1
The Logic of Political Monopolies
DESPITE MANY MEANINGFUL differences between machine and reform cities, they share political patterns. The logic of political monopolies successfully explains these similarities. Threats to political goals inspired the development of monopolies in both types of cities. Politiciansâ desire to ensure reelection undermined the ability for all constituents to be represented in both types of cities. Choices incumbent coalitions made to preserve their power ultimately led to their defeat in both types of cities. This chapter explores these arguments in detail, describes how they apply in different contexts, and makes predictions that will inform the analysis of data in later chapters.
Political scientists argue that elections play a fundamental role in keeping representatives accountable. Reelection-seeking politicians attempt to maximize votes (Downs 1957) and organize institutions to allow them to do so (Mayhew 1974). When political elites, worried about reelection, mobilize large numbers of voters, turnout is high because groups are encouraged to enter the political process (Jackman 1987). Thus, participation in governance is broad, and the government provides benefits to many different groups in the community (Morton 1987; Bueno de Mesquita et al. 2003; Keiser 1997; Uhlaner 1995). Competitive elections provide an incentive for politicians to serve these goals. In short, the quest for reelection is the heart of representative democracy.
But not all elections motivate leaders in this way. Once elected, rational leaders not only have an incentive to win reelection but also have an incentive to reduce uncertainty about their future. In other words, they have an incentive to try to guarantee themselves reelection. Whatâs more is that they may have the power to do so. We can expect that once a coalition wins election to office, it will take steps to increase the probability that it will retain control over government (Cox and Katz 2007). Such safety in office has the potential to affect the quality and effectiveness of governance. Similarly, a substantial literature in economics analyzes the relationship between the safety of incumbent firms and market performance (Posner 1975; Harrington 1984).
We do not find perfect competition in either economic or political markets. However, exploring ideal types allows us to understand the ways in which reality deviates from these models, and can offer insight into interventions that might improve social welfare. Economists have developed a set of conditions that describe perfectly competitive markets in which products are offered for the lowest possible price in the long run. One of the most (some say the most) important condition is that entry into and exit from the market be free and instantaneous for both buyers and sellers. This limits the ability of firms to raise the price of products above marginal cost. Additionally, buyers and sellers must have perfect information about the quality and price of the product. In a seminal book, Baumol, Panzar, and Willig (1982) argue that the threat of entry can be enough to prevent firms from extracting monopoly profits. They find that such markets remain âcontestableâ and that in these cases even firms with large market shares are compelled to keep prices competitive. Thus, the degree to which the market remains contestable determines the degree to which we can expect social welfare to be maximized.
Analogously for political incumbents, the threat of removal may be enough to inspire responsiveness and maximization of social welfare.1 We can draw on this framework for thinking about political markets in a relative way, arranging systems along a spectrum of contestability. Contestability can be described by three indicators of bias: the degree to which constituentsâ information about governmental performance or available alternatives is limited; the degree to which the entry, exit, or behavior of candidates or voters is restricted; and the degree to which incumbentsâ control over government seats is insulated from their vote share.
In the (unrealizable) political market with perfect information, free entry and exit of voters and candidates, and a perfectly responsive and neutral translation of votes to seats, incumbents can only win by being responsive to voters and elites.2 This idealized version of the world represents one end of the contestability spectrum. Institutions or strategies (e.g., bias) that move the political market away from this end increase the probability of reelection for incumbents without regard to the effectiveness of representation or the quality of the candidate. Thus, the opposite end of the spectrum can be described as an uncontestable political market in which reelection of incumbents is wholly disconnected from quality of candidates and representation of voters. These biased systems reduce electoral risk for the incumbent coalition, increasing its probability of retaining power at any given level of performance.
It is important to note that coalitions at both ends of this spectrum are likely to dominate government for multiple successive terms. Hence the difference between monopolies and responsive dominant coalitions can not be empirically detected through an analysis of reelection rates alone. It requires evidence of the presence of bias. Only then is there an incentive for the coalition to concentrate municipal resources on a minority of the population. Like economic monopolies, these coalitions can earn positive profits for the members of the coalition at the expense of the broader public.3 In the absence of bias, dominant coalitions can only remain in power by being responsive to the electorate and elites. Thus, it is bias that sets political monopolies apart from responsive government and leads to the possibility of extortion by political leaders.
All political systems are biased to some degree. They feature barriers to entry and exit by making it difficult for new contestants to enter and by regulating who can vote, they feature voters and candidates with incomplete information, and they demonstrate varying degrees of neutrality and proportionality in the allocation of seats. Yet, some electoral arenas are more biased than others, creating more or less contestable markets. A system in which incumbents routinely raise more money than challengers, while perhaps not free, is closer to competitive than one in which the incumbent jails potential opponents. The latter is an uncontestable political market. The greater the degree of bias in the system, the greater incentive there is for the coalition to respond to a narrowly defined constituency.
However, bias is not sufficient for establishment of monopoly control. Political scientists have determined that gerrymandered districts have reduced competition for congressional seats, biasing the system in favor of incumbents (see, for example, Cain 1985; Cox and Katz 1999). But no monopoly exists so long as no single faction gains control over the entire government. This suggests a second component to the maintenance of a political monopolyâcoordination. Coordination refers to the extent to which individuals and/or groups work together to nominate, elect, and control politicians in public offices in order to influence âgovernment policies, policy stands, projects, graft, appointive government jobs, and other valued things officials. . . . have access to and give outâ (Mayhew 1986, 4). A higher degree of coordination in the system translates to fewer groups competing for power and access to policy making. The coordination spectrum runs from totally atomistic, with each person for himself or herself, to a single dominant coalition managing the political system. A high degree of coordination requires organization in both the electorate and among elites. In other words political monopolies need to assemble a regime.
The concept of an organized regime is a common theme in studies of cities (Shefter 1976; Brown and Halaby 1987; C. Stone 1989; McCaffery 1992). A regime can be thought of as the collaborative arrangements through which the local governments and private actors assemble the capacity to govern. We can think of the organizational components for a regime as Stone does, as composed of the electoral coalition and the governing coalition.
The individuals (or groups) included in the electoral coalition might be determined by a range of factors including historical circumstances; economic considerations; race, class, or ethnic divisions; and compatibility with other membersâ preferences (Axelrod 1970; but cf. Wright and Goldberg 1985). The highest degree of coordination would mean that all individuals in the electorate belong to the same coalition.
The governing coalition is composed of a set of elites who control the resources necessary for maintaining power. Some elites control resources by virtue of being elected to office. Monopolies need power over a sufficient number of office holders in both the executive and legislative branches to make policy, direct the distribution of benefits, and manage the appointment of government officials. Controlling nominations and maintaining cohesiveness of elected officials are crucial components to preserving this level of coordination. The highest degree of coordination would mean that all elective office holders in the executive and legislative branches belong to the same organization. Other elites control resources exogenous to the cityâs political system but that are nonetheless necessary to make and implement governance decisions. These may be private actors or, in a federal system, public officials at higher levels of government. Private actors controlling important resources for governance might include business elites, labor leaders, religious leaders, or nonprofit organizations. Collectively, these elites make up the governing coalitionâthose people whose agreement is required for a change of the status quo or the implementation of policy. According to regime theorists, without the cooperation of these actors, either tacit or overt, elected officials remain severely constrained in their ability to govern (C. Stone 1989; Hunter 1953, 1980; Keiser 1997). For this reason, simply winning election to office, even winning a majority of the city council and the mayoralty, does not necessarily translate into control over city policymaking, and so does not equate to monopoly.
Thus, we can categorize political systems along two continuous dimensions: the degree of coordination and the degree of bias. A high degree of coordination and bias are jointly sufficient conditions for a single coalition to dominate government for multiple terms. Figure 1.1 shows how different systems might be categorized according to these concepts.
Fig. 1.1: Categorization of political systems
As figure 1.1 indicates, monopolies and responsive regimes are both examples of single dominant coalitions, but only the former are characterized by strongly biased systems. The degree of bias in a system determines to how large a proportion of the population a coalition must be responsive in order to win reelection. The degree of coordination in the system determines how effective winners will be in implementing their preferred policies.
In arenas where multiple groups compete for power, the degree of bias in the system reflects the difference between âresponsibleâ party government praised by 1950s political scientists and factional politics.4 In both of these situations, divided government or turnover in office means that no one group dominates; the political market remains contestable. The term factional politics refers to systems like some of the states in V. O. Keyâs one-party South, where suffrage restrictions eviscerated the Republican Party, but a lack of organization limited the dominance of any one faction of the Democratic Party (though dominance by whites was ensured). This category also includes factional machine-style politics, in which patronage is used to organize voters and suppress opposition, but no single city-wide machine controls government. According to Brown and Halaby (1987) this was the most common form of machine-style politics to have existed in the United States. Politicians, working largely independently, engaged tactics like vote fraud or bribery, but failed to combine their efforts into a single organization that could dominate government. Powerful figures like Bostonâs James Michael Curly or New Yorkâs Robert Moses also represent this type of system in which one or a small number of leaders dominate areas of the city or portions of the city government. Similarly a powerful but uncoordinated bureaucracy as in historical Los Angeles (Erie 2004) might also lead to a factionalized system in which no centralized organization controls city policy. Finally, friends-and-neighbor politics is the kind of political system that emerges when organizations are ephemeral; no stable coalitions exist. Friends-and-neighbor systems may be characterized by bias, advantaging some political actors over others, but they need not be.
Given that machine and reform monopolies appear in only about 30 percent of cities, it is these other patterns of coordination and competition that define the remaining 70 percent. Although this project can not speak directly to these other forms, it is likely that the factional form is most pervasive. Gilbert and Clague (1962) find that not a single city in their twenty-four-city analysis could be described as having a truly organized and competitive system. Because a fractured political system, either one defined by factionalism or friends-and-neighbor politics, is characterized by a lack of control over the entire city or the entire city government, it is highly likely that voters are not provided responsive government in these scenarios either. In fact gridlock might be the overriding feature of policy making. It is also plausible that in cities where a politician or coalition has captured a geographically or functionally defined portion of the city government a similar pattern of narrowing the constituency will occur within that realm. Thus, a responsive government is not implied by the lack of monopoly, but it might be. This suggests that maintaining competition that even begins to approach the idealized structure discussed above is likely to be quite rare. However, only in the case of monopolization, where the dominant coalition achieves success through bias, should we see the governing coalition have the opportunity and incentive to be responsive to a narrow portion of the electorate at the expense of the rest of the community on a citywide scale. While fractured systems might not be responsive to the broadest electorate, they are unlikely to have the extensive, negative consequences ascribed here to political monopolies.
Knowing that there are different ways that we might categorize political systems is helpful, but falls short of delineating where along the coordination and bias continua a city needs to fall in order to qualify as a monopoly. Theoretically we should be able to distinguish between levels of bias and degrees of coordination, and to use both to predict the presence or absence of monopoly. To carry out such an analysis we would need precise, continuous measures of bias and coordination as well as the degree of monopolization. While this would clearly be the ideal approach, a lack of specificity in the measures available leads me to take a less elegant tack. First, I use electoral dominance of a single coalition as an indicator of monopoly and show that at high levels of bias and coordination, the incumbentsâ reelection prospects improve. Then, to distinguish between monopolies and responsive regimes, I analyze changes in participation and policy outputs. If it were true that a coalition in power for a decade represented a responsive regime, declines in participation would likely be driven by those most content with the regime and we should see no systematic skew in policy output. Instead I show that where bias and coordination were present, groups peripheral to the monopoly were most likely to cease participating and government benefits to these nonmembers declined relative to expenditures in cities that lacked either a dominant coalition or severe bias. Thus monopolies are identified here as having a high degree of coordination and bias, which increases the incumbent coalitionâs probability of maintaining power, and leads to suboptimal government responsiveness. Where access to the political process was limited for some groups by bias, the result was reduced responsiveness.
This practice of identifying monopolies in part by their tactics and in part by their effects has precedent in both economics and politicalscience literatures. The legal determination of a monopoly firm takes into consideration the firmâs market share and its ability to set profitable prices due to barriers to entry. Similarly, urban-politics scholars have defined political machines by their electoral success, their high degree of organization, and reliance on a strategy of rewarding supporters with material benefits in order to secure their loyalty. Erie (1988) argues that in mature machines âpower was centralized in a single party boss,â they âexhibited staying power, winning several municipal elections and remaining in power for at least a decade,â and âtrafficked primarily (but not exclusively) in divisible material benefits . . .â (19). Brown and Halaby (1987) specify âthat to qualify as a dominant machine a political organization had to have controlled the election of the municipal executive and a majority of the legislative branch for an uninterrupted series of three elections,â have been âsustained by patronage and favors,â be âheaded by a boss,â and âbased on identifiable grass roots organizationâ (597). In sum, these authors use bias, coordination, and dominance to define a machine. Instead I show that bias and coordination lead to dominance and then, like Erie and Brown and Halaby, use the combined presence of these factors to signify monopolies. For the remainder of this book, the term monopoly will refer to only those dominant coalitions that achieve success through bias.
No doubt some readers will object to the hyperbole encapsulated in the term monopoly. No mayor or manager in this analysis ever governed as a dictator. Incumbents always faced challengers and voters continued to have the right to select their leaders. Violence was never used on a regular basis or a large scale to control political outcomes. Furthermore, no coalition included in this study was able to rely on biased reelection strategies alone. To the extent that monopoly means controlling the market without competition, it is indeed an overstatement in the political context. Like a competitive market, a political monopoly is an ideal type that rarely exists in the real world. For this reason it might be more appropriate to think of the coalitions under study here as relatively more like monopolists than compet...