University of Vermont
Preliminary Draft: Do not quote or cite.
December 9th, 2010
In this paper, we use the familiar strategy of exploiting the quasi-experimental distribution of political institutions in the Americas, caused by variations in European colonial experience, to examine the role of institutions on urban and local development in the Americas. Political institutions—whether unitary or federal, or parliamentary or constitutional—matter for urban and local development because they define the lines of authority among national, regional and local governments. We argue in this paper that in Latin America, the centralization of political power in federal governments at the expense of local governments contributed to the rise of urban primacy whereas in Canada and the US, political decentralization fostered more balanced patterns of urban and local development. Yet, even between US and Canada, we find that the differences in the levels of political centralization at the state or provincial levels led to important divergences in the organization and number of cities and local governments.
Within a relatively short period of time, a significant literature has emerged that links history, institutions and economic growth, especially for the European colonial economies (see, for instance, Engerman and Sokoloff, 1997, 2002; Acemoglu et al., 2001). History matters because variations in historical conditions at the time of European conquest—whether factor endowments, indigenous pre-colonial development, climate, the disease environment facing European settlers, or the conditions of European technological and political development at the time of conquest—played vital roles in influencing the transplantation of a variety of colonial institutions. And even though subsequent changes in economic and political conditions have modified in important ways these initial institutions, early institutions, by placing constraints on future institutional change, significantly shape the form of today’s institutions, which, in turn, may help explain why some countries are rich and why others are poor.
The urban economics literature, on the other hand, has been much slower to examine the role of institutions in influencing urban or spatial development. With some exceptions, the literature treats the growth, location and size distribution of cities as being determined solely by the economic forces of agglomeration and congestion. This lack of interest among urban scholars in studying the role of institutions is surprising since cities are often viewed as “engines of economic growth.” It is also puzzling because cities are foremost “creatures of states,” meaning that a higher-level of government usually has the power to determine how and whether local governments can be established, their geographic boundaries, the authority that local governments may possess, and the revenue sources they may tap, unless explicitly constrained by constitutional provisions or home rule legislative enactments.
In this paper, we explore the complex linkages among history, institutions and urban development in the New World. In the Americas, the exogenous role played by European colonists in influencing the path of institutional development has not been questioned (North 1990). Yet, except for the early literature on Latin American urban primacy (see, for instance, Morse 1971 and McGreevey 1971), relatively few scholars have seriously explored the effect of institutions on urban or spatial development. Because political institutions fundamentally define the rights of property and the lines of political authority among national, state, and local government units, institutions are likely to play a key role in determining the type of fiscal federalism that emerges and the patterns of local economic development.
The impact of institutions on economic development of North and Latin America are especially sharp. In Latin America, conditions of inequality led to poor, un-democratic institutions that hindered economic growth (Engerman and Sokoloff 1997). The concentration of political power in a handful of elites also favored the rise of powerful centralized national governments and politically and fiscally weak regional and local governments (Sokoloff and Zolt 2006). Not surprisingly, this imbalance of political power between various political jurisdictions led not only to the rise of urban primacy among national and regional capital cities (Galiani and Kim 2010), but is also believed to have contributed to poor long-run economic performance.
In North America, on the other hand, initial conditions of equality led to the rise of democratic institutions that are believed to have fostered innovation and economic growth. In addition, egalitarianism and democracy were accompanied by political decentralization at the state and local levels.1 In the US, New England towns became symbols of local autonomy.2 In the South, counties became important units of local government. In the Middle Atlantic, local governments consisted of counties and towns. In the new states of the West, counties became the most dominant form of local government that sometimes included townships.3 The status of cities varied by region and over time. In general, cities were initially recognized as public corporations, not unlike other corporations at the time, but with the rise of Dillon’s rule, cities became “creatures” of state governments. Yet, even before the rise of the municipal home-rule movement, American cities possessed considerable political autonomy.
In Canada, unlike the US, towns and counties were never as important local government units.4 The most important unit was the municipality. Scholars generally believe that, from the colonial years until the late nineteenth century, local municipalities in Canada had considerable local autonomy (Crawford 1954). Indeed, in both the US and Canada, local governments were the most important units of government at the turn of the twentieth century as evidenced by the fact that local government expenditures exceeded the expenditures of other levels of government in both nations. Nevertheless, there is considerable evidence that the political autonomy of local governments diverged between US and Canada, especially over the second half of the twentieth century (Kim and Law 2010). Whereas US local governments became more independent of state governments, the provincial governments in Canada, especially in Ontario, began to exercise greater political authority over local governments.
While establishing a link between institutions and urban development is challenging, we believe that variation in urban development between Latin America and North America, between the US and Canada, and even among US states and Canadian provinces sheds considerable information. In this paper, we explore variations in urban primacy and the size distribution of cities among the countries in Latin America, between the United States and Canada, and among regions of Canada and the US to show the role of political centralization in determining the pattern of urban development. Additionally, we explore the role of political institutions between US and Canada and among Canadian and US regions by examining differences between Canada and the US, as well as among US states, in the number and scope of local government units.
While the institutional rules guiding the power of national, regional and local governments change and evolve over time, we find that these changes are conditioned upon institutional rules adopted during the colonial period. In Latin America, recent moves toward greater political decentralization are unlikely to affect the pattern of urban development. While there is evidence of a greater degree of political decentralization in several Latin American countries, the legacy of highly centralized government in terms of the distribution of city sizes and the extent of urban primacy persist to this day. In Canada, centralization of political power within provincial governments was made possible by the adoption of the parliamentary form of government at the provincial level. The absence of provincial constitutions combined with powerful provincial parliaments provided little political authority for Canadian local governments. In the US, despite the changing roles of state and federal governments and their relations to local governments, state governments are constrained by constitutions and local representation within state legislatures. Local governments have therefore always been able to exercise considerable political autonomy. Indeed, the proliferation of local governments, all with taxing powers, is a uniquely American development.
The paper is organized as follows. In section II, we outline the history of political institutions in the Americas as it relates to the rights and authority of local governments. In section III, we examine urban primacy and the size distribution of cities for countries in the Americas as well as regions in the US and Canada over time. In section IV, we analyze variations in the number, size and scope of local governments in US and Canada over time and use differences in the degree of political centralization between the two countries to examine the relationship between political centralization and the number of local government units across Canadian provinces and US states, as well as between Canadian and US metropolitan areas. In section V we exploit variation across US states and metropolitan areas to explore how differences in the extent of political decentralization (measured by municipal home rule status) across US states influenced the number of local government units and the fragmentation of metropolitan areas. Finally, section VI concludes.
II. The Institutional Rules of Local Governments in the Americas
In this section, we examine the institutional rules that determine the political and fiscal autonomy of local governments in the Americas from a historical perspective. Institutions, whether formally defined by constitutions, informally by norms and precedents, or by the sheer force of military rule, define the lines of authority among national, state and local governments. In the Americas, due to differences in colonial histories, the distribution of powers granted to the different levels of government varied significantly. In Latin America, political power has been historically centralized within national governments, usually at the expense of regional (provincial or state) and local governments. To this day, notwithstanding recent movements toward greater local decentralization in Latin America, taxing authority continues to remain highly centralized at the national level (Diaz-Cayeros 2006). In the US, colonial conditions led to the creation of a federalist form of government that granted significant political autonomy to states and their local governments. In Canada, despite the attempts by the founders of Canadian confederation to form a more centralized federal union, federalism came to favor decentralized provincial rights, but these rights were not extended to local governments.
A. Local Governments in Latin America
In Spanish America, just as in the Iberian Peninsula, political power was highly centralized in the King, his council and the Viceroy. Unlike in North America, no charters were granted to establish colonists’ rights of self-government. Local assemblies possessed few political rights. Because the main objective of the Spanish crown was to mine silver using native labor, there was little incentive for the Spanish to establish institutions that provided colonists and natives with powers of local autonomy (Sokoloff and Engerman 1997). Thus, local governments in Latin America were under the direct control of the King and his advisors.
Even after independence, civil wars and the establishment of constitutional governments in many countries, most Latin American nations remain highly centralized at the national level. Formally, countries such as Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Panama and Paraguay are highly centralized, as they possess a unitary form of government; Bolivia, Chile, Honduras, Nicaragua, Peru and Uruguay are moderately centralized whereas Argentina, Brazil, Colombia, Mexico and Venezuela are decentralized, possessing federalist forms of government (see Table 1).
Nevertheless, throughout Latin America, central governments wielded considerable authority over local governments until very recently, even in countries with a federalist form of government. Table 2, based on Myers (2002), presents information showing how national governments centralized political power over capital cities and other municipalities in Latin America during the period between 1944-1962 and 1978-1982. In the mid-twentieth century, the most centralized countries were Mexico, Colombia and Peru. In these countries, the national government appointed mayors for the capital city and other cities; in the next group of countries, Brazil and Argentina, the central government appointed capital city mayors but used a mixture of elections and appointment for mayors of other cities; in Venezuela, the national government appointed its capital city mayor but elected other city mayors; and, in the least centralized countries (Cuba, Guatemala, and Chile) mayors of all cities (capital or otherwise) were elected. By the second half of the twentieth century, however, most Latin American countries moved toward political decentralization, allowing the election of mayors for capital and other cities. The major exceptions were Havana and Buenos Aires whose mayors continued to be appointed. The national governments of Cuba and Venezuela also appointed the mayors of other cities. These patterns of central government control over municipal affairs contrast sharply with the experience in Canada and the US, where the mayors of all cities, including national and sub-national capital cities, have always been elected.
Yet, despite these recent moves toward greater formal political decentralization in Latin America, there remains considerable evidence that Latin American countries are still highly centralized. Despite the fact that Mexico has a federal form of government, Nickson (1995) argues that Mexico is still an extremely centralized state, even by the standards of Latin America. In Mexico and other Latin American nations, political power was originally centralized in the national capital, which in turn received the lion’s share of the national government’s revenues. Similar histories can be told for Argentina, Colombia, Chile and elsewhere in Latin America (Nickson 1995). Political centralization in capital cities was aided by a policy of general neglect of other municipalities.
Sokoloff and Zolt (2007) and Willis et al (1999) note that local governments in Latin America possessed little political autonomy and few fiscal resources. While local governments in the US and Canada relied heavily on property taxes to fund local public goods such as roads, infrastructure, and education, municipalities in Latin America were prevented from raising revenues by their national and state governments. Many countries in Latin America also relied heavily on regressive taxes on consumption such as excise taxes and taxes on foreign trade—taxes that are usually collected by central governments--rather than on personal or corporate income taxes. In 1930, taxes on international trade accounted for 44% of central government in Brazil, 48% in Argentina, 54% in Chile, 55% in Colombia, 41% in Mexico, and 51% in Venezuela (Sokoloff and Zolt 2007). Diaz-Cayeros (2006) notes that tax collection in Latin America remains highly centralized in the national government and that state and local governments rely primarily on transfers from the central government.
B. Development of Local Governments in the US
When British settlers arrived in North America, they attempted to establish local institutions that were similar to those of their mother country. Contemporaneously in England, local governments consisted of first-tier counties that had political jurisdiction over smaller second-tier townships or manors. The King and Parliament had, by the late seventeenth century centralized authority over most these local governments through the appointment of county sheriffs and justices but their control was not highly systematic as local officers were chosen from the local rural gentry. The larger cities, on the other hand, possessed royal corporate charters that granted them some measure of autonomy such as commercial privileges and the right to elect local officials.
Due to initial differences in the degree of inequality income and social status, the American colonies varied in their adaptation of English local institutions.5 In Virginia and other southern colonies where inequality was high, the elites closely adopted English local government institutions and established counties as the local unit of government. Similar to England, the governor of the colony appointed county officers in these southern colonies. In more egalitarian New England, however, local institution diverged from that of the mother country from the outset. Favoring compact settlements, the town became the primary unit of local government in New England, although counties were also organized for judicial purposes. Most importantly, however, the town officers in New England were elected. In the Middle Colonies, the local institutions combined New England towns and southern counties but with democratic features as both town and county officials were elected.
From the War of Independence to the Civil War period, the US county-town model of local government evolved toward its modern form and gradually became more uniform across the country. While local institutions continued to vary across the states, the hybrid system developed in the Middle Colonies eventually spread to large parts of the US and, with the rise of Jacksonian democracy, local officials even in the South were chosen by elections. As cities increased in number, the status of municipalities in relation to their respective states also came into question. While few cities possessed government charters as in England, Frug (1999) argues that most courts treated cities initially as public corporations. However, as cities and other forms of local government became integrated into state and national economies, the state legislatures and courts began to define more clearly the lines of political authority between states and their localities.
Under US federalism, the federal-state relationship is federal but the state-local relationship is unitary (Elazar, 1972). Since the federal Constitution is silent on matters of local governments, local governments became “creatures” of states because states possessed residual rights not specified in the Constitution. Thus, according to Justice Taney, the “[c]ounties are nothing more than certain portions of the territory into which the state is divided for the more convenient exercise of powers of government” of the state. With the adoption of Dillon’s rule, the cities also became “creatures” of states, but legislatures and courts made a distinction between municipalities and counties as municipalities, unlike counties, were created mainly for the interest of the locality and its people.6 While municipalities generally possess governmental, corporate, and proprietary powers, counties usually lack corporate and proprietary powers.7
Yet, despite the fact that US local governments became “creatures” of state governments, for a variety of reasons local governments continued to possess considerable political and fiscal autonomy. First, many Americans, especially those in New England, informally viewed states as a federation of localities. This gave popular credence to the notion that local governments should have their own sphere of authority. Second, many states began to recognize counties as municipal corporations through constitutional or statutory provisions. Third, over time, state constitutional reforms placed strict limitations on state legislatures, prohibiting them from creating or destroying counties or cities or changing their boundaries, and outlawing special legislation for cities. Fourth, many state legislatures granted home rule to cities and, in some cases, to counties.8 Finally, the geographic diffusion of political power within states made it costly for members of state legislatures to coordinate in the exercise of their authority. While strong party discipline could at times induce state legislatures to intrude on local matters, most state legislation was routine and was initiated by localities (Burns and Gamm 1997).
C. The Development of Local Governments in Canada
From the beginning of European settlement until the late eighteenth century, local government scarcely existed in French and British Canada. While there were a handful of cities with charters (e.g. Quebec City and Halifax), settlement was scattered and most settlers had no influence over local affairs. Authority over local matters was vested in colonial governors and their appointed councils. After the 1780’s, however, the influx of American migrants from New York and New England increased the demand for local government autonomy. However, neither the governors of British Canada nor the Canadian elites who held appointed offices had an incentive to extend autonomy to local governments. Counties were created but with no municipal functions. The 1793 Parish and Town Officers Act established the election of local officers, but assessment and rates of taxation were set by acts of the colonial legislatures rather than by local officers. In cities, however, the increased demand for local public goods eventually led to incorporation and greater local autonomy (Crawford 1954).
In Upper Canada, the first formal system of local government was established in 1841 with the District Councils Act. Under this act, the inhabitants of each district constituted a body corporate and its powers were exercised by a warden appointed by the Governor and a council elected by inhabitants. The Governor also appointed the district clerk, treasurer and surveyors, although they were accountable to the council. In 1849, the Municipal Act, also known as the Baldwin Act, established the local municipal system as it largely exists in Ontario today. Unlike in the US, the county, which replaced the district, became the upper-tier of municipal government but with no true municipal functions. Instead, rural townships, villages, towns and cities all became independent units of municipal government.
In Canadian federalism, residual rights not specified in the federal constitution belong to the federal rather than provincial governments. Nevertheless, under the terms of the 1867 British North America Act (Canada’s original constitutional document) local governments, became “creatures” of provincial governments. Similar to the US, the provincial-local government relationship is unitary. But, unlike in the US, local governments in Canada occupy a very weak legal/constitutional position. First, provincial governments, unlike US states, are not bound by written constitutions.9 There are no constitutionally entrenched documents that limit the power of provincial governments over their respective localities. Accordingly, in Canada, there is no home-rule for local governments. Second, in contrast with the US, the parliamentary form of government and the accompanying notion of “parliamentary sovereignty” have historically facilitated the centralization of provincial power over local governments. Third, the development of significant provincial government revenues from a variety of sources allowed centralization of provincial authority over its local governments (see Kim and Law 2010 for a more detailed discussion of these trends).
Finally, unlike in the US, the Canadian federal government does not deal directly with local governments. Whereas in US federal government funds are often distributed directly to local governments for disbursement, the federal government deals directly only with provincial governments, which in turn disburse funds to local governments (usually for very specific purposes). Thus, in Canada, intergovernmental transfers, which are very significant, re-enforce the fiscal authority of provinces over their local governments (Goldberg and Mercer 1986).
III. Institutions and Urban Primacy in the Americas
In this section, we examine the broad correlations between institutions and urban primacy in the Americas. Because political centralization at the national or sub-national (provincial or state) level is more likely to lead to an unbalanced urban or local development as compared to political decentralization at the local-level, we examine urban primacy over time (Galiani and Kim 2010). While there is no single widely accepted definition of urban primacy, there are two general types of measures. The first, motivated by Jefferson (1939), examines the population of the largest city or the n-largest cities as a percentage of either the urban population or the total population. The second, motivated by the rank-size literature, estimates the size distribution of cities using the Pareto distribution.10
Scholars generally believe that urban primacy is a salient character of urban development in Latin America but not in North America. While opinions vary, many believe that urban primacy arose in Latin America in the early nineteenth century or even earlier. Morse (1971), using the share of the population of the largest city as a measure of primacy, finds that urban primacy emerged in Argentina and Cuba around 1800, in Colombia, Mexico, and Peru in 1850, and in Brazil and Venezuela by 1900 (see Table 3). In all of these cases, the primate city was also the national capital. McGreevey (1971), however, using a measure based on the Pareto distribution of city sizes, dates the rise of primacy in Mexico to as early as 1750, Cuba to 1825, Chile to 1830, Argentina to 1850, Brazil to 1880, Peru to 1925, and Venezuela and Colombia to 1950. Portes (1976) argues that, by 1970, most Latin American countries, except for perhaps Brazil and Colombia, exhibited significant urban primacy characteristics.
In Tables 4 and 5, we present estimates of urban primacy based on the share of the urban population accounted for by the largest city and the Pareto coefficient, respectively. The data are based on municipalities with population greater than 25,000.11 In Table 4, we find that the largest city in the US, New York City, accounted for 17.4% of the urban population greater than 25,000 in 1900 but that figure gradually declined to 6.8% by 2000. By this measure, Toronto, Canada, comprising 18% of the Canadian urban population, was much more primate in 2000. In Latin America, the largest cities in Bolivia, Chile, Colombia, Ecuador, Honduras, Nicaragua, Panama, Peru and Uruguay contained a much higher share of their urban population than their counterparts in Anglophone North America. Importantly, in contrast with the US and Canada, the largest cities in each Latin American country, except for Brazil, were national capitals.
We find that the Pareto coefficient for the U.S. rises from 1.06 to 1.27 between 1900 and 2000 (see Table 5), suggesting a shift in the skewness of the size distribution of cities toward smaller-sized cities. While the Pareto coefficient for Canada and the Latin American countries were close to one, suggesting a more balanced distribution of city sizes, the coefficients are relatively much smaller than those of the US. The Latin American countries, however, exhibited much greater primacy when we defined cities using a lower population threshold of 2,500. As expected, the Pareto coefficients for Argentina, Bolivia, Chile, Ecuador, Mexico, Panama, Peru, Uruguay and Venezuela were less than 0.9 suggesting a significantly skewed distribution toward the largest cities in those countries.
The data shown in Tables 4 and 5 generally support the view that urban development in Latin America has been characterized by more primacy than urban development in North America. At any point in time, these measures indicate that the distribution of cities sizes in the US is less primate than the distribution in Canada, which in turn is less primate than the distribution for Latin America; over the twentieth century, the data also indicate that the city-size distributions in the US and Canada have become less primate whereas the pattern is reversed for those in Latin America. Thus, the difference in the level of primacy between Latin America and English-speaking North America seems to have widened over time.
Because we have consistent population data on city sizes and for Canada and the US, we also computed various measures of urban primacy for US and Canadian cities during the twentieth century. A comparison of the patterns of primacy between these two countries is informative because the two countries share much in common (language, British colonial institutions and legal institutions, a common border, etc.) but, as noted earlier, the degree of political centralization of state or provincial government authority over local governments varied between the two countries. Accordingly, if institutions matter, we may observe different trends in the degree of primacy in Canada and the US. Additionally, the richness of our Canadian and US data sets allow us to investigate whether the extent of primacy varies by region and whether there has been convergence or divergence in primacy within Canadian and US regions.
Figure 1 plots the percentage of the total population residing in the largest city and largest metropolitan area in Canada and the US for each decade during the twentieth century. Regardless of which measure of primacy we use, Canada is more primate than the United States. Additionally, while the US measures of primacy have either remained fairly constant or declined over time, there is evidence of an increase in the extent of primacy in Canada, particularly in recent decades. Tables 7 and 8 show estimates of the Pareto exponent for US and Canadian cities with different population cutoffs, and for different portions of the city-size distribution. We report estimates using both log(rank – ½) and log(rank) as the dependent variable.12 While the overall distribution of city sizes in both countries appears to be become less primate over time, regardless of how we slice the data, the degree of primacy is less in the US, and the trend toward less primacy more pronounced for the US than for Canada.
Tables 8 and 9 show the trends in primacy, as measured by the Pareto exponent, for US and Canadian regions. Again, we report estimates using log(rank – ½) and log(rank) as the dependent variable. Several interesting patterns emerge from an analysis of these tables. First, for each US region (with the exception of East South Central), the Pareto coefficient becomes larger over time, indicating greater equality in city sizes over time (i.e. less primacy). Accordingly, the trend away from primacy appears fairly uniform across US regions, which is consistent with the perspective that political decentralization at the local level has contributed to greater equality in city sizes across most US regions. Second, at any point in time, New England has a larger Pareto coefficient (i.e. is less primate) than other regions. This is perhaps a long-lasting consequence of the early diffusion of local political authority in New England within towns and villages.
Very different patterns are apparent from an analysis of the data from Canadian regions. Interestingly, while the estimates of the Pareto distribution for Canada as a whole indicated declining primacy over time, the estimates for most Canadian regions suggest an increase in primacy (i.e. a smaller estimate of the Pareto coefficient) over time. While the city-size distribution became more equal for Quebec, it became less equal for other regions. The experience for most Canadian regions is thus consistent with the idea that political centralization within Canadian provinces has contributed to greater urban primacy in Canada.
In recent years, a number of cross-country studies of urban primacy have appeared in the urban economics literature.13 These papers generally find that urban primacy is not mono-causal but that economic, demographic and geographic factors all seem to contribute to explaining primacy. For example, GDP per capita and total population seem to increase primacy whereas total land area, share of trade in GDP, and transportation density decrease primacy. The most important factor, however, seems to be political as the concentration of government expenditures, unitary political systems, political corruption, and dictatorships all seem to raise primacy significantly. In addition, when capital city dummies are included, the correlation with primacy was always strongly positive.
It is well known that urban primacy in Latin America is caused by disproportionately large national and provincial capital cities relative to their non-capital counterparts (Myers 2002; Portes 1976). By contrast, the national capitals in the U.S. and Canada are not their largest cities and, in the US, state capitals are the largest city in their respective states in only half of all states. While economic forces alone may explain urban primacy, many suspect that the excessively large capital cities in Latin America are caused by political forces such as the centralization of political power, corruption, and rent-seeking. On the other hand, the fact that US state capital cities are relatively small probably reflects political decentralization within states.
In two related papers, Galiani and Kim (2010) and Kim and Law (2010) have investigated the effect of national and sub-national capital city status on city size, using data from Latin America and North America. Controlling for various geographic and economic factors that are likely to affect city population, Galiani and Kim (2010) find that for Latin American countries in 1900, national capital status increased population by 523% relative to non-national capital cities and that this effect increased to 919% by 1990. The corresponding estimates for sub-national (state or provincial) capitals were 70% in 1900 and 232% in 1990. These estimates suggest substantial political centralization within Latin American countries. Kim and Law (2010) estimate similar regressions for Canada and the US. While the impact of capital city status on population for the two North American countries is smaller than for their Latin American counterparts, there is nevertheless considerable evidence that political centralization influences city size. Specifically, Kim and Law (2010) find that Washington, DC’s status as the national capital increased its population by 68% relative to other cities in 1900 and 447% in 2000. In Canada, Ottawa’s capital city status increased its population by 159% in 1920 and the effect increased to 589% by 2000. These trends reflect political centralization at the national level in both countries. Within US states and Canadian provinces, Law and Kim (2010) find a divergence in the effect of sub-national capital city status on city population. While Canadian provincial capital city status increased city population by 48% in 1900 and over 100% in 2000, the US state capital city effect increased only modestly, from 30% to 49%. Law and Kim (2010) interpret these different patterns as evidence of diverging trends in Canadian and US federalism, specifically, the centralization of political power within Canadian provinces, and a continued emphasis on localism within US states.
IV. The Number, Size and Scope of Local Governments: the US versus Canada
Institutions not only affect urban primacy or the size distribution of cities across nations. They also influence the number, size and scope of local governments. Due to data limitations, we focus exclusively on the differences in the number and composition of local governments between US and Canada as well as between US states and Canadian provinces. In the US, the institutional rules that provide local governments with significant autonomy have also led to the proliferation of the number and types of local governments to a bewildering degree. In Canada, by contrast, the centralization of political authority by provincial over their local governments has led to the physical and functional consolidation of cities within metropolitan areas. Thus, the number, size and functions of cities in Canada sharply differ from the US.
Local governments vary by size, function and political hierarchy. There is no widely accepted definition of local government, but, for Anderson (1949) the key characteristics of a “unit of government” is that it has authority over persons within its territory, has a legislative body composed of elected or appointed officials, has a separate legal identity, a degree of autonomy, is empowered by law to perform governmental service, and has the power to raise revenue by taxation, by special assessment, or by other means. By this definition, the generally recognized units of local government in the US are counties, cities, villages, boroughs, incorporated towns, towns, townships, school district and other special districts for schools, parks, sanitation, drainage, and so forth. In Canada, they are the urban incorporated city, town and village, rural district, county or municipality, as well as various agencies, boards and commissions (ABCs) for education, parks, utility, and transportation, among other functions. In general, election districts, judicial administration areas, civil administration districts are not considered to be a unit of local government.
In the US, counties are considered as a major unit of local government because they are the principal or the largest territorial division of state for purposes of local government. Counties were originally designed to serve functions of general- and state-wide interest that the state legislature desired to have performed uniformly throughout the state area. The second category of local units is incorporated places. These are generally urban municipalities but could also include towns, villages or boroughs. As a rule, these units are part of the county or counties in which they are situated. The third class is rural government units, which include towns or townships. Like the second class these units belong to a county. The fourth and final class consists of special districts that are organized for a single special purpose such as schools, drainage, or sanitation.
There are three striking differences in the structure of local governments between US and Canada. First, Canada does not possess a US-style county government system. In Canada, counties are rural units of local government in Nova Scotia and New Brunswick, not much different from other types of rural districts. In Ontario and Quebec, county boundaries are similar to the US in that they encompass other local governments, but unlike those in the US, Canadian counties are not technically governmental units as they do not have direct tax levying and tax collection functions. Rather, county councils allocate amounts to municipalities based on equalized assessments. In other provinces, counties simply do not exist. Second, Canada does not have US-style special purpose districts. While ABC’s in Canada perform specialized functions, they seem to be substantially different from US special purpose districts. Whereas US special purpose districts are locally created and controlled, the creation and control of ABC’s in Canada seems to be in the hands of municipal representatives or the provincial government.14 Given the influence of provincial governments over their operation, it is therefore not obvious whether ABC’s in Canada should be counted as units of “local” government. Third, Canada’s federal government, unlike its US counterpart, is restricted in its interaction with local governments. In the US, starting with Richard Nixon’s general revenue sharing and community programs in 1968, the federal government began to distribute funds to local governments directly with few strings attached. In Canada, on the other hand, federal intergovernmental transfers are primarily sent to provincial governments. Provinces, in turn, distribute funds to local government units, generally for very specific purposes. The nature of Canadian fiscal federalism has therefore increased provincial governments’ authority over local governments whereas, in the US, fiscal federalism has strengthened the authority and autonomy of local governments with respect to state governments (Goldberg and Mercer 1986).
The divergent trends in centralization/decentralization between Canadian provinces and US states is likely to have different implications, not only with respect to urban primacy, but also concerning the autonomy, scope, and number of local government units in Canada and the US. In the US, the continued emphasis on localism within states should result in more autonomous local government units, and also more local governments (normalized by population or land area). One might expect local governments within the US not only to retain significant autonomy, but also to have broader scope of jurisdiction. In Canada, on the other hand, political centralization within Canadian provinces should give rise to less autonomous local government units and also fewer local governments. Greater political centralization may manifest itself in fewer local government units because centralization raises the cost of establishing new units of local government, because uniform standards of regulation or public services are easier to establish when there are fewer local governments, or because economies of scale in administration and public service delivery are more easily achieved when there are fewer local governments.15 Comparing metropolitan areas in Canada and the US, we might also expect greater metropolitan fragmentation (i.e. more local government units per metropolitan area) in the US than in Canada.
While it is difficult to compare quantitatively the differing scope, functions and autonomy possessed by the various types of local government units in Canada and the US, it is possible, at least to some degree, to compare the number of local government units in each country over time. Unfortunately, data limitations prevent us from comparing at the broadest level the number of local government units in the two countries. While for the US, the decennial Census of Government provides a very comprehensive portrait of all types of government within the US (at the local level, the Census of Government provides information on counties, sub-county general purpose local government units, school districts, as well as special districts), we only have data on the number of general-purpose local government units for Canada.16Accordingly, for most of subsequent comparative analysis, we will restrict our attention to sub-county general-purpose local government units (i.e. municipalities) within the two countries, although we will also present some data on US special districts.
Tables 10 through 12 show, for each Canadian province and US state the number of local government units by decade from 1950 to 2000. As noted, Canadian figures include only general-purpose local government units (i.e. municipalities). For the US we report the number of sub-county general-purpose local governments (cities, towns, townships, villages, boroughs) as well as the number of special districts (not including school districts). We ignore counties because counties in the US and Canada are not comparable in terms of form or function. The raw data show growth in the number of local government units in both countries over time, although the growth would seem to be more dramatic for the US than for Canada, and within the US, much of that growth is due to increases in the number of special districts. In some Canadian provinces (Ontario, Quebec and Nova Scotia) the total number of local government units has even declined in the last two or three decades, largely as a result of municipal amalgamations imposed by provincial governments. In contrast, few US states have experienced declines in the number of local government units.
The tables also present information on the “density” of local government units within each state and province. We compute “density” in two ways. The first is by normalizing the number of government units by population (per 1,000 persons) within each state or province. The second is by normalizing by the land area (per 1,000 square miles) of each state or province. We remain agnostic as to which normalization is more appropriate. On the one hand, we might expect the number of local government units to increase with population since economies of scale in local government public service delivery may depend on population size. On the other hand, we might also expect the number of local government units to increase with land area since economies of scale may also be a function of geographic size.
Restricting attention to general-purpose local governments in Canada and the US, the tables reveal several interesting facts about trends in the density of local government units. First, while there is substantial cross-sectional variance in the density of local governments within Canadian provinces and US states (regardless of whether density is measured in terms of population or land area), the variance would seem to be higher for the US than Canada. Second, in both countries, density measured in terms of population has fallen while density in terms of land area has increased. Third, while the number of local government units per 1,000 persons is roughly comparable for Canada and the US as a whole, the number of local governments per 1,000 square miles is significantly higher for the US than for Canada.
Tables 13 and 14 present data on the number of local government units within the largest metropolitan areas in Canada and the US. For Canada we show data on the number of municipalities for each of the largest twenty metropolitan areas from 1960 to 2000. For the US we present data on sub-county general-purpose local government units as well as for the special districts (exclusive of school districts) for the top 50 metropolitan areas in 1997. While the US and Canadian figures exhibit substantial variance, the US numbers are generally larger, indicating greater political fragmentation within metropolitan areas and greater decentralization.
While these patterns are roughly consistent with our hypotheses regarding the effects of political centralization/decentralization on the number of local government units, there are obviously many other geographic and economic factors that may influence the number of local government units within a political jurisdiction. To obtain sharper estimates of the effect of political centralization on the number of local governments within US states (N = 50) and Canadian provinces (N = 9 for 1960 and N = 10 for 1970 onward), we pooled the Canadian and US data to estimate the following regression:
where LGUidenotes the number of sub-county general purpose local government units in state or province i; POPiis the population of state or province i; Areaiis land area (in square miles) of state or province i; Canada is an indicator variable equal to 1 for Canadian provinces and 0 for US states; Xi is a vector or state or province specific control variables; andi is an error term. As control variables we include the latitude and longitude of a province or state’s geographic centre; an indicator equal to 1 if the province or state has access to a major port and 0 otherwise; an indicator equal to 1 if the province or state has civil law legal origins17; and the number of years since a state or province joined the US or Canada to approximate the number of years of settlement within a state or province. We expect the number of local government units within a state or province to be increasing in the population and land area, decreasing with civil law legal origins (it is usually argued that civil law legal origins give rise to more centralizing political institutions), and increasing with years of settlement. Finally, if political centralization within Canadian provinces reduces the number of local government units, the coefficient on the Canada indicator should be negative. We estimate equation (1) separately for each decade from 1960 to 2000.
Coefficient estimates are displayed in Table 15. Each column represents a separate regression. We estimate the regression equation with and without the civil law and years of settlement control variables. We find that the Canada indicator is negative and statistically significant in all regressions from 1970 onward. Additionally, in absolute value, the size of the Canada indicator has also increased over time. These findings suggest that diverging patterns of centralization within Canadian provinces and US states have contributed to differences in the number of local government units. Other things held constant, greater political centralization within Canadian provinces (relative to US states) has resulted in relatively fewer local government units in Canada, an effect that has been increasing over time.
The coefficients of the other variables generally significant and have the expected signs. Having access to a major port has a negative and statistically significant effect on the number of local governments. States and provinces with a major port have fewer local governments, perhaps because ports are natural sources of agglomeration and concentrate economic and political power. In terms of magnitude, the effect of having a port on reducing the number of local governments is smaller, however, than the effect of being in Canada. States and provinces with more people and greater land area also have more local governments. The coefficient on latitude is positive and significant, implying that the number of local government units increases as one moves northward, while the coefficient on longitude is negative and significant, indicating fewer local government units as one moves west. The latter effect reflects denser settlement in eastern states and provinces. Finally, the civil law indicator and years of settlement are not statistically significant.
Pooling data on metropolitan areas for Canada in 2000 (N = 138) and the US in 1997 (N = 314), we can also estimate the effect of divergences in the degree of political centralization within US states and Canadian provinces on metropolitan fragmentation (i.e. the number of sub-county general local government units in each metropolitan area) within Canada and the US. To do this we estimate a regression similar to equation (1). In these regressions, the dependent variable is either (i) the number of sub-county general-purpose local governments within each metropolitan area, or (ii) the number of sub-county general-purpose local governments per 100,000 persons. Population is the number of people residing within a metropolitan area (we omit this variable as a control if we use governments per 100,000 as the dependent variable); land area is the number of square miles of the metropolitan area, and the Canada dummy is an indicator equal to 1 for Canadian metropolitan areas. As before, we predict that the Canada indicator should be negative. Greater political centralization within Canada should result in fewer local government units within Canadian metropolitan areas (i.e. less metropolitan fragmentation), controlling for other factors.
As additional controls, we include the geographic coordinates (latitude and longitude) of each metropolitan area, a port indicator equal to 1 if the metropolitan area includes a major port, and a civil law indicator equal to 1 if the metropolitan area is located in a jurisdiction with civil law legal origins. We also include an indicator equal to 1 if the metropolitan area is located on a major river, and climate controls (average daytime temperature and annual average precipitation). Finally we include an indicator variable that is equal to 1 if the metropolitan area includes a national capital (i.e. Washington, DC and Ottawa) and another indicator equal to 1 if it includes a sub-national (state or provincial) capital city.
Panel A of Table 16 shows the coefficient estimates when the dependent variable is the log of local government units. Panel B shows the results when the dependent variable is the number of local government units per 100,000 persons. Regardless of which dependent variable we use, the coefficient on the Canada indicator is negative and statistically significant. Other things equal, Canadian metropolitan areas have far fewer local government units than their US counterparts. Divergences in the degree of political centralization may have therefore also contributed to a divergence in the extent of metropolitan fragmentation across the two countries. The effect is also economically significant and robust to the inclusion of additional control variables. When the dependent variable is the number of local governments per 100,000 and all controls are included in the regression (see Panel B, column 3), the estimates indicate that Canadian metropolitan areas have eight fewer local government units per 100,000 than US metropolitan areas, which represents over two-thirds of the average number of local government units per capita (the sample mean is between 11 and 12 local governments per 100,000 within Canadian and US metropolitan areas).
Metropolitan areas with more population and larger land area have more local government units. As before we find that the number of local government units increases with latitude, and decreases as one moves west. When the dependent variable is the log number of local government units, we find that metropolitan areas that include a major port have fewer local governments, which is consistent with our earlier findings using data on states and provinces. When the dependent variable is the number of local governments per capita, the coefficient on the major river indicator is negative and highly significant. Metropolitan areas located on major rivers have fewer local government units per capita, perhaps because rivers provide natural political boundaries. Temperature and precipitation do not have robust effects on the number of local government units. Finally, the coefficient estimates suggest that capital city status reduces the number of local government units, most significantly when the dependent variable is the number of local governments per capita. These results are consistent with the idea that concentrations of political power result in less metropolitan fragmentation and fewer local government units within a metropolitan area.
V. Home rule status and the number of local governments within the US
The data on local government reveals substantial variation in the number and density of local government units even within US states and metropolitan areas. In this section we use variation in municipal home rule status—a proxy for the degree of political decentralization—to explore the effect of political decentralization on the number of local government units within US states, as well as the degree of metropolitan fragmentation across US metropolitan areas.
Home-rule grants the citizens an explicit right to local self-government where they can alter, abolish and create new governments of their choice. Without home-rule, the structure of local governments can only be altered through the actions of state legislatures or constitutional amendments. Because home-rule significantly lowers the costs of creating local governments, the diversity and density of local government have increased greatly in states in which home-rule has been adopted (Ostrom, Bish and Ostrom, 1988).
Because there are many dimensions to municipal home rule, there is no single, widely agreed-upon measure of home rule status. Various studies have categorized states according to the type and form of municipal home rule that is present. At one extreme, home rule status could imply significant local government autonomy in terms of form, function, and finances. At the other, home rule status may be met if Dillon’s rule does not apply within a state. For our purposes, we use an intermediate definition of home rule. We classify a state as having home rule status by a given year if the municipalities within the state have some degree of formal and functional home rule. Our information on municipal home rule within each state comes from Krane, Rigos, and Hill (2000).18
To investigate the effect of home rule status on the number of local government units within US states (N = 50), we estimated regressions similar to equation (1) except we omit the Canada indicator and replace it with a municipal home rule status indicator equal to 1 if the state has home rule status by 1950 or by 1970, and 0 otherwise. We estimate the equation separately for each decade from 1960 to 2000. For 1960 through 2000 we use home rule status in 1950 as our measure of political decentralization. For 1980 through 2000 we also use home rule status in 1970 to proxy for political decentralization. Our identification strategy is to use home rule status in earlier years since this provides a more plausibly exogenous source of variation in the extent of political decentralization.
Coefficient estimates are displayed in Table 17. Each column represents a separate regression. The estimates show that for all years from 1980 to 2000, states that had implemented municipal home rule by 1950 or 1970 had more sub-county general-purpose local government units, controlling for other factors. This finding is consistent with the view that the decentralization of political power away from states and toward municipalities increases the number of local government units. As before, we find the number of local governments is increasing in latitude, decreasing in longitude, and increasing in both population and land area. Additionally, there is evidence that states that include a major port have fewer local governments, although the effect is not robust across years. Finally, years of settlement and civil law status do not have consistent, statistically significant effects on the number of local governments.
In Table 17 we report coefficient estimates on the effect of home rule status on metropolitan fragmentation within US metropolitan areas (N = 314) in 1997. In Panel A the dependent variable is the log of the number of general-purpose local governments within each metropolitan area. In Panel B, the dependent variable is the number of local governments per 100,000 persons. These regressions are identical to those estimated using Canadian and US metropolitan except that we replace the Canada indicator with the home rule indicator for either 1950 or 1970. In panel A, the coefficient on the home rule indicator is positive and significant in most regressions. In panel B, the coefficient is positive and significant in the baseline regression, and loses significance when more controls are included. Accordingly there is some evidence that home rule, by decentralizing political power away from state governments and toward local authorities, has contributed to greater metropolitan fragmentation within the US.
The other covariates generally have the predicted signs and are, mostly, statistically significant. Metropolitan fragmentation is increasing in population and land area. As one moves north, fragmentation increases. As one moves west, fragmentation falls. Rivers increase metropolitan fragmentation; the presence of a major port decreases it. Finally, the degree metropolitan fragmentation within US metropolitan areas declines for political (national and state) capitals.
In the Americas, the diversity of European colonial experience and local conditions contributed to the emergence and persistence of a variety of institutional structures. These initial sets of institutions not only shaped the patterns of long-run development in the Americas, but also the structure of their regional and local economies. In much of Latin America, the Spanish and Portugese rule, in conjuction with factor endowments which fostered income inequality, has led to the emergence of non-democratic institutions where political power was centralized at the federal-level at the expense of regional and local levels. In North America, by contrast, British rule and factor endowment conditions which fostered relative equality led to the rise of democratic institutions and a federalist form of government which granted significant autonomy to regional and local governments. Yet, even between US and Canada, their forms of federalism diverged causing significant differences in their patterns of urban and local development.
In this paper, we present a variety of evidence that suggests that differences in the extent of national or regional political centralization, which has strong colonial roots, affects the degree of urban primacy, the size distribution of cities, and the number and mix of various local governments. First, political centralization at the national or regional levels is generally associated urban primacy or a skewed distribution of city sizes toward the largest cities. Second, political centralization is associated with a greater use of general general purpose (cities, counties, towns, etc.) as compared to special purpose (school districts, water, sewage districts, etc.) governments. Third, political centralization limits the number, density and the variety of local governments. Fourth, political centralization generally increases the jurisdictional land area of local governments.
Because it is difficult to estimate the productivity of local governments, it is also challenging to determine the economic impact of institutions on local governments and their national economies. In comparing North America and Latin America, however, there is a general belief that political institutions mattered. In Latin America, political centralization at the national level often went hand in hand with the absence of voting rights, lower levels of investment in public schooling, fewer public goods, as well as weak local governments. In these countries, political centralization contributed to the rise of urban primacy, or unbalanced urban development. While estimates of the static costs of urban primacy might not be large, there are reasons to believe that dynamic costs might be fairly high, especially from the standpoint of (absent) innovations in political institutions and policies. For instance, in North America, gasoline taxes were first tried and implemented in Oregon but quickly spread to the rest of the states, as well as to Canadian provinces. Likewise, the modern income tax was successfully implemented in Wisconsin before it spread elsewhere. Excessive political centralization may impede the ability of local or regional authorities to experiment with new institutions and new public policies.
For the US and Canada, the difference, while significant, is one of degree rather than kind in terms of the level of decentralization between the federal and regional (state or provincial) governments, and especially between regional and local governments. For economists, the general benefits of centralization or consolidation include economies of scale and administration and tax collection, the capture of external economies, and perhaps a more equitable distribution of taxes and public spending. The costs are lack of competition, less “voting with one’s feet,” a weaker matching of preferences for local public goods, and less policy innovation and experimentation. Whether the proliferation of local governments in the US as compared to Canada is seen to increase welfare by facilitating the matching of local preferences to the provision of local public goods (Ostrom, Bish and Ostrom, 1988) or is seen to increase inequality and ultimate reduce welfare (Burns, 1994), it is likely that institutions play a non-neutral role in the long-run development of local and national economies.
Acemoglu, Daron, Simon Johnson and James A. Robinson. 2001. “The Colonial Origins of Comparative Development: An Empirical Investigation,” American Economic Review 91: 1369-1401.
Ades, A.F. and E.L. Glaeser. 1995. “Trade and Circuses: Explaining Urban Giants,” Quarterly Journal of Economics 110: 195-227.
Anderson, William. 1949. The Units of Government in the United States: An Enumeration and Analysis. Chicago: Public Administration Service.
Berkowitz, Daniel and Karen Clay. 2006. “The Effect of Judicial Independence on Courts: Evidence from the American States.” Journal of Legal Studies 35: 399-440.
Bollens, John C. 1957. Special District Governments in the United States. Berkeley: University of California Press.
Brownstone, Meyer and T. J. Plunkett. 1983. Metropolitan Winnipeg: Politics and Reform of Local Government. Berkeley: University of California Press.
Bulmer-Thomas, Victor. 2003. The Economic History of Latin America Since Independence. Cambridge: Cambridge University Press.
Burkholder, Mark A. and Lyman L. Johnson. 2008. Colonial Latin America. 6th Edition. New York: Oxford University Press.
Burns, Nancy. 1994. The Formation of American Local Governments: Private Values in Public Institutions. New York: Oxford University Press.
Burns, Nancy and Gerald Gamm. 1997. “Creatures of the State: State Politics and Local Government, 1871-1921,” Urban Affairs Review 33 (1): 59-96.
Carroll, Glenn R. 1982. “National City-Size Distributions: What Do We Know After 67 Years of Research?” Progress in Human Geography 6: 1-43.
Carroll, Glenn R. and John W. Meyer. 1983. “Capital Cities in the American Urban System: The Impact of State Expansion,” American Journal of Sociology 88 (3): 565-578.
Chase-Dunn, Christopher. 1985. “The Coming of Urban Primacy in Latin America,” Comparative Urban Research 11: 14-31.
Cortés Conde, Roberto. 2008. “Spanish America Colonial Patterns: The Rio de la Plata” Departamento De Economia, Universidad de San Andrés.
Crawford, Kenneth G. 1954. Canadian Municipal Government. Toronto: Toronto University Press.
De Luca, Miguel, Mark P. Jones and Mariá Inés Tula. 2002. “Buenos Aires: The Evolution of Local Governance,” in Capital City Politics in Latin America: Democratization and Empowerment, edited by D. Myers and H. Dietz. Boulder: Lynne Rienner Publishers.
Diaz-Cayeros, Alberto. 2006. Federalism, Fiscal Authority, and Centralization in Latin America. Cambridge: Cambridge University Press.
Eisenstadt, Shmuel N. and A. Shachar. 1987. “Urbanization in Colonial Latin America,” in Society, Culture and Urbanization, edited by S. Eisenstadt and A. Schachar. Newbury Park: Sage Publications.
Elazar, Daniel J. 1972. American Federalism: A View from the States. Second Edition. New York: Harper & Row.
Engerman, Stanley L. and Kenneth L. Sokoloff. 1997. “Factor Endowments, Institutions, and Differential Paths of Growth Among New World Economies: A View From Economic Historians of the United States,” in How Latin America Fell Behind: Essays on the Economic Histories of Brazil and Mexico 1800-1914. Stanford University Press.
Engerman, Stanley L. and Kenneth L. Sokoloff. 2002. “Factor Endowments, Inequality, and Paths of Development Among New World Economies,” Economia 3: 41-102.
Fairlie, John A. 1908. Essays in Municipal Administration. New York: McMillan Company.
Fairlie, John A. and Charles M. Kneier. 1930. County Government and Administration. New York: Century Company.
Frug, Gerald E. 1999. City Making. Princeton: Princeton University Press.
Gabaix, Xavier and Rustam Ibragimov. 2007. “Rank 1 – ½: A Simple Way to Improve the OLS Estimation of Tail Exponents.” NBER Technical Working Paper No. 342.
Gabaix, Xavier and Yannis M. Ioannides. 2004. “The Evolution of City Size Distributions,” in Handbook of Regional and Urban Economics, Volume 4, J. V. Henderson and J-F. Thisse, eds. Amsterdam: North Holland.
Goldberg, Michael A. and John Mercer. 1986. The Myth of the North American City: Continentalism Challenged. Vancouver: University of British Columbia Press.
Hartog, Hendrik. 1983. Public Property and Private Power: The Corporation of the City of New York in American Law, 1730-1870. Ithaca: Cornell University Press.
Henderson, J. Vernon. 2002. “Urbanization in Developing Countries,” World Bank Research Observer, vol. 17, no. 1: 89-112.
Hill, Melvin B. 1978. State Laws Governing Local Government Structure and Administration. Athens: Institute of Government, University of Georgia.
Kim, Sukkoo. 2008. “Urbanization,” New Palgrave Dictionary of Economics, 2nd edition.
Kim, Sukkoo. 2009. “Institutions and U.S. Regional Development: A Study of Massachusetts and Virginia,” Journal of Institutional Economics 5 (2): 181-205.
Kim, Sukkoo and Marc Law. 2010. “Political Institutions, Federalism and Urban Development: Evidence from the US and Canada,” mimeo.
Krane, Dale, Platon N. Rigos and Melvin B. Hill Jr. 2001. Home Rule in America: A Fifty-State Handbook. Washington DC: Congressional Quarterly Inc.
Krugman, Paul and Raul. E. Livas. 1996. “Trade Policy and the Third World Metropolis,” Journal of Development Economics 49: 137-150.
Lattimer, John. 1990. “Intergovernmental Cooperation,” in Illinois Local Government: A Handbook, edited by J. Keane and G. Koch. Carbondale: Southern Illinois University Press.
Lockhart, James and Stuart B. Schwartz. 1983. Early Latin America: A History of Colonial Spanish America and Brazil. Cambridge: Cambridge University Press.
Lyman, Brad. 1992. “Colonial Governance in the Development of Urban Primacy,” Studies in Comparative International Development 27: 24-38.
Mancke, Elizabeth. 2005. The Fault Lines of Empire: Political Differentiation in Massachusetts and Nova Scotia, CA. 1760-1830. New York: Routledge.
McBain, Howard L. 1916. The Law and the Practice of Municipal Home Rule. New York: Columbia University Press.
McGoldrick, Joseph D. 1933. Law and Practice of Municipal Home Rule 1916-1930. New York: Columbia University Press.
McGreevey, William P. 1971. “A Statistical Analysis of Primacy and Lognormality in the Size Distribution of Latin American Cities, 1750-1960,” in Morse, Richard, editor, The Urban Development of Latin America 1750-1920. Stanford: Stanford University.
Miller, Gary J. 1981. Cities by Contract: The Politics of Municipal Incorporation. Cambridge: MIT Press.
Moomaw, Ronald L. and Mohammed A. Alwosabi. 2004. “An Empirical Analysis of Competing Explanations of Urban Primacy Evidence from Asia and the Americas,” Annals of Regional Science 38: 149-171.
Morse, Richard M. 1971. “Latin American Cities in the 19th Century: Approaches and Tentative Generalizations,” in Morse, Richard, editor, The Urban Development of Latin America 1750-1920. Stanford: Stanford University.
Myers, David J. 2002. “The Dynamics of Local Empowerment: An Overview,” in Capital City Politics in Latin America: Democratization and Empowerment, edited by D. Myers and H. Dietz. Boulder: Lynne Rienner Publishers.
Nickson, R. Andrew. 1995. Local Government in Latin America. Boulder: Lynne Rienner Publishers.
North, Douglass C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press.
Ostrom, Vincent, Robert Bish, and Elinor Ostrom. 1988. Local Government in the United States. San Francisco: ICS Press.
Portes, Alejandro. 1976. “The Economy and Ecology of Urban Poverty,” in Urban Latin America, edited by A. Portes and J. Walton. Austin: University of Texas Press.
Rajan, Raghuram G. and Rodney Ramcharan. 2008. “Landed Interest and Financial Development in the United States,” NBER Working Paper #14347.
Rose, Albert. 1972. Governing Metropolitan Toronto: A Social and Policy Analysis, 1953-1971. Berkeley: University of California Press.
Rosen, Kenneth T. and Mitchel Resnick. 1980. “The Size Distribution of Cities: An Examination of the Pareto Law and Primacy,” Journal of Urban Economics 8: 165-86.
Sancton, Andrew. 1985. Governing the Island of Montreal: Language Differences and Metropolitan Politics. Berkeley: University of California Press.
Siavelis, Peter M., Esteban Valenzuela Van Treek, and Giorgio Martelli. 2002. “Santiago: Municipal Decentralization in a Centralized Political System,” in Capital City Politics in Latin America: Democratization and Empowerment, edited by D. Myers and H. Dietz. Boulder: Lynne Rienner Publishers.
Smith, Carol A. 1985. “Class Relations and Urbanization in Guatemala: Toward an Alternative Theory of Urban Primacy,” in Urbanization in the World-Economy, edited by M. Timberlake. New York: Academic Press.
Soo, Kwok Tong. 2005. “Zipf’s Law for Cities: A Cross-Country Investigation,” Regional Science and Urban Economics 35: 239-63.
Sokoloff, Kenneth L. and Eric. M. Zolt. 2006. “Inequality and the Evolution of Institutions of Taxation: Evidence from the Economic History of the Americas,” mimeo.
Sproule-Jones, Mark. 1993. The Pragmatics of Special Purpose Bodies: A Public Choice Analysis of Special Purpose Bodies in Canadian Urban Public Economies,” in Agencies, Boards and Commissions in Canadian Local Government, edited by D. Richmond and D. Seigel. Toronto: The Institute of Public Administration of Canada.
Sublett, Michael D. 2004. Township: Diffusion and Persistence of Grassroots Government in Illinois, 1850-2000. New York: Peter Lang Publishing.
Tremblay, Jean-François. 2006. “Fiscal Federalism and Public Service Provision in Canada,” mimeo.
Wallis, John J. 2000. “American Government Finance in the Long Run: 1790 to 1990,” Journal of Economic Perspectives 14: 61-82.
Wheaton, W. and H. Shishido. 1981. “Urban Concentration, Agglomeration Economies and the Level of Economic Development,” Economic Development and Cultural Change 30: 17-30.
Wiesner, Eduardo. 2003. Fiscal Federalism in Latin America: From Entitlements to Markets. Washington DC: Inter-American Development Bank.
Willis, Eliza, Christopher da C.B. Garman, and Stephen Haggard. 1999. “The Politics of Decentralization in Latin America,” Latin American Research Review 34: 7-56.
Wilkie, James, Eduardo Aleman and Jose G. Ortega. 2000. Statistical Abstract of Latin America. Volume 36. Los Angeles: UCLA Latin American Center Publications.
Winters, John M. 1961. State Constitutional Limitations on Solutions of Metropolitan Area Problems. Anne Arbor: Michigan Legal Publications.
Zimmerman, Joseph F. 1981. Measuring Local Discretionary Authority. Washington DC: U.S. Advisory Commission on Intergovernmental Relations.
Political Centralization in Latin America circa 1995
Country Constitutional Structure Selection of Executive Democratic Formal Overide