Which Policy Issues Matter in Canadian Municipalities? 2019 PDF

PUBLICATIONS SPP Research Paper Volume 12:8 March 2019 WHICH POLICY ISSUES MATTER IN CANADIAN MUNICIPALITIES? A SURVEY OF MUNICIPAL POLITICIANS Jack Lucas and Alison Smith SUMMARY Whether it’s a big city or a small town, all Canadian municipalities have core issues that the...

PUBLICATIONS SPP Research Paper Volume 12:8 March 2019 WHICH POLICY ISSUES MATTER IN CANADIAN MUNICIPALITIES? A SURVEY OF MUNICIPAL POLITICIANS Jack Lucas and Alison Smith SUMMARY Whether it’s a big city or a small town, all Canadian municipalities have core issues that their elected politicians are concerned about. Regardless of size, the daily business of a municipality must be managed and policies determined about such bread-and-butter issues as garbage collection, snow removal, wastewater and sewage, fire protection, economic development and fixing potholes. However, when size increases, so do the layers of issues that engage municipal politicians. This paper examines the results of a cross-Canada survey of more than 1,000 mayors and councillors from communities ranging in population size from 5,000 to more than two million. With an increase in population size, the numbers and complexity of issues creep up as well. Tiny municipalities typically aren’t concerned with issues such as immigrant settlement, homelessness and public transit. Those issues are much more pressing for larger municipalities. A focus on some types of issues, such as public transit, grows right alongside population growth. The physical size of large municipalities means they contain a population whose needs are naturally more diverse than they are in smaller cities, towns and villages, thus shifting politicians’ concerns to such things as homelessness and climate change. However, issues such as relations with Indigenous people and climate change also tend to hold regional, not just municipal, importance. They may be extremely important to a small municipality because of its geographic location and less important in a larger municipality located elsewhere. For example, municipal politicians in British Columbia reflect regional concerns with their emphasis in the survey on the importance of tackling homelessness, affordable housing, climate change and Indigenous relations. Yet, next door in Alberta, Indigenous relations and climate change ranked in the survey as being of low importance, along with climate change, despite the presence of two cities in the province with populations hovering around the million mark. http://dx.doi.org/10.11575/sppp.v12i0.56964 w w w. policyschool.ca  The number one issue for municipalities regardless of size is economic development, since job creation and attracting investment are key for a healthy municipality regardless of its location or size. And nearly every politician surveyed listed planning, water supply and transportation infrastructure (roads, highways and bridges) as being of deep importance to their communities. Of almost equal importance in the survey were a second slate of issues including emergency planning, parks and recreation, public health, solid waste removal and policing. The results of this survey are intended to lay the groundwork for future researchers who want to focus on specific problems in the area of urban policy-making. Those who want to study the bread-and-butter issues can do so among a wide range and size of municipalities, knowing that these issues are vital to all. Those with an interest in homelessness and immigrant populations can focus on the big cities while being assured they are not missing out on key points among smaller communities. This survey will be highly beneficial for researchers in urban policy issues as it will help them to decide where to look and exactly what to look for. 1  Which policy issues matter to local politicians? Canadian researchers increasingly argue that urban policy-making today is often as much about “wicked problems” of climate change and social policy as it is about snow clearing and potholes (Bradford, 2002; Bradford, 2005; Hughes et al., 2018). But do municipal politicians agree – do they think that traditional municipal issues have been overlaid with new policy concerns? And how do these opinions about municipal issues vary across Canadian cities? Earlier this year, in a survey of Canadian mayors and councillors from municipalities across the country, we sought to answer these questions. Across 18 areas of public policy, more than 1,000 mayors and councillors told us which issues they felt were important in their communities.1 Their answers provide us with a new and systematic look at the policy issues that Canadian municipal leaders feel are important across municipalities of widely varying sizes, regions and provinces. We hope you’ll read through to the end of this report, but here is the punch line: size matters. In Canadian cities, a long list of issues are important in every municipality – issues like policing, waste management and water supply – which we call the bread-and-butter issues of Canadian municipal policy-making. But as cities get bigger, an additional list of issues ranging from homelessness to immigrant settlement is layered atop the bread-and-butter ones. We call these the big city issues. These findings are important both for policy-makers and for researchers. For policy-makers, our findings lend weight to the argument that big cities really are different from other municipalities. Big city politicians must continue to deal with the bread-and-butter issues of municipal policy- making, but they must also address a suite of additional policy challenges that are distinctive to their big city context. The Federation of Canadian Municipalities acknowledges this through the Big City Mayors’ Caucus, which lobbies specifically for transit, housing, climate policy and partnerships. Our findings suggest that there are a few other big city issues, and some that are in fact only relevant to the biggest cities. This may or may not make the job of the big city politician harder than in smaller places – big city politicians also tend to have more staff support – but it certainly makes it more complex. Our findings strengthen the argument for distinctive statutory provisions for big cities – so-called city charters – in the Canadian context (see also Garcea, 2014; Sancton, 2016; Smith and Spicer, 2018a) As for researchers, our findings provide a useful recipe for refining research goals and case selection plans in future studies of urban policy-making in Canada. Researchers who are interested in bread-and-butter policy issues can study a wide range of Canadian municipalities, from small to large, knowing that such issues are important to politicians in nearly every municipality in the country (see also Goodman and Lucas, 2016). Those who are interested in the big city issues, like homelessness and immigrant settlement, can focus on big Canadian cities with less worry that they are overlooking these issues in much smaller municipalities (Good, 2009). And researchers who are interested in the relationship between population size and public policy challenges might choose to focus on a particular set of issues – most notably public transit – whose importance grows steadily in relation to a municipality’s population size. In other words, understanding how perceived issue importance varies across municipalities in Canada will help researchers understand where to look and what to look for in future studies of urban policy-making. 1 Please see Appendix 1 for technical details and question texts. 2  POLICY ISSUE IMPORTANCE: A PRELIMINARY OVERVIEW Later in this report, we will provide evidence to show that there are real differences in the perceived importance of policy issues across municipalities of different sizes. But before we get into the details, let’s begin with the big picture: What are the most important policy issues in Canadian municipalities? To answer this question, we drew up a list of 18 policy issues that we thought were probably at least somewhat important in some Canadian municipalities. We developed the list using a simple approach: we went to the library, checked out all the textbooks we could find on municipal politics and government, and read through them. If a textbook mentioned a policy issue, we added it to our list.2 Once we had compiled a long list of possible issues to include, we narrowed things down to cut down on overlap and to allow for broad coverage across issues without overwhelming our survey respondents. We ran this list past a long-time city councillor in a medium-sized Canadian city; he agreed that our list was thorough. While there were some hard choices to make in compiling the list – we wish we could have included fire protection, sewers and liquid waste, water treatment, flood mitigation, public libraries and many other issues – we felt that the 18 items we had selected would provide a broad first look at how Canadian municipal politicians think about a diverse array of local policy issues. Having developed our list, we sent a survey to every councillor and mayor in every Canadian municipality with a population above 5,000 – a total of 4,578 individuals. We received a total of 1,077 responses, for a very respectable response rate of 24 per cent – much higher than many surveys of political elites and the general public. While our respondents were not selected randomly and may not be fully representative of the entire population of municipal politicians, we have compared our respondents to observable characteristics of the full population (gender, region, population size), and our results give us confidence that our respondents are broadly reflective of elected municipal officials.3 Our question was very general: we simply asked our respondents how important each policy area was to life in their community. Interpreting importance was therefore up to our respondents: some may have emphasized importance for quality of life, others for economic growth, others for cultural vibrancy and so forth. Our aim was to capture very general sentiment, rather than constrain our respondents’ thinking by framing “importance” in one or another of these senses. So which issues do municipal politicians think are most important for their communities? The answer is in Figure 1, which summarizes our respondents’ average answer for each policy issue. The black circles in the figure represent mean estimates, and the black lines on either side of the circles are 95 per cent confidence intervals – meaning that we can be reasonably confident that if we gathered all of Canada’s municipal politicians in a room and asked them these questions, their answers would fall somewhere within the territory marked out by the confidence intervals on either side of the mean estimates. 2 This approach means that we did not include issues involving no real role for local governments, such as defence or pension policy. 3 See Appendix 1 for more detail. 3  FIGURE 1 ISSUE IMPORTANCE: OVERALL RESULTS Indig. Relations Immigrant Settlement Homelessness Climate Change Poverty Reduction Public Transit Culture / Arts Electricity and Gas Housing Emergency Planning Parks and Recreation Public Health Waste Police Planning Water Supply Roads Highways Bridges Economic Development Slightly Moderately Very Extremely The winning issue, as Figure 1 makes clear, is economic development, whose average importance score for municipal politicians is somewhere in the range between “very” and “extremely” important; in fact, more than 90 per cent of our respondents marked economic development as either “very” or “extremely” important. Much the same is true of the other policy issues at the bottom of the figure: planning, water supply and transportation infrastructure (roads, highways and bridges). Nearly every municipal politician considers these issues – which we call the “big four” – deeply important for their communities. Below the “big four” issues at the bottom of the figure, a second suite of issues – what we call the “next five” – falls close behind, all of which municipal politicians tend to consider very important: emergency planning, parks and recreation, public health, solid waste removal and policing. These five issues are essentially indistinguishable from one another in our data: they are all, on average, roughly equal in importance. After we move past the big four and next five issues, we start to see a distinct decline in perceived importance, beginning with issues like housing and electricity and gas, which fall somewhere between moderately and very important, all the way down to Indigenous relations and immigrant settlement, whose importance barely reaches the realm of moderate importance to municipal politicians. We will soon see that some of these issues enjoy a much higher prominence in big cities, but when we pool all of our responses into a single overall score, these issues are not considered important with anything like the same consistency as the big four. In the case of homelessness, for instance, nearly a third of our respondents marked the issue as either “not at all” or “slightly” important in their communities. 4  POPULATION AND POLICY ISSUE IMPORTANCE Everyone knows that a town of 5,000 in rural Nova Scotia faces very different policy challenges than a city in Alberta with more than a million residents. So how does a municipality’s population size affect the issues that politicians think are important? To provide a first look at this question, we tested the relationship between respondents’ assessments of the importance of each issue and a variable that captures each municipality’s population size.4 If the relationship is positive, respondents from big municipalities tend to think an issue is more important than do respondents from small municipalities. If the relationship is negative, the opposite is true. Figure 2 provides a plot of these relationships across each of the 18 policy issues. If the relationship is statistically significant and positive, the coefficient and error bars will be to the right of the dotted vertical line; if the relationship is statistically significant and negative, the coefficient and error bars will be to the left of the dotted line. Any coefficient whose error bars cross the dotted vertical line has no statistically significant relationship with population size. FIGURE 2 ISSUE IMPORTANCE AND POPULATION SIZE: BIVARIATE ASSOCIATIONS Elec / Gas Emergency Water Supply Econ. Dev. Public Health Policing Waste Roads Planning Indig. Parks and Rec. Arts and Cul. Housing Poverty Clim. Change Immigrant Homeless Transit -.2 0.2.4.6 Two patterns are important to notice in the figure. First, none of the coefficients is statistically significant and negative; not one of the 18 policy domains in our study is less important for municipal politicians in big cities when compared to small ones. Instead, error bars for a number of issues clearly cross the vertical line: electricity and gas, public health, economic development, emergency management, planning, policing, roads, solid waste, Indigenous relations and water supply. However important these issues are to municipal politicians – and 4 In keeping with the standard approach, we use logged population to adjust for the highly skewed distribution of population sizes in Canada. 5  as figure 1 suggests, some are very important and others less so – that importance tends to be consistent across small and large municipalities. The second important pattern to notice in Figure 2 is the cluster of issues whose coefficients are very distant from the dotted vertical line: homelessness, housing, immigrant settlement, public transit, climate change and poverty reduction. These issues have strong and statistically significant relationships with municipal population size, meaning that higher importance scores on these issues are very clearly associated with higher population cities. Canadian municipal politicians from big cities are much more likely to consider these issues important than politicians from small municipalities. Combined with the overall ranking in Figure 1, the relationships we have found between issue importance and municipal size suggest that issue importance in Canadian cities can be characterized by a set of bread-and-butter policy issues, whose importance is recognized everywhere, alongside a basket of big city issues whose perceived importance is likely to be much higher in big places than in small ones. Interestingly, however, we have found no issues with a negative relationship to municipal size, meaning that no issues recede in importance as municipalities get bigger. We will return to these findings in more detail below. REGION AND POLICY ISSUE IMPORTANCE Before we dive deeper into our findings on population size and issue importance, we need to consider another possible variable that could affect perceived importance: region. Given historical, economic and political variation across Canadian regions, along with variation in patterns of local policy autonomy across provinces, we might expect that municipal politicians in some regions rank certain issues more highly than municipal politicians in other regions (Sancton and Young, 2009; Smith and Spicer, 2018b). To explore this possibility, we have used an additional regression analysis to estimate the likely importance score for each issue in each region; since we know from the analysis above that population size is related to importance for some issues, we control for population size in this analysis. The results are displayed in Figure 3; these results can be interpreted as our best guess at average importance scores for each issue in each region while controlling for municipal size. Here, too, it is important to attend to both the grey bars and the black interval lines; if the bars are different heights but their interval lines overlap, the difference in issue importance from one region to another is not statistically significant. 6  FIGURE 3 ESTIMATED ISSUE IMPORTANCE BY REGION Electricity and Gas Homelessness Parks and Recreation Public Health Culture / Arts Economic Development 4 4 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic Emergency Planning Housing Immigrant Settlement Planning Police Public Transit 4 4 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic Roads Highways Bridges Waste Indig. Relations Water Supply Climate Change Poverty Reduction 4 4 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic West Ontario QuebecAtlantic The central lesson that we learn from Figure 3 is that only a select few policy issues appear to vary in perceived importance by region. In some cases, issue importance is distinctly lower in certain regions, such as homelessness and Indigenous affairs/relations in Quebec – a finding that may reflect a sense that these issues are the responsibility of the provincial and federal governments respectively, rather than municipal governments in Quebec. Similarly, climate change and poverty reduction appear to be considered less important in Western Canada than in other parts of Canada. In the case of climate change policy, this regional effect hides substantial variation within Western Canada; in British Columbia, climate change generally receives a higher importance score from municipal politicians than in Alberta, Saskatchewan and Manitoba. We will discuss these differences in more detail below. In other cases, one region gives distinctly high importance scores to particular issues, such as immigrant settlement, poverty reduction and Indigenous relations in Atlantic Canada. These patterns reflect a general tendency among our respondents in Atlantic Canada to rank all policy issues slightly higher in importance than their counterparts in other regions; for reasons that are not entirely clear, about 35 per cent of Atlantic Canadian respondents rank issues as “extremely important” compared to about 26 per cent of respondents in other parts of Canada. Do these regional patterns hold at the provincial level? We received sufficient responses in Canada’s four most populous provinces – British Columbia, Alberta, Ontario and Quebec – to enable us to at least partially answer this question. The results are displayed in Figure 4, with the same basic setup as Figure 3 above. Several patterns are immediately noticeable in the figure, such as distinctly high scores for homelessness, housing, Indigenous relations and climate change in British Columbia – a clear indication that regional analyses which combine B.C. with Canada’s Prairie provinces probably obscure as much as they reveal about 7  issue importance across Canadian municipalities. Similarly, the figure reveals a number of distinctly low-importance issues in Alberta, such as public transit, Indigenous relations and climate change. Overall, however, the message from Figure 4 is similar to that of Figure 2: on the bread-and-butter issues, we see very little variation in issue importance across the provinces. Issues like economic development, planning and water supply are uniformly high in municipalities across Canada. FIGURE 4 ESTIMATED ISSUE IMPORTANCE BY PROVINCE Electricity and Gas Homelessness Parks and Recreation Public Health Culture / Arts Economic Development 4 4 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC Emergency Planning Housing Immigrant Settlement Planning Police Public Transit 4 4 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC Roads Highways Bridges Waste Indig. Relations Water Supply Climate Change Poverty Reduction 4 4 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 1 1 1 1 1 1 AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC AB BC ON QC In both of these instances – issues like climate change that are less important in some regions and provinces than others, and issues like Indigenous affairs that are more important in some regions and provinces than others – additional research would be valuable. Our data do not provide us with a sufficiently detailed look at policy-making processes across Canada to identify the factors that are driving these regional differences. In some cases, these differences probably reflect genuine regional variation in policy priorities across Canada. In other cases, they may reflect perceptions about the relative importance of municipal governments in complex, multilevel policy files (Horak, 2012; Lucas and Smith, forthcoming). More focused case study research, guided by the general patterns we have identified here, will help us understand how and why municipal policy issue importance varies across Canada’s regions. MORE DETAIL ON POPULATION AND ISSUE IMPORTANCE Thus far, we have identified a clear relationship between issue importance and population size for a handful of big city issues in Canada. We have also noted that some patterns of regional variation, while more subtle and less substantively important than variation by population size, are also present in our data. What remains is to investigate the relationship between population 8  size and issue importance in a bit more detail in order to more precisely understand how the two are related. While we have tentatively referred to immigrant settlement and public transit as big city issues in our discussion above, it is possible that this description is misleading. Issue importance might be related to population size in a non-linear fashion, such that an issue is not considered important until a city is very large. But it might also be related to population size in a more straightforward linear relationship, such that the importance of an issue gradually increases as we move from smaller to larger municipalities. To understand the nature of the big city issues we have identified, we need to better understand if the relationships we have uncovered are linear or nonlinear. Figure 5 allows us to probe this question of linearity across each of the policy issues in our survey. To construct the analysis, we first divided our responses into population quartiles: the first consisting of very small municipalities (5,000 to 8,999 population), the second consisting of small to medium-sized municipalities (9,000 to 17,999 population), the third consisting of medium-sized municipalities (18,000 to 59,999), and the fourth consisting of medium-sized to very large municipalities (60,000 to more than two million). These quartiles are not intended to reflect any claims about how to divide Canadian municipalities by population; they simply reflect the division of our responses in the dataset into four equally sized categories to enable comparison across population sizes (we provide an additional analysis using more fine-grained population categories in Appendix 2). For each issue, the figure reports an estimated issue importance score for each of these population quartiles. If the relationship between population size and issue importance is linear, we should see a gradual increase in importance as we move from the first to the fourth quartile. If the relationship is one in which the issue surges to importance only in very big cities, we will see issue importance leap upward only in the fourth quartile. To aid in interpreting the figure, we have included only the issues that are positively related to population size in our regression analysis above.5 5 The seven issues in the figure are consistently related to issue importance across a range of specifications and models. See Appendix 2 for more detail. 9  FIGURE 5 ESTIMATED ISSUE IMPORTANCE BY POPULATION QUARTILE Homelessness Culture / Arts Housing 4 4 4 3 3 3 2 2 2 1 1 1 1 2 3 4 1 2 3 4 1 2 3 4 Immigrant Settlement Public Transit Climate Change 4 4 4 3 3 3 2 2 2 1 1 1 1 2 3 4 1 2 3 4 1 2 3 4 Poverty Reduction 4 3 2 1 1 2 3 4 Figure 5 suggests that the nature of the relationship between population size and issue importance does indeed vary by policy issue. Some issues, such as homelessness, public transit and climate change appear to be related to population size in a fairly straightforward linear relationship. Immigrant settlement and poverty reduction are more clearly big city issues, with flat importance scores until the top population quartile, when those scores leap upward. Culture and arts also have something of the same character, though the upward surge appears further 10  down the population ladder in mid-sized municipalities. The upward surge is particularly noticeable in the case of immigrant settlement, whose average importance is below “slightly important” in the three lowest population quartiles and then surges to “very important” in large and very large cities. This relationship holds even within the fourth population quartile: among our respondents from Canada’s biggest cities (those above the 90th percentile, larger than 240,000 in population), the importance of immigration settlement is fully 0.67 points higher than among respondents in the top population quartile who are not in very big cities (that is, those between the 75th-90th percentile). On a five-point scale, this is a substantial difference, and it reflects the distinctive character of immigrant settlement as an issue of substantial importance in Canada’s big cities. Perhaps the most remarkable sub-figure within Figure 5, however, is the summary of issue importance scores across population quartiles for public transit. As we have noted above, this relationship is linear, growing steadily through all of the quartiles. But the slope of that increase is remarkable: it grows from an average score below “slightly” important in small municipalities to a score well above “very” important in Canada’s biggest cities. While we see clear relationships between population quartile and perceived importance in the other issue areas, only in public transit does an issue rise so sharply through the quartiles. In Canada’s big cities, public transit ranks among the pre-eminent policy issues, comfortably within the same range as the big four issues we described in section 1 above. CONCLUSION Which local policy issues matter according to Canadian municipal politicians? In this report, based on a pan-Canadian survey of mayors and councillors, we have identified a few clear patterns, which we summarize in Table 1 below. First, a number of policy issues are considered important in all Canadian municipalities regardless of population size. We call these the bread- and-butter issues of local policy, and we have divided them into the most important issues (the “big four”) and a cluster of issues that are somewhat less important than the big four but which municipal politicians still consider very important (the “next five”). TABLE 1 ISSUE TYPES IN CANADIAN MUNICIPALITIES "Bread and Butter Issues" "Big City Issues" Big Four Growers Planning and Land Use Homelessness Water Supply Culture and the Arts Roads, Highways, and Bridges Housing Economic Development Public Transit Next Five Climate Change Emergency Planning Parks and Receation Surgers Public Health Immigrant Settlement Solid Waste Poverty Reduction Policing 11  We have also found a number of issues whose importance is positively related to population size. Some of these issues are “growers”, related to population size in a straightforward linear fashion: as population goes up, perceived importance also goes up. Among these, public transit is distinctive; in small municipalities, transit is generally unimportant, but in Canada’s largest municipalities, its importance rivals the big four bread-and-butter issues. A final set of issues are the “surgers” – poverty reduction and immigrant settlement – whose importance is low in all municipalities except the very largest. These issues are related to population size in a less linear fashion, remaining low in perceived importance across the three lowest population quartiles and then surging higher in importance only in Canada’s largest cities. To our knowledge, these findings represent the first systematic pan-Canadian analysis of perceived issue importance in Canadian municipalities. Of course, the analysis is not without limitations – the most important of which is that our respondents were elected politicians, whose views may not always align with perceived issue importance among residents, local businesses or other local organizations. They are, however, on the front lines of local politics and policy-making, and thus represent an important perspective in the study of policy-making in Canada. We hope that future surveys will fill in our analysis in this report with additional survey and case study data from these and other local actors. Despite the inevitable limitations, however, we believe that both practitioners and researchers have much to learn from our findings. We have found that Canadian municipalities share many important issues in common: the bread-and-butter issues listed in Table 1 are consistently important across Canadian municipalities. In these areas of public policy, pan-Canadian research and discussion across municipalities of all sizes are likely to be very informative and valuable, and standardized statutes which address these issues uniformly across all of a province’s municipalities may also be entirely appropriate. We have also found, however, that Canada’s big cities are different, not because some issues are less important in big cities, but rather because a cluster of important big city issues are layered atop the bread-and-butter issues. These issues include climate change, homelessness, housing, immigrant settlement, poverty reduction and – above all – public transit. This longer list of important issues in Canadian big cities suggests that recent calls for distinct big-city legislation – often called city charters – reflect real differences in the array of issues that policy-makers in Canada’s large municipalities feel they must address. Understanding the specific funding instruments, policy capacities and intergovernmental relationships that emerge as a result of this additional layer of big city issues must continue to be an important research focus for Canadian urban policy researchers. At the same time, scholars of Canadian municipal policy should not neglect the bread-and-butter issues of pre-eminent importance across municipalities of all sizes. 12  REFERENCES Bradford, Neil. 2002. Why Cities Matter: Policy Research Perspectives for Canada. Ottawa: Canadian Policy Research Networks. Bradford, Neil. 2005. Place-Based Public Policy: Towards a New Urban and Community Agenda for Canada. Ottawa: Canadian Policy Research Networks. Garcea, Joseph. 2014. “The Empowerment of Canadian Cities: Classic Canadian Compromise,” International Journal of Canadian Studies 49 (19): 81-104. Good, Kristin. 2009. Municipalities and Multiculturalism: The Politics of Immigration in Toronto and Vancouver. Toronto: University of Toronto Press. Goodman, Nicole, and Jack Lucas. 2016. “Policy Priorities of Municipal Candidates in the 2014 Local Ontario Elections,” Canadian Journal of Urban Research 25 (2): 35-47. Horak, Martin. 2012. “Multilevel Governance in Toronto: Success and Failure in Canada’s Largest City.” In Sites of Governance: Multilevel Governance and Policy Making in Canada’s Big Cities, 228-62. Montreal and Kingston: McGill-Queen’s University Press. Hughes, Sara, Eric K. Chu, and Susan G. Mason, eds. 2018. Climate Change in Cities. New York: Springer. Lucas, Jack, and Alison Smith. Forthcoming. “Multilevel Policy from the Municipal Perspective: A Pan-Canadian Survey,” Canadian Public Administration. Sancton, Andrew. 2016. “The False Panacea of City Charters? A Political Perspective on the Case of Toronto.” The School of Public Policy SPP Research Papers 9 (3). Sancton, Andrew, and Robert Young, eds. 2009. Foundations of Governance: Municipal Government in Canada’s Provinces. Toronto: University of Toronto Press. Smith, Alison, and Zachary Spicer. 2018. “The Local Autonomy of Canada’s Largest Cities,” Urban Affairs Review 54 (5): 931-61. 13  APPENDIX 1: MUNICIPAL SURVEY: ADDITIONAL DETAIL Please note: Some of the text below is drawn from the appendices in Lucas and Smith (forthcoming), which uses the same survey in an analysis of multilevel policy involvement in Canadian cities. The data discussed above are drawn from a survey of 4,578 municipal mayors and councillors from municipalities with populations over 5,000 in Canada.6 When we were unable to locate email addresses online, we requested them directly from city clerks. In the end, we estimate that our invitation list included more than 95 per cent of the full population of municipal mayors and councillors across Canada. The survey was in the field from Jan. 16, 2018 until April 10, 2018, with a total of 1,084 complete responses. Our overall response rate among elected officials was 24 per cent, with a 30 per cent response rate among mayors and reeves and a 23 per cent response rate among councillors. Mayors and reeves are thus slightly over-represented in our sample relative to the population; however, analyses of our findings using only the councillors in our dataset generate the same substantive results. To ensure a satisfactory response rate, we asked our respondents about a random draw of half of the 18 policy issues on our list. This reduced our total response numbers for each issue but allowed us to gather data about a much larger list of policy issues while maintaining a high response rate. Our total number of respondents for each policy issue is summarized in Table 2 below. TABLE 2 RESPONDENTS BY POLICY ISSUE Policy Area Respondents Electricity 484 Homelessness 474 Parks and Recreation 477 Public Health 490 Arts and Culture 479 Economic Development 448 Emergency Planning 451 Housing 485 Immigrant Settlement 458 Planning and Land Use 456 Policing 464 Public Transit 473 Roads, Highways, Bridges 491 Waste Disposal 471 Indigenous Affairs 453 Water Supply 478 Climate Change 488 Poverty Reduction 476 6 Some portions of our survey description in this appendix are based on Lucas and Smith (2018). 14  The survey’s primary focus concerned multilevel governance and the role of different levels of government and public/private actors in the urban policy process (Lucas and Smith, forthcoming). However, the survey began with a simple question: How important are each of these policy areas to life in your community? ((1) Not at all important; (2) slightly important; (3) moderately important; (4) very important; (5) extremely important). Because this question was the first in the survey, we have little reason to be concerned about the influence of the multilevel governance questions which followed on responses to this question. To develop our list of eligible municipalities for the survey, we began by downloading a full list of Canadian census subdivisions by 2016 population from Statistics Canada (Table T301EN).7 Seven hundred and thirteen census subdivisions in this table are listed as having a population above 5,000. We initially excluded three census subdivision types from our email collection process due to the absence of elected municipal officials in those types: First Nations communities (one in dataset), New Brunswick parishes (five in dataset) and unorganized areas (three in dataset). Unfortunately, we discovered after the data collection process was complete that we had also excluded electoral areas in British Columbia; these emails were not available in the B.C. municipal directory and we only later discovered that they were available on the websites of the relevant regional districts. Statistics Canada’s Census Subdivisions table contains more than 5,000 distinct CSDs; after eliminating the three CSD types listed above from the table, nearly 4,000 CSDs remain. Thus our sampling frame captures only a small proportion (about 16 per cent) of all census subdivisions in Canada. From a population perspective, however, our sampling frame includes 31,059,089 individuals, more than 88 per cent of Canada’s total 2016 population (35,151,728). While we acknowledge that the 5,000 population threshold is arbitrary, we felt that it struck a good balance between capturing a wide range of urban, suburban and rural municipalities while also avoiding the immense practical challenges involved in collecting email addresses for very small municipalities. Table 3 provides an overview of the total number of census subdivision types in our database, and the number of municipalities in each type for which we were able to locate email contact information, either from online sources or from city clerk requests. Given that the missing municipalities are overwhelmingly very small rural councils, we estimate that we collected email addresses for more than 95 per cent of the local elected councillors and mayors in municipalities above 5,000 population. 7 This table is available here: https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/hlt-fst/pd-pl/comprehensive.cfm. 15  TABLE 3 CENSUS SUBDIVISION TYPES AND EMAILS COLLECTED Total Collected Percentage Canton / Canton Unis 2 2 100.0% City / Ville 297 257 86.5% District Municipality 24 23 95.8% Indian Reserve N/A Municipal District 43 43 100.0% Municipality 76 56 73.7% Parish N/A Regional District Electoral Area 20 0 0.0% Regional Municipality 4 3 75.0% Rural Community 1 1 100.0% Specialized Municipality 4 4 100.0% Subdivision of County Municipality 4 4 100.0% Town 137 117 85.4% Township 69 44 63.8% Unorganized N/A Total Overall 681 554 81.4% Total Excluding Electoral Areas 661 554 83.8% Table 4 outlines our respondents’ demographics relative to the population from which they are drawn. Aside from regional distributions – Quebec politicians are somewhat overrepresented and Ontario politicians are somewhat underrepresented – we see little reason for concern in these distributions. While caution is always in order in non-random elite survey research, these distributions suggest that our findings are broadly representative of the population. Please see the next appendix for additional robustness tests. 16  TABLE 4 COMPARISON OF SAMPLING FRAME AND RESPONDENTS Sampling Frame Respondents Gender Men 67% 64% Women 33% 36% Province Alberta 14% 16% British Columbia 11% 8% Manitoba 4% 3% New Brunswick 3% 2% Newfoundland 2% 3% Nova Scotia 5% 5% Nunavut 0% 0% Northwest Territories 0% 0% Ontario 30% 26% Prince Edward Island 1% 0% Quebec 28% 35% Saskatchewan 2% 2% Yukon 0% 0% Population 1st Quartile 24% 26% 2nd Quartile 25% 25% 3rd Quartile 25% 25% 4th Quartile 26% 24% For the complete survey, along with the dataset and replication files used to produce the analysis in this paper, please visit Jack Lucas’s dataverse site at the following link: https://dataverse.scholarsportal.info/dataverse/jacklucas. 17  APPENDIX 2. REGRESSION ANALYSIS: ADDITIONAL DETAIL In our analysis in the main text, we visually reported a number of regression coefficients to describe the associations between population size, region and issue importance. In this appendix we discuss those coefficients in more detail and provide a number of alternative analyses to demonstrate the robustness of the findings in the main text. Table 5 provides full results for a series of regression analyses using two distinct models and several specifications. What is crucial to note about the table is that each coefficient represents a distinct regression analysis. For instance, in the top left corner of the table, the first cell reports the coefficient for a regression of electricity and gas importance scores on logged population size; the next cell to the right reports the coefficient for a regression of electricity and gas importance scores on logged population size, with region-fixed effects; and so on. Each issue importance variable is a distinct dependent variable in a distinct regression analysis. To aid in interpretation, we have highlighted in orange all coefficients that are both statistically significant and positive (there are no statistically significant and negative coefficients). TABLE 5 – POPULATION AND ISSUE IMPORTANCE: DETAILED RESULTS Bivariate OLS OLS w/ Region FEs OLS w/ Province FEs Bivariate Oprobit Oprobit w/ Region FEs Electricity -0.0600 (0.086) -0.0408 (0.254) -0.0438 (0.224) -0.0538 (0.095) -0.0338 (0.314) Homelessness 0.216*** (0.000) 0.237*** (0.000) 0.248*** (0.000) 0.209*** (0.000) 0.236*** (0.000) Parks and Recreation 0.0474* (0.025) 0.0384 (0.076) 0.0393 (0.070) 0.0852* (0.017) 0.0710 (0.056) Public Health -0.0217 (0.410) -0.0165 (0.541) -0.0167 (0.537) -0.0334 (0.316) -0.0266 (0.436) Arts and Culture 0.0875** (0.001) 0.0783** (0.004) 0.0753** (0.005) 0.103** (0.001) 0.0948** (0.005) Economic Development -0.0292 (0.202) -0.0337 (0.154) -0.0308 (0.196) -0.0473 (0.202) -0.0554 (0.148) Emergency Planning -0.0329 (0.286) -0.0289 (0.363) -0.0298 (0.349) -0.0384 (0.299) -0.0321 (0.400) Housing 0.124*** (0.000) 0.128*** (0.000) 0.127*** (0.000) 0.146*** (0.000) 0.152*** (0.000) Immigrant Settlement 0.213*** (0.000) 0.239*** (0.000) 0.242*** (0.000) 0.218*** (0.000) 0.258*** (0.000) Planning and Land Use 0.0456 (0.056) 0.0512* (0.036) 0.0487* (0.046) 0.0683 (0.060) 0.0766* (0.039) Policing -0.00200 (0.944) 0.0340 (0.237) 0.0332 (0.249) 0.000721 (0.984) 0.0515 (0.180) Public Transit 0.390*** (0.000) 0.357*** (0.000) 0.342*** (0.000) 0.439*** (0.000) 0.416*** (0.000) Roads, Highways, Bridges 0.0344 (0.140) 0.0364 (0.130) 0.0388 (0.108) 0.0600 (0.089) 0.0614 (0.092) Waste Disposal 0.00443 (0.873) 0.0204 (0.472) 0.0208 (0.463) 0.00788 (0.834) 0.0293 (0.449) Indigenous Affairs 0.0565 (0.135) 0.0957* (0.012) 0.0806* (0.034) 0.0481 (0.154) 0.0863* (0.014) Water Supply -0.0293 (0.287) -0.0196 (0.495) -0.0210 (0.467) -0.0326 (0.346) -0.0208 (0.563) Climate Change 0.175*** (0.000) 0.154*** (0.000) 0.150*** (0.000) 0.179*** (0.000) 0.162*** (0.000) Poverty Reduction 0.153*** (0.000) 0.145*** (0.000) 0.146*** (0.000) 0.162*** (0.000) 0.156*** (0.000) The central question to answer using Table 5 is whether the findings we have reported in the main text are robust to alternative models and specifications. More specifically, we might worry that the relationship between population size and issue importance is (a) actually driven by regional or provincial variation, given differences in the number of large and small municipalities in different parts of the country, or (b) poorly captured in a linear regression model, given the ordinal character of the dependent variable (“not at all important”, “slightly important”, and so on). To test for the first possibility, the first three columns in the table compare the simple bivariate OLS coefficients to two alternative specifications: the first with dummy variables for each region (West, Ontario, Quebec, Atlantic), and the second with dummy variables for each province. These coefficients with regional and provincial fixed effects focus the analysis on within-region and within-province variation, and as is clear from the patterns of highlighted cells, most coefficients are stable across all three specifications. In a few cases, restricting the regression to within-region or within-province variation tips a coefficient out of statistical significance (parks and recreation) or into statistical significance (Indigenous affairs, planning and land use), but these changes are minor and are not a point of emphasis in the main text. 18  The second potential challenge is that our dependent variables – which are responses to a series of questions about the importance of particular issues for municipal politicians using a five- point scale – may not be appropriately treated as interval. To account for this possibility, we repeat the analysis using ordered probit, a model designed for ordinal-level dependent variables. The patterns of statistical significance in these models (the two rightmost columns in Table 5) are identical to the OLS models. We find seven policy issues that are consistently statistically significant and positive across all models and specifications: homelessness, arts and culture, housing, immigrant settlement, public transit, climate change and poverty reduction. In these issue areas, we are confident that there is a real relationship in our data between population size and perceived issue importance. This is the reason for our focus on these six as the big city issues in the main text above. ADDITIONAL POPULATION ANALYSIS In our analysis in the main text, we also divided the population in our sample into four quartiles to understand the linearity of population relationships to issue importance. Because of the uneven distribution of population sizes among Canadian municipalities, this had the unfortunate consequence of lumping an enormous range of population sizes (from 60,000 upward) into the final quartile. In Figure 6 below, we provide a somewhat more fine-grained analysis, in which population is divided into those in the bottom 10 percentiles (below 6,291), 10th-25th percentile (6,291 to 9,041), 25th-50th percentile (9,041 to 17,396), 50th-75th percentile (17,396 to 55,648), 75th-90th percentile (55,648 to 239,700), and 90th-100th percentile (above 239,700). This comes at the cost of a less even distribution among the categories, but allows us to confirm that our findings are robust to a division of population sizes that has more face validity. As the figure suggests, the basic patterns identified in the main text are replicated in this more fine-grained analysis. 19  FIGURE 6 POPULATION AND ISSUE IMPORTANCE: ADDITIONAL RESULTS Homelessness Culture / Arts Housing 4 4 4 3 3 3 2 2 2 1 1 1 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 Immigrant Settlement Public Transit Climate Change 4 4 4 3 3 3 2 2 2 1 1 1 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 Poverty Reduction 4 3 2 1 1 2 3 4 5 6 20  SENSITIVITY ANALYSIS We conclude with a test of the sensitivity of our findings to the inclusion of particular observations in the dataset. In this analysis, we drew 50 subsamples from the dataset, each of which consisted of a random sample of half of the responses, and then estimated the relationship between issue importance and (logged) population for each subsample. Figure 7 summarizes the coefficients for each of the subsamples, capturing the general direction of coefficients and the consistency of coefficients. Table 6 then summarizes the results, recording the most common direction of the coefficient, the proportion of subsamples in which coefficients were that direction (“Consistency”) and the proportion of subsamples in which coefficients were that direction and were also statistically significant at p