https://orcid.org/0000-0002-3035-6206
Abstract
In most federations, the division of power between central and subnational governments represents an important cleavage dividing voters and structuring party systems. Yet we lack a robust body of research regarding individuals’ preferences for different forms of devolved decision-making such as decentralization and asymmetrical federalism. This article contributes to this research agenda by analyzing the effects of identity and grievances on public opinion toward the division of powers in Canada. Leveraging four waves of the Confederation of Tomorrow survey, we find that respondents who identify predominantly with their province are more likely to prefer decentralization and asymmetrical federalism, whereas those who hold grievances against the federation prefer decentralization. Studying provincial variations in the impact of our main variables, we point to the role of the political context by showing that in certain provinces, the political mobilization of grievances strengthens the relationship between provincial identity and support for decentralization.
Introduction
The politics of federalism is characterized by a continual tension between orders of government who seek to increase their own powers, sometimes by encroaching on the constitutional division of powers afforded to other governments (Bednar 2009). Debates about the pertinence of further devolving policymaking to regional governments are raging in several multinational federations, such as Canada, the UK, and Spain. Proponents of decentralization put forward principles of subsidiarity, while supporters of centralization contend that a strong central government promotes equality between regions of a country.
Demands for decentralization most commonly emerge from subnational governments of regions with cultural, linguistic, or economic distinctiveness. Since the other regions of a country may not share the same appetite for decentralization, central governments may respond with some form of asymmetry, a limited transfer of the central government’s authority to the subnational jurisdictions that demand it, allowing the central government to respond differently to regions with distinct needs and demands. Yet, support for decentralized––and even asymmetrical––decision-making in federal systems, is not unique to minority nations and devolution of powers has been observed in most democracies (Hooghe et al. 2016).
Despite increased attention to the issue of decentralization, we lack a robust body of public opinion research on the allocation of authority and responsibilities between national and subnational governments (Henderson and Medeiros 2021). Public opinion toward the division of power is a key variable explaining the dynamics of (de)centralization (Dardanelli et al. 2019), but the conditions leading to support for the devolution of power still need to be identified. Furthermore, research has not yet fully examined public preferences toward different dimensions of the division of powers in federal systems; not only one’s preference for (de)centralization of power, but the possibility that decentralization could be asymmetrical between subnational units.
To address this gap, we pool four waves of the Confederation of Tomorrow survey, conducted annually from 2019 to 2023 in Canada. Canada offers an informative case to explore public opinion toward the federal division of powers. The Canadian federation is characterized by strong subnational identities and grievances against the central government (Gibbins and Berdahl 2003; Breton, Jacques, and Parkin 2022). For most of the country’s history, the division of power between the provinces and the federal government has been in flux as Canada is characterized by deep disagreements between competing narratives about the meaning of federalism, its objectives, the forms of its operations, and the degree of autonomy granted to different orders of government (Hueglin 2021).
We make several contributions to the study of public attitudes toward federalism. First, we compare the correlates of public preferences for decentralization and asymmetry, a subject that has been neglected in the literature. We find that Canadians have different views on these federal dynamics. Whereas a strong Canadian identification is associated with opposition to asymmetry, a strong subnational identity and grievances against the federation are both associated with increased support for decentralization. As such, the correlates of preferences for the division of power are firmly grounded in individuals’ identities and go beyond more fluid factors such as support for the party in power at the provincial and federal levels.
We also observe that in some instances, subnational identities can be mobilized politically. Previous research finds that territorial identification influences preferences for the (de)centralization of power, both in Canada (Medeiros and Gauvin 2021) and elsewhere (Galais Herrera, and Porta 2014; Guinjoan and Rodon 2014; Reuchamps et al. 2021; Verhaegen, Dupuy, and Van Ingelgom 2021). Our analysis advances this research by illustrating the scope conditions structuring the relationship between subnational identification and support for the devolution of powers in Canada. Pooling several annual surveys asking a common set of questions affords us the statistical power to study contextual variation in the effect of grievances and territorial identification across provinces. We show that those identifying most strongly with their province do not necessarily become more supportive of decentralization as their grievances against the federal government increase. Strong grievances do enhance the association between provincial identity and support for decentralization in Alberta, Saskatchewan, and Quebec, but not in the province of Newfoundland and Labrador, despite its residents typically identifying strongly with their province and holding clear grievances against the federal government. Our second contribution is therefore to argue that grievances can be used to mobilize support for further decentralization among those with strong provincial identities, but absent political mobilization, grievances alone do not condition the relationship between identity and preferences toward the division of powers.
Theory and hypotheses
Explaining preferences for the division of power
Group dynamics rooted in subnational identities and grievances have played a key role in explaining attitudes toward the division of powers between central and subnational governments in federal systems. Distinctive subnational identities are seen as the main catalyst motivating decentralization (Sorens 2005; Henderson, Jeffery, and Liñeira 2015). Rooted in theories of social identification and its consequences for ingroup favoritism (e.g., Tajfel and Turner 1986), we expect that citizens identifying primarily with their province as opposed to their country will prefer decision-making powers to be devolved from the central government toward their province. Furthermore, given the prominent role of group-based perceptions of fairness in shaping political attitudes (e.g., Mutz and Mondak 1997), grievances against the central government for the treatment of one’s province are expected to work in conjunction with provincial identification to mobilize support for the devolution of decision-making from central governments toward subnational units. Political actors may build on these dynamics, mobilizing subnational identities and perceived unfairness to gather support for their proposals to devolve power toward subnational units (Dardanelli et al. 2019). We posit that both provincial identification and grievances, alongside contextual political factors, explain public attitudes toward the division of power in federal systems.
To illustrate the importance of subnational identities in explaining support for devolved decision-making toward one’s province (or any subnational unit), consider the relationship between subnational identification and the dynamics of redistribution. Citizens with a predominant subnational identity generally perceive their community of solidarity to be within their region, and are therefore less likely to support redistribution toward other provinces. Citizens with an identity tied primarily to the central state, however, extend their community of solidarity across provincial boundaries (Balcells Fernández-Albertos, and Kuo 2015). Likewise, Canadian surveys on the federal equalization program, the most explicitly redistributive intergovernmental transfer to poorer provinces, reveal that respondents who identify predominantly with the central state are more likely to support equalization (Jacques, Béland, and Lecours 2022). Interprovincial redistribution in Canada does not only take place via equalization, but also via taxes collected by the central government (which are higher in richer provinces) to fund federal programs (which tend to be more important for poorer provinces). Centralization facilitates redistribution from rich regions to poor regions, whereas decentralization reduces the amount of money to redistribute between regions. Hence, the association between a pan-Canadian identity and support for centralization may be explained by a willingness to ensure that all citizens receive a similar level of public services.
Canadians’ strong provincial identities have restrained the process of centralization seen in other federations and helps to explain why Canada remains one of the most decentralized federations in the world (Dardanelli et al. 2019; Lecours 2019). Francophone residents of Quebec generally identify strongly with the Quebec nation, and Canadian decentralization can be partly explained by demands from Quebec for greater recognition and autonomy (Lecours 2019). However, Quebec is not the only province to be characterized by a strong provincial identity. Residents of Newfoundland and Labrador, the last province to join Canada in 1949 through a narrow referendum (see Blake and Baker 2019), generally identify first-are-foremost with their province before their country. In short, our theoretical discussion, as well as the results of previous studies (Galais Herrera, and Porta 2014; Guinjoan and Rodon 2014; Medeiros and Gauvin 2021; Reuchamps et al. 2021; Verhaegen, Dupuy, and Van Ingelgom 2021) lead us to pose our first hypothesis:
H1a. A stronger provincial (Canadian) identity is associated with support (opposition) for decentralization.
Another component of the division of power concerns the notion of asymmetry, which happens when the central government grants greater autonomy or authority to at least one subnational unit in a specific domain. Centralization and symmetry represent a centripetal view of power-sharing strategies toward diverse regions that impose limits to the accommodation of their demands, whereas decentralization and asymmetry are closer to a consociational view of federalism which favors the autonomy of the constitutive units (O’Leary 2013). Asymmetry is generally a response to plurinational diversity and to demands for autonomy made by a “mobilized” national minority group (McGarry 2007). Yet, asymmetry and decentralization do not necessarily go hand in hand (Brie and Mathieu 2021). Most federations tend to be symmetrical because subnational units in which the majority nation live may resist asymmetrical federalism if they perceive that it relegates them to a second-class status relative to the “privileges” granted to the minority nation (McGarry 2007). Hence, the clearest examples of asymmetry occur in unitary states where central authorities can act unilaterally without the consent of subnational authorities (McGarry 2007). Indeed, of the six federations analyzed by Dardanelli et al. (2019), Canada is the only one to practice some form of asymmetry in recent decades.
Asymmetrical federalism in Canada is political rather than constitutional, stemming from negotiations between provinces and the federal government, generally resulting from repeated demands for autonomy by successive governments in Quebec (Gagnon 2008). Negotiations allowed the province to opt out of federal programs in favor of provincial initiatives and to enjoy special authority in several policy areas, for example, by managing its own pension plan and opting out of the recent national childcare program since it was the only province that already had a quasi-universal childcare program (Arsenault Jacques, and Maioni 2018). Yet, Quebec’s demands for asymmetry have become de facto symmetrical, as similar arrangements have been offered to other provinces, though they have not always chosen to use them (Hueglin 2021). Increasingly, however, provincial demands for asymmetry do not only stem from Quebec’s calls for greater autonomy. Western provinces such as Alberta and Saskatchewan have also demanded special powers to protect and promote their distinct economies based on oil production (Gibbins and Berdahl 2003; Hueglin 2021; Medeiros and Gauvin 2021), while the Atlantic provinces of Newfoundland and Labrador and Nova Scotia have also strongly advocated for an asymmetrical approach to natural resource management, developing unique co-management bodies for offshore resources with the federal government (Dunn 2003).
To our knowledge, only two studies relate to individuals’ preferences for asymmetry, and both are in the Canadian context surrounding debates over national unity. Johnston and Blais (1996) study public opinion about the referendum on the Charlottetown agreement of 1992, which was about entrenching asymmetrical federalism in the constitution, while Ferland and Turgeon (2020) focus on Canadians’ attitudes toward “doing more for Quebec,” which may necessitate asymmetry to accommodate provincial demands. Ferland and Turgeon (2020) find that survey respondents with stronger Canadian identities want to “do less” to accommodate Quebec’s demands. Hence, in seeking to shed more light on public attitudes toward asymmetry, we expect that a stronger Canadian identification will be negatively correlated with preferences for asymmetrical federalism. In part, we believe this to be the likely case because the notion of equality between provinces has been mobilized by opponents of the recognition of the special status of Quebec and has been at the core of political discourse elsewhere in the country (Gagnon and Laforest 1993). Moreover, the association between Canadian identity and willingness to reduce inequality between Canadian citizens may tie into support for the notion of equality between the provinces, whereas asymmetry tolerates inequalities in the devolution of power. In contrast, provincial identifiers may be more inclined to support asymmetry if they want their own province to have more power. This leads us to pose the following hypothesis:
H1b. A stronger provincial (Canadian) identity is associated with support (opposition) for asymmetry.
Our second explanatory variable of interest, grievances stemming from the treatment of one’s province in the federation, are a separate but related factor in the literature on public opinion toward the division of power in federal systems. Grievances may be related to concerns about the unequal distribution of economic gains or losses between groups; that is, a lack of “socio-tropic fairness” (Mutz and Mondak 1997). Mutz and Mondak (1997) showed that voters react to this perceived group disadvantage by punishing the incumbent when groups with which they identify are losing economic grounds. Economic grievances influence preferences toward federal arrangements as poorer regions protest not receiving enough, while richer regions object to paying too much (Massetti and Schakel 2015). If one perceives that the subnational group with which they identify is treated unfairly and is not well represented in the federation, one may seek to punish the central government deemed responsible for this injustice or take action to promote greater input from subnational authorities by devolving powers from the central government to their province. Consequently, grievances against the central government may be associated with a willingness to decentralize power toward the subnational government. Indeed, Henderson, Jeffery, and Liñeira (2015) have shown that the belief that one’s nation does or does not receive its fair share of the benefits in return for their contribution shapes constitutional preferences.
Such grievances against the federal government are prominent in the Canadian political landscape and are central to Quebec’s desire for increased autonomy as well as growing calls for more decision-making authority from other provincial jurisdictions. The concept of Western alienation has been central in explaining the institutional and economic grievances in the Prairie provinces of Alberta and Saskatchewan, whose governments lament they do not receive their fair share of resources in the federation (Gibbins and Berdahl 2003). In Newfoundland and Labrador, the results of a provincial Royal Commission into the province’s relationship with Canada also highlighted strong grievances against the federal government (Dunn 2003). These grievances are attributed to the perception that Newfoundland and Labrador brought extensive natural resource wealth into Canada with confederation to the benefit of others, yet the province still does not see the promised benefits to the economy and living standard, which are further undermined by the associated negative consequences of outmigration (Dunn 2003; Delisle 2013). Hence, we hypothesize:
H2. Grievances related to the treatment of one’s province are associated with support for decentralization.
We do not have clear expectations regarding the relationship between grievances and asymmetry. On the one hand, if respondents associate asymmetry with the accommodation of Quebec’s demand, it is likely that grievances will correlate negatively with support for asymmetry. Indeed, grievances and resentment in the federation are often related to Quebec and its perceived lack of contribution to the common good or as being favored with asymmetrical powers (Breton, Jacques, and Parkin 2022). Support for “doing more for Quebec” is lower if respondents believe that their province does not receive its fair share of the resources of the federation, which is an attitude related to regional grievances (Ferland and Turgeon 2020).
On the other hand, the era during which the accommodation of Quebec’s minority nationalist movement was at the center stage of Canadian politics is over (Hueglin 2021), and other provinces have started demanding asymmetrical arrangements. Respondents may therefore not be thinking about Quebec when they respond to a general survey question on asymmetry. Also, their grievances may be geared toward the federal government and as such, they may prefer if Ottawa gave more autonomy to the provinces that demand it, rather than refusing their requests. This would involve a positive association between grievances and support for asymmetry. Since we have theoretical expectations for the association between grievances and asymmetry to be either positive or negative, we refrain from posing a formal hypothesis and instead explore the relationship between the two variables.
The political activation of grievances and identity
While we hypothesize that identity and grievances are both crucial factors explaining support for devolution in federal systems, we contend that provincial identification has a particularly strong impact on preferences for decentralization if it becomes politically activated by salient grievances against the federal government and, critically, by a political context that mobilizes public sentiment in favor of further devolution of power toward the province. The political activation of subnational identification occurs when place-based resentment develops (Munis 2022). Place-based resentment follows from the perception that political elites and current institutions produce representational, cultural, and distributive inequities against the region with which people identify. Hence, grievances can help to politicize identification to one’s province when provincial politicians champion further decentralization by presenting devolved decision-making as a solution to provincial grievances against the federal government.
We therefore expect provincial identity to have a stronger association with public preferences for decentralization when grievances against the federal government over the treatment of one’s province are present, and these grievances have been mobilized by political actors to demand additional autonomy. While one could contend that identity and grievances reinforce each other regardless of the political context, we argue that a political context mobilizing these feelings is a necessary condition to any interaction between provincial identity and grievances. To make this argument, we start from the premise that political context and socialization influence preferences and generate attitudinal divergences in divided societies like Canada (Stiers and Hooghe 2023). In their study of the impact of ethnocentrism, Kinder and Kam (2009) argue that attitudes such as ethnocentrism must be politically activated by priming and framing effects by the media and political actors to influence policy attitudes. In our case, citizens are more likely to connect their provincial identification and grievances to demands for additional provincial autonomy if provincial political actors attribute problems to the federal government and as a result, demand greater local decision-making authority to solve these problems.
These arguments are supported by several previous studies suggesting that the political activation of subnational identity and grievances increases the effect of these variables on citizens’ preferences. Bond and Rosie (2010) show that territorial identity influences support for constitutional change in Wales, Scotland, and Northern Ireland, but not in England, where political actors are less likely to connect Englishness to demands for autonomy. Some argue that the presence of a minority nationalist movement is necessary to activate the impact of subnational identity on preferences (Guinjoan and Bermúdez 2020). In contrast, Medeiros and Gauvin (2021) show that identity influences preferences for decentralization not only in Quebec, but also in the Prairies, where economic grievances, not nationalism, have led to regional tensions and resentment against perceived federal encroachment. This phenomenon is best exemplified by the emergence of the Reform party in the late 1980s who put the federal government on notice that, “the West wants in.” Since then, Western alienation has remained a salient feature of party politics in the Prairie provinces, most notably in Alberta and Saskatchewan. Medeiros and Gauvin (2021) show that identity does not influence preferences for decentralization outside Quebec and the Prairies but do not study interprovincial variations in the impact of grievances.
We seek to build on this work to illustrate a more nuanced relationship between subnational identification and grievances in explaining public opinions toward federalism. Such an analysis, we hope, might serve our understanding not only of the Canadian case, but yield insights that are relevant for comparative public opinion research in other federal systems as well. We argue that grievances strengthen the relationship between provincial identification and opinions about decentralization in three provinces where provincial political parties regularly call for increased autonomy, Alberta, Saskatchewan, and Quebec. Indeed, the Parti Québécois, an independentist party, has been one of the two governing parties with the Quebec Liberal Party from 1976 to 2018, and support for independence remains comparatively high in the province. According to the Confederation of Tomorrow surveys, which we use in this research, 22.2 percent of respondents in Quebec consider themselves to be sovereigntists, against 33.5 who identify as federalists. Even though independence has never been an option proposed by a popular party in Alberta or Saskatchewan, 49 percent of Albertans and 48.3 percent of Saskatchewanians strongly or somewhat agree with the following statement: “Western Canada gets so few benefits from being part of Canada that they might as well go it on their own.” These three provinces are also where federalism is seen as the least advantageous (Brie and Mathieu 2021).
While challenges certainly arise between provincial and federal governments outside of these three provinces, many provinces have often advocated for greater provincial consultations and input into federal decision-making, not necessarily the devolution of powers from the federal government to the province (e.g., Dunn 2003). Elsewhere in the country, while provincial identification and grievances are expected to be predictive of support for decentralization, the association between provincial identification and preferences for different federal arrangements should be less impacted by grievances against the federal government as there are no major political parties mobilizing grievances to champion further decentralization or asymmetry. Such dynamics are noticeably absent in the politics of Newfoundland and Labrador. A 2003 survey finds that despite residents of Newfoundland and Labrador having among the strongest levels of provincial identification in the country and reporting frustration with the federal government, only about 12 percent of survey respondents oppose Confederation between Newfoundland and Canada.
What explains the absence of political mobilization for greater devolved decision-making in Newfoundland and Labrador? Newfoundland and Labrador has a small population (about 540 000) living across a very large territory, which increases the cost of public services and the dependence on federal transfers. Access to federal social transfers was a core factor affecting voters’ decisions to join Canada in 1949 (Blake and Baker 2019). While the province did not receive equalization from 2009 to 2023, Newfoundland and Labrador maintains one of the largest negative fiscal balances of the federation, receiving significantly more federal expenditures and transfers than what it pays in tax revenue to Ottawa (Tombe 2018). This contrasts with Alberta and Saskatchewan, the two richest provinces per capita that are net contributors to the central state, and with Quebec, which, despite being a poorer province and a recipient of significant equalization payments, has developed a stronger state capacity to bolster its demands for decentralization.
Hence, we suspect that despite holding strong grievances against the federal government and identifying strongly with their province, Canadians are still not all likely to share the same preferences for decentralization. We contend that grievances and provincial identification reinforce each other, and fuel demands for decentralization in provinces with party systems that mobilize feelings of alienation against the central government, such as in Quebec and the two Western Prairie provinces of Alberta and Saskatchewan. However, grievances and provincial identity do not interact when the party system does not mobilize this alienation against the central government to fuel demands for decentralization. This situation can happen in provinces that maintain a fiscal dependence on the central state (such as in Newfoundland and Labrador and potentially other Atlantic provinces), or in provinces where grievances are weak and provincial identification quasi-inexistent (such as in Ontario). This leads us to the following hypothesis:
H3. Identity and grievances interact to generate stronger demands for decentralization in Alberta, Saskatchewan, and Quebec.
In contrast to support for (de)centralization, we do not expect that the effect of identity and grievances on asymmetry differs by provinces since asymmetry is not a salient and divisive issue in most provincial party systems. Considering the absence of public opinion research on the role of subnational identity and grievances on support for asymmetry, we refrain from making any predictions about the nature of this relationship but present the data here for exploratory purposes and to inform future work.
Data and methods
Dependent variables
To test these expectations, we use pooled data from the Confederation of Tomorrow Surveys from the years 2019 to 2023,1 which include repeated questions across survey waves providing sizeable samples from each of Canada’s provinces.2 Each question offered respondents the possibility to choose that they “cannot say,” in which case they were excluded from the analysis. As outcome variables, we focus on two survey items measuring public preferences for the division of powers in the Canadian federation. First, to measure support for decentralization, respondents were asked, “Which of the following statements is closest to your own view of how the government should work in Canada?” Respondents indicated whether “the government of my province should take charge of many of the things the federal government does right now,” an indicator of support for increased decentralization, or whether, “the federal government should take charge of many of the things the government of my province does right now,” an indicator of support for increased centralization. Respondents could also indicate whether, “things should be left pretty much as they are,” which we set as our reference category.
Our second outcome measure captures a specific form of support for decentralization: support for asymmetrical powers among subnational units within the federation. “Thinking about the distribution of powers between the different levels of government in Canada,” respondents were asked to indicate, “which of the following do you prefer?” Those favoring an asymmetrical distribution of powers could indicate, “the federal government should be allowed to offer more powers to those provinces that want them, so that the federal system can respond to the different needs that some provinces may have.” On the other hand, those opposed to an asymmetrical distribution of power could indicate, “the federal government should treat each province the same, so that no province has any more powers than another.” Our operationalization about support for asymmetry captures general support for asymmetrical federalism within Canada.
Independent variables
Our independent variables of interest measure respondents’ subnational identity and their grievances against the federal government. To measure subnational identity, we draw on the well-known Linz-Moreno question, which measures whether respondents identify primarily with their country, primarily with their province, or with both equally. Respondents were told that, “people have different ways of defining themselves,” before being asked, “do you consider yourself to be: A Canadian only; A Canadian first, but also [provincial identity; for example, a Quebecer]; equally a Canadian and a [provincial identity]; A [provincial identity] first, but also a Canadian; or A [provincial identity] only.” This variable was recoded on a five-point scale ranging from 0 to 1, with higher scores indicating a primarily provincial identification.
To measure provincial grievances, we construct an additive index from three survey questions that tap into representational and distributive considerations relating to place-based resentment (Munis 2022). One item asked, “in your opinion, is [province] treated with the respect it deserves in Canada?” Respondents that indicated yes are coded 0, while those indicating a perceived lack of respect are coded as 1. A second item asked, “in your opinion, how much influence does [province] have on important national decisions in Canada?” Respondents that indicating their province has “more than its fair share” or “about its fair share” of influence are coded 0, while those indicating their province has “less than its fair share” of influence are coded as 1. A third item instructed respondents to think, “about all the money the federal government spends on different programs and transfers to the provinces and territories,” and asked whether their province receives “more than its fair share,” “about its fair share” (coded 0) or “less than its fair share” (coded as 1). These indicators were summed together into a 0 to 1 scale with higher scores reflecting greater grievances (α = 0.78).
Control variables
In addition to our independent variables, our models include several control variables. We control for political ideology since left- (right-) wing respondent may prefer (de)centralization to increase (decrease) redistribution between unequally endowed provinces (Borwein et al. 2023). Moreover, outside Quebec, the political left is more likely to identify with Canada (Borwein et al. 2023). Political ideology may also be related to grievances since at the time the surveys were conducted, most provinces were governed by right-of-center parties, whereas the federal government was left-of-center. We control for political ideology with a question asking respondents about their preferences regarding the size of government (higher scores indicate a preference for smaller government).3 We also control for partisan alignment between respondents’ stated vote choice and the incumbent party forming government at the provincial and federal levels. Citizens may be more inclined to devolve power to the provincial government if they support the incumbent provincial government, while voters of the incumbent federal government or of the provincial opposition may prefer increased centralization (Breton, Lucas, and Taylor 2022). As voting for the federal incumbent may also correlate with grievances, we control for incumbent vote choice in previous provincial and federal elections. We also control for respondents’ gender (dummy variable, with female scored 1), level of education (dummy variable, with those who attended a post-secondary institution scored 1), income (measured on a six-point scale, ranging from under $30,000 to $150,000 or more and recoded to range from 0 to 1), language (a dummy variable capturing whether French is the main language spoken at home), and a dummy variable reflecting whether respondents are born outside of Canada. We present the descriptive statistics of all the variables included in the models below in Supplementary Appendix Table A1.
Methods
When predicting support for (de)centralization, we use multinomial logistic regressions. While it may be argued that our measure of decentralization is an ordinal measure, it is not clear that respondents perceive centralization versus decentralization as a matter of degree, or as opposite poles on the same continuum, with support for the status quo between the two. It may be the case that these reflect qualitatively different modes of governance, and therefore better reflected as a nominal variable. Moreover, most respondents are split between preferring the status quo or further decentralization, with relatively few respondents favoring greater centralization (see fig. 1). Given the heavily decentralized nature of the Canadian federation, supporting the status quo is closer to support decentralization than supporting centralization. The relationship between variables across levels of our dependent variable is not expected to be consistent as a respondent judges the implications of more centralization or more decentralization in relation to the status quo, violating the parallel trend assumption of ordered logistic regression. The multinomial logistic regression allows us to estimate support for either greater centralization or greater decentralization in relation to the status quo. In all models, we include provincial and survey years fixed effects and cluster the standard errors by province and survey years. To model support for asymmetry, we dichotomize the measure and use logistic regression to model support for greater asymmetry relative to preferences for symmetrical powers across the provinces.
Distribution of support for decentralization by province. Confederation of Tomorrow pooled survey data, 2019–2023.
Results
Public support for (de)centralization and asymmetry in Canada
We begin our analysis with an examination of the distribution of our two dependent variables, public support for decentralization and asymmetry. Figure 1 shows that Canadian public opinion toward the division of power varies across the provinces. Support for a more decentralized federation is highest among respondents from Quebec (42 percent), Alberta (42 percent), Saskatchewan (39 percent), and Newfoundland and Labrador (35 percent). Decentralization is also supported by a plurality of respondents from British Columbia (32 percent) and Nova Scotia (29 percent), though in both cases support for further decentralization is nearly indistinguishable from the proportion of respondents preferring the status quo. In all other cases, support for more decentralization is equal to, or less than, support for the status quo. It is noteworthy that support for more centralization remains weak across the country, not only in regions that traditionally demand more autonomy like Quebec and the western Prairie provinces of Alberta and Saskatchewan. Less than a quarter of respondents (and in some provinces, as low as 15 percent) support the federal government taking charge of many responsibilities handled by the provinces.
Figure 2 shows that in all ten provinces, a plurality of respondents oppose asymmetrical federalism which suggests the doctrine of the equality of provinces is strongly entrenched in the country. The idea that the federal government should treat all provinces equally is especially popular among respondents from the Atlantic region, including New Brunswick (47 percent), Nova Scotia (56 percent), Newfoundland and Labrador (54 percent), and Prince Edward Island (54 percent). It is also supported by 50 percent or more of the respondents in Canada’s three Prairie provinces, Alberta, Saskatchewan, and Manitoba. Even in Quebec, where asymmetrical federalism has received the most attention, asymmetry is favored by only 28 percent of respondents. We also note a larger proportion of respondents did not know what to say to the question on asymmetry compared to the question on decentralization, suggesting that respondents do not have as well-defined opinions on the issue.
Distribution of support for asymmetry by province. Confederation of Tomorrow pooled survey data, 2019–2023.
Explaining public support for decentralization and asymmetry
To test our expectations about the role of provincial identification and grievances on support for decentralization and asymmetry, we regress public support for decentralization and asymmetry on measures of provincial identification and grievances, controlling for different political indicators and personal background characteristics (see Supplementary Appendix Table A2 for full model results). We also present models without controls in the Supplementary Appendix (Supplementary Appendix Tables A3 and A4). In Table 1, we report the average marginal effects (AMEs) of interest capturing the average change in the probability of a response to our two outcome measures, across a unit shift on a given independent variable, controlling for other variables included in our models. We visualize changes in the predicted probability of a given response in figure 3.
Predicted probability of supporting decentralization, the status quo, centralization, and asymmetry, as a function of provincial identification and grievances.
Note: 95% confidence intervals shown. Calculated using the models in Supplementary Appendix Table A2.
AMEs of provincial identification and provincial grievances (from minimum to maximum values) on support for centralization, status quo, decentralization, and asymmetry.
| Centralization | Status quo | Decentralization | Asymmetry | |
|---|---|---|---|---|
| Provincial identification | −0.111*** | −0.121*** | 0.232*** | 0.123*** |
| (0.023) | (0.026) | (0.031) | (0.025) | |
| Provincial grievances | 0.013 | −0.234*** | 0.221*** | 0.051† |
| (0.022) | (0.016) | (0.022) | (0.027) | |
| Controls | Yes | Yes | Yes | Yes |
| Observations | 9,880 | 9,880 | 9,880 | 10,097 |
| Centralization | Status quo | Decentralization | Asymmetry | |
|---|---|---|---|---|
| Provincial identification | −0.111*** | −0.121*** | 0.232*** | 0.123*** |
| (0.023) | (0.026) | (0.031) | (0.025) | |
| Provincial grievances | 0.013 | −0.234*** | 0.221*** | 0.051† |
| (0.022) | (0.016) | (0.022) | (0.027) | |
| Controls | Yes | Yes | Yes | Yes |
| Observations | 9,880 | 9,880 | 9,880 | 10,097 |
Note: Robust standard errors, clustered by province and survey year, in parentheses. Calculated using the models in Supplementary Appendix Table A2.
P < .10,
P < .05,
P < .01,
P < .001.
AMEs of provincial identification and provincial grievances (from minimum to maximum values) on support for centralization, status quo, decentralization, and asymmetry.
| Centralization | Status quo | Decentralization | Asymmetry | |
|---|---|---|---|---|
| Provincial identification | −0.111*** | −0.121*** | 0.232*** | 0.123*** |
| (0.023) | (0.026) | (0.031) | (0.025) | |
| Provincial grievances | 0.013 | −0.234*** | 0.221*** | 0.051† |
| (0.022) | (0.016) | (0.022) | (0.027) | |
| Controls | Yes | Yes | Yes | Yes |
| Observations | 9,880 | 9,880 | 9,880 | 10,097 |
| Centralization | Status quo | Decentralization | Asymmetry | |
|---|---|---|---|---|
| Provincial identification | −0.111*** | −0.121*** | 0.232*** | 0.123*** |
| (0.023) | (0.026) | (0.031) | (0.025) | |
| Provincial grievances | 0.013 | −0.234*** | 0.221*** | 0.051† |
| (0.022) | (0.016) | (0.022) | (0.027) | |
| Controls | Yes | Yes | Yes | Yes |
| Observations | 9,880 | 9,880 | 9,880 | 10,097 |
Note: Robust standard errors, clustered by province and survey year, in parentheses. Calculated using the models in Supplementary Appendix Table A2.
P < .10,
P < .05,
P < .01,
P < .001.
We find strong associations between both provincial identification and grievances against the federal government with preferences toward the division of powers in Canada. Compared to someone who identifies exclusively with Canada, those reporting the strongest level of provincial identification are an estimated 23 percentage points more likely to support further decentralization (AME = 0.232, SE = 0.031, p < .001), 12 percentage points less likely to support the status quo arrangement (AME = −0.121, SE = 0.026, p < .001), and 11 percentage points less likely to desire more centralization (AME = −0.111, SE = 0.023, p < .001). A primary provincial identification is associated with an estimated 12-percentage-point increase (AME = 0.123, SE = 0.025, p < .001) in support for asymmetrical federalism. In line with our first set of expectations, strong provincial identities are associated with preferences for more decentralization and asymmetry in the federal-provincial power sharing relationship, independent of other variables controlled for in our models.
Turning to the role of grievances, we again observe a similar estimated increase in support for decentralization, in line with our second hypothesis. Grievances against the treatment of one’s province are positively associated with support for decentralization (AME = 0.221, SE = 0.022, p < .001) and negatively with support for the status quo (AME = −0.234, SE = 0.016, p < .001). All else equal, grievances are uncorrelated with preferences for further centralization. Although the estimated association between grievances and asymmetry is marginally statistically significant (p = 0.055), the shift in predicted probability comparing those with highest and lowest levels of grievances in the sample is small (AME = 0.051, SE = 0.027).
We observe other noteworthy results pertaining to political attitudes and support for devolved powers (see Supplementary Appendix Table A2). Compared to the status quo arrangement, respondents who prefer a smaller government tend to prefer decentralization and oppose asymmetry. This goes in line with previous studies showing that support for “doing more for Quebec” (Ferland and Turgeon 2020) and support for a strong central government (Borwein et al. 2023) is lower among the right. Voting for the federal government is associated with lower support for decentralization and for asymmetry, but it is not associated with support for centralization, whereas voting for the provincial incumbent reduces support for centralization and increases support for decentralization and for asymmetry.
These results illustrate the direct, independent associations of provincial identification and grievances against the federal government with support for further decentralization and asymmetrical federalism. It may be the case, however, that among those harboring strong grievances against the federal government, provincial identification is an especially strong predictor of preferences for devolved decision-making. To test the possibility that the estimated relationships between subnational identity and support for decentralization and asymmetry are conditioned by grievances, we conduct moderation tests using two-way interaction terms. The results are reported in Supplementary Appendix Table A5. We find no evidence that grievances condition the relationship between provincial identification and support for further decentralization or asymmetry. Interestingly, grievances moderate the relationship between provincial identification and support for centralization; a primary provincial identification reduces support for centralization more when grievances are low and the effect of provincial identification on support for centralization gets closer to zero when grievances are high (see Supplementary Appendix Figure 1A). Hence, grievances and provincial identity substitute each other in their impact on support for centralization. Nonetheless, few respondents favor centralization (see fig. 1) and the combination of identification and grievances does not result in increased support for decentralization or asymmetry, indicating that the effect of identity remains regardless of changing levels of grievances.
Interprovincial variation in the role of subnational identification
The results presented so far illustrate the influential role of subnational identification and grievances in explaining support for different forms of devolved decision-making in federal systems. It is important to note, however, as we described above, that there is significant variation in the extent to which provincial identities and grievances are expressed throughout the country. Data from the Confederation of Tomorrow surveys further illustrate this point in figure 4. In this section, we exploit the pooling of multiple survey waves to leverage sizeable samples even in Canada’s smallest provinces. This approach affords us a nuanced look at how the salience of provincial identities and grievances varies across the country, and whether this variation leads to more pronounced support for decentralization and asymmetry uniformly across the country when grievances are high, or whether the influence of identity and grievances in mobilizing public opinion is limited to certain regions.
Provincial identification, measured on a scale from 0 (exclusively Canadian) to 1 (exclusively provincial), and grievances, measured on a 0 (low) to 1 (high) scale, by province of residence. 95% confidence intervals shown.
As can be seen from figure 4, with the notable exceptions of Quebec and Newfoundland and Labrador, most Canadians identify more strongly with Canada than with their province of residence. Grievances against the federal government, however, are quite pronounced in nearly all regions of the country but are most notable in the provinces of Newfoundland and Labrador, Saskatchewan, and Alberta but remain lower in Ontario and Quebec.
In the previous section, we determined there was no evidence that grievances moderate the relationship between provincial identification and support for decentralization or asymmetry (see Supplementary Appendix Table A5).4 In this section, we take this analysis a step further to test whether there is a moderating effect of grievances on the relationship between provincial identity and public opinion, but only in provinces that lobby for greater provincial autonomy: Alberta, Saskatchewan, and Quebec.5 To test this expectation, we calculate the predicted probabilities of supporting decentralization, centralization, and asymmetry within each province, among respondents with a primary provincial identity at relatively high or low levels of grievances concerning the way their province is treated in the federation.6 Supplementary Appendix Table A6 presents the full regression results, with the three-way interaction between provincial identification, grievances, and province of residence visualized in figure 5. We refer the reader to Supplementary Appendix Figure 5A in the supplementary materials for full results at high and low levels of both provincial identification and grievances.
Predicted probabilities of supporting decentralization, centralization, and asymmetry within each province, among respondents with a primary provincial identity at relatively high or low levels of grievances.
Note: 95% confidence intervals shown. Calculated using the models in Supplementary Appendix Table A6. See Supplementary Appendix Figure 2A for full results from multinomial modelling.
The results confirm our third hypothesis that grievances do in fact moderate the relationship between identity and support for decentralization, but only in the provinces of Alberta, Saskatchewan, and Quebec, where provincial politicians frequently mobilize grievances against the federal government to call for greater provincial self-rule. On average, outside of these three provinces, we observe consistent associations between provincial identification and support for decentralization, independent of their levels of grievances. Particularly noteworthy is the lack of variation in the association between identification and decentralization among respondents from Newfoundland and Labrador at different levels of grievances. Despite strong grievances and provincial identification, we do not observe increased support for decentralization among Newfoundlanders and Labradorians. In Alberta, Saskatchewan, and Quebec, however, a primary provincial identification results in significantly more support for decentralization when grievances are high. In contrast, at relatively low levels of grievances, respondents identifying with their province from these three provinces report comparable levels of support for decentralization as respondents from the rest of Canada.7
Finally, with respect to support for asymmetry, our results highlight inconsistent effects across the country. While in most provinces, grievances do not moderate the relationship between provincial identity and preferences for asymmetry, support for asymmetry is stronger among provincial identifiers who have more grievances against the federal government in Ontario and Quebec but weaker in Prince Edward Island.
Discussion and conclusion
This article studies individuals’ preferences for two crucial elements of the division of power in a federation: (de)centralization and asymmetry. Pooling four waves of the Confederation of Tomorrow surveys, we find strong public support for decentralization in Canada, already a highly decentralized federation. Survey results indicate Canadians’ are generally satisfied with the status quo power-sharing relationship. In some provinces, a plurality of respondents supports further decentralization. This result echoes previous research showing citizens generally want to devolve authority closer to where they live (Breton, Lucas, and Taylor 2022). Previous studies have shown that Canadians support centralization to manage certain policy issues (Brie and Mathieu 2021). However, using generalized measures of support for (de)centralization and asymmetry that do not specify any particular policy domain, we find that Canadians have little appetite for increased centralization.
What explains public preferences toward decentralization in a multinational federal system characterized by significant regional diversity? Building on research by Medeiros and Gauvin (2021), we show that a greater identification to one’s province (as opposed to Canada) is indeed associated with a preference for decentralization across the country. We also find that place-based grievances correlate with greater support for decentralization. However, because of the cross-sectional nature of our analysis, we cannot confirm if grievances precede dissatisfaction with the current division of power in Canada or if opinion toward the division of power generates grievances. To solve this problem and to better assess causal relationships, one should be using panel surveys or leverage experimentation, which is unfortunately uncommon in the study of preferences regarding federalism.
We make two main contributions pertinent to researchers interested in public opinion toward federalism beyond the Canadian case: we advance not only decentralization, but also asymmetry, as an important dimension of attitudes toward federalism, illustrating the extent of support for these two dimensions differ. We also shed light on the complex relationship between provincial identification, grievances against the central government, and political context in mobilizing public support for decentralization. Indeed, we find that subnational identification can be politically activated by grievances in certain instances to produce especially strong demands for changes to the existing power-sharing arrangement to afford greater decentralization of powers from the federal government to the provinces. This is the case in Alberta, Saskatchewan, and Quebec, but not in Newfoundland and Labrador. It is noteworthy that provincial identification does not interact with grievances in Newfoundland and Labrador, despite residents of the province typically reporting a very strong provincial identification and grievances against the federal government.
While still limited by the correlational analysis, this finding allows us to reflect on the possible scope conditions around which subnational identification strengthens support for devolved decision-making in federal systems. Possible scope conditions, we believe, may relate to the political activation of identity and grievances. In a small province like Newfoundland and Labrador, political parties do not tap into citizens’ strong identity and high grievances to demand self-rule, as it may be considered that further decentralization would be detrimental to the economic interests of the province. Indeed, even though grievances against the federal government are particularly acute in the province, as highlighted not only by the Confederation of Tomorrow survey data, but also in the province’s Royal Commission (2003) into its relationship with Canada, or during the “Anything but Conservatives” campaign led by former provincial Premier Danny Williams against the incumbent Conservative federal government, the government of Newfoundland and Labrador advocated for further engagement and negotiation with the federal government to resolve the issues at hand, seeking to re-establish positive relationships between the two levels of government. It is quite possible the fiscal dependency of a subnational unit on the central government, or the relatively weak bureaucratic capacity of a small and sparsely populated province may represent a factor that dampens the political activation of identity and grievances. Provinces must be able to make good use of additional autonomy to demand it. Future research, however, is needed into this issue to refine the conditions and mechanisms that lead to the political activation of identity and grievances.
Our arguments about the political mobilization of subnational identity raise considerations about the generalization of our findings to other federal systems. We believe that here, too, the Canadian case can be instructive. Provincial political actors have strong incentives to mobilize public support for decentralization because changes to the division of power are inherently political in Canada, as formal constitutional change remains unlikely. The federation is characterized by conflicts about the division of power, competition, and encroachment of the powers between orders of governments. Moreover, since Canada has weak formal institutions of shared rule between provincial and federal governments, provincial political actors may therefore be more incentivized to mobilize public opinion in favor of self-rule (see also Schakel and Brown 2021). The political mobilization of identity and grievances may be less likely in a more cooperative and less conflictual federation with more shared rule institutions. Furthermore, it is likely that identity has stronger effects in a multinational federation like Canada than in a territorial federation with weaker subnational identities like Australia or the United States, for example. In such a federation, the debate about the division of power is potentially more likely to be structured by a left-right conflict about redistribution. Future research should compare the impact of identity and grievances on demands for self-rule between countries and types of federation.
While significant effort has been made to examine support for decentralization, research has been notably absent regarding public attitudes toward asymmetrical federalism. As such, our analysis of the determinants of preferences for asymmetry remained exploratory as we have few theoretical guidance to explain preferences regarding this issue. Still, in line with our expectations, provincial identity is associated with support for asymmetry, whereas grievances are positively correlated with support for asymmetry, but the association does not reach statistical significance. Our results suggest that grievances reflect dissatisfaction related to the federation and are associated with demanding more flexible federal arrangements that respond to the different needs of distinct communities. Asymmetrical federalism should therefore be considered not just as a tool to satisfy Quebec, but also as a policy solution to assuage the high level of grievances against the federation found in several provinces.
However, we do not find that grievances and identity are mobilized by political actors to demand asymmetry in certain provinces, even in Quebec. Indeed, the topic is not particularly salient, nor popular as we find little support for asymmetrical federalism. This suggests that the doctrine of the equality of provinces is well entrenched in Canada. Hence, Canada tends to be conceived by its population as a territorial federation, composed of ten equal units that should be treated symmetrically. This diverges with the multinational view of federalism, which promotes asymmetrical federalism to respond to the distinct needs of the different nations that compose the federation like Quebec and the Indigenous communities (Gagnon 2008; Brie and Mathieu 2021).
We recognize that our question on asymmetry suffers from several drawbacks which represent a limitation of the study. First, the question is complicated and almost a third of respondents do not express an opinion, suggesting that many Canadian citizens are not holding strong opinions about asymmetrical federalism, or perhaps not having enough information to respond to the question. Second, we do not know if respondents have a particular province in mind when answering the question. Since the question asked in the Confederation of Tomorrow surveys is about asymmetry in general, it may also suggest that some individuals need to know which provinces want more powers to form an opinion on the topic. Third, our measure does not specify whether this asymmetry would be entrenched in the constitution or granted by more a political negotiation. Still, we believe that this question is useful to study the individual determinants of support for asymmetry, which remains a blind spot of the literature. Further research should refine the question about which region of the country would be accommodated by asymmetry and compare average support for the demands of different regions.
In conclusion, while we believe that public opinion toward the distribution of power is an important issue, as politicians often claim to have a mandate from their constituents to alter the division of power (Breton, Lucas, and Taylor 2022), future research should investigate the extent to which politicians are responsive to public opinion when it comes to the “high politics” of the distribution of power in a federation.
Acknowledgements
The authors would like to thank the four anonymous reviewers and the editors for their comments on the article. Earlier versions of this article were presented at the Canada Research Chair in Electoral Democracy in Montréal, the 2023 Canadian Political Science Association Annual Conference in Toronto, and the 2024 Swiss Political Science Association Conference in St Gallen. The authors thank Jack Lucas, Sean Mueller, and the participants of these conferences for their feedback. The authors wish to thank the Centre d’Analyse Politique-Constitution et Fédéralisme as well as Andrew Parkin for putting together and sharing the survey data. All authors have contributed equally, and their names are listed alphabetically.
Supplementary material
Supplementary data can be found at www.publius.oxfordjournals.org.
Funding
Philippe Chassé was supported by funding from the Social Sciences and Humanities Research Council of Canada and Sciences Po, Olivier Jacques by the Fonds de Recherche du Québec-Société et Culture, NP—327434, and Colin Scott by the Social Sciences and Humanities Research Council of Canada.
Conflicts of interest. None declared.
Notes
The Confederation of Tomorrow surveys are administered by the Environics Institute for Survey Research in partnership with various public policy organizations across the country. Data are collected through online web panels to national samples of Canadians. The 2019 survey took place online between December 2018 and January 2019; the 2020 survey took place online in the provinces and, in the territories, over the phone between January 13 and February 20; the 2022 survey was conducted online in the provinces between January 18 and February 10 and, in the territories, over the phone between January 6 and 30; and, the 2023 survey was conducted online in the provinces between January 26 and February 9 and, in the territories, over the phone between January 24 and February 26. We pool the surveys from the years 2019, 2020, 2022, and 2023. While the survey ran in 2021, a different question format was used to measure territorial identification, forcing us to exclude data from this survey wave from the analysis.
We do not study the territories because of their distinct relationship with the federal government.
In the 2019 and 2021–2023 survey waves, our proxy for political ideology was a single question asking respondents whether they generally favor a smaller government with fewer services (scored 1), a larger government with more services (scored 0), or neither in particular (scored 0.5). In the 2020 survey wave, a different question formulation was used. In this survey wave, respondents were asked to indicate whether it was a “high,” “medium,” or “low” priority, or “not a priority at all” to minimize government interference in the free market. Responses were recoded to range from 1 to 0, with higher scores indicating it is a priority to minimize government interference.
As shown in the appendix (Supplementary Appendix Tables A7 and A8, see also Supplementary Appendix Figure 3A and Figure 4A), models interacting identity or grievances with a dummy for each province suggest that the effect of provincial identity and grievances on support for decentralization is consistently strong in Alberta, Saskatchewan, and Quebec.
It is worth noting that results presented in the appendix (Supplementary Appendix Table A9) show that the relationship between provincial identity or grievances and support for decentralization is weaker but remains positive and statistically significant in the seven other provinces when we exclude observations from Alberta, Saskatchewan, and Quebec.
Grievances are considered relatively low or high by computing scores one standard deviation below and above the mean, respectively.
The results are similar if we compare the interaction effects in two groups of provinces (Alberta, Quebec and Saskatchewan on one hand and the seven others on the other hand, see Supplementary Appendix Figures 6A–8A). Moreover, the substitution effect discussed earlier, in which grievances (provincial identity) have stronger effects when provincial identity (grievances) is low, exists only in the bloc of seven provinces.
References
Henderson, Ailsa, and Mike Medeiros.
