOur research results that proved to be significant are re-

ported in Table 9. These are very strong results given thesmall sample. The coefficient for any one of these variablesis indeed large enough to suggest that there are strong rela-tionships between these factors and the odds of profitability.

According to the equation estimated in Table 9, in the ab-sence of other factors, the odds are 1.3 to 1 that a firm willbe profitable. The odds contribution of each variable isgiven in the final column and is calculated directly from thecoefficient by raising e to the value given by the coefficient,as explained above. Since the constant term is measuredwith considerable uncertainty, initial success may vary, butnearly even odds is a reasonable outcome.

As predicted by theory, the separation of day-to-day busi-ness decisions from political decisions is important to theprofitability of enterprises — in this case, strongly so. Thisis consistent with results from the Harvard Project con-ducted in the US and supports recommendations in a num-ber of reports on the necessity of keeping business separatefrom politics (McBride 2001; Office of the Auditor General2003). This does not mean that business is completely di-vorced from politics, as politics are an inescapable fact oflife, as noted by Merkel (2003). However, excluding politi-cal considerations from day-to-day operations can increasethe certainty over the environment within which the business

Trosper et al. 233

# 2008 NRC Canada

operates, contributing to its potential profitability. If theodds are originally 1.3 to 1 for profitability, the addition ofthe separation of business from politics raises the odds ofprofitability to 120 to 1.

Staggered terms are also important, but with a coefficientthat has a weak significance level. Our problem here is thatcommunities can use a variety of techniques to achievestability, such as establishing longer terms for chief andcouncil or taking measures to ensure accountability. Wefind that for this and one other measure of stability there isa significant relationship. (We report below on duration ofthe current chief in office, which did not give a significantrelationship.) Here, only 5 out of the 40 communities sur-veyed reported using staggered terms. However, they con-tribute significantly to the odds of profitability, most likelythrough increasing political stability and continuity. Waito(2004), reviewing factors that contribute to the success ofFirst Nation and industry forest sector partnerships in Can-ada, says that First Nations have learned the advantages ofstability and predictability and maintaining continuity evenwhen government changes. Just adding staggered termsraises the odds of profitability to 21 to 1.

Our results also show that if the band was or had beenunder third party management, the odds of profitability arelower, although the coefficient was significant at only a10% level. This is consistent with the idea that communitiesthat have had trouble developing effective economic gover-nance mechanisms for their own internal affairs will alsoface this difficulty in regards to operating business activities.Having been under third party management reduces the oddsof profitability to 1 to 7.

The positive effect of forest management planning sug-gests that the broader the scope of management responsibil-ity, the greater the likelihood of profitability. We interpreteda positive answer in two ways. First, we felt that it providedgreater information such that firms involved in logging (aswas most of our sample) would be better able to identifyhigher quality timber. Indeed, all firms that carried out for-est management planning activities were involved in logging(Table 8). Second, First Nations’ involvement in forest man-agement planning activities indicates a greater scope for de-cision making within the enterprise and, in the case ofcollaboration, may provide more autonomy to a band-ownedenterprise within a partnership. The increase in the likeli-hood of profitability may then flow from the fact that theFirst Nation has more ability to make decisions that have animpact upon the enterprise. Finally, it may also be the casethat the ability to carry out forest management planning ac-

tivities is an indicator of capacity (e.g., expertise). The oddsof profitability with only forest management planning activ-ities added to the background odds is 163 to 1.

If there is a formal role for elders and hereditary chiefs inthe decision-making process, then the likelihood of profit-ability is significantly reduced. The most obvious explana-tion for this is that hereditary chiefs or elders may placegreater significance upon environmental and cultural valuesrelative to the economic benefits of business. In terms ofthe enterprise, this could mean that the odds of profitabilityare reduced either through constraints placed upon opera-tions or in requirements for the enterprise to deliver otherbenefits. Alternatively, profitability could be compromisedif the enterprise is operating in an environment with con-flicting objectives (this may be either because of a lack ofconsensus within the community as to the appropriate bal-ance or because the enterprise may be responding to differ-ent sets of authorities). The formal role of elders andhereditary chiefs may extend beyond the governing structureand allow interference in the activities of the businesses.Perhaps the entire community is less business orientedwhen traditional leaders have formal roles. The explanationmay be a combination of these factors. Whatever the explan-ation, the impact of having elders or hereditary chiefs for-mally involved in decision making reduces the odds ofprofitability to 1 to 15.

As discussed in the methods section, one concern in usingthe logit techniques is a small sample size limits the validityof estimates of standard errors. This concern does not applyto ordinary least squares estimation, but that specificationcreates biased coefficients when the dependent variable isdichotomous. Table 10 presents the ordinary least squaresresults for three specifications. The first specification con-tains the three variables that are significant at a 5% level inthe logit model. Two of the three are also significant at thatlevel in the ordinary least squares specification. The secondspecification adds the two variables that were not as stronglysignificant in the logit specification. In ordinary leastsquares, both are significant at better than 10%; they in-crease the R2 value from 0.36 to 0.46 and the strong separa-tion variable becomes significant at p = 0.015. The thirdspecification tests the importance of the type of business; asreported above, whether or not the firm is band owned orjointly owned has no significant effect on the results.

The ordinary least squares tests show that the variables af-fect each other to a degree; the correlation matrix for these var-iables is shown in Table 11. It shows that none of thevariables are highly correlated, although some are corre-

Table 9. Probability of profitability in the previous 3 years.

Variable Coefficient (bi) SE Significance Contribution to odds (ebi )Separation 4.556* 1.956 0.020 95.2Staggered rerms 2.180{ 1.488 0.143 8.84Third party management –2.463{ 1.501 0.101 0.085Management planning 4.858* 1.661 0.003 128.7Formal role for elders or hereditary chiefs –2.914* 1.329 0.028 0.054Constant 0.233 0.778 0.764 1.26

*Significant at the 5% level.{Significant at the 10% level.{Significant at the 15% level.

234 Can. J. For. Res. Vol. 38, 2008

# 2008 NRC Canada

lated with coefficients >0.5. One way to judge the qualityof the model is to examine its success in predicting thecases used to estimate it. Table 12 reports that data, show-ing that on average the model correctly predicts the obser-vations for 84% of the cases.

One question that we asked was to what extent the sizeand type of activity might determine profitability. We testedwhether or not diversification (was an enterprise engaged inlogging, silviculture, and other activities) was important butcould not establish any effect upon profitability. We alsoconsidered whether or not the enterprise was involved inmanufacturing, but again could not establish whether manu-facturing had a positive or negative effect on profitability.Where enterprises were engaged in multiple activities, wewere unable to identify their relative importance as we didnot ask for the contribution of revenues from the differentactivities. Therefore, we were unable to control for differen-ces in the scale of the various activities.

We also examined the importance of size where we usedinvested capital as our measure (Table 4). Here, we first ranour model using our three strongest results (managementplanning, separation, and a formal role for elders and heredi-tary chiefs).8 We then added three additional dummy varia-bles so we had four categories of invested capital (<$1million; $1–$2 million; $2–$10 million; and >$10 million).We then compared the log-likelihood function between thetwo specifications and found that the improvement in thelog likelihood function was not significant.9

We also investigated several other hypotheses. One factorthat has been proposed as contributing to the success of for-est sector partnerships and enterprises included having abusiness plan. We asked the enterprises whether there wereset business objectives. Our results were not significant andtherefore we could not establish the importance of havingset business objectives. We also considered whether priorbusiness experience would be important in increasing the

Table 10. Ordinary least squares results for three models.

Unstandardized coefficients

R2 Variable bi SE t test Significance0.36 Constant 0.600 0.123 4.866 0.000

Strong barriers between business and politics 0.321 0.185 1.737 0.092Management planning 0.519 0.144 3.595 0.001Formal role for elders and hereditary chiefs –0.391 0.146 –2.673 0.012

0.46 Constant 0.519 0.129 4.017 0.000Strong barriers between business and politics 0.498 0.193 2.575 0.015Management planning 0.598 0.142 4.222 0.000Formal role for elders and hereditary chiefs –0.342 0.142 –2.414 0.022Third party management –0.310 0.182 –1.702 0.099Staggered terms 0.360 0.199 1.805 0.081

0.47 Constant 0.619 0.183 3.375 0.002Strong barriers between business and politics 0.503 0.195 2.583 0.015Management planning 0.591 0.143 4.137 0.000Formal role for elders and hereditary chiefs –0.358 0.144 –2.483 0.019Third party management –0.294 0.185 –1.593 0.122Staggered terms 0.361 0.201 1.797 0.082Type of business (1 = joint venture or other) –0.120 0.155 –0.770 0.448

Table 11. Correlation matrix.

ConstantForest managementplanning Separation

Formal role for eldersand hereditary chiefs

Staggeredterms

Third partymanagement

Constant 1.000Management planning –0.102 1.000Separation –0.147 0.597 1.000Formal role for elders and

hereditary chiefs–0.439 –0.601 –0.432 1.000

Staggered terms –0.264 0.255 0.276 –0.194 1.000Third party management 0.007 –0.482 –0.483 0.071 –0.136 1.000

8 We only used the three variables. as using all five would reduce the degrees of freedom too much when combined with our capital dummyvariables.

9 The value for the –2 log-likelihood function with just the three institutional variables was 31.917; after adding in the three additionaldummy variables to control for differences in invested capital, the value for the –2 log-likelihood function was 27.546. The differencebetween the two was a w2 = 4.371 with two degrees of freedom, with p = 0.224.

Trosper et al. 235

# 2008 NRC Canada

chances of success. Here we expected to find a positive rela-tionship, as the First Nation would presumably have gainedexperience and expertise. However, our results were againnot significant and we could not establish the importance ofprior business experience.

We also looked at the duration of the term of the currentchief. In some cases, this may be a means of providing po-litical stability and certainty. The terms served by the chiefvaried significantly, from 2 to 22 years in one case. Again,we could not establish the importance of this, as our resultswere not significant. It may be the case that the internecinepolicies that can plague First Nations communities in whichrent-seeking behaviour dominates may result either in en-trenched (but not necessarily effective) leadership or on-going political strife, leading to a constant turnover inpolitical office. We also looked at whether chief and councilhad ever been removed from office. We treated this as ameasure of accountability. Again, we could not establish astrong relationship.

Further discussion of selected resultsWe explored these results further in a follow-up workshop

held in February 2006, to which we invited survey partici-pants as well as researchers in the field. Although one fourthof our responding participants (10) agreed to attend, in theend 7 of the survey participants participated. Representativesfrom three of our four case studies were present. Six of the7 participants represented were from British Columbia; a to-tal of 10 persons represented the seven entities. At ourworkshop, participants generally agreed with our interpreta-tion of these variables. They paid particular attention to twoof our variables: first, the idea of separation, and second, therole traditional leaders may play.

They all agreed that political stability and keeping politicsout of business is important. They pointed out, however, thatthe word ‘‘separation’’ might not adequately describe the sit-uation. The relationship between political leaders and man-agers of firms can be complex, with several subtleties thatwere not captured by our questionnaire. Adequate communi-cation is particularly important. Too much separation couldlead to poor decisions by outsiders who do not understandthe complexities of regulations affecting businesses on re-serves. Political leaders need to understand business realitiesand be informed about what businesses are doing to assist inkeeping politics and management separate. If separation

were taken to mean reducing communication, our resultswould be seriously misunderstood.

Government stability has more dimensions than thosemeasured by the two variables we used, staggered terms andthird party management. Staggered terms, while a good ideato offset the political instability caused by a mandatory FirstNations election every 2 years (Indian Act requirement),may also have to be augmented by longer terms. When com-menting on our draft checklist, many workshop participantsrecommended omitting staggered terms as a proxy for bandstability. They preferred that we emphasize the length ofterms, accountability, economic development planning, ca-pacity building, and the creation of a stable land base. Theyalso felt that self governance (getting out of the Indian Act),having a constitution, and the legitimacy of the businesswithin a larger strategic direction were important.

The negative relationship between a formal role for eldersand hereditary chiefs and profitability produced considerablediscussion at our workshop, indicating that care must betaken interpreting its meaning. Our survey asked whether aformal role exists for elders in the governmental structure,but not in the structure of the business. The question, then,is how is that role connected to a firm’s profitability? Dis-cussion at our workshop indicated that we should examineconditions that provide both a formal role for elders andthat indicate situations that can reduce profitability. Profit-ability could be compromised if the enterprise is operatingin an environment with conflicting objectives, resultingfrom a lack of consensus within the community as to the ap-propriate compromises among competing values. The lackof consensus could mean that managers of the businesswould be asked to respond to conflicting sets of authorities.The negative role could also be related to a lack of culturalmatch between the requirements of provincial forest tenuresand business partners and the views of elders and hereditarychiefs regarding the balance between extraction and protec-tion of the forest.

ConclusionsIn this paper we have focused on the institutions estab-

lished by First Nations related to forest sector enterprises.Our results show the importance of institutions as explana-tions for successful economic performance in terms of profit.Firms that reported a separation from the political process,for instance, had better profitability. We found that onlyone variable that could be given a direct economic explan-ation, participation in forest management planning activ-ities, had an impact on profitability. Neither business tsize,nor management skill, business planning, or type of prod-uct was important. These perhaps surprising results estab-lish the importance of institutional factors in explainingthe profitability of these enterprises.

However, our study does not address why these institu-tions may exist. Why are they present for some First Nations?Why have some chosen staggered terms? How importantare leaders in initiating or developing these institutions?10

Another set of questions relates to how institutions are re-

Table 12. Percentage of cases correctly predicted by the logit model.

Predicted

Profitable on averagelast 3 years

Observed No Yes Percent correctProfitable on average

last 3 yearsNo 11 4 73.3Yes 2 20 90.9

Overall percentage 83.8

10 ‘‘Strong Leadership is the backbone of the corporation. . .The driving force behind the formation of the NEDC (Nuu-Chah-Nulth Eco-nomic Development Corporation) was George Watts who was chairman of the tribal council when NEDC was formed’’ (McBride 2001,p. 52).

236 Can. J. For. Res. Vol. 38, 2008

# 2008 NRC Canada

lated to the organization of the enterprise and the activitiesin which it is engaged. Most of the enterprises surveyedwere in some kind of formal partnership agreement with anindustrial partner. These agreements differ in their formalityand their features. It may be that business activities in part-nerships require the establishment of more elaborate organi-zations and agreements (i.e., development of explicitconflict resolution procedures, communication strategies,etc.), depending either upon the institutional ‘‘infrastructure’’within the First Nation or the nature of the business activity(i.e., more capital-intensive manufacturing). This is an areawe intend to pursue further as we have survey data that canbe analyzed to offer insight into this question.

Finally, we need to consider the causality of our results.We have identified institutional criteria that indicate in-creased chances of profitability. However, it may be thatthe presence of those institutions reflects the capacity ofthat particular community (defined broadly here to includenot only the human but also the social capital within thatcommunity). Therefore, while we may observe a ‘‘snapshot’’of such enterprises and see some that are profitable andsome that are not, we may not be able to capture the devel-opment of capacity that may be accompanying the enter-prise. There is a possibility that a First Nation may indeedimplement the institutions represented by the variables weobserve (separation, staggered elections) as capacity in-creases and enlarge its scope of activities (move into forestmanagement planning) and develop the internal financialcontrols such that third party management is no longerlikely. For example, staggered elections reflects a consciousdecision to not use the Indian Act, but to choose customelections and then to choose staggered terms, implying care-ful consideration and the development of a political consen-sus. Therefore, while we observe strong relationshipsbetween the institutional variables, we also observe increasedodds of profitability. We should therefore realize that a FirstNation may not be able to simply turn an unprofitable enter-prise into a profitable one by moving to staggered terms. Ad-ditional research is required to understand what otherinstitutional changes may accompany that move (what kindof capacity is required) and to what extent various institu-tional features (separation) may reinforce one another.

AcknowledgementsWe would like to thank the Sustainable Forest Manage-

ment Network for financial support and Sarah Allen, JeremyBoyd, Holly Mabee, and Katja Pecarevic for research assis-tance. We also thank our participants in the survey for tak-ing the time needed to answer our questions.

ReferencesAnderson, R.B. 2002. Aboriginal entrepreneurship and business de-

velopment. Captus University Press, Toronto, Ont.Atlantic Canada Opportunities Agency. 2003. Aboriginal eco-

nomic development in Atlantic Canada [online]. Availablefrom www.acoa-apeca.gc.ca/e/library/reports/aboriginal/index.shtml [accessed July 31 2005]. Atlantic Canada OpportunitiesAgency, Moncton, N.B.

Boyd, J.J. 2006. Aboriginal economic development by two Caroboo–Chilcotin forestry joint ventures. M.Sc. thesis, The Universityof British Columbia, Vancouver, B.C.

The Canadian Royal Commission on Aboriginal Peoples. 1996. Re-port of the Royal Commission on Aboriginal Peoples. Vol. 5.Renewal: a twenty year commitment [online]. Available fromwww.ainc-inac.gc.ca/ch/rcap/sg/cg_e.html#access [accessed 31July 2005]. Minister of Supply and Services, Ottawa. Ont.

Cornell, S., and Kalt, J.P. 1992. Reloading the dice: improving thechances for economic development on American Indian reserva-tions. In What can tribes do? Strategies and institutions inAmerican Indian economic development. Edited by S. Cornelland J.P. Kalt. American Indian Studies Center, University of Ca-lifornia, Los Angeles, Calif. pp. 1–59.

Cornell, S., and Kalt, J.P. 1995. Where does economic develop-ment really come from? Constitutional rule among the contem-porary Sioux and Apache. Econ. Inq. 33: 402–426.

Cornell, S., and Kalt, J.P. 1997. Successful economic developmentand heterogeneity of governmental form on American Indian re-servations. In Getting good government: capacity building in thepublic sectors of developing countries. Edited by M.S. Grindle.Harvard Institute for International Development, Cambridge,Mass. pp. 257–296.

Council for the Advancement of Native Development Officers(CANDO). 1994. Annual Report 1994–5 [online]. Availablefrom www.edo.ca/meet/agm.htm [accessed 22 June 2002].

Department of Indian Affairs and Northern Development. 2002.Words first: an evolving terminology relating to Aboriginal Peo-ples in Canada [online]. Available from www.ainc-inac.gc.ca/pr/pub/wf/index_e.html [accessed 22 June 2005]. Government ofCanada, Ottawa, Ont.

Duhaime, G., Searles, E., Usher, P.J., Myers, H., and Frechette, P.2004. Social cohesion and living conditions in the CanadianArctic: from theory to measurement. Soc. Indic. Res. 66: 295–317. doi:10.1023/B:SOCI.0000003726.35478.fc.

Elias, P.D. 1995. Northern economies. In Northern Aboriginal com-munities: economies and development. Edited by P.D. Elias.Captus Press, Concord, Ont. pp. 3–31.

Graham, J., and Edwards, H. 2003. Options for commercial enter-prises in First Nations [online]. Available from www.iog.ca/publications/FNenterprises.pdf [accessed 30 April 2005]. Insti-tute on Governance, Ottawa, Ont.

Graham, J., and Wilson, J. 2004. First Nation and forest industryrelationships: some perspectives from British Columbia [online].Available from www.nafaforestry.org/docs/bc1.pdf [accessed 30April 2005]. National Aboriginal Forestry Association, ForestProducts Association of Canada, First Nations Forestry Program,Institute on Governance, Ottawa, Ont.

Grant, K., and Taylor, J. 2007. Managing the boundary between busi-ness and politics: strategies for improving the chances for successin tribally owned enterprises. In Rebuilding native nations: strate-gies for governance and development. Edited by M. Jorgensen.University of Arizona Press, Tucson, Ariz. pp. 175–196.

Gysbers, J., and Lee, P. 2003. Aboriginal communities in forest re-gions in Canada: disparities in socio-economic conditions. Glo-bal Forest Watch Canada, Edmonton, Alta.

Harris, R.R., Blomstrom, G., and Nakamura, G. 1995. Tribal self-fovernance and forest management at the Hoopa Valley IndianReservation, Humboldt County, California. Am. Indian Cult.Res. J. 19: 1–38.

Hickey, C., and Nelson, M. 2005. Partnerships between First Na-tions and the forestry sector: a national survey. Sustainable For-est Management Network, Edmonton, Alta.

Hindle, K., and Landsdowne, M. 2005. Brave spirits on new paths:toward a globally relevant paradigm of Indigenous entrepreneur-ship research. J. Small Business Entrepreneurship, 18: 131–141.

Hindle, K., Anderson, R.B., Giberson, R.J., and Kayseas, B. 2005.

Trosper et al. 237

# 2008 NRC Canada

Relating practice to theory in Indigenous entrepreneurship: a pi-lot investigation of the Kitsaki partnership portfolio. Am. IndianQ. 29: 1–23. doi:10.1353/aiq.2005.0050.

Indian and Northern Affairs Canada. 2004. Financial managementand accountability [online]. Available from www.ainc-inac.gc.ca/ps/lts/fng/prev/JMACFR_fma_e.html. [accessed 27 July 2005].Government of Canada, Ottawa, Ont.

Institute of Governance (IOG). 1998. Exploring the relationship be-tween Aboriginal Peoples and the Canadian forest industry:some industry perspectives [online]. Available from www.iog.ca/publications/foreste.pdf [accessed 27 July 2005]. Institute ofGovernance, Ottawa, Ont.

Jorgensen, M. 2000. Bringing the background forward: evidencefrom Indian country on the social and cultural determinants ofeconomic eevelopment. Ph.D. dissertation, Kennedy School ofGovernment, Harvard University, Cambridge, Mass.

Leamer, E.E. 1978. Specification searches: ad hoc inference withnonexperimental data. John Wiley and Sons, New York.

Long, J.S. 1997. Regression models for categorical and limited de-pendent variables. Sage Publications, Thousand Oaks, Calif.

McBride, J. (Editor). 2001. Our own vision — our own plan: whatsix First Nations organizations have accomplished with theirown economic development plans [online]. Available from www.sfu.ca/cscd/aboriginal%20ced/Our_Own_Vision8.pdf [accessed27 July 2005]. Community Economic Development Centre, Si-mon Fraser University, Burnaby, B.C.

Merkel, G. 2003. Keeping politics out of business: Aboriginalonly? Not even! Presentation made at First Nations and Sustain-able Forestry: Institutional Conditions for Success, Faculty ofForestry, The University of British Columbia, 24 October 2003.

Myers, H. 1996. Neither boom nor bust: The renewable resourceeconomy may be the best long-term hope for northern commu-nities. Altern. J. 22: 18–23.

Natcher, D.C., and Hickey, C.G. 2002. Putting the community backinto community-based resource management: a criteria and indi-cators approach to sustainability. Hum. Organ. 61: 351–363.

National Aboriginal Forestry Association. 2002. First Nations gov-ernance and forest management. National Aboriginal ForestryAssociation, Ottawa, Ont.

National Aboriginal Forestry Association. 2003. Aboriginal-held

forest tenures in Canada 2002–2003. National Aboriginal For-estry Association, Ottawa, Ont.

National Aboriginal Forestry Association and the Institute of Govern-ance (NAFA-IOG). 2000. Aboriginal-forest sector partnerships:lesson for future collaboration. NAFA and IOG, Ottawa, Ont.

Office of the Auditor General. 2003. Report of the Auditor Gen-eral: Chapter 9. Economic development of First Nations commu-nities: institutional arrangements [online]. Available from www.oag-bvg.gc.ca/domino/reports.nsf/html/03menu_e.html#03nov[accessed 20 June 2006]. Office of the Auditor General, Ottawa,Ont.

Peredo, A., and Chrisman, J.J. 2006. Toward a theory of community-based enterprise. Acad. Manage. Rev. 31: 309–328.

Ross, M., and Smith, P. 2002. Accommodation of Aboriginalrights: the need for an Aboriginal forest tenure. Synthesis report[online]. Available from www.ualberta.ca/sfm [accessed 27 July2005]. Sustainable Forest Management Network, Edmonton,Alta.

Waito, B. 2004. Creating a New Climate for Aboriginal forestry. InProceedings of the National Conference on Aboriginal Forestry,11–13 May 2004, Thunder Bay, Ontario. pp. 52–53. First Na-tions Forestry Program of Natural Resources Canada and Indianand Northern Affairs, Government of Canada, Ottawa, Ont.

White-Harvey, R. 1994. Reservation geography and the restorationof native self-government. Dalhous. Law J. 17: 587–611.

Whiting, K. 2001. Indigenous–industry partnerships and capacitybuilding for Aboriginal economic development: the case of for-est industry joint ventures in North Central British Columbia.M.Sc. thesis, University of Northern British Columbia, PrinceGeorge, B.C.

Wilson, J., and Graham, J. 2005. Relationships between First Na-tions and the forest industry: the legal and policy context. Insti-tute on Governance, Ottawa, Ont.

Wuttunee, W. 1992. In business for ourselves: northern entrepre-neurs. McGill–Queen’s University Press and the Arctic Instituteof North America, Montreal, Que., Kingston, Ont., and Calgary,Alta.

Wuttunee, W. 2005. Living rhythms: lessons in Aboriginal eco-nomic resilience and vision. McGill-Queen’s University Press,Montreal, Que., and Kingston, Ont.

238 Can. J. For. Res. Vol. 38, 2008

# 2008 NRC Canada