Abstract
This article argues that American federalism led both to a greater national role in rail promotion and more centralized railroads in the antebellum period. Local competition among states led Congressional representatives from state unable to build local railroads to turn to federal assistance. Early support for railroads came from representatives in the South and frontier, who were primarily drawn into rail coalitions because of their own inability to build local rail networks. However, over time, competition among states within the coalition as well as concerns about federal power led many initial members of the coalition to drop out. In their place, states that favored a stronger federal state stepped into the coalition and subsequently built a more nationally oriented rail system. This analysis argues that the shifting of policies from local control to national oversight due to local resource shortages is an important aspect of American states building.
Through the 19th and 20th centuries, the federal government gained responsibility over a range of new policy areas, each of which greatly expanded the national state’s power to control its territory. Significantly, this growth in national power was not solely the result new state powers being invented from whole cloth. Rather, the expansion of national power consistently resulted from shifts “in governing authority” (Orren & Skowronek, 2004, p. 123). Specifically, the 19th-century state building resulted from changes in American federalism, with local issues shifting upward to become congressional responsibilities that increasingly served national ends. 1
A critical example of 19th-century American state building resulting from local issues shifting arenas to the national level is infrastructure development, and, in particular, early railroad construction. Infrastructure, the material and interpersonal resources required for communication as well as the transport of goods, people, and ideas, is a critical component of state building (Mann, 1986). Roads, waterways, and railroads provide important links within a state, tying together political institutions, sites of economic production, and citizens. Without effective transportation, a state cannot enforce its own policies, distribute goods, regulate its economy, or defend its borders (Harvey, 2000). In the end, a state lacking infrastructure cannot develop the autonomy to pursue its own goals or compete on the international stage (Evans, Rueschemeyer, & Skocpol, 1985). In effect, a state cannot be a state without a strong infrastructural backbone. More specifically, a strong state requires not only national control over infrastructure but also a transportation system that serves the central state’s interests by effectively connecting key political and economic sites within a state’s territory. However, despite infrastructure’s salient role in governance, the antebellum American state was almost entirely inactive when it came to internal improvements. Although there was interest in a greater national role in infrastructure construction, federalism’s checks and balances system stymied most national action on infrastructure during the pre–Civil War period.
However, the national abstention on infrastructure policy changed in 1850, with a series of land grants to state governments. The grants turned federal public lands over to state governments which were then free to sell or lease the land as a mechanism to fund rail construction (Stover, 1975). These 1850 land grants were the first step in building a truly centralizing infrastructure system. Yet, the grants themselves were perplexing for two critical reasons. First, the timing of the federal land grants is puzzling. In the wake of Jacksonian Democracy, the Federalist and Whig parties, which had both favored a more active national state, were soundly defeated. Furthermore, through a combination of state governments assisting railroad projects and private capital, the continental rail system was being constructed through local action. Thus, congressional intervention was surprising in the absence of a national emergency and in the presence of robust local actions. Second, the earliest land grants reflected particularistic and local concerns rather than serving national connectivity. However, with time, the grants become increasingly focused on enhancing political centralization. 2 Given this context, the question arises, “Why did Congress not only intervene in the American rail system in 1850, but how did Congress come to specifically support railroads that furthered American political centralization?” 3
The development of a national infrastructure system that enhanced political concerns, over local concerns, occurred because of two inherent features of American federalism: local competition and debates over the scope of the federal state. Local competition resulted in state-level issues rising to the national state, while struggles over federal power furthered the centralizing nature of these formerly local policies once they transitioned to congressional oversight. The 1850 federal action on railroad development began as a result of competition among states, which drove states that were unable to launch successful rail programs to seek federal intervention on internal improvements. Once at the national level, rail policies went through two phases of bargaining, which ultimately resulted in a national infrastructure system. In the first stage, representatives from states seeking federal aid on railroads constructed coalitions that balanced local and national goals through forming partnerships with Southern members, thereby minimizing the centralizing nature of these early grants. In the second period of greater national focus, the rail coalition broke down as members from states that received aid along with members concerned about growing federal power exited the coalition. Then, members seeking rail aid were increasingly forced to turn to members favoring greater national authority to secure land grants. This change in the rail coalition produced railroads that were increasingly attuned to national, rather than local, needs. As a result of these federal conflicts, the American rail system shifted from being a regional system into a more continental design.
The Federal Government and Transportation Infrastructure Before 1850
Given the limited character of the antebellum American state, the congressional federal land grants provided to the states through the 1850s were an unexpected interjection into the economy. Of course, the national state was by no means powerless or wholly inactive during the early 19th century. The federal state controlled the military and possessed considerable taxation authority (Edling, 2008). The national state also established and maintained hospitals for merchant marines who provided critical commercial services in the 1800s (Rao, 2008). Quieter federal actions, such as the Louisiana Purchase, also underscore the reach of the early American state (Balogh, 2009). However, despite these activities, the federal state’s scope during this period was still truncated. The 19th century was an era of dual sovereignty federalism, with states and the national government often operating in largely distinct policy spheres (Conlan, 1998; Elazar, 1962; Nagel, 2001; Peterson, 1995; Peterson, Rabe, & Wong, 1986; Walker, 2000). Compared with the modern state, in an era of cooperative federalism, the antebellum national state’s reach was shorter and its policy responsibilities fewer. This constrained reach of national power certainly applied to transportation infrastructure construction.
In terms of internal improvements, demand for national action on infrastructure programs was a common political cry in the antebellum period. As early as 1807, Congress drafted a resolution that charged Treasury Secretary Albert Gallatin with developing a plan for a national transportation infrastructure system. Gallatin set about identifying all of the internal improvement programs currently being considered in the early 19th century United States. After identifying these key routes, Gallatin evaluated the routes based on technical difficulty, overall value to the nation as a whole, and political conflict over regional advantage. In the end, Gallatin produced a volume that recommended national action on a key series of roads and waterways as a means to enhance national connectivity (Larson, 2001). Ultimately, Gallatin’s survey aimed to secure congressional funding for an extensive national internal improvements plan. While his recommendations were sound, Gallatin’s plan ultimately fell victim to regional conflict and partisan politics in Congress.
After the failure of the Gallatin Plan, the Bonus Bill of 1817 attempted to use profits from the Second Bank of the United States to fund national transportation projects. Though the bill nimbly attempted to balance regional conflicts and Constitutional concerns, it was vetoed by James Madison and never successfully revived (Larson, 2001). Despite demands on the national state from both citizens and political leaders to enhance the national roads and waterways, efforts at building a national transportation infrastructure network were repeatedly defeated by regional conflicts.
Political conflicts that impeded federal action on internal improvement revolved around two issues. First, congressional members were reluctant to repurpose federal funds for local construction efforts, afraid of the advantage it would give to other states regardless of the national benefit it could provide (Larson, 2001; Wallis & Weingast, 2005). Second, there were ongoing legal concerns that federal action on internal improvements was unconstitutional. It was unclear whether Congress possessed the authority to build roads or fund infrastructure projects. The lack of any unified agreement on what a national transportation infrastructure system should be in the antebellum period, including inadequate support from a military that was originally wary of railroads’ value, doomed national action on railroads (Angevine, 2004). Thus, federal action, even in desperately needed areas like transportation infrastructure, was stalled prior to 1850 as a result of conflicting regional agendas.
This regional and partisan conflict frustrated most attempts at national infrastructure building, outside of a handful of projects. For instance, the federal government was involved in dredging rivers and maintaining harbors (Hill, 1957; Maass, 1951). In addition, numerous postal roads were funded to facilitate mail delivery (John, 1995). There were a handful of sporadic grants for canals and roads during the antebellum era, but most of those projects were eliminated during Andrew Jackson’s presidency (Larson, 2001; Wallis & Weingast, 2005). 4 Many of these early efforts were evenly spread across the nation, with benefits shared by most if not all congressional districts. The small-scale and widespread nature of these early infrastructure projects’ benefits was intentional. Congressional support could only be wrangled when infrastructure boons were highly inclusive and incorporated most districts in a log-roll. Although valuable to some degree, these smaller infrastructure actions failed to cohere into anything even approaching a national infrastructure system (Wallis & Weingast, 2005).
In addition, while these ongoing political debates stifled federal action on infrastructure, there was also no singular crisis that accounts for why the national state finally took action on infrastructure in 1850. The period following the Nullification Crisis, when South Carolina attempted to void the Tariff of 1828, and building up to the Civil War was tumultuous (Cooper & Terrill, 1990). However, there was no major war, economic depression, or technological shift that would suddenly prompt national action on rail development (Skowronek, 1982). In fact, the antebellum land grants occurred a full decade before the Civil War and in the early days of American industrialization, prior to the exponentially greater social complexity of the late-19th-century economy (Skowronek, 1982). In 1850, there was neither a preestablished record of federal action on transportation infrastructure nor an outside event that encouraged federal action on railroads. 5 Given this context, the question of how a properly national rail system, which directly connects distant sites to centers of economic and political power, arose persists.
Federal Railroad Aid and Intergovernmental Relationships
To some degree, the federal government avoided risk when it supported Western railroads by supporting those routes through land grants. Land grants are an example of taxless finance, where a currently held resource is distributed to support a project without raising taxes. The goal of land grants was to turn over land with little value and increase its value through development. Given its low risk, support for this type of internal improvement support was easier to muster in Congress (Wallis & Weingast, 2005). However, the land grant process still remained controversial, as states without public lands preferred to sell federal lands to generate revenue (MacGill, 1917). Hence, a series of critical questions remain in regard to the antebellum land grants, notably, “Why did Congress become involved in rail promotion and how did that congressional action increasingly contribute to political centralization?” Picking up these questions, an analysis of how congressional land grants for railroads both emerged and evolved reveals the role of federalism in American state building.
American federalism, despite appearing stable, is actually always reconfiguring itself into newer, more centralized forms due to its inherent tensions. Therefore, the replacement of classical dual sovereignty federalism with modern cooperative federalism is not the result of a new political ideology or a crisis but rather an accumulation of many smaller political choices that are part and parcel of American political institutions. The first step in the inherent pressures of federalism expanding federal power, and thereby producing national state building, is state-level policy shortcomings increasing the number of policy areas for which the federal state is responsible. In practice, increased federal authority over new policy areas results from subnational governments abdicating control of local policies after their own failures and shifting responsibility to national institutions. The migration of responsibility for specific policy issues among levels of government is a common occurrence in federal systems (Schattschneider, 1960). Especially in a democratic context, elected officials have an interest in both shirking costs onto other levels of government as well as in expanding their own sphere of influence as a means to appease constituents (Bednar, 2004; Gerber & Kollman, 2004; McDonagh, 1992; Rodden, 2006). In fact, local governments lobbying the national state for funds is a common feature of contemporary federalism (Haider, 1974). Furthermore, federalism, with its numerous sites of governance providing multiple opportunities for elected officials to influence policies, was an exceptionally strong force in shaping American railroads in particular (Dunlavy, 1994). Given early railroads’ significance to the American economy and military might, combined with the openness of American federalism, it is not surprising that the rail issues very quickly came to be directly addressed by Congress.
The national state’s involvement with railroad development begins with state politics and regional economic rivalries. Although a great deal of antebellum rail development was driven by private capital, and centered on conflicts between urban centers like St. Louis and Chicago, state governments were nonetheless significant actors in antebellum railroad development (Belcher, 1947; Cronon, 1992). In addition, although corporate managers were also instrumental in shaping rail coordination and policy, these financial interests did not dominate rail planning, especially in the antebellum era (Chandler, 2003; Kolko, 1965). Instead of being directed entirely by private interests, during the antebellum years, many railroads emerged due to support from state governments for local rail development. State actions supporting railroads varied greatly and ranged from using public funds to build railroads to providing private rail companies with banking rights to assist them in raising capital (Goodrich, 1960; Larson, 2001; MacGill, 1917; Pisani, 1987; Rubin, 1961; Scheiber, 1975; Taylor, 1951; Usselman, 2002). Furthermore, all railroad charters were approved by state legislatures, granting state governments significant influence over the route and shipping rates of private rail enterprises (Burgess & Kennedy, 1949). Therefore, state governments were an important, but more rarely considered, element of antebellum American railroad construction.
Whether merely approving railroad routes or actively promoting local railroads through government intervention, state legislatures used rail development as an engine for local economic growth. Economic development is especially salient for local policy makers. At the state and local level, governments possess little control over their borders, allowing for the free flow of capital and population across state lines. This free movement of people and capital leads states to constantly pursue economic growth as a means to generate revenue, serve citizens’ demands, and guarantee incumbents’ reelection (Hwang & Gray, 1991; Peterson, 1981; Pisani, 1987). By the mid-19th century, the economic necessity of railroads was becoming clear: railroads connected markets as well as producers and were a vital part of economic development. 6 Intense local competition among states resulted in legislatures rushing into rail promotion programs and routing local railroads to frustrate regional rivals (Rubin, 1961; Scheiber, 1975). Citizens looking to travel, conduct business, be connected to political events in Washington or feel assured about the availability of security all placed demands on state governments for improved railroads.
However, not all states were in an equal position to meet those requests. Frontier states, notably those west of Ohio in the mid-19th century, had an especially difficult time in developing local railroads. 7 Financial support for railroads did not have to arise solely from local sources; private capitalists from other states often invested in nonlocal railroads (Thies, 2002). By doing so, capitalists gained access to new markets and materials otherwise not available. Despite the role of private capital in promoting much of the American rail system, there were simply too many barriers confounding frontier states’ ability to construct railroads. Frontier states were limited by their own shortcomings in bureaucratic organization, state revenues, locally available private capital, or even by difficult terrain that vexed the technical skills of local engineers. In addition, Western states were often large and spatially isolated. However, if a state could not meet the demands for improved rail infrastructure, businesses would relocate, new settlers would choose a different homestead, and voters would punish their elected officials. When state agencies were unable to provide sufficient railroad promotion, members of the state legislature would redirect citizens’ demands from local political institutions upward to the state’s congressional delegation to lobby for federal assistance.
Shifting policy responsibility from the local level to the federal government increases political centralization, as it expands the power and authority of the political core. However, centralization itself exists on a continuum, with more and less centralization being possible. At the same time that American federalism begins political centralization through pushing state-level issues onto the national scene, competitive pressures and ideological tensions within federalism further enhance that political centralization over time. This centralization passes through two phases, best characterized by initial coalitions on the issue and, later, stabilized coalitions. In the first stage of seeking aid, members pushing for federal intervention aim to develop a coalition which primarily serves local needs with as minimal compromise as possible (Berk, 1994; Mueller, 2003; Usselman, 2002). However, not surprisingly, achieving support for localist networks was difficult in Congress: members from other states perceived little advantage in supporting a railroad that offered advantages to only one state.
To obtain federal support for local rail projects, states seeking aid needed to expand their coalition. However, while seeking adequate votes to pass their legislation, members pursuing federal intervention will still strive to maximize their own benefits while offering as few incentives as possible to other members from competitor states (Krehbiel, 1998; Lee, 2000). The most direct mechanism for expanding a railroad coalition was to increase railroad aid by incorporating projects from other states into early rail grant bills. During this process, as much as possible, members seeking congressional support will partner with members that also share a preference for localist railroads. Essentially, members from states lacking capital and resources to support indigenous rail construction form a broad coalition to spread benefits only to their own states, placing special emphasis on local concerns in their rail plans. In the case of antebellum railroads, this process plays out with frontier states pushing rail aid into Congress and turning to members elected from similarly less developed states in the West and South as coalition partners.
While there was a strong preference in these initial rail coalitions for localist roads, the pressures of coalition politics still ensures that federal rail aid produced roads that balanced local interests alongside regional as well as national concerns. As a result, early federal aid still produced centralizing rail networks. Centralizing railroads meandered less than more parochial routes, tying together only the economically most important parts of the country. 8 In practice, coalition-prompted compromises on rail planning lead local rail plans toward greater connectivity, including rail plans that reach across state lines and connect into regional thoroughfares. Thus, through the realities of coalition politics, local rail concerns that were lifted to Congress began the incremental process of building a more developed, robust federal state.
However, while still shifting more toward national centralization than the local interests which states preferred, the initial rail grants still generally reflected parochial concerns. Thus, early railroads often detoured from direct ties between major sites of economic importance or defense to connect cities that were of local or regional, but not necessarily national, significance. However, with time, rail aid became increasingly geared toward political centralization. Centralization increased due to the unstable nature of early rail coalitions, which broke down and later reconfigured into a form that primarily favored greater political centralization. Initial coalitions that pull the federal government into new policy areas are susceptible to two pressures, both of which increase future centralization. First, local competition pressures members from states that received aid and were interested in localist policies to slowly exit the coalition. Once a state has received its own aid, there is an interest in that state’s members refusing aid to future states also seeking assistance. In seeking to maximize their own advantage, representatives who obtained grants for their state would not wish to offer support to potential competitors. Thus, we should anticipate that grant-recipient states should be fickle coalition partners, prone to not supporting future states’ attempts to secure their own railroad grants.
The second force upsetting the initial aid coalition, and contributing to greater political centralization, is some members’ concern over the nature of American federalism. When the federal government begins to act in a new policy sphere, such as railroads, the relationship between state and federal governments is altered. At the most basic level, the federal state expands its power when it increases its policy responsibility. However, the process of expanding federal power is not as smooth as the national state merely staking out new policy territory. There is an antagonistic relationship between state and federal powers, as states aim to maintain as much local authority as possible while federal legislators often feel an electoral pressure to involve themselves in major issues to please constituents. At a deeper level, there can also be profound concerns about self-governance and liberty that lead some representatives to pause when faced with greater federal authority (Conlan, 1998; Nagel, 2001). As a result of this long-standing conflict in American politics, representatives who seek federal action on a new policy area must overcome this barrier by partnering with representatives who favor political centralization and a stronger national state. At the same time, members of the initial coalition who are worried about growing federal power will leave the coalition as the federal state continues to exert ongoing influence in this new policy area. 9 That would lead to the hypothesis that Southern states, who were wary of growing federal power as it related to the issue of slavery in the antebellum United States, would slowly resign from the railroad coalition.
However, as members with concerns about federal power withdraw from the coalition, members seeking aid still find themselves under local electoral pressure, which itself stems back to interstate competition, to obtain federal assistance to supplement resource shortages in their home states. To compensate for losing support of both grant recipient members as well as representatives wary over federal power, members still needing grants for their states must turn to other members who are keenly interested in increasing political centralization for support. In terms of antebellum railroads, this suggests that later grants should be increasingly centralized, directly connecting major cities and rarely swerving to connect more parochial interests.
Method
To analyze how local railroad development projects became a national issue, a multimethodological approach is used. First, a statistical analysis is performed to provide a broad overview of the congressional legislative process at work. The statistical analysis examines what factors, notably region represented and state’s grant status, influenced congressional members’ voting on railroad grants. Alongside the statistical analysis, a historical comparative analysis focusing on when local rail policies were moved to the national stage and how states were transformed by that transition is conducted as a means of providing context to the statistical results.
The statistical analysis is composed of two discrete components. In the first set of analyses, bivariate regressions display notable relationships between congressional voting on railroads and key independent variables. In the second set of analyses, multivariate analyses with controls are used. Given the range of social forces shaping legislative decision making, inclusion of these controls is necessary to the analysis. Through these controls, the influence of key explanatory variables on congressional voting is more accurately ascertained.
More broadly, in the statistical model, both the characteristics of member i and his home state at time t are used to predict his probability of voting in favor of congressional railroad aid. The dependent variable is binary and is a measure of congressional votes on railroad aid packages for between the years 1838 and 1863 (Poole & Rosenthal, 1997). Railroad votes were identified by searching each Congress’s codebook, as constructed by Keith Poole, for bills that provided railroad support. In some cases, vote codes were reversed, to label all supporting votes consistently. Specifically, votes against tabling railroad grants were recoded positively as supporting railroads. The time period analyzed captures the breadth of antebellum railroad policies, beginning with the first attempts to legislate railroad aid, and ending with the passage of the transcontinental railroad bill. As the focus is on the evolution of the railroad coalition across multiple Congresses, only final bill roll call votes were included and amending votes were not incorporated into the data set. Finally, voting on all rail proposals in a congressional session, rather than specific proposals, is conducted to illuminate broad patterns in voting.
For the analysis, separate models for both the House and Senate were estimated. As the analysis’s emphasis is on how local economic conditions within states affected members votes on railroad aid, Senate votes are the more obvious data to analyze. However, Senate votes provide less insight as there are fewer senators, fewer votes on railroad aid in the Senate, and senators tended to vote in favor of rail packages when compared with House members (Riegel, 1926). Thus, Senate votes alone offer less insight than initially anticipated on analyses of how congressional members used federal railroad aid to further the interest of their home state.
Therefore, although a model for Senate votes is also estimated, the primary focus is how House members voted on railroad legislation. House members’ votes are also an imperfect measure of state-level railroad preferences, as members represent districts and not entire states. However, state delegations regularly share information among themselves and often vote together (Deckard, 1972). In that light, and given both the greater number of House members and railroad packages voted on by those members, analyzing House voting patterns provides more traction on the questions at hand.
The independent variables were developed from both historical census data, as well as Keith Poole’s congressional roll call data set (Haines, 2005; Poole, 2007). Using the historical census data, values for each state’s population, manufacturing capital, and agricultural output were collected from every decennial year from 1830 through 1860. Then, values for each discrete year were interpolated by measuring the rate of change between decades. Finally, representatives were assigned their state’s census characteristics for every session they served. The assigned census values were always the state-level totals and do not reflect district-level features. In addition, as congressional sessions span multiple years, members were linked with their state’s interpolated economic and demographic information from the median year of each session.
In the analysis, there are two key independent variables. As my theory argues that regional competition among states shaped rail grants, the first independent variable of interest is whether a congressional member was from the North or the South. The region measure is a binary measure, with Northern states marked as 1 and Southern states as 0. 10 The second independent variable of interest is a binary measure of whether a state had received a railroad grant in a previous congressional session. For the grant variable, members from states which previously received a grant were coded as one. By including the grant measure, I can track whether members from states that received grants continued to support future rail aid for other states or if members from grant recipient states dropped out of the rail aid coalition.
In addition to the primary independent variables, several controls were also included in the analysis. First, a measure of total population was included. In addition, manufacturing capital and agricultural output measures for each state were also included in the multivariate models. Given the direct role of railroads in the local economy, controlling for the economic context of each member’s state is crucial. 11 Prior to analysis, all demographic measures were divided by their mean to produce more sensible coefficients. Furthermore, a measure of party was included. Republicans, Oppositions, and Whigs were coded as 1, whereas all other parties (notably Democrats) were coded as 0.
Finally, in the bivariate analyses, the impact of each state’s year of entry in the Union on its vote on railroad land grants is analyzed. The year measure offers insight into how frontier states, versus colonial states, approached the rail issue. The year variable, however, is not included in the multivariate analyses since year of entry in the Union covaries with numerous other measures, such as state population or industrialization.
For the regression analyses, a conditional logistic regression was used in all cases, with the data pooled by each individual floor vote. 12 The conditional logistic regression was selected because the data are binary but aggregated by vote. In addition, the statistical analyses are broken into discrete temporal groupings. The first model incorporates the entire set of railroad votes between the years 1837 and 1863. However, a single model that combines all of the period’s railroad votes fails to address the issue of how rail aid shifted from being localist to nationalist as well as the changing roles of members from both manufacturing states and grant recipient states in that process. To address this issue, a series of four additional conditional logistic regressions was performed. Each of these additional models examines a smaller selection of the entire data set and represents a discrete period of time. Each model used the same data set as the overall model but isolated the congressional sessions of the specific land grants selected for case study. The supplemental models estimated congressional voting prior to the first grant to Illinois as well as the sessions where railroad grants for Illinois (31st Congress), Missouri (32nd), and Iowa (34th) were passed. 13
In addition to the statistical models, historical case studies were incorporated into the analysis. Qualitative institutional analysis relies on the historical record and controlled comparison as another means of examining the impact of independent variables on the dependent variables. By limiting the study to a handful of states with similar economies and geographies during the same period, numerous confounding factors are effectively controlled. At the same time, through secondary historical sources, a more comprehensive picture of both how local policies were lifted to the federal level as well as the impact of shifting coalition politics on railroad packages is ascertained (Amenta, 2003; Pierson, 2004).
Three railroad aid packages are specifically singled out for more in-depth historical analysis. The historical case studies focus on land grants in Illinois, Iowa, and Missouri. The Illinois and Missouri cases are ideal, as Illinois was the first federal land grant in 1850 and the Missouri grant in 1853 came shortly after Illinois’s success. The Iowa grant in 1856 occurred later and is included as an instance of how changes in the railroad coalition altered land grant policies toward a nationalist form.
Results
First, Figures 1 and 2 capture the descriptive patterns of railroad legislation in both the House and Senate in the antebellum period. As indicated by Figure 1, rail legislation became a national issue slowly but peaked quickly in the 1850s. However, after this initial flurry of rail legislation, the number of bills introduced in either chamber quickly dropped. In addition, the House clearly displays more action on railroad land grants than the Senate. Figure 2 summarizes the proportion of votes for and against each railroad bill by chamber. The figure indicates that, especially for early legislation, the Senate was markedly more supportive of railroad assistance than the House. Yet toward the end of the antebellum period, both chambers display a distinct preference for approving railroad legislation. The figures suggest both an early reluctance to introduce local rail issues into Congress as well as significant conflict over railroad land grants. However, congressional rail coalitions eventually coalesced into a stable form, and rail promotion bills began to pass with greater frequency.
Number of railroads votes in the House and Senate, 1838-1862
Proportion of votes in favor of railroad grants in House and Senate, 1838-1862
To expand on these initial findings, it is useful to consider overall support for rail promotion though an analysis of all votes on rail support from 1838 through 1863. Table 1, which summarizes bivariate regressions between each relevant independent variable and the dependent variable of rail vote, highlights House members from the North opposed federal action on railroads. At the same time, it reveals that House members from grant recipient states consistently supported rail promotion. In the overall analyses from Table 2, the impact of House members being from the North was small, as there is 50% probability (when looking at just the single coefficient) that Northern members voted for railroad programs. Hence, Northern members were no less likely to vote against rail promotion programs than Southern members. Given that Northern members shifted from opposition to support for railroads through the 1850s, this small effect is not surprising. In contrast, there is a 65% probability that House members from grant recipient states voted for rail promotion programs, which suggests considerable supported across time. Hence, there was some stability in the rail grant coalition, with many members from early grant recipient states continuing to support rail legislation for future requesters. In addition, Table 1 also examines the impact of when a state joined the Union on rail aid. Table 1 highlights that, in both the House and Senate, more recent additions to the Union, which tended to be frontier states strapped for resources, consistently voted in favor of rail promotion.
Bivariate Conditional Logistic Regression Coefficients, House (H) and Senate (S) Votes on Railroad Aid, 1838-1863
The dependent variable is dichotomous and measures how members of Congress voted on railroad support, developed from Keith Poole’s congressional Roll Call data set. A value of one is assigned to votes that support railroad grants, which required reverse coding in some instances. For each regression, the sole independent variable was regressed against the dependent variable.
Significant at .10 level. **Significant at .05 level. ***Significant at .01 level.
Conditional Logistic Regression Coefficients, House of Representatives Votes on Railroad Aid, 1838-1863
The dependent variable is dichotomous and measures members of Congress votes on railroad support, developed from Keith Poole’s congressional Roll Call data set. Table 2 is the results for the House, and Table 3 summarizes the Senate analysis. A value of 1 is assigned to votes that support railroad grants, which required reverse coding in some instances. The independent variables were developed from the Poole data set and historical census data. The model is a conditional logistic regression grouped by floor vote.
Significant at .10 level. **Significant at .05 level. ***Significant at .01 level.
Moving on to the multivariate analysis of overall voting patterns, Table 2 confirms when other variables are controlled, House Republicans favored rail aid while representatives from manufacturing states were more wary. Interestingly, region is no longer a significant predictor of voting on railroads when other major demographic characters are incorporated into model. 14 Table 3, which summarizes multivariate analyses on Senate votes, reveals voting patterns statistically similar to the House. As a broad first cut, these overarching findings weakly support the proposed hypotheses. In particular, the bivariate findings suggest that Southern House members as well as members from states that initially received federal rail assistance favored federal action on railroads. Table 1 also stresses that members from frontier states, in both the House and Senate, favored rail promotion. The multivariate analyses offer less support to the hypotheses, though the House findings do underscore that members from grant recipient states continued to support future rail aid. While useful in offering an overall view of rail promotion voting, these initial analyses do not indicate why Congress became involved with rail grants or how congressional land grants shifted with time.
The evolution of rail grants coalitions are best revealed by examining rail votes across time, beginning with the congressional votes from the four Congresses prior to that first land grant in 1850. The second model in Table 1 stresses that neither Northern nor Southern members were especially in favor of early rail grants, though it does reinforce the claim that frontier members were advocates of federal action on railroads. Furthermore, the second model in Table 2 argues that, after controlling for a range of factors, the only major predictor of House members’ voting on the first attempts at federal rail aid was membership in the Republican Party. Table 3 displays that Senate Republicans also voted in favor of rail aid. Taken as a whole, these results suggest that ambivalence about early rail grants was sufficient to prevent a coalition from cohering around the rail promotion issue.
Conditional Logistic Regression Coefficients, Senate Votes on Railroad Aid, 1838-1863
Significant at .10 level. **Significant at .05 level. ***Significant at .01 level.
This initial failure of members seeking federal rail aid stemmed from the localist emphasis of their early rail grant proposals. A closer study of the first attempts to secure rail aid in Illinois illuminates both how the parochial nature of early rail proposals doomed them as well as how the notion of federal rail intervention itself stemmed from localized rail promotion failures. While Illinois’s congressional delegation was successful in securing a federal rail grant in 1850, the actual origins of that project began nearly two decades earlier as a local project, supported by the state (Stover, 1975, pp. 8-12).
The Illinois legislature began its experiment with state-sponsored rail construction in 1836, when the state legislature voted to incorporate the Central Illinois Railroad (Corliss, 1950, pp. 6-10). To aid in funding such a significant undertaking, the Illinois congressional delegation sought the right-of-way through federal public lands for the fledgling endeavor. However, congressional assistance was not immediately obtained. Despite the lack of federal support, the state of Illinois pressed forward with the plan for the north-to-south Central Illinois and passed the Internal Improvements Act in 1837 (Stover, 1975, p. 9). The plan involved the state taking on not only the burden of the north-to-south Central Railroad but also constructing four other railroads that would cross the state in an east-to-west pattern and intersect with the Central (Corliss, 1950). Essentially, the state was funding a massive, dense railroad network that touched all major locales within the state.
Despite high ambitions and the best intentions, the network was never completed. The railroad network was supported through borrowed funds, and the debt was ultimately too large for the state to handle, especially following the Panic of 1837 (Donald, 1996). Despite spending upward of a million dollars on surveys and construction, all the state accomplished was 40 miles of embankment for the Central and a single working railroad that traversed the 24 miles between Meredosia and Jacksonville (Corliss, 1950; Stover, 1975). This original attempt at rail promotion in Illinois failed to accomplish much of anything substantive.
Illinois, like many Western states, was comparatively lacking in access and economic resources. Faced with this setback, Illinois’s political leaders turned once again to the federal government. Beginning in 1846, Senator Sidney Breese attempted to obtain federal support for the Illinois Central Railroad (Corliss, 1950, pp. 13-14). Unfortunately, Breese’s early efforts to obtain federal assistance for local railroads met with no real success. However, the notion of obtaining federal aid for Illinois’s railroads remained a recurrent political call within the state. During the 30th Congress, Stephen Douglas, the newly elected Illinois Senator, proposed another rail bill that differed significantly from Breese’s plan. Breese’s bill had proposed a single railroad to Galena, in the Western part of the state. In contrast, Douglas’s proposal established both Chicago, on the eastern side of the state, and Galena, as northern branches, producing a Y-shaped rail system that traversed the state. Though Douglas’s bill passed through the Senate, it was defeated in the House (Stover, 1975, pp. 15-18). This failure to develop a rail plan that incorporated other states, and merely proposed a local Illinois road, explains the voting patterns revealed in Tables 1, 2, and 3. In effect, no regional or economic constituency perceived sufficient benefit in Breese’s or Douglas’s proposals to support the Illinois grant.
Douglas was ultimately successful in obtaining rail aid for Illinois, largely by expanding the coalition interested in rail promotion through offering benefits to other representatives. Seemingly nonplussed by his initial failures, Douglas introduced another provision for the Illinois Central in the 31st Congress. To pass the bill, Douglas agreed to an amendment proposed by Senator William R. King of Alabama that extended the land grants into Alabama and Mississippi. The goal of the additional grants was to establish a single railroad line from the Ohio River to Mobile. In addition, Senator George W. Jones of Iowa pressed for an amendment shifting the northwestern terminus from Galena, Illinois to Dubuque, Iowa. Aware of the need to make his rail proposal more attractive to a greater number of interests, Douglas agreed to both amendments. The final bill granted the Illinois Central a right-of-way through all public lands as well as a gift of alternate sections of land, to a distance of 6 miles wide on each side of the road. The bill also came with the provision that work must begin within 2 years and be completed in 10 years, or else the land reverted back to the government (Corliss, 1950; Stover, 1975). The decision of how to deal with the land and route was ultimately left up to the state of Illinois. In the end, Springfield tightly directed the route of the railroad, ordering a fairly direct line from Dubuque and Chicago south to Cairo (Corliss, 1950; Stover, 1975).
To obtain the successful Illinois rail grant in 1850, Douglas was forced to give up somewhat on his localist focus and propose a rail plan that spoke to broad regional even national interests (Stover, 1975, p. 19). The impact of this change is evident in the voting patterns exhibited by members of Congress. Supporting the aforementioned hypotheses, the third model in Table 1 reveals overwhelming support from the South. From the bivariate models analyzing the House, it is evident that there was only a 31% probability of Northern members voting in favor of the Illinois rail grants. Similarly, there was only a 25% probability among Northern Senators. In addition, Table 1 highlights that members from frontier states also supported the Illinois Central grant. However, the substantive impact of frontier states’ support was small during this period, as there was only an approximately 51% probability of frontier members in the House or Senate supporting Illinois rail aid according to the bivariate estimates Tables 2 and 3 confirm Southern support for the initial Illinois rail grant.
While Western interest in federal action on railroads is clear, as frontier states often lacked the capital and administrative capacity to nurture rail construction, Southern support for these early rail grants is surprising at first glance. 15 However, it is important to recall the South had significant interests in shipping, notably in Charleston and New Orleans, that actively competed against Northern shipping centers in New York, Boston, and Baltimore. Furthermore, while many Southern states were wealthy and supported railroads, there were also poorer Southern states that lacked the capital to successfully launch their own rail promotion programs (Cline, 1997; Griffith, 1972; Mills, 1978). Thus, Southern representatives shared many frontier members’ concerns about local development and capital shortfalls.
Just as critically in terms of Southern rail support for Western grants was the escalating conflict over slavery. By the 1850s, when the rail grant process was just beginning, the conflict over slavery was increasingly shaping American politics (Einhorn, 2006). By this point in antebellum history, the South was increasingly worried about the fate of slavery in the United States. Southern interests were concerned that growing Northern power, combined with an increase in Western free states, would end slavery as an institution. In practice, Southern representatives feared being outvoted in Congress. Therefore, the South had a real interest in building political, economic, and even cultural ties with Western states as a means to court Western support for slavery (Bensel, 1995). Thus, the early rail grants were propelled through Congress both by members that shared localist concerns as well as by members that were courting broad cultural ties with one another.
At the same time that interstate rivalries over shipping and slavery pushed Southern interest in Western rail grants, those same rivalries also dissuaded Northern support for early rail grants. While there were Northern members that supported the Illinois grant, Northern support was fairly weak in comparison with members from the West and the South. In large part, this stemmed from the North’s history with internal improvements. In the early 19th century, Northern states invested heavily in their own canal and railroad projects, without any federal assistance (Larson, 2001; Scheiber, 1969). Hence, many Northern interests were wary of using federal dollars to aid the economic development of Western states and opposed this de facto redistribution.
To summarize the findings at this point, from both the statistical analyses and historical cases, the initial hypotheses that flow from the broad theoretical claims are supported. Notably, representatives raised local issues to the federal level only after state governments failed to provide services or goods demanded by citizens. However, to receive federal support, members needed to expand the reach of their rail proposals from localist to somewhat more regional in scope.
Moving forward, while the Illinois grant created a prorailroad coalition, that coalition was not stable as evidenced by future railroad grants in Missouri and Iowa. Missouri, like many of the new frontier states faced the problem of a large territory with few resources when attempting infrastructure projects. Aware of the competition among states to establish a viable transportation network, Missouri vigorously pursued internal improvements. Even before 1850 and the Illinois Central grant, Missouri sought and failed to obtain federal assistance (Million, 1896). Pushing onward despite these setbacks, Missouri began its own program of internal improvements. Missourians initially refrained from building an entirely state funded railroad system. As seen in Illinois, such programs often resulted in high taxes, waste, and failure at establishing a quality transportation network. In the end, states with failed transport networks became burdened with high levels of debt, after their internal improvement project failed to produce tolls or higher land values along its route. However, given that Missouri did not have a large population or great wealth, some form of state aid was a necessary requirement of jump-starting an internal improvements program.
In 1851, Missouri’s solution was to seek a balance between the desire for the efficiency and innovation of private enterprise along with the reality that state aid was necessary. The actual construction and operation of the railroads was handled by private companies, who were charted by the state legislature. However, the funding for the railroads was obtained by the state, using its own credit, to guarantee that adequate funding was received at a lower interest rate. However, this state funding was not freely given. At first, Missouri supported a limited number of roads and each railroad receiving aid had to demonstrate a minimum number of stock subscriptions (Million, 1896, pp. 63-68).
Ultimately, Missouri would spend more on internal improvements than any other state in the antebellum era (Goodrich, 1960). However, state promotion was not always successful in Missouri. Whereas early efforts produced numerous quality railroads, later rail promotion programs were mired in waste and corruption. These failures significantly slowed the extension of Missouri’s rail network. Aware of the inability of its internal tax revenue and industrial capital to build a complete statewide rail system, Missouri’s congressional delegation also submitted a request for federal aid (Belcher, 1947; Million, 1896).
Missouri’s congressional delegation successfully obtained two federal land grants in 1853, the second of which included a similar award in Arkansas. The Missouri grants, though more local than the far-reaching Illinois Central grant, critically extended Western railroads from St. Louis. By doing so, the grant opened up the Western territory in a significant manner, through offering a better connection to a major port city (Million, 1896, pp. 73-95). Much as in the case of Illinois, local failures lead Missouri’s rail boosters to turn to the national state for support. Furthermore, again much like the Illinois case, a railroad grant was obtained by Missouri in 1853 through expanding the rail coalition by including other, notably Southern, states in the aid package.
Turning to the data and broader patterns in the 32nd congressional session, Table 1 stresses that Southern support in the House remained important to landing rail grants in Missouri. Furthermore, Table 1 indicates House members from grant recipient states as well as frontier members in both chambers supported Missouri’s grant program. Table 2 also suggests that even after controlling for other factors, House members from the South continued to vote for rail aid, with a 66% probability of support. Grant recipient states similarly continued to vote in favor of railroad bills, with an 80% probability of members from states that received their own grant voting for the Missouri program according to Table 2. In addition, states with less manufacturing capital, many of which would have been frontier states, also supported rail assistance in Missouri. Table 3 highlights, as expected, less consistent voting patterns among Senators, outside ongoing Republican support. However, broadly, the prorailroad coalition continued to hold for the Missouri grants. 16 In addition, as this federal action on rail power was still novel, we would expect that concerns over growing federal power would not be a major concern at this time. As predicted, the Missouri case once again highlights how members from states with failed rail programs raised the rail issue the federal level, as well as how those members slightly expanded their localist proposals to incorporate other states and thereby increase congressional support. In particular, the Missouri grant continues the pattern of Southern and Western states collaborating to further a localist rail promotion agenda. Thus, the Missouri grant proposal closely resembled the Illinois grant in content and voting patterns.
The major change to the railroad coalition becomes evident by 1854, when we examine voting patterns on the Iowa rail grant. Much like other Western states, Iowa faced problems of settlement and transportation. Iowa was a resource-rich state, with great agricultural potential. Yet with Iowa’s small population, taking full advantage of those resources was difficult. In addition, without a robust natural river system, actually reaching the interior of the state to further populate it or develop its farmland was difficult (Larson, 1984). Furthermore, Iowa’s internal improvements were hampered by a state constitution, developed by Jacksonian Democrats reacting to failed internal improvement projects across the country, that severely limited the state’s capacity to take on debt. As capital was still rare in the state, government support was needed to encourage rail development. Stepping into the gap left by the state’s fiscal caution, counties and municipalities offered incentives to railroads (Goodrich, 1960; Hofsommer, 2005b; Larson, 1984). Although these efforts did produce railroads, the result was less than desirable. Cities competed for railroads, and the result was slowly built lines that often changed direction to obtain a better subsidy from a competing bid. Rather than a coherent statewide infrastructure system, Iowa’s program was haphazard and inefficient.
Faced with the necessity of infrastructure and the realization that its current approach was not working, Iowa appealed for congressional land grants, attempting to shirk the cost of local endeavors onto the federal government much as its Western neighbors had done. Iowa’s initial requests for federal land grants, which involved a complex rail network that would develop local farm-to-market roads as well as be an integral leg of a national route, were refused (Larson, 1984). After applying for several years, Iowa finally received its federal land grant in 1856. The grant was given to the state, to be distributed among five railroads (Hofsommer, 2005b, p. 12). It is important to note that all of Iowa’s land grants were in an east-to-west orientation (Russel, 1948). In effect, the Iowa grants built railroads that served national interests, especially Atlantic manufacturing concerns, rather than fulfilling the requests of Iowans. Significantly, the east–west orientation was not the only viable route. The Iowa Central operated for many years on a local route, connecting St. Louis and Minneapolis via central Iowa (Hofsommer, 2005a). Despite the viability of a dense local network, Iowa’s grant were for a much more meager and direct rail system. This shift in rail policy directly resulted from ongoing changes in the coalition pressing for federal land grants to assist in local rail development.
The results for the analyses of the Iowa rail grant votes in the 34th Congress suggest significant changes in which representatives were supporting rail grants. Table 1 shows a significant leap in House members from the North supporting Iowa’s rail grants, as Southern states fell out of the rail coalition. At the same time, Table 1 supports the hypothesis that members from grant recipient states were becoming less supportive of railroad grants, though frontier state members in both chambers were still generally voting for rail aid. Echoing these findings, with the inclusion of additional measures, Table 2 strongly asserts the support of Northern House representatives for the Iowa grant. There was a 69% probability that Northern members voted for the Iowa grant, which is a sizable increase from Northern members’ support for the Missouri package. In addition, Table 2 shows strong support from members in more industrial states. However, Table 2 also suggests that House members from grant recipient states still supported the Iowa grants but only after controls are introduced. 17 According to Table 3, region was not a major predictor of how members voted on the Iowa rail grant in the Senate. However, across the board, the data analysis suggests that the rail coalition shifted by 1856, with Northern manufacturing states playing a larger role in shaping rail grants while Southern and grant recipient states declined in influence.
By the time of the Iowa grants, members from grant recipient states became less interested in offering aid to other states. Once a state received a grant, it was in a position to build its railroads and enhance its economic advantage vis-à-vis its neighbors. If a state remained in the coalition, that state’s representatives were voting for offering aid to rivals. Hence, there was a real incentive for representatives to leave the rail coalition once they received aid for their own state.
At the same time, the ongoing federal rail grants were changing the nature of federal power. By the mid-1850s, greater federal power was increasingly problematic for Southern representatives. Federal rail grants established railroads built with those grants as federal highways, accessible to the national government for little or no cost (Bensel, 1995; MacGill, 1917). Bit by bit, federal aid to railroads was expanding the national state’s power to reach out from the center and into the hinterlands. Both of these processes suggest a significant political centralization. This increasing centralization led Southern representatives, who worried that more federal power would endanger slavery, to withdraw from the rail coalition (Mulcare, 2008).
Faced with a shrinking support from Southern and grant recipient state members, Iowa’s representatives were forced to turn to Northern states for support. As suggested by Table 2, House members from states with larger manufacturing sectors became supporters of the Iowa rail grants. Northern members and members serving from industrial states were primarily interested in railroads that provided quick access from the frontier to the coast. Hence, the meandering, localist railroads of earlier grants was replaced with the very nationally oriented and minimalist roads of the 1856 Iowa grant. The Iowa grant experience boils down to representatives that were concerned about growing federal power being unwilling to join future rail coalitions, as voting for those rail grants threatened to make the national state stronger. As a consequence, the only support Iowa could gather was from members that either were indifferent to growing national power or actually sought a heftier national state. In essence, the coalition became more nationally focused, due in part to both local competition and the conflicts over state/national spheres of influence that are each inherent to American federalism. 18
Discussion
The emergence of a national rail system, and the modern state it supported, occurred as a response to state-level projects being elevated to Congress by local congressional delegations. When states were unable to adequately fund local railroad systems, they turned to the federal government to fill the gaps in their transportation infrastructure. However, federal action alone does necessarily build a stronger central state. For instance, Congress could have chosen to continually honor local preferences in land grants, as it did initially. However, due to competition and uncertainty over the growth of federal power, the coalitions pushing for federal aid changed. Coalition changes produced comparable transformations in the actual land grants offered to the states. As the land grant process evolved, the grants increasingly served centralizing, rather than local, demands. The evolution of national rail policy in the United States underscores how competition in American federalism leads local concerns to percolate upward and become nationally salient issues.
The role of local actors, as well as the impact of early federal decisions on later policies, is not unique to railroad land grants. National infrastructure development regularly occurs because local actors shift burdens onto other levels of government. Modern flood control is the result of local economic elites building national coalitions over decades and pulling the federal government into river management (O’Neill, 2006). Furthermore, much like railroads, initial river policies altered and then cemented relationships among discrete regions. Early flood control policies resulted from a coalition between the lower Mississippi Valley and the Sacramento Valley in California. This early coalition, formed at the end of the 19th century continues to direct present day river policies (O’Neill, 2006). Similarly, the American highway system was developed through a series of bills, beginning with local interests demanding funds for farm to market roads, interurban connections, as well as better roads within urban areas (Rose, 1990). Although the Interstate Highway Act may have been a single bill with massive repercussions for defense, the economy, urban development, and national spatial organization, it was not an isolated action. Instead, it was the culmination of decades of local road bills finally resulting in a national highway network.
More broadly, the case of antebellum rail grants highlights an important process of institutional change in American federalism. American political development theories strive to understand the shift in American political power from local to national governments (Orren & Skowronek, 2004). Explanations for the growth in national power often cite industrialization or war as key contributors (Bensel, 1995; Skowronek, 1982). However, antebellum rail grants suggests an alternative, concurrent cause for a larger national state in American politics: an inherent tendency towards the national state located within the workings of American federalism. The realities of subnational state interaction produce a powerful compulsion toward the national state becoming involved in new policy areas. Local competition and local inequalities will lead frustrated local institutions to turn to the national state and ironically expand federal power in the service of more parochial needs. Once under congressional authority, new policy areas will increasingly serve national rather than local interests as congressional coalitions undergo transformations driven, once again, by local competition coupled with unease over federal growth. Whereas antebellum railroad grants are only one case, the history of rail policy development suggests a compelling alternative explanation for American state building. Policies that favor national interests over local aims result from the confluence of local failures and coalition politics that compromise American federalism.
Footnotes
Acknowledgements
I would like to thank Eric Oliver, Elisabeth Clemens, Mark Hansen, William Novak, and three anonymous reviewers for comments that greatly improved this essay. In addition, participants at the University of Chicago American Politics Workshop also provided numerous insights into my analysis. Earlier versions of this paper were presented at the Western Political Science Conference as well as the Midwest Political Science Conference.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
The author(s) received no financial support for the research, authorship, and/or publication of this article.