A new model for making aboriginal policy? Evaluating the Kelowna accord and the promise of multilevel governance in Canada.

Alcantara, Christopher ; Spicer, Zachary

An enduring theme among Canadian academics and public commentators is that federal and provincial governments in this country have not been able or willing to properly accommodate the demands of Indigenous communities. As a result, Indigenous leaders have resorted to a range of tactics to protest the lack of effective government action and public policy for their constituents (Alcantara 2010; Wotherspoon and Hansen 2013). At the core of these conflicts is a fundamental incongruence between Canadian and Aboriginal worldviews. On the one hand, the federal and provincial governments view their relationship with Indigenous communities through the lens of Canadian federalism. That is, each level of government believes it has been assigned a particular set of Aboriginal responsibilities and they alone have the authority to manage those responsibilities. Sometimes those responsibilities require intergovernmental coordination and in those instances, governments rely on the familiar processes of intergovernmental relations (Jhappan 1995; Smith 2004: 55, 80; Papillon 2014). Aboriginal communities in this model are limited to acting as stakeholders rather than as equal government partners (Abele and Prince 2003).

In contrast, many Indigenous communities, leaders, and scholars see the Aboriginal-Crown relationship very differently. Rather than viewing themselves as stakeholders or junior governments, they argue that Aboriginal communities should be treated as full and equal partners in the federation, operating on a nation-to-nation basis with the Crown. Scholars of treaty federalism suggest that Canadian politicians and policymakers must take into account the constitutional status of the many treaties Indigenous governments have signed with the Crown; these treaties guarantee that Aboriginal communities are equal, yet independent partners in the policymaking process (Henderson 1994; RCAP 1994).

In reality, Aboriginal policy-making very much adheres to the Canadian federalism model (Jhappan 1995). Evidence of treaty federalism is limited and scattered. As Martin Papillon (2012) has argued, the structures of Canadian federalism remain fundamentally unchanged despite the emergence of new spaces for Aboriginal-Crown interaction. The traditional federalism model remains dominant and entrenched despite demands for more treaty federalism.

Given the lack of progress in transforming Canada into a system of treaty federalism, some scholars have turned to the concept of multilevel governance as a possible alternative (Ladner 2010; Rodon 2014; Wyatt and Nelson 2014). This concept first emerged in the early 1990s to describe a broad trend in Europe in which power seemed to be moving away from a system of government hierarchy towards "a system of continuous negotiation among nested governments at several territorial tiers" (Marks 1993: 392). Scholars in Canada have drawn on this literature to capture a range of developments, including instances of Aboriginal actors challenging the hierarchical structure of the Canadian system (Smith 2004: 80; Papillon 2012; Alcantara and Nelles 2014).

Students of Aboriginal politics have tended to favour a descriptive approach to using multilevel governance and so few scholars have empirically explored the potential of this concept for restructuring the Aboriginal-Crown relationship. In this paper, we begin to address this lacuna by providing a preliminary analysis of the promise and potential pitfalls of multilevel governance as an alternative model to Canadian and treaty federalism. To do so, we examine the policymaking processes used to generate the 2005 Kelowna Accord, an agreement that promised to spend $5.1 billion over five years on a variety of Aboriginal issues. In many ways, the Kelowna Accord was a textbook example of multilevel governance (see Alcantara and Nelles 2014). Unfortunately, the Accord was never implemented; Paul Martin's Liberal Party lost the 2006 federal election and was replaced by a Conservative government that had no interest in the Accord. (1)

Yet the Accord is still a useful case study for assessing multilevel governance as a potentially new and transformative model of Aboriginal policymaking in Canada. It was also hailed by a number of commentators and leaders as a significant milestone for Canadian-Indigenous relations (Patterson 2006; Markwick 2008: 2; Martin 2013). To support their arguments, they point to the unprecedented infusion of money that was to be spent on Indigenous issues and the unique processes used to negotiate the agreement as some of its most notable features. As well, a number of side commitments still came into effect after the demise of the Accord, such as the Aboriginal Health Human Resources Initiative and the Aboriginal Health Transition Fund.

Our main goal in this paper is to assess the concept of multilevel governance as a potential alternative model of Aboriginal policy-making in Canada. Our core findings suggest that multilevel governance may have the potential to produce highly desirable outcomes in Aboriginal policy, all without needing to dramatically alter the institutional structures of the country. This potential, however, is tempered by the fact that the emergence of multilevel governance processes seems to require significant political agency. To make these arguments, we focus on the process that generated the 2005 Kelowna Accord and draw some lessons from those experiences. Unfortunately, assessing the aftermath of the Accord is beyond the scope of our paper. Instead, we hope interested readers will read our paper alongside those that focus exclusively on the events following the demise of the Accord (see Larocque and Noel 2015).

The organization of this paper is as follows. We begin by describing three models of Aboriginal policymaking in Canada. Next, we briefly describe the negotiation of the Kelowna Accord and its aftermath. Finally, we offer some lessons regarding what multilevel governance may offer as a new model of Aboriginal policy-making. Throughout the paper, we rely on primary and secondary literature, but mostly on fourteen elite interviews that we conducted in 2013 with leaders and advisers who were involved in the Kelowna Accord. None of the interviewees requested anonymity and so their details are recorded in the bibliography and referenced throughout the paper using in-text citations.

Federalism and multilevel governance

According to s. 91 (24) of the Constitution Act, 1982, jurisdiction over "Indians, and Lands reserved for the Indians" falls exclusively to the federal government (Smith 2004: 55). Indeed, the original architecture of the Canadian federation left no room for Aboriginal governments and organizations to play any meaningful role in the creation of Aboriginal policy in Canada. It is no coincidence, then, that Aboriginal actors were also left out of the constitutional conferences in the mid 1860s and the constitutional rounds up until 1983 (Russell 2004). During that time period, Aboriginal actors were generally viewed and treated as stakeholders, rather than governments equal in status to the federal and provincial governments of Canada (Alcantara 2013).

This conception of Aboriginal inclusion in the Canadian federation as stakeholders is at odds with Aboriginal understandings of the Aboriginal-Crown relationship. Many Aboriginal communities insist they have never relinquished their right to self-government and sovereignty. Instead, these Aboriginal rights continue to exist today and are protected by the many treaties signed with the Crown. Contrary to the Crown's belief that such treaties extinguished Aboriginal title, many Indigenous communities argue the treaties established a nation-to-nation relationship between their governments and the Crown. The treaties codify a spiritual and legal arrangement in which the Aboriginal and non-Aboriginal governments of Canada agree to work together to pursue common goals while respecting the right of each community to govern themselves as they see fit (Henderson 1994). To achieve this relationship, scholars argue federal and provincial governments must fundamentally restructure their institutions and ideas to embrace an intergovernmental policy-making dynamic that is more respectful and accommodating of Aboriginal rights, title, and interests (RCAP 1994).

Despite repeated calls for this fundamental transformation, the Canadian federal system remains highly resistant to major institutional change (Papillon 2012). Although Indigenous communities have negotiated self-government agreements and modern treaties and have achieved legal victories and other policy changes, the fundamental logic of federal-provincial dominance remains (Jhappan 1995; Abele and Prince 2003). As a result, Indigenous communities continue to express their frustration over the inability of the Canadian system to accommodate their demands for greater autonomy, influence and participation.

One potential solution to this impasse is multilevel governance. At its core, multilevel governance (MLG) "is a process of political decision making in which governments engage with a broad range of actors embedded in different territorial scales to pursue collaborative solutions to complex problems" (Alcantara and Nelles 2014:185, emphasis added; see also Bache and Flinders 2004; Peters and Pierre 2004). Typically, Canadian scholars have used the term as a "catch-all" label for describing a set of trends in which Aboriginal actors create new spaces for dialogue and shared decision-making with the Crown (Bakvis, Brown, and Baier 2009; Papillon 2012; Rodon 2014). Some scholars have expressed discomfort with this approach, noting that such a definition does not allow for a clear conceptual separation between multilevel governance and federalism (Rouillard and Nadeau 2013). Indeed, Hooghe and Marks' (2001) influential work is somewhat puzzling in that they conceive of multilevel governance and federalism as variants of a broader category of arrangements called multilevel governance.

In response to this frustration, Alcantara and Nelles (2014) have tried to draw a much narrower boundary around the concept of multilevel governance by emphasizing three crucial components. First, and most importantly, multilevel governance involves the existence of a negotiated order in which "decision-making is the result of bargaining and negotiation between actors rather than top-down or hierarchical determined relationships" (see also Piattoni 2010: 90; Bache and Flinders 2004). Second, "at least one actor is embedded at a different political/territorial scale from the others in the partnership" (also Peters and Pierre 2004, emphasis added). Finally, the process or partnership must involve "at least one constitutionally recognized government in partnership with nongovernmental and/or quasi-governmental actors" (Alcantara and Nelles 2014: 186; see also Faludi 2012). Although all three components are important, they suggest that the involvement of actors from multiple jurisdictional scales and the existence of negotiated processes are the crucial defining features of multilevel governance partnerships (Alcantara and Nelles 2014).

In addition to these criteria, Alcantara and Nelles (2014; see also Alcantara, Broschek, and Nelles 2016) conceive of multilevel governance not as a system or regime of power allocation and decision-making akin to federalism, but instead as an instance of decision-making that can emerge intermittently or more regularly in a variety of different settings and countries (for example, unitary/federal/confederal). This conception is consistent with Papillon (2012) who argues that Indigenous multilevel governance can emerge alongside of and parallel to existing federal structures without transforming the institutional architecture of a political system. In that sense, multilevel governance is dynamic and fluid, capable of emerging suddenly and unexpectedly without having to wait for or transform the underlying logic of federal systems.

As a potential model for restructuring the Aboriginal-Crown relationship, multilevel governance offers an appealing vision. One the one hand, its criteria demands a more-inclusive and less-hierarchical approach to Aboriginal policy-making. Instead of participating as stakeholders in the policy-making process, multilevel governance requires that Indigenous governments and organizations participate as co-producers of public goods (for example, policy and political decisions, among other things) (see Bache and Flinders 2004; Piattoni 2010: 90). Multilevel governance has the additional advantage of fluidity in that major institutional reform is unnecessary for its emergence. Whereas treaty federalism demands a fundamental restructuring of the Canadian state, something that the federal and provincial governments have yet to embrace, multilevel governance conceived of as an instance is something that these governments can embrace more readily and use strategically without changing the foundations of the Canadian federal system. If the goal is to create a policy-making process that respects the underlying logic of the Canadian federation but creates space for nation-to-nation interactions within a policy-making process, then multilevel governance may be an ideal model for achieving these outcomes in Canada.

Some readers may object to a number of our assumptions regarding the nature of Canadian federalism and how multilevel governance fits with those assumptions. Perhaps, for instance, Canadian federalism is not as hierarchical as we suggest, especially if we consider: a) it is our Westminster system and its tendency to concentrate power in the hands of First Ministers that may be to blame for hierarchy rather than Canadian federalism itself; b) that hierarchy is weakened by the fact that Canada is supposedly a highly decentralized federation (although see Broschek and Turgeon 2015); and c) certain trends and models relating to Canadian federalism and social policy (see Banting 2012) share a number of similarities with some of the assumptions underpinning Alcantara, Broschek, and Nelles's (2016) definition of multilevel governance. Although we agree that it is somewhat misleading to paint all aspects of Canadian federalism as hierarchical, the literature is very clear that when it comes to how the federal, provincial, and territorial governments of Canada treat Indigenous governments and communities hierarchy is the appropriate characterization (Abele and Prince 2003; Ladner 2010; Moore, Walker and Skelton 2011; Papillon 2012). As well, we agree there is terminology within the Canadian federalism literature (for example, shared cost and joint-decision federalism, as well as the other terms provided by Abele and Prince 2003) that captures some of the particular dynamics of multilevel governance that we use in this paper. The advantage of our approach, however, is that we use a term already anchored in the Aboriginal federalism literature yet which also has the potential to travel much more widely within and outside of Canada. A concept like "shared-cost" federalism, for instance, is foreign to the Aboriginal federalism literature and does not travel well to other federal and non-federal systems. Multilevel governance, on the other hand, conceived of as an instance of multilevel politics (see Alcantara and Nelles 2014; Alcantara, Broschek and Nelles 2016), does not suffer from such restrictions. Overall, we believe that the advantages of our approach outweigh the advantages of potential alternatives.

The Kelowna Accord

The Kelowna Accord was the result of many months of multilevel negotiations. (2) According to one estimate (Patterson 2006), the entire process involved over 1,000 invitees. However, most negotiations involved smaller groups of officials from the federal, provincial and territorial governments and the five main national Aboriginal organizations in Canada: the Assembly of First Nations (AFN), the Inuit Tapiriit Kanatami (ITK), the Metis National Council (MNC), the Native Women's Associations of Canada (NWAC), and the Congress of Aboriginal Peoples (CAP). Individual Indigenous governments were not directly involved in the process. Instead, they worked through the Aboriginal organizations that ostensibly represented their interests. Given the terms of the Accord and the sheer number of communities affected by it, there was simply no other way at the time to conduct these negotiations. As well, Aboriginal communities chose this model and believed it was legitimate because their negotiators were required to continually consult and gain the consent of their constituents. According to Paul Martin (2013) and others, the five Aboriginal organizations regularly took breaks in the negotiations to engage with their constituents.

The Accord process began in April 2004 with the Canada-Aboriginal Peoples Roundtable on Strengthening the Relationship. This meeting progressed without any set topics, agendas or timelines (Mitchell 2013). Instead, the goal was to create an open dialogue and identify priorities for more focused, future discussions (Copenace 2013). Aboriginal leaders did receive specific commitments at the conclusion of the roundtable, including further discussions on six thematic areas: health, lifelong learning, housing, economic opportunities, negotiations, and accountability and a future policy retreat involving Aboriginal leadership and the Cabinet Committee on Aboriginal Affairs. (3)

The next stage involved sectoral meetings between November 2004 and January 2005 followed by a May 2005 bilateral policy retreat. At this meeting, Aboriginal representatives expressed their opposition to a pan-Aboriginal approach because they believed their unique constituencies required a unique policy response (MNC 2006; Patterson 2006: 5). An asymmetrical approach to policy-making would acknowledge the economic, geographic, social and cultural differences inherent in each community. This approach would also solidify the political and institutional differences between each constituency and provide each with a measure of independence in policy formation and funding distribution.

The policy retreat was pivotal because it produced the federal government's long-term commitment to continue the process of cooperative policy development (Copenace 2013; Watson 2013). A meeting of a Multilateral Indicators Working Group, which was composed of provincial and federal officials along with representatives from the AFN, ITK, MNC, CAP and NWAC, helped move the process forward. This group was responsible for developing guiding principles and indicators for further discussions (Patterson 2006: 7). Aside from this group, there was also a small informal "oversight committee," which was composed of staff from the Privy Council, the Prime Minister's Office, several key Minister's offices and representatives from the five major Aboriginal groups. This committee met once a month and was designed to work through "back channels" to keep the process moving (Copenace 2013; Martin 2013). The resulting agreement launched a five-year, $5.1 billion effort to close the gap in the quality of life between Aboriginal people and other Canadians. The specific initiatives and financial commitments of the 2005 Kelowna Accord are shown in Table 1.

The 2005 Kelowna Accord was the first phase of what was conceived to be a multi-phase effort to improve the lives of Aboriginal Canadians. This first phase was to focus entirely on improving living standards and reducing poverty levels, while subsequent phases would focus on land and treaty issues. Respondents agreed the entire process was consistent with existing treaty relationships and that this first phase was an important expression of the federal government's responsibility to respect the treaty process, especially in terms of improving living standards and honouring past obligations. The second phase was meant to build upon these beginnings by furthering the treaty relationships to a greater degree. According to interviewees, the rationale for using this multi-phase approach was that negotiating new self-government and other structural-related changes and arrangements would be difficult for many Indigenous communities because of their persistent and significant poverty and capacity issues.

Once these issues were addressed, the parties would then be able to negotiate a second five-year agreement to address structural reforms (Martin 2013).

According to Martin (2013) and others, the federal government's plan was for Indian and Northern Affairs Canada to transfer money to the band councils and other Indigenous governments to address their various priority areas as identified in the Accord. Provincial and territorial governments would handle the money and related programs and services for the constituents of the National Women's Association of Canada, the Congress of Aboriginal Peoples and the Metis National Council (MNC). Before the agreement could be fully implemented, however, Martin's minority government fell in November 2005 and subsequently lost the ensuing election to Stephen Harper's Conservative Party. After the election, Conservative officials claimed that the Kelowna Accord "did not exist" because funds had not been formally budgeted for its implementation. (4) Additionally, the Conservatives claimed that the Accord did not represent a national consensus because Aboriginal groups from Quebec did not take part in the negotiations, although they did attend as observers (Martin 2013). Harper also argued that the agreement was the result of election-driven bargaining and was unfeasible from a financial perspective.

In the wake of the Accord's demise, several provinces took actions that were in the general spirit of the Accord. Manitoba and Quebec introduced documents that were similar to the goals and aims of the Kelowna Accord (Manitoba 2013; Quebec 2006). Ontario signed several bilateral agreements with the Metis and Treaty 3 First Nations in 2008 and 2009 (Ontario 2009), while British Columbia took the largest step by implementing the tripartite agreement signed with the federal government and First Nations in the province. Called the "Transformative Change Accord," the agreement adheres to the same policy areas, measures, targets and funding as allocated in the original Accord (British Columbia 2005).

Multilevel governance as a new and preferred model?

The Kelowna Accord is an excellent example of multilevel governance (MLG) and fits very nicely with Alcantara and Nelles's (2014) criteria. In terms of the actors involved, the process included the federal, provincial and territorial governments of Canada negotiating with five major national Aboriginal organizations. These actors came from a variety of territorial scales and the process itself was relatively non-hierarchical, with a strong emphasis on negotiations, in contrast to previous instances of Aboriginal policy-making (see Papillon 2012; also Copenace 2013; Martin 2013). Indeed, it is worth emphasizing that the Kelowna process was unlike any other previous negotiation or accord dealing with Indigenous issues, at least when compared against examples similar in scale. During the intergovernmental meetings leading up to the Constitution Act, 1982, for instance, federal and provincial governments excluded Aboriginal participants almost completely from the negotiation process. Subsequent rounds involved Aboriginal representatives as stakeholders and observers in constitutional discussions about Aboriginal rights and self-government, but none of these negotiations involved the same number and range of participants, issue areas and resources as were discussed during Kelowna. The Meech Lake Accord completely ignored Aboriginal actors (Cairns 1988). In the Charlottetown Accord negotiations, select Aboriginal leaders were engaged on a greater level, but were overshadowed by provincial delegations (Russell 2004).

In our view, the Kelowna process was unique in that the actors, scales, decision-making processes, and the scope of the issues and concerns to be addressed were unparalleled. It was also the first time Aboriginal actors sat at the table as equal partners and had the authority to determine the issues to be addressed. No other Aboriginal-related accord has ever featured these characteristics. In the next section, we provide some lessons that policymakers and scholars might draw from the Kelowna experience.

Creating multilevel governance

One of the potential advantages of multilevel governance is that it does not require a fundamental restructuring of the Canadian federal architecture to be implemented. Instead, instances of multilevel governance, which promise Indigenous groups a stronger role and voice in the decision-making process, can emerge quite readily and suddenly without the need for major institutional reforms (Papillon 2012). The experience of the Kelowna Accord confirms this assumption. Although the scope of the policy problem and the sheer number of actors involved were quite large, the Accord came together without needing any formal institutional or constitutional changes. As well, although the time to complete the Accord might seem at first to be rather lengthy, it was accomplished relatively quickly. According to Paul Martin, the average intergovernmental agreement takes about two years to complete; the Kelowna agreement, on the other hand, was completed in 18 months, and involved many more actors (Martin 2013).

Of course, the completion of the Accord was facilitated by the fact that the stakes for the provinces and territories were relatively low; the federal government was providing all of the money to the provinces and territories (and the Aboriginal groups), who were then responsible for spending the money on Aboriginal education and health care, among other things. Indeed, Martin assured the Premiers they would not be responsible for funding any of the policy areas discussed during the Accord negotiations (Campbell 2013), and the Premiers believed him. Martin had slowly earned the trust of the Premiers over several First Ministers meetings, earning a reputation as a "straight shooter" with the group (Calvert 2013).

The lack of engagement from the provinces was an incredibly striking feature about the Accord negotiations. Provincial governments generally play a large role in Canadian federalism, with their actions routinely dictating the course of intergovernmental discussions. Kelowna was different, however. The provincial Premiers had several concerns going into the discussions, mainly centered on jurisdictional encroachment and long-term policy funding, but Martin worked hard to alleviate these fears by promising that only federal funding would be devoted to fulfilling the Accord's mandate (Calvert 2013; McGuinty 2013). While the Premiers acknowledged there were varying degrees of enthusiasm about the Accord, they all agreed that the Accord was addressing the most important policy areas, thus ensuring a basic "buy-in" among the Premiers (Campbell 2013). The Premiers largely observed the proceedings without providing much input into the content or the negotiations, preferring instead to work through their ministers and civil servants when necessary. One consequence of this passivity is that it provided more space for the Aboriginal actors to influence and shape the process, which, in turn, led to a more open and honest dialogue with government representatives.

It should also be noted the Kelowna Accord was very much the result of the attention given to the issues by then-Prime Minister Paul Martin. In fact, the process that brought the Accord into existence began well before he became Prime Minister. In 1997, while Minister of Finance, Martin began discussions with Aboriginal leadership about creating a "big" process to address the various problems plaguing their communities (Fontaine 2013). Between 1997 and 2000, Martin had conversations with Aboriginal representatives regarding education and made it clear he wanted to create a process that would fundamentally change the relationship between Aboriginal communities and the federal government (Fontaine 2013; Martin 2013).

When Martin finally won the leadership of the Liberal Party and became Prime Minister, he began an informal process of engagement with Aboriginal leaders (Copenace 2013). Martin recounts how he simply walked over to the AFN headquarters shortly after being sworn-in as Prime Minister to discuss policy issues, much to the surprise of AFN staff members who were not expecting him (Martin 2013). Martin also took the opportunity to describe the relationship between his new government and the leadership of the AFN as "equals," which signaled that his government's approach to Aboriginal issues would be different than what it had been in the past (Fontaine 2013; Martin 2013). When he selected his first Cabinet, Martin made it clear to his Minister of Aboriginal Affairs, Andy Mitchell, that he wanted to change the Indigenous-Crown relationship to one that better reflected a nation-to-nation logic (Mitchell 2013), something that is echoed today by Prime Minister Justin Trudeau and his Liberal government.

From his discussions with First Nations leadership, Martin knew whatever process he entered into would need to be open, collaborative and transformative (Martin 2013). Martin believed that the historical relationship with First Nations and the Crown was one of the primary contributors to the awful living conditions found in many Aboriginal communities (Martin 2013). If was going to be able to achieve real progress, his government would need to adopt a new process.

Crucial to the success of the Kelowna Accord, therefore, was the fact that Martin was personally invested in the process and wanted the Accord discussions to be more open and Aboriginal-driven than what had occurred in the past. Without his sustained interest in the process, the Accord would have never have happened. Individual agency in the form of Prime Ministerial interest, then, was a crucial factor for the emergence and success of this MLG partnership. Martin combined a passion for this policy area with a set of previously established relationships with key Aboriginal actors, such as Phil Fontaine, Jose Kisugak and Dwight Dorey (Martin 2013). On the basis of this evidence, it seems that while MLG may be easier to establish (for example, because major institutional reforms do not need to precede its emergence), it still requires a significant amount of agency from federal and Aboriginal political elites.

Meaningful multilevel governance

A second advantage of multilevel governance is that it supposedly creates a decision-making process that is inclusive and empowering of Indigenous actors to meaningfully influence the co-production of public goods. Our analysis of the Kelowna Accord suggests there is some merit to this assumption. At its beginning, the Kelowna process involved broad consultation and dialogue, aimed not only at identifying policy areas of concern for Canada's Aboriginal population, but also changing the relationship between the Crown and its Aboriginal communities. The initial Accord meeting, on 19 April 2004, set the tone going forward. The meeting was open and designed to facilitate frank discussion about the types of issues facing Aboriginal communities in Canada. Throughout the meeting, a large focus was given to the relationship between Aboriginal Canadians and the Crown. Consensus from Aboriginal representatives was that the current and largely historic--relationship between the federal government and Aboriginal Canadians had largely hindered progress in the past (Mitchell 2004). Aboriginal leaders contended that for this process to be demonstrably different, the federal government had to be willing to listen to their concerns and take them seriously (Fontaine 2013).

At most of the negotiation sessions that followed the initial April 2004 meeting, there was a great deal of consensus amongst the different Aboriginal groups, the provinces and the federal government. Each actor knew that the living conditions and the health and education levels of Aboriginal Canadians were well below the national average. There was also consensus about the solutions to these policy problems (Fontaine 2013). Division was really only found at the margins of the process, centering mainly around accountability and the role of the private sector. During the housing roundtable, Minister Joe Fontana pushed the idea for greater private sector involvement in the creation of on-reserve housing (Fontana 2013). Improved measurement techniques was another issue addressed by the Premiers, mainly in the accountability sessions (McGuinty 2013). The idea of an Aboriginal auditor general was even discussed (Goodale 2013). Aboriginal representatives disagreed with both of these latter initiatives, but greater measurement tools were included in the final agreement. Even in these instances, however, it was relatively easy to reach agreement (Fontaine 2013).

The process did leave some groups on the outside looking in. For instance, CAP had deep and lingering concerns that the Accord was too focused on programming for those living on-reserve. CAP also had concerns about the process itself, arguing that they were not seen as an "equal partner" (Dorey 2013). The organization believed that they were not offered enough information throughout the negotiations and that the other major Aboriginal organizations (for example, the AFN, ITK and Metis Nation) were allowed to drive the process (Dorey 2013).

The National Association of Friendship Centres (NAFC) was another group that felt ostracized. The NAFC believed they should have played a central role in the Accord process (Hill, Lynn and MacFarlane 2007). However, key federal bureaucrats working with the Cabinet Committee on Aboriginal Affairs decided it would be best to work with the five national Aboriginal organizations that have "traditionally" been included in negotiations with the federal government, despite the insistence from staff in the PMO that NAFC had something substantial to offer the process (Copenace 2013). As a result, the NAFC protested the final Accord signing, arguing that their exclusion delegitimized the negotiations (Copenace 2013).

Finally, the Aboriginal leadership from Quebec withdrew from the discussions. The Quebec Chiefs wanted to hold separate meetings with the Province of Quebec and the federal government. This request was rejected by those organizing the meetings and roundtables, which prompted the Quebec Aboriginal leadership to remove themselves from the process (Copenace 2013). They did, however, attend the meetings as observers (Watson 2013).

The process was designed to have breadth. At some instances, hundreds of participants were involved in setting the agenda for discussion. As such, it is natural for some divisions to occur. Nonetheless, the success that those involved with Kelowna had with balancing the competing interests involved with the process was remarkable. Much of this success can be attributed to the open nature of the negotiations and its non-hierarchical nature. As such, the Kelowna model of MLG may be normatively appealing for those interested in better engaging Aboriginal actors in public policy design and implementation. The challenge, however, is in designing a process that does a better job of incorporating the various actors who were left out of the original Kelowna Accord experience.

Despite these challenges, it is perhaps surprising that the decision-making process produced a set of outcomes at all (Scharpf 1994), let alone one that produced a high level of agreement among the various participants and affected parties. According to the interviewees, when discussions began, it was not difficult to identify the core problems that needed to be addressed nor was it difficult to identify potential solutions. For instance, nearly everyone involved knew that housing, health and education needed to be improved. In fact, most of those involved knew that these three issues were intimately connected (Fontaine 2013; Martin 2013). As Phil Fontaine (2013) explained, Aboriginal communities could each have two new schools, but if there were inadequate housing and unclean water, drop-out rates would remain high. This is precisely why the vast majority of those involved knew that Martin's holistic approach was unique and appropriate (Chartier 2013; Fontaine 2013).

While the policy issues at hand were treated holistically, the outcomes designed for each organization were not. One of the more interesting components derived from the roundtables process was the desire to avoid pan-Aboriginal initiatives. During discussions, the MNC, AFN and ITK emphatically rejected any "pan-Aboriginal approaches," with the MNC describing such processes as "designed by bureaucrats and delivered by organizations that are not accountable to Metis, Inuit and First Nation peoples" (MNC 2006). There was a recognition very early on that policy could not be uniformly applied across the country. The differing geography and economic conditions that certain groups found themselves in needed to be taken into account. While this did provide some challenges, it was addressed very early on in the negotiations, allowing further discussions to take into account these differences.

The government designed an open process and placed relatively few barriers upon it. Even the final cost of the Accord was not pre-determined, meaning that there was broad flexibility built into the process from the very beginning (Goodale 2013; Martin 2013). Those opposed to the process, such as the Quebec representatives and CAP, simply voiced their dissatisfaction and removed themselves from the discussions. According to Phil Fontaine (2013), despite the opposition of these groups, there was a lot of momentum moving through the process, and it seemed unlikely that any one group would have been able to derail it.

Enduring and resilient multilevel governance

In many ways, the experiences associated with the Kelowna Accord seem to suggest that multilevel governance may be a preferable model for producing Aboriginal policy in Canada. Instances of multilevel governance seem to combine the inclusive and empowered nature of treaty federalism with the desire among government actors to protect the status quo of Canadian federalism, all the while producing more effective and legitimate (for example, Aboriginal-supported) policy decisions. Indeed, although Kelowna was never implemented, it did continue to have some, albeit limited effect on Aboriginal policy (see British Columbia 2005; Quebec 2006; Ontario 2009; Campbell 2013; Manitoba 2013). For the Metis, the Accord was important for establishing the principle that their organization should have the same status as the AFN and ITK (Chartier 2013; Chartrand 2013). At one point during the negotiations, the drafters of the Accord had removed the word "Nation" from the word "Metis" (Chartrand 2013). Prime Minister Martin intervened on their behalf to ensure that the text read "Metis Nation," an important and symbolic recognition of their place within the Canadian constitutional order (Chartrand 2013). Metis representatives believe that this change is permanent and that although Kelowna never became a reality, it did finally legitimize them as equal, constitutional actors.

On the other hand, the Kelowna Accord may have been a once-in-a-lifetime instance of multilevel governance. We have yet to see another Kelowna Accord style process emerge in Canada despite the continued and intractable presence of political and economic poverty among many Indigenous communities. Although a number of small-scale instances of multilevel governance have emerged across Canada (see Papillon 2012; Alcantara and Nelles 2014; Wilson, Alcantara, and Rodon 2015), none have approached the sheer scale and scope of the Kelowna process. Indeed, the Kelowna Accord experience demonstrates that unless there is strong political will and agency, especially from the Prime Minister of Canada, large-scale instances of multilevel governance are unlikely to emerge. In that sense, major institutional reforms to the federal system may be necessary for the creation of enduring MLG structures and incentives that can produce meaningful and inclusive Aboriginal policy-making.

Conclusion

In this paper, our primary goal was to investigate the potential of multilevel governance as a new model for Aboriginal policy-making in Canada. The Kelowna Accord experience suggests that large-scale instances of multilevel governance may be a double-edged sword. On the one hand, multilevel governance has the potential to produce highly desirable outcomes without first needing to dramatically alter the institutional structures of a country. These advantages, however, are also its greatest weakness; our findings suggest that the emergence of meaningful instances of multilevel governance may also depend heavily on political agency. Without that agency, the institutions of federalism seem to suppress the desire among government actors to pursue macro-level multilevel governance processes. Long-lasting and meaningful change to Aboriginal policy-making practices, therefore, will most likely require significant institutional adjustments to the Canadian federal system.

The multilevel governance research agenda is still in its infancy in Canada, and there is ample room for future research. In particular, new research is needed to examine the full range of multilevel governance instances in Canada, comparing variations on the three criteria identified by Alcantara and Nelles (2014). Different configurations of actors, scales, and decision-making processes should produce different types of outcomes compared to what we found with the Kelowna Accord. In particular, different power configurations among Aboriginal and non-Aboriginal actors are likely to produce significant variation in the types of multilevel governance instances produced (Bakvis 2013). A second research agenda might also investigate what kinds of factors produce different types of multilevel governance arrangements. Multilevel governance, in our view, is a concept, and not a theory, and so scholars should look to apply existing social science theories to explain variation in terms of the number and types of multilevel governance configurations in Canada (Alcantara, Broschek and Nelles 2016).

Notes

(1) Interestingly, Prime Minister Justin Trudeau's 2016 federal budget seems to have resurrected some of the commitments originally included in the Kelowna Accord, promising over $8 billion dollars over five years to address a variety of Indigenous policy issues and problems.

(2) An excellent history of the Accord negotiation process can be found in Patterson (2006). Our goal with this paper is to analyze the process through the lens of the MLG. As such, our discussion of the history of the Accord is short. We urge any readers wanting a broader history of the Accord to read Patterson (2006).

(3) A final commitment was the creation of an Inuit Secretariat within the Department of Indian Affairs and Northern Development (ITK 2004).

(4) Paul Martin argues that Ralph Goodale had already transferred the financial commitments into the federal budget and that the incoming Conservative government simply scooped that money out for other purposes.

(5) Adopted from Patterson (2006).

References

Abele, Frances, and Michael J. Prince. 2003. "Aboriginal governance and Canadian federalism." In New Trends in Canadian Federalism, edited by Francois Rocher and Miriam Smith. Toronto: Broadview Press, pp. 135-65.

Alcantara, Christopher. 2010. "Indigenous contentious collective action in Canada." Canadian Journal of Native Studies 30 (1): 21-43.

--. 2013. "Ideas, executive federalism, and institutional change." Canadian Journal of Political Science 46 (1): 27-48.

Alcantara, Christopher, and Jen Nelles. 2014. "Indigenous peoples and the state in settler societies." Publius journal of Federalism 44 (1): 183-204.

Alcantara, Christopher, Broschek, Jorg, and Jen Nelles. 2016. "Rethinking multilevel governance as an instance of multilevel politics: A conceptual strategy." Territory, Politics, Governance 4 (1): 33-51.

Bache, Ian, and Matthew Flinders. 2004. Multi-Level Governance. Oxford: Oxford University Press.

Bakvis, Herman. 2013. "In the shadows of hierachy: Intergovernmental governance in Canada and the European Union." Canadian Public Administration 56 (2): 203-18.

Bakvis, Herman, Gerald Baier, and Douglas Brown, eds. 2009. Contested Federalism: Certainty and Ambiguity in the Canadian Federation. Toronto: Oxford University Press Canada.

Banting, Keith. 2012. "The three federalisms revisited." In Canadian Federalism: Performance, Effectiveness, and Legitimacy, edited by Herman Bakvis and Grace Skogstad. Don Mills: Oxford University Press, 3rd ed., pp. 261-81.

British Columbia. 2005. Transformative Change Accord between Government of British Columbia and Government of Canada and the Leadership Council Representing the First Nations of British Columbia. Victoria: Government of British Columbia.

Broschek, Jorg, and Luc Turgeon. 2015. "Embracing Complexity." Paper presented at the Annual Conference of the Canadian Political Science Association. Ottawa: ON, pp. 1-20.

Cairns, Alan C. 1988. "Citizens (outsiders) and governments (insiders) in constitution-making." Canadian Public Policy 14 (Supplement): S121-45.

Calvert, Lorne, (Former Premier, Saskatchewan). 2013. In conversation with the author, February.

Campbell, Gordon. (Former Premier, British Columbia). 2013. In conversation with the author, March.

Chartier, Clement. (Former President, Metis National Council). 2013. In conversation with the author, March.

Chartrand, David. (President, Manitoba Metis Federation). 2013. In conversation with the author, April.

Copenace, Jeffrey. (Former Senior Special Assistant, Aboriginal and Northern Affairs--Office of the Prime Minister of Canada). 2013. In conversation with the author, February.

Dorey, Dwight. (Former National Chief, Assembly of First Nations). 2013. In conversation with the author, February.

Fontana, Joe. (Former Minister of Housing and Infrastructure). 2013. In conversation with the author, March.

Fontaine, Phil. (Former National Chief, Assembly of First Nations). 2013. In conversation with the author, February.

Goodale, Ralph. (Former Minister of Finance). 2013. In conversation with the author, March.

Henderson, James Youngblood. 1994. "Empowering treaty federalism." Saskatchewan Law Review 58: 241-329.

Hill, Carey, Trevor Lynn, and Jonathan MacFarlane. 2007. "Dangling participants: Is the Kelowna Accord constitutionally binding as a federal-provincial agreement?" Presented at the Annual Meeting of the Canadian Political Science Association.

Hooghe, Lisbet, and Gary Marks. 2001. "Types of Multi-Level Governance." European Integration Online Papers, No. 11. Available at http://eiop.or.at/eiop/pdf/2001-011.pdf.

Jhappan, Radha. 1995. "The Federal-Provincial Power-grid and Aboriginal Self-Government." In New Trends in Canadian Federalism, edited by Francois Rocher and Miriam Smith. Peterborough: Broadview Press, pp. 155-84.

Ladner, Kiera. 2010. "Colonialism isn't the only answer." In Federalism, Feminism and Multilevel Governance, edited by Marian Sawer Hausmann and Jill Vickers. London: Ashgate, pp. 67-82.

Larocque, Florence, and Alain Noel. 2015. "Kelowna's uneven legacy: Aboriginal poverty and multilevel governance in Canada." In Aboriginal Multilevel Governance, Canada: The State of the Federation 2013, edited by Martin Papillon and Andre Juneau. Montreal-Kingston: MQUP, pp. 237-58.

Manitoba. 2013. All Aboard: Manitoba's Poverty Reduction and Social Inclusion Strategy. Winnipeg: Government of Manitoba.

Marks, Gary. 1993. "Structural policy and multilevel governance in the EC." In The State of the European Community, edited by A. Cafruny and G. Rosenthal. New York: Lynne Rienner, pp. 391-410.

Markwick, Michael. 2008. The Land God Gave to Cain: Refuge and Canada's Security Regime. Centre for Canadian Studies, University of Edinburgh. Available at http://www.cst.ed. ac.uk/Events/Conferences/documents/MarwickMPaper.pdf. Accessed 12 June 2013.

Martin, Paul. (Former Prime Minister of Canada). 2013. In conversation with the author, January.

McGuinty, Dalton. 2013. (Former Premier of Ontario). In conversation with the author, May.

Metis National Council (MNC). 2006. First Ministers' Meeting on Aboriginal Issues: An Overview for Metis Citizens. Ottawa: Metis National Council.

"Mitchell, Andy. 2004. Transcription - Speech by Andy Mitchell, Minister of Indian Affairs and Northern Development. Ottawa: Indian Affairs and Northern Development.

--. (Former Minister of Aboriginal Affairs and Northern Development). 2013. In conversation with the author, April.

Moore, R. Walker, and I. Skelton. 2011. "Challenging the New Canadian Myth." Canadian Journal of Native Studies 31 (1): 17-42.

Ontario. 2009. Results-Based Plan Briefing Book 2009-2010: Ministry of Aboriginal Affairs. Toronto: Government of Ontario.

Papillon, Martin. 2012. "Adapting federalism." Publius: The Journal of Federalism 42 (2): 289-312.

--. 2014. "Framing self-determination." In Comparing Canada, edited by Luc Turgeon, Martin Papillon, Jennifer Wallner, and Stephen White. Vancouver: UBC Press.

Patterson, Lisa. 2006. Aboriginal Roundtable to Kelowna Accord: Aboriginal Policy Negotiations, 2004-2005. Ottawa: Library of Parliament.

Peters, Guy, and Jon Pierre. 2004. "Multi-Level Governance and Democracy" In Multi-Level Governance, edited by Ian Bache and Matthew Flinders. Oxford: Oxford University Press, pp. 75-92.

Piattoni, Simona. 2010. The Theory of Multi-level Governance. Toronto: Oxford University Press Quebec. 2006. Acting Now ... For the Future: First Nations Socioeconomic Forum Report. Quebec City: Government of Quebec.

RCAP. 1994 Report of the Royal Commission on Aborignial Peoples, Volume II. Ottawa: Supply and Services Canada.

Rodon, Thierry. 2014. "Working together: The experience of multilevel governance in Nunavut." Arctic Review of Law and Politics 5 (2): 250-270.

Rouillard, Christian, and Genevieve Nadeau. 2013. "Recasting and reframing a polymorphous concept." In Making Multilevel Public Management Work, edited by Denita Cepiku, David K. Jesuit and Ian Roberge. CRC Press, pp. 185-202.

Russell, Peter. 2004. Constitutional Odyssey. Toronto: UTP.

Scharpf, Fritz. 1994. "Games real actors could play." Journal of Theoretical Politics 6 (1): 27-53.

Smith, Jennifer. 2004. Federalism. Vancouver: UBC Press.

Watson, John. (Former Cabinet Secretary, Cabinet Committee on Aboriginal Relations). 2013. In conversation with the author, January.

Wilson, Gary, Christopher Alcantara, and Thierry Rodon. 2015. "Multilevel governance in the Inuit regions of the territorial and provincial north." In Aboriginal Multilevel Governance, Canada: The State of the Federation 2013, edited by Martin Papillon and Andre Juneau. Montreal-Kingston: MQUP, pp. 43-64.

Wotherspoon, Terry, and John Hansen. 2013. "The 'Idle No More' movement." Social Inclusion 1(1): 21-36.

Wyatt, Stephen, and Harry Nelson. 2014. "Aboriginal engagement in Canada's forest sector." institute of Intergovernmental Relations Working Paper. Kingston: Institute of Intergovernmental Relations.

Christopher Alcantara is Associate Professor, Department of Political Science, Western University, London, Ontario. Zachary Spicer is Assistant Professor, Department of Political Science, Brock University, St. Catharines, Ontario.

Table 1. Targets and Financial Allocation by Policy Area (5)

Policy area      5-Year target(s)      10-Year        Total financial
                                       target(s)        allocation

Education        --22,000 more      --High school     $1,800,000,000
                 high school        graduation
                 graduates          rate equal to
                 --14,800 more      other Canadians
                 post-secondary     --Improve post-
                 graduates          secondary
                                    completion
                                    rate by 50%

Housing and      --Reduce the on-   --Reduce on-      $1,600,000,000
Infrastructure   reserve housing    reserve housing
                 shortage by 40%    shortage by 80%
                 --Close gap in     --Eliminate gap
                 access to off-     in access to
                 reserve housing    off-reserve
                 by 50%             housing
                 --Reduce housing   --Reduce
                 gap in North       housing gap in
                 by 35%             North by 70%

Relationships    --Not specified    --Not specified    $170,000,000
and
Accountability

Economic         --Increase         --Increase         $200,000,000
Opportunities    employment         employment
                 levels by 30%      levels by 50%
                 --Narrow median    --Broadband
                 employment         access in an
                 income gap by      additional 250
                 half               communities

Health           --Reduce infant    --Reduce infant   $1,315,000,000
                 mortality, youth   mortality,
                 suicide and        youth suicide,
                 childhood          and child-hood
                 diabetes by 20%    diabetes by 50%
                                    --Double health
                                    professionals

Total                               --Equal living    $5,085,000,000
                                    standard