Abstract
In the past 20 years, Political Science research has increasingly focused on urban and local politics. We systematically review this literature and find that smaller cities are disproportionally underrepresented, particularly outside the United States. Smaller cities exhibit economic, social, and political patterns that differ from those in large metropolitan areas. Using administrative data and existing research, we show how cities of different sizes vary in their demographic characteristics; citizens’ preferences; resources and capacity; intergovernmental relationships; and electoral politics. These patterns indicate the potential to update existing theories in Political Science, including those about gender and political participation, second-order elections, and intergovernmental relationships. We suggest that scholars consider how smaller cities might differ from larger cities and include smaller municipalities in their case study research. We also highlight political issues unique to small cities as new areas of inquiry.
Introduction
Political scientists are increasingly studying urban and local politics. Researchers demonstrate that a significant amount of policymaking and implementation occurs at the local tier of government (Hoffmann-Martinot and Sellers 2005; Trounstine 2009; Post 2018). Additionally, due to the relative ease of study implementation in urban relative to rural areas, cities are also frequently the sites of in-depth data collection supporting more general theories, particularly in low- and middle-income countries. Figure 1 shows a steady increase in the fraction of articles either about urban politics or situated in cities in major general interest and comparative politics journals over a 15-year period. This growth reflects the importance of studying urban areas, where the United Nations estimates more than half of the global population currently lives. Understanding the political dynamics of urban areas will be even more critical in the future, as 68% of the world's population is projected to live in urban areas by 2050 (United Nations 2018).
Fraction of articles about urban politics in major general interest (GI) and comparative politics (CP) journals, 2005–2019.
A large portion of this growing urban population lives outside of major metropolitan areas. Table 1, for example, examines the populations of the ten countries with the largest urban populations in the world. It shows that aside from two outliers (Japan and Pakistan), well over half of the urban population lives in cities outside of each of the countries’ ten largest urban agglomerations. We see similar trends when examining cities with fewer than 1 million residents. These data are complementary to work by Clancey (2004) who notes that – at the time – half of urban residents worldwide lived in cities of less than 500,000 people, and two-thirds of urban residents in developing countries lived in cities of less than 1 million.
Percent of Urban Population Outside of 10 Largest Urban Agglomerations in Countries with the Largest Urban Populations, 2010.
Source: The World Bank (https://data.worldbank.org/indicator/SP.URB.TOTL).
Source: The United Nations World Urbanizations Prospects, Population Division (https://population.un.org/wup/Download/). Full data on population totals for these countries is available in Appendix Table A1.
In neighboring disciplines, scholars have increasingly recognized the importance of these smaller cities where so many live. In Urban Studies and Geography, Bell and Jayne (2009), Ofori-Amoah (2007), and Véron (2010) outlined a research agenda for smaller cities about one decade ago, and these disciplines have moved forward with both case studies of small cities and explicit analyses of how their outcomes of interest vary with city size. This literature has continued to develop, with recent edited volumes (Ruszczyk, Nugraha and de Villiers 2020), interdisciplinary digital magazines (Neves and Ruszczyk 2021), and special issues (Grossmann and Mallach 2021; Mukhopadhyay, Zérah and Denis 2020; Ocejo, Kosta and Mann 2020) focusing on smaller, overlooked cities. The studies coming out of this agenda have improved both empirical understanding and theoretical conceptualizations of cities in these disciplines.
In Political Science, however, much of our knowledge about urban politics comes from the largest cities. We systematically review all empirical studies of urban politics or studies taking place in cities in seven Political Science journals for 20 years, from 2000 to 2019. We consider two ways in which small cities may appear in an empirical analysis, namely through large N studies of intermunicipal variation and case studies of intramunicipal variation. About 36% of the studies in our dataset are large N studies. Smaller cities are less well represented in in-depth case studies of one or a few cities. Overall, we see potential to include more small cities in the empirical analysis of urban politics, particularly in case studies focusing on intramunicipal variation. We see similar trends in five interdisciplinary journals focusing on Urban Studies, yet the volume of scholarship on cities in these journals suggests that there is at least some theory-building and empirical research on small cities in disciplines outside of Political Science.
Beyond the normative importance of better investigating the needs and behavior of hundreds of millions of underexamined citizens, we argue that it is important for political scientists to include smaller cities in their empirical analyses because large cities and urban agglomerations are not representative of patterns of politics in their smaller counterparts. We draw upon administrative data and studies from Urban Studies, Geography, and Economics, which suggest that, along with other empirical trends, smaller cities vary in at least four fundamental ways that affect politics. First, smaller cities and large metropolitan areas have different populations with respect to variables such as race, gender, and age. These differences can affect citizens’ political preferences and behavior. Second, smaller cities vary systematically in their access to resources, which shape political outcomes. Third, smaller cities have fundamentally different vertical and horizontal intergovernmental relationships. Finally, smaller cities exhibit different rules and patterns of electoral politics than large cities. City size does not necessarily cause these differences in trends; our point is rather that smaller cities may exhibit distinct patterns of politics that are important to understand.
Throughout these illustrations, we highlight how studying smaller cities may change existing theories in political science. For example, we show that unlike large metropolises, small cities in South Asia sometimes have more women than men, suggesting that households here have different patterns of bargaining, which has been found to affect women's political participation (Cheema et al. 2022; Khan 2021; Prillaman 2021). In another example, we examine the applicability of theories of participation in national vs. local elections (i.e., “second order elections” (Reif and Schmitt 1980)) to smaller cities, where there is greater participation in local politics (Dahl and Tufte 1973) and are fewer candidates affiliated with national parties (Kjær and Elklit 2010b). More importantly, we highlight how little is actually understood about the politics of smaller cities, which indicates that the largest theoretical contributions from studying them remain a priori unknown.
Next, we discuss political phenomena and policy problems that may be unique to smaller cities. We include examples, including strategies to navigate vertical relationships in a country with high levels of urban primacy, the symbiotic nature of urban–rural and core–satellite relationships typical in regions with central towns or cities, and the politics of towns with a single major industry or employer. These are new and promising directions of research on smaller cities, but many more themes are likely to emerge as smaller cities become better understood.
We conclude with low-cost suggestions for how to study smaller cities, particularly in developing contexts where access to administrative data can be scarce. Potential strategies include digital surveys and interviews, crowdsourced data, and remote-sensing applications. Concerns about convincing editors and reviewers about the representativeness and importance of these smaller population cases may be allayed by multiple-case studies and/or randomly sampling cases from a set of cities deemed to be similar on a set of relevant parameters.
Overall, we believe that expanding Political Science's research focus on smaller cities will improve our understanding of politics by better capturing the full diversity of political and policy outcomes in the cities in which a large portion of the world's population lives. Moreover, we also believe that policymakers will benefit from our improved understanding of problems or solutions that might be unique to smaller cities and thereby directly impact the lives of this considerable, yet underexamined, group of citizens.
Defining Small Cities
Our analysis considers two measures of city size. First, we employ a relative measure of city size, namely a city's population rank within a country. For example, any city outside a country's top ten most populous cities might be considered relatively small. This measure provides information about a city's primacy or relative importance within a country and can shed light on intermunicipal and intergovernmental relationships. Second, we also measure the total number of residents in a metropolitan area. Cities are classified as small if their population is less than 1 million residents, regardless of where they may fall on their national ranking. We use 1 million as our threshold because major cross-national data sources such as the OECD classify metro areas with a population greater than 1 million as “large” (OECD and European Commission 2020). As we illustrate, the measure of total residents is correlated with other characteristics of the city itself, such as demographics, local tax revenue per capita, or municipal-level electoral politics. We argue throughout the paper that these characteristics vary with the absolute number of individuals or citizens living in an agglomeration, regardless of its relative position in a particular country.
These two different measures often classify different cities as “small,” depending on the distribution of city populations within a country. In high population, high urbanization countries such as the U.S., there are many major metropolitan areas with populations easily exceeding 1 million residents that nevertheless rank outside the top 10 largest in the country. However, many smaller countries may only have zero or one city meeting a 1 million population threshold. As such, the second largest city would be classified as large by a relative measure but small by an absolute measure. Using Sweden as an illustration, UN statistics show Stockholm's metropolitan area population to be 1,360,000 in 2010. Götenborg, the second largest metro area, had a population of only 546,000. As such, Sweden's second-largest city is counted as a large city in our relative measure but not in our absolute one (for related discussion in postcolonial urban theory, see Fonseca Alfaro 2021). The same is true for many smaller countries following Zipf's Law (Arshad, Hu and Ashraf 2018), which predicts that a country's second-largest city has roughly half the population of its largest city.
Studies of Cities in Major Political Science Journals
To more systematically understand how often Political Science studies smaller cities, we conduct a thorough literature review of several major and specialized Political Science journals following the model of Wilson and Knutsen’s (2022) analysis of geographic bias in Political Science publishing. We look at seven major journals for the 20-year period between 2000 and 2019: three top general interest journals (American Journal of Political Science, American Political Science Review, and Journal of Politics); three top subfield journals in Comparative Politics (Comparative Politics, Comparative Political Studies, and World Politics); and Urban Affairs Review, the discipline's leading urban politics subfield journal. 1
Across these journals, we first identify all articles focusing on cities and local politics. We include studies that focus on the lowest relevant tier of government. 2 We also include studies that do not specifically focus on local government but feature data collection (such as a survey) or experiments that occurred exclusively in urban areas.
We do not include articles that the authors self-identify as focusing on exclusively “rural” jurisdictions or “villages,” as these studies do not focus on communities that could credibly be called “cities.” Any study that features both urban and rural local government is included. Additionally, articles that focus on task-specific jurisdictions that cannot be clearly associated with one city (see Hooghe and Marks 2003), such as school districts or fire protection districts, are excluded from our analysis. Finally, we exclude purely theoretical articles without any empirical cases.
We next code articles as focusing on large versus small and medium-sized cities (defined using both the relative and absolute measures). Studies focusing on metropolitan areas rather than individual municipalities are associated with the characteristics of their core cities for ease of comparison.
The results of this data collection effort are presented in Table 2. We include 658 studies in our dataset. 3 An immediate observation is that the vast majority of the articles are about the U.S. There are almost twice as many studies on the U.S. as there are on all other countries combined, suggesting that the urban politics of American cities are a distinct area of study. 4 For this reason, we examine the study of small cities in the U.S. and non-U.S. contexts as separate phenomena.
Summary of Studies of Urban Politics in Seven Political Science Journals, 2000–2019.
Studies using statistical analysis on intermunicipal data and/or more than 10 cases.
Percentage of total urban politics studies.
Number of articles including at least one small city. Small cities are defined as those outside the top 10 most populous metropolitan areas based on data from The UN World Urbanizations Prospects, 2010. Hong Kong is considered administratively separate from China in UN population rankings.
Articles including small cities as a fraction of articles classified as small-n case studies.
Small cities defined as those with a population of less than 1 million based on data from The UN World Urbanizations Prospects, Population Division, 2010.
Studies in which all cases are from the United States.
Studies including at least one case outside the United States.
American Political Science Review, American Journal of Political Science, and Journal of Politics.
World Politics, Comparative Politics, and Comparative Political Studies.
Urban Affairs Review.
There are two ways in which a small city can appear in an empirical article. The first is as an observation in a large-N (larger than 10 units) study making intermunicipal comparisons. On the whole, smaller cities are fairly well-represented in this way, as about 36% of the studies in our dataset are of this type. A few large-N studies focus exclusively on relatively small- and medium-sized cities (e.g., Anderson et al. 2019; Weitz-Shapiro 2012) and provide considerable insight into politics in this category of municipalities.
Smaller cities are less well represented in articles situated in a single city or a small number (fewer than 10 units) of cities. Here, we define these articles as case studies; by case studies, we mean studies that focus on a smaller number of specific cities in some detail. These often explore intramunicipal variation, are important for theory building, and comprise well over half of the sample. At first glance, small cities (using the relative measure, namely any city not in a country's top 10 most populous urban agglomerations) seem well represented in the literature. About 43% of these case study articles include at least one small city. Yet this figure is driven primarily by studies of the U.S., and only about 28% of case studies outside the U.S. include a single small city. This trend is particularly striking for the discipline's top-tier general interest journals. In the past 20 years, only four case studies of urban politics outside the U.S. include a single small city. We see similar trends for the American Political Science Association's flagship journal on Urban Politics, Urban Affairs Review. About 41% of case studies include at least one small city, but this percentage shrinks to 22% for countries outside of the U.S.
We see slightly different trends when using the absolute definition of small cities. Here, we define a city as small if it has a population of less than 1 million. Previous analysis of articles in Urban Studies journals in 2016 showed that 72% of articles about cities in the developing world focused on cities of more than 1 million (Post and Kuipers forthcoming). Similarly, we find that just 29% of case studies include at least one small city by this definition, with better representation for non-U.S.-based cities (37%) than U.S.-based cities (24%). Representation is less uneven in Urban Affairs Review, with 25% of all case studies, 22% of U.S.-based case studies, and 33% of non-U.S.-based case studies including a small city in absolute terms. Yet the overall representation is low, especially in the discipline's top tier general interest journals. Over a 20-year period, just 17 articles in these three journals include a single city with a population of less than 1 million, and just six of these are located outside of the U.S.
The previous analysis specifically discusses publishing in Political Science journals, but relevant articles could also be published in journals specifically focusing on cities and urban dynamics. To examine if this is the case, we review all articles published in five interdisciplinary Urban Studies journals (Cities, the International Journal of Urban and Regional Research, the Journal of Urban Affairs, Local Government Studies, and Urban Studies) in 2019. We follow the same rules for inclusion as in our review of articles in Political Science journals and find that the trends are notably similar. Table 3 summarizes the extent to which these journals focus on small cities. Empirical articles in Urban Studies are not very likely to feature small cities as part of a large-N study, as fewer than 20% of empirical articles use such methodologies. Relatively small cities are somewhat more likely to find their way into case studies, with 28% of case studies including at least one city outside of a country's top ten most populous metropolitan areas. As in Political Science, this figure is much higher at 67% of case studies, for U.S.-focused research. Also, only about 25% of case studies feature a city with a population of less than 1 million. Nevertheless, due to the large overall volume of empirical research on cities in Urban Studies, small cities receive more attention here than in Political Science journals. Furthermore, some of the work by Political Scientists on small cities finds a home in these interdisciplinary journals, as we find six large N studies, three case studies including relatively small cities, and three case studies of absolutely small cities with a Political Science author across the five journals in 2019. Broadly, however, small cities remain disproportionately underrepresented in this interdisciplinary research.
Empirical Studies in Five Interdisciplinary Journals of Urban Studies, 2019. 1
Cities, International Journal of Urban and Regional Research, Journal of Urban Affairs, Local Government Studies, Urban Studies.
Studies using statistical analysis on intermunicipal data and/or more than 10 cases.
Percentage of total empirical studies.
Number of articles including at least one small city. Small cities are defined as those outside the top 10 most populous metropolitan areas based on data from The UN World Urbanizations Prospects, 2010. Hong Kong is considered administratively separate from China in UN population rankings.
Articles including small cities as a fraction of articles classified as small-n case studies.
Small cities are defined as those with a population of less than 1 million based on data from The UN World Urbanizations Prospects, Population Division, 2010.
Studies in which all cases are from the United States.
Studies including at least one case outside the United States.
In this paper, we focus on Political Science because the discipline's nascent-yet-growing interest in cities provides ample room to develop rich and broad-reaching theories about cities of all sizes. We see opportunities for future research in small cities in all countries. Scholars of American Politics have diversified their cases outside the ten largest metropolitan areas in the country but could still study more metropolitan areas with fewer than 1 million people. There is greater room for improvement for scholars of Comparative Politics; both general interest and subfield journals show a clear bias toward articles focusing on larger (in both relative and absolute terms) metropolitan areas. This bias has substantive implications for our understanding of politics.
Why the Politics of Smaller Cities May Be Different
We contend that it is important to include smaller cities in empirical studies because they are likely to exhibit economic, social, and political patterns that differ from those found in large metropolitan areas. We illustrate patterns using administrative data and research from Urban Studies, Geography, and Economics, fields that have already begun focusing on smaller cities. 5 We also rely on Political Science examples from the United States, which, as we have seen, is currently where small cities are most represented in empirical studies of Urban Politics.
In the sections that follow, we show how cities of different sizes vary in their demographic characteristics, citizens’ preferences, access to resources, intergovernmental relationships, and electoral politics. We further discuss how and why these variables are important to the study of politics.
Variation in Demographic Characteristics and Preferences
Hajnal and Trounstine (2014) ask “What is urban politics really about?” In their analysis of local elections in the United States’ 25 largest cities, they find that many variables shape the urban vote, including class, religion, sexuality, age, gender, and most of all, race. In other words, they find that individual and aggregate demographic characteristics play an important role in how citizens make political choices at the local level in cities.
Consider, for example, the study of gender and urban politics. This is still a nascent field of study, but recent research finds key differences between men's and women's political behavior in urban India and Pakistan, where gender is an extremely salient aspect of citizens’ day-to-day lives. Women are less likely to express their own opinions (as opposed to those of their spouses) (Khan 2021), to vote because they lack information or their male family members will not allow them to do so (Cheema et al. 2022), or to be recruited or contacted by party organizations because it is deemed inappropriate for them to interact with men outside the family (Goyal 2020). In other words, women are less involved in local politics, in large part because of norms about women's behavior and intrahousehold bargaining power.
Yet it is possible that these patterns vary with sex ratios. Sex ratios may affect norms and intrahousehold bargaining due to their direct effects on a number of outcomes, including household income and female labor force participation (Angrist 2002), attitudes about gender roles (Grosjean and Khattar 2019), and the lived experience of being a woman (see Dyson 2012 for a review).
Furthermore, sex ratios can vary quite a bit by city size, as shown in the Indian case (Figure 2). The largest cities, particularly those with populations over two million, tend to have more men than women, with most sex ratios lying between 0.8 and 1.0. However, we also zoom in on the 486 cities in the dataset with populations below two million. These cities exhibit considerably greater weight in the distribution of sex ratios between 0.9 and 1. In fact, 62 of India's top 500 cities by population have a sex ratio greater than 1, where there are more women than men. 6
Sex ratios for cities in India: sex ratio for the top 500 cities by population, zoomed into cities with a population under two million.
More generally, this variation in sex ratios suggests that norms surrounding women's behavior and patterns of intrahousehold bargaining look fundamentally different in smaller cities where there is a relatively higher sex ratio for the whole year or at least part of it. It is further possible, therefore, that patterns of women's political participation, expressions of women's preferences, and men's preferences surrounding policies salient to women are also different in smaller cities.
And indeed, some studies on gender, particularly those based in the U.S. or other OECD countries, do capture intramunicipal variation related to gender and urban politics. (e.g., Holman 2014; Langarita Adiego 2020). More broadly, the discussion of gender serves as an example that simple variation in demographics (which can include all of the variables that Hajnal and Trounstine (2014) find to be important) may be an indication of deeper differences in the local-level allocation of power between smaller and larger cities.
Variation in the Levels of Resources and Local Capacity
A substantial portion of the literature on urban politics and the politics of development focuses on resource allocation. Research on the distributive politics of service provision, for example, examines who gets access to basic services such as water, electricity, sanitation, and even land use along with predictors of quality (e.g., Auerbach 2016; Farole 2021; Golden and Min 2013; Habyarimana et al. 2007; Holland 2017; Ichino and Nathan 2013; Kumar et al. 2022; Trounstine 2020). These studies essentially identify patterns of resource allocation within cities.
Yet it is likely that the aggregate level of resources and local capacity themselves vary by city size. Sahasranaman and Bettencourt (2021) find that while slums in India typically offer worse access to services than non-slums within a city, slums in smaller cities overall tend to have lower levels of services than those in larger cities. Ferré, Ferreira and Lanjouw (2012) find that in Morocco, 80% of citizens have access to sewerage in cities greater than 1 million but less than 50% do so in smaller cities. Conversely, Post and Kuipers (forthcoming) find that smaller cities offer better access to basic health and education services than their larger counterparts in Brazil, Indonesia, and India. It is likely that patterns of political contestation and mediation look substantially different in places where resources are more (or less) scarce. Two reasons that smaller cities might have less access to resources are variation in revenues from taxation and foreign direct investment.
Tax revenue. The resources available for public services may vary with city size in part because smaller cities collect less tax revenue per capita. Figure 3, for example, shows local tax revenue per capita for cities in Hungary in 2010. Aside from a few outliers (which themselves are smaller cities worthy of study), the graph shows a clear positive trend, or that local tax revenue per capita increases with city size. This trend could hold for many reasons, including that wages tend to be higher in larger cities (see Rosenthal and Strange 2004 for a review of the empirical literature), that incentives for taxation and administrative capacity vary with city size (Afonso, Araujo and Nóbrega 2012), or that larger cities are more likely to have implemented meritocratic civil service reforms, resulting in more competent employees (Ruhil 2003).
Local tax per capita by city size in Hungary, 2010 in thousands of HUF.
City size may also systematically impact the types of taxes that cities rely upon. Larger cities benefit from a more varied tax base. Moreover, different tax tools are likely to be more effective in larger, more central cities (Bourassa 1990). Some small cities are disproportionately reliant on taxes related to tourism or visitors from nearby areas, which may lead them to use a different tax mix (i.e., more sales than property tax) to outsource tax burden to non-residents (Snodgrass and Otto 1990). Property tax revenue is another area where we expect significant variation between larger and smaller cities. Given that housing prices (Ellis and Andrews 2001) and rent prices (Eaton and Eckstein 1997) are both typically higher in larger cities and that housing demand in major cities tends to be less elastic (Zhang et al. 2016), we can implicitly expect property and land values to be higher in larger cities, thereby creating greater potential for property tax revenue. A broad analysis of property tax revenue in developing countries shows considerable within-country variation on the share of tax revenue from property taxes (Dillinger 1988).
Foreign direct investment. Smaller cities may also have fewer resources because they receive less foreign direct investment (FDI), which heavily concentrates in major global cities (Belderbos, Du and Slangen 2020; Friedmann 1986; Taylor and Derudder 2015). 7 Many of these same patterns in investment concentration in global cities are replicated on regional levels. In knowledge-intensive industries, van ‘t Hoff and Wall (2020) find a clear concentration of FDI in the global cities of Western Europe. In Eastern Europe in the 1990s, national capitals received the largest share of incoming FDI, with some capitals getting more than 60% of total incoming investment (Hamilton et al. 2005). In sub-Saharan Africa, not only is more FDI attracted to larger cities, but more productive firms are located in larger cities as well (Sanfilippo and Seric 2016), a finding which builds on a large literature of finding increasing firm productivity in larger cities (e.g., Segal 1976).
There are several reasons why FDI may tend to flow to the largest cities. The clearest path to attracting investment is being a part of an agglomeration economy (Guimarães, Figueiredo and Woodward 2000). 8 Proximity to other potential investment sites can matter not only for the agglomerative effects, but simply being near other potential investment sites could allow cities to be alternate recipients (Blanc-Brude et al. 2014). Secondary characteristics of smaller cities also may matter more for attracting investment. Su and Liu (2016) find that the impact of human capital on attracting FDI is greater in sub-provincial level cities than in larger ones, suggesting that larger cities are less reliant on having a highly educated workforce than smaller ones in trying to attract investment. Goerzen, Asmussen and Nielsen (2013) further argue that multinational corporations disproportionately invest in these larger global cities because of the reduced liabilities of “foreignness” due to their greater cosmopolitanism and integration into the world economy. Smaller cities may instead need to market themselves based on their perceived “authenticity” (Banks 2022) or rely on “place-sensitive strategies” for economic development (Rodríguez-Pose and Griffiths 2021).
The considerable concentration of FDI resources in the largest cities is important to our understanding of urban politics because it has economic and policy effects that further differentiate large and small cities. For example, the environmental impacts of FDI differ according to city size, with FDI having smaller impacts on environmental pollutants in larger cities (Zheng, Kahn and Liu 2010). This could partially account for Pisani et al. (2019) finding that cities pursuing service-sector related FDI can actually participate in a “race to the top,” wherein they implement stricter environmental policies to attract greener firms with higher “environmental capabilities.” This differs from the standard narrative of a “race to the bottom” to pursue manufacturing FDI, and this strategy may be especially reliant on the higher average state capacity of larger cities to successfully implement.
Intergovernmental Relationships
Taxation and FDI, however, are not the only factors affecting local government revenues. Intergovernmental fiscal relations and the receipt of grants and transfers from higher tiers of government also matter a great deal. These systematically affect smaller cities differently than larger cities in both unintentional and politicized ways. For example, as part of the dominant party in Hungary's attempts to consolidate power during the COVID-19 pandemic, fiscal transfer rules were altered. Cities with populations under 25,000 would be reimbursed for lost revenue by right, while larger cities, where the opposition is stronger, would only be compensated on a case-by-case basis (Stenberg, Rocco and Farole 2022). In another example, eligibility for federal Community Development Block Grants in the United States is based on city size and/or designation as a principal city in a metropolitan area.
These types of differentiated intergovernmental fiscal relationships are only one way that intergovernmental relations affect small and large cities differently. We consider how vertical relationships with other tiers of government and relationships with other cities vary by city size.
Vertical relationships. First, small cities may have different relationships with other tiers of government. Fundamentally, we miss particular intergovernmental political dynamics at the second-tier of government by focusing solely on large cities, and the placement of cities in larger multilevel contexts matters (Sellers 2005; Zhang 2020). This can be especially true in more rural states and provinces, which are likely to have different political dynamics than the most urbanized ones. Table 4 illustrates this point for countries with the largest urban populations. For each country, we show the percentage of second-tier subnational units (i.e., states in the United States) represented in lists of cities by the population at three different thresholds: the top 10, 30, and 50 municipalities by population. 9 Here, we focus on municipalities rather than urban agglomerations because agglomerations can often cross state lines and involve multiple municipal jurisdictions.
Percentage of Meso-Level Subnational Jurisdictions (e.g., States or Provinces) with Cities in a Given Ranking. 1
Nigeria is excluded from our data because the Nigerian government does not have formally incorporated cities, and existing data sources have inconsistent data on Nigerian city population. This paucity of accurate population data for Nigerian cities outside of the top 10 further points to the need to study smaller cities more.
Jakarta is considered as five separate cities administratively.
Russian sovereignty over Crimea is not recognized, thereby resulting in 83 administrative jurisdictions.
We can see a great deal of variation. In some countries, such as Russia or Brazil, a focus on the top 10 cities will capture dynamics in ten separate provinces or oblasts. In Indonesia or the United States, this would capture much less of the diversity in intergovernmental politics. The differences become starker as one considers the top 50 cities. Although countries with a greater number of second-tier jurisdictions will always have lower percentages by this measure, they show major disparities. In India, the top 50 cities are concentrated in just 19 of 36 second-tier jurisdictions (states and Union Territories). This means that the intergovernmental political dynamics in about half of states would be missed in a study that focuses on even the top 50 cities by population. Similarly, in Japan, more than half of prefectures are not represented by the top 50 cities, and in the U.S., more than 20 states would not feature.
There are theoretical reasons to believe that intergovernmental relationships look substantially different in subnational units with and without major cities. Consider, for example, the concept of urban bias, or what some scholars argue is the tendency of governments to disproportionally favor the interests of the urban elite over rural populations (Lipton 1977). To the extent that this pattern does occur, we might expect subnational units without major cities to receive different treatment from central governments. The reverse pattern may appear in countries with a rural bias. Even where policymaking and implementation occur at the subnational level, the relationship between rural and urban elites looks different in states and provinces with smaller cities, particularly as smaller towns and cities are more likely to be destinations for rural and circular migrants (Christiaensen, De Weerdt and Todo 2013). Different political and institutional features in intergovernmental relations can even impact the ways that cities work together in intermunicipal relationships (Weir, Wolman and Swanstrom 2005; Meza et al. 2019). We, therefore, expect different patterns of central–subnational and subnational–local relationships in subnational units with and without major cities.
Intermunicipal relationships. Smaller cities also exhibit different patterns of intermunicipal relationships. Although municipalities of all sizes often cooperate with others, the types of cooperation vary with city size. Over the past thirty years, cities have begun working together to advocate for global political change (Acuto and Rayner 2016). This form of intermunicipal cooperation is a logical extension of the “global cities” discourse (King 1990; Sassen 2001), as cities begin to seek to exercise the power afforded to them by their economic importance. 10 These configurations have different focuses but are typically driven by large cities. The Eurocities network, for example, emphasizes large cities that are not national capitals (Payre 2010). The C40 network has seen 40 large cities working in conjunction to promote urban responses to climate change (Davidson, Coenen and Gleeson 2019). We also see some larger national-level networks of cities, such as the U.S. Conference of Mayors or the Deutscher Städtetag, advocate for local government interests in national politics. 11 The primary trend across these cases is large cities working to combine their influence to effect national or global policy change.
Smaller cities exhibit different patterns of intermunicipal cooperation. Teles (2016) has highlighted the cooperation with other nearby cities to improve public goods provision and policy implementation. Such cooperation is usually pursued for efficacy and efficiency gains in public goods provision (Hulst and van Montfort 2007). Using evidence from El Salvador, Muraoka and Avellaneda (2021) find that these cooperative relationships create spillovers that have broader positive impacts on local government performance. Soukopová and Vaceková (2018) find that the costs to municipalities of intermunicipal cooperative relationships vary according to city size. The gains, furthermore, are particularly large for smaller cities, as in a metastudy, Bel and Sebő (2021) likewise find that the greatest cost benefits are for smaller municipalities. 12
Electoral Politics
As discussed, there are often systematic differences between voters in smaller and larger cities, leading to different preferences visible on election day. However, the party system itself can be different in smaller and larger municipalities. These findings from the literature discussed below suggest a complicated relationship between city size and electoral politics that is neither captured by studying only the largest municipalities or by a purely linear relationship where partisan strength in local politics increases as city size increases.
In some cases, electoral systems differ by law in cities beneath a certain population threshold. For example, in municipal elections in Poland prior to recent reforms, cities under 20,000 residents used a direct electoral system rather than a proportional electoral system for municipal council elections (Gendźwiłł and Żółtak 2017). In many Indian states, such as Maharashtra, cities meeting different population thresholds have a different ratio of citizens to local representatives. The incumbency advantage is also higher in cities with larger populations (Freier 2015).
In other cases, variation in patterns of electoral politics is more a result of intergovernmental relationships and strategic choices by parties on how to allocate resources. Kjær and Elklit (2010a; 2010b) and Ennser-Jedenastik and Hansen (2013) find that there is greater nationalization of local political elections – meaning greater contestation by national political parties instead of parties specific to a municipality or region – in larger municipalities. 13 Similarly, Parchure, Phadke and Talule (2016) find that from 2004 to 2008 in Maharashtra, the proportion of independent candidates winning seats in Municipal Councils, the institutions governing Tier-II cities (those with a population of 50,000–99,000), is about 28%. This is roughly double the percentage of independent party-held seats in Municipal Corporations, the institutions that govern Tier-I cities (those with a population of over 100,000). Otjes (2018) finds that local parties, which do not contest elections nationwide but focus on a particular city or region, are more successful in smaller municipalities. Bolgherini, Grimaldi and Paparo (2021) further find that national and regional politics more impact the conduct of local politics in larger municipalities.
The nationalization of local politics and its relationship to city size is also related to another long-standing theory of multilevel politics: the second-order election. Largely popularized by Reif and Schmitt (1980), the theory of second-order elections presumes voters participate in elections at supranational and subnational levels with national issues and national party performance in mind. They do this because they typically see elections at the supra- and subnational levels as less important, presuming that most of the power lies in the core national elections. As such, voters participate in those elections at lower rates. Since the theory's development, many scholars have applied it to discussions of local and subnational elections (Freire 2004; Hijino 2013; Miller 1988). Patterns of politics in smaller cities, however, illustrate two complications to the application of the second-order election model to municipal politics.
First, a key feature of second-order elections, namely lower turnout relative to national elections (Reif and Schmitt 1980; Schakel 2015), may be found less in smaller cities. This is because we generally see greater political participation in smaller cities (McDonnell 2020), particularly with respect to electoral turnout in local elections (Dahl and Tufte 1973; Frandsen 2002; Gendźwiłł and Kjaer 2021).
Even if there is lower turnout in local elections relative to national elections in smaller cities, the mechanism discussed in the literature on second-order elections may not hold. Local elections are “second order” as they are often portrayed as referenda on national-level issues. To a certain extent, this mechanism requires voters in local elections to assign municipal candidates to corresponding national parties. However, as discussed above, smaller cities have a larger proportion of independent candidates and local parties that may be difficult to consider as proxies for national actors. Hijino (2013) discusses the decision to list independently as a strategic one based on a candidate's desire (or not) to be associated with a national party; however, if elections in smaller cities are simply contested less by national parties, then this strategic decision is moot. Kjær and Steyvers (2019) note this tension in their reimagining of local elections from second order to “second tier.” Given that there is systematically less party participation in elections in smaller cities, they may be unlikely to function as referenda on national political performance.
Theory and Policy Issues Unique to Smaller Cities
The numerous ways in which smaller cities differ from their larger counterparts suggest that empirical studies of smaller cities will be fruitful for updating or refining existing theories of urban politics and comparative politics. There may, furthermore, exist policy problems and political phenomena unique to smaller cities. These issues are potential areas for the generation of novel theories and policy recommendations. We discuss promising areas of inquiry as examples, namely issues related to urban primacy, urban–rural relationships, intermunicipal relationships, and the politics of concentrated industrial power.
As discussed above, smaller cities may exhibit vertical relationships that are different from those of their larger counterparts. These relationships are likely to look distinct in cases of extreme urban primacy, or where a large percentage of a country's urban population is concentrated in one or two cities (see e.g., Bertinelli and Strobl 2007; Cuervo and Cuervo 2013; Ding, Fu and Jia 2019; Frey, Dietz and Marte 1986; Henderson 2002; Short and Pinet-Peralta 2009). The phenomenon is common in countries in Africa, with Bujumbara, Burundi; Asmara, Eritrea; and Monrovia, Liberia containing over half of their respective countries’ urban population as of 2012 (Hartley 2015). Djibouti City, in fact, is home to over 75% of the country's urban population!
Urban primacy is likely to have important implications for urban citizens and politicians who live outside of the primate city. Henderson (2002) finds that non-primate cities receive a considerably smaller share of public investment from central governments. A fruitful area of research, therefore, is investigating how leaders from non-primate cities negotiate access to this public investment and how central governments strategically allocate funds to these smaller cities. As primacy is hypothesized to first increase and then decrease with a country's GDP per capita over time (Moomaw and Alwosabi 2004), this type of research is relevant in countries wherein non-primate cities are beginning to grow.
Small cities have a widely varying geographic distribution, and understanding the different implications of a city's relative geographic location is worth further study. Small cities are often not major industrial centers but rather single company towns or administrative, political, and trading centers for more rural areas (Ullman 1941). The nature of urban–rural relationships is uniquely relevant to a smaller city. Small towns and cities typically have more fluid relationships with rural communities than larger metropolitan areas, as they are centers of trade connecting agriculture to markets and often the first stop for rural–urban migrants. Vaishar and Zapletalová (2009) argue, for example, that small towns are important centers in rural regions in Czechia because they provide services, jobs, contacts, and local identity to the surrounding rural areas. Agergaard et al. (2019) further argue that smaller cities are important organic parts of rural areas and rural economies in sub-Saharan Africa.
These observations suggest that the politics of small cities are often more intimately connected to those of the countryside, and this relationship presents opportunities for research and theory-building. In a study of Anand, Gujarat in India, for example, Verstappen and Rutten (2015) find that the regionally dominant Patel caste uses this city to bolster and establish its power, while an oppressed Muslim community relies upon it as a refuge from rural violence and marginalization. More broadly, there are many open questions along this theme, including the nature of strategic relationships between rural and urban local leadership, the extent to which smaller cities rely on the rural hinterland economically and politically, and the role these nodes play in political or party systems with a rural bias.
Although large metropolitan areas often have potential employers spanning a range of industries, smaller cities may have one dominant industry or company. Understanding the conditions under which these small cities do or do not grow into larger agglomeration economies is one fruitful area for research. Exploring the power dynamics within such cities is another. Consider, for example, Bentonville, Arkansas in the United States, which is home to Walmart, the largest retail company in the world. The city has a population of just under 50,000. At the same time, over 14,000 people work for the company's corporate headquarters (Adler and Florida 2020). Given a worker's set of skills, a company like Walmart may be the sole or best employer in a small labor market. Such a monopsony is likely to give the company and its leadership disproportionate power in local and regional politics. This is perhaps made even more true by the fact that small business jobs outside major metropolitan areas in the U.S. can often be less stable and lower paying (Malizia 1986). Research on the politics of the “company town” (Porteous 1970) and cities with high levels of industrial concentration more generally is likely to generate rich new theories and evidence of how different actors gain and exercise power in a smaller city. As an illustration, Neumann (2016) discusses how dependence on a large metallurgical complex might have prevented the residents of a small mining town in Peru from pursuing collective action against environmental harm.
Conversely,other small cities are located near major metropolises and can fall into a core-satellite city relationship. Typically, the academic analysis of these relationships exhibits a bias toward the larger cities (Servillo, Atkinson and Hamdouch 2017, 372–73). Satellite cities have a diverse range of origins, from industrial suburbs of major urban centers (Taylor 1915), to regional cities brought into the orbit of a larger city as commuter towns (Salamon 2003), to recently constructed New Towns as planned urban expansion (Van Leynseele and Bontje 2019). Given that satellite cities have a wide variety of functions (Davidovich 1962, 8–10), this conceptual bias toward the regional center can lead researchers to misunderstand dynamics within smaller cities. Li, Zhang and Liang (2010) observe that, while growth in satellite cities is no doubt influenced by core cities, satellite cities also individually exhibit independent urbanization processes. Research is also beginning to examine patterns in where satellite cities emerge in larger urban regions (e.g., Ikeda, Aizawa and Gaspar 2021). Altogether, research on the varying ways that small cities interact with their surrounding regions is worthy of further study, and there are key potential insights that may come from overcoming a conceptual bias toward core cities when studying urban regions.
How to Study Smaller Cities
There are practical reasons why typical research methods are less tractable for small cities. Fieldwork is often constrained by real-life logistical concerns. In some countries, limited transportation connections may make it difficult to access lower-population cities. Furthermore, the municipal-level administrative data required for quantitative analyses of intra-municipal variation may simply be unavailable, especially in developing contexts. We suggest the use of digital surveys and interviews, crowdsourced data, and remote-sensing data as a first step for circumventing some of these problems. 14
First, digital communications technology allows researchers to conduct qualitative and quantitative interviews with officials and residents of small cities that can be difficult to access in person. The telephone numbers and email addresses of lower-level officials are frequently listed on public websites, and researchers may be able to secure phone or video interviews simply by reaching out to them. Citizens can be recruited and interviewed through targeted Facebook ads or Mechanical Turk (Boas, Christenson and Glick 2020). Geocoded social media posts can provide data about a wide range of social phenomena that can be relevant to urban politics. Beyond these ideas, we anticipate that the coming years will reveal a great deal of innovation in digital interview and recruitment techniques spurred by the travel restrictions imposed by the COVID-19 pandemic.
Second, researchers can also use the crowdsourced initiatives upon which cities increasingly rely to collect and provide information about service delivery and other government functions (Ashton, Weber and Zook 2017; Bannister and O’Sullivan 2021; Post, Agnihotri and Hyun 2018). These data can be used to estimate outcomes relevant for political behavior. Fałkowski and Kurek (2020), for example, use mapped data on the number of streets with religious names and the number of religious monuments to approximate church attendance in Polish local governments. Technology company databases can potentially provide novel data about many variables; for example, Zillow on home prices (for one Political Science example, see Canes-Wrone and Park 2014); Moovit or the Transit App on public transportation (Basu and Ferreira 2021); restaurant reviews on Dianping (Williams et al. 2019); and even city subreddits (Lê et al. 2020). Crowdsourced data is also relevant to the study of urban politics. Examples include public complaint systems about water and air quality (Sachdev 2017), problems with service delivery (Iliffe et al. 2014), and the collection of bribes (Ang 2014; Ryvkin, Serra and Tremewan 2017). Researchers should, however, be aware of inferential problems posed by crowdsourced data, such as those related to selection and data continuity, and develop strategies to address them (Post, Agnihotri and Hyun 2018).
Finally, researchers have begun to use remote sensing data and digital data (see Gatrell and Jensen 2008) to approximate outcomes of great interest to political scientists. Yang et al. (2021) use satellite nighttime light data to examine shrinking cities in China; Duque et al. (2015) use land cover data to study intra-urban poverty in Medellín, Colombia; Rains, Krishna and Wibbels (2019) use roof shape data to measure slum cover in India. When made available by telecom companies, mobile phone use data can also provide powerful information about household-level wealth and behavior (Blumenstock, Cadamuro and On 2015).
Even once data have been collected and analyzed, studies of smaller cities may be less well received by editors and reviewers because they examine a smaller population than studies of larger metropolitan areas. Indeed, the case selection in studies of large metropolitan areas is often motivated by the sheer size of the population under scrutiny. Researchers focused on small cities will want to emphasize the theoretical contributions of their work, as doing so is likely to increase its appeal to editors and reviewers. Nevertheless, studies of smaller cities may be more likely to be met with questions about the external validity or substantive importance of findings. To minimize such concerns, researchers might consider study designs that randomly sample one or several cities for inclusion, much like the standard practice for studies of rural villages. Bashar and Bramley (2019) offer another strategy, deliberately picking cases in both large and secondary cities to have a variation in city size. Even studies of single cities can be motivated by data on how comparable the case is to other cities in the given country or region.
More generally, this article has highlighted the limited external validity of studies conducted in large metropolitan areas as well. There is likely substantial heterogeneity in patterns of urban politics across all cities. We, therefore, advocate a clearer articulation of scope conditions, particularly with respect to city-level variables, in all papers about urban politics, even those focusing on large agglomerations.
In spite of the challenges for studying smaller cities, there are some advantages for researchers who study smaller cities. Costs for room and board are likely to be lower upon arrival. Local officials might be less burdened by interview requests and more likely to be amenable to answering questions. These practical considerations can be real benefits for scholars short on resources as they go into the field.
Conclusion
We have demonstrated that cities that are “small” in both relative and absolute terms are understudied in Political Science relative to the large share of urban residents that live in them. Using a wide range of data and insights from literature in Urban Studies, Economics, Geography, and the research in Political Science that does consider smaller cities, we highlight how these smaller cities have different patterns of urban politics than their much larger counterparts. In particular, we focus on demographic characteristics, citizens’ preferences, access to resources, intergovernmental relationships, and electoral politics. We argue that both focusing specifically on smaller cities and including them in larger empirical analyses will broaden and deepen our theories while simultaneously lending greater credibility to our findings. We also highlight political phenomena and policy problems that are unique to small cities, and therefore, present promising avenues for new research here. Our hope is that this is an agenda-setting piece that not only motivates special issues and symposia on smaller cities but also highlights their importance to general interest research in Political Science and Urban Politics.
A better understanding of the politics of smaller cities will have important policy implications, particularly in urbanizing countries. In a study of Albania, Brazil, Kazakhstan, Kenya, Mexico, Morocco, Sri Lanka, and Thailand, Ferré, Ferreira and Lanjouw (2012) find that the incidence of poverty is greatest in the countries’ smaller cities. These poverty rates are compounded by the lower access to services and resources that we have highlighted. It is, therefore, important to understand the policies that may improve the lives of these citizens as well as the politics of implementing those policies. Additionally, high values of urban primacy have been found to be detrimental for a country's overall economic development (Castells-Quintana 2017; Rondinelli 1983), further indicating the importance of policymakers’ focus on smaller cities. Smaller cities may also present solutions to broader problems. For example, while cities have generally been acknowledged as leaders in the fight against climate change (Rosenzweig et al. 2010), smaller cities may serve as important nodes for technology and policy diffusion to the countryside (Wisner et al. 2015).
Considering the whole range of cities in which people live will improve our overall understanding of urban politics. Both case studies of intramunicipal variation and large N analyses of intermunicipal variation are of value, but Political Science's turn toward design-based inference suggests that case studies may be more tractable and strategic for researchers. Moreover, research that generates better data about small cities can facilitate future intermunicipal comparisons. A more representative empirical base for our findings is likely, furthermore, to alter theories about a range of outcomes, a few of which we have discussed. Our understanding of political behavior, electoral politics, intergovernmental relations, and subnational economics will also benefit from a greater breadth of focus, and the findings from this research can improve outcomes for a considerable fraction of the world's population that live in smaller cities. Overall, we believe that cities are important laboratories to study the world's most pressing policy problems, such as inequality, climate change, and infectious disease, and our current global environment demands a more comprehensive understanding of their politics.
Supplemental Material
sj-docx-1-uar-10.1177_10780874221124610 - Supplemental material for Why Political Scientists Should Study Smaller Cities
Supplemental material, sj-docx-1-uar-10.1177_10780874221124610 for Why Political Scientists Should Study Smaller Cities by Tanu Kumar and Matthew Stenberg in Urban Affairs Review
Footnotes
Acknowledgments
The authors thank Andrew Tanner, Sania Shahid, and Isha Patel for their valuable research assistance on this project. The authors thank Adam Auerbach, Shelby Grossman, Alison Post, Margaret Weir, and five anonymous reviewers for their comments.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article
Supplemental Material
Supplemental material for this article is available online.
Notes
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
