Presidential incentives, bureaucratic control, and party building in the Republican Era.

Rogowski, Jon C.

Scholars dating back at least to Neustadt (1960) have distinguished modern presidents from their predecessors. According to Rossiter (1960, 106), the crises posed by two world wars and the Great Depression combined with public demand for energetic leadership to provide the presidency with greater influence over the nation's affairs than ever before. Though scholars disagree somewhat on exactly when the modern presidency was born, considerable research reflects Greenstein's (1982, 3) observation that "the transformation of the office has been so profound that the modern presidencies have more in common with one another in the opportunities they provide and the demands they place on their incumbents than they have with the entire sweep of traditional presidencies from Washington to Hoover." Focusing on modern presidents, scholars have made great headway in understanding a range of presidential behaviors, including exercising veto powers (e.g., Cameron 2000) and unilateral prerogatives (e.g., Howell 2003; Lowande 2014), capitalizing on expertise (e.g., Howell, Jackman, and Rogowski 2013), appealing to the public (e.g., Canes-Wrone 2006), managing executive departments (e.g., Rudalevige 2002), and centralizing and politicizing bureaucratic structures (e.g., Moe 1985).

It is not immediately clear, however, whether the insights from the rich theoretical and empirical literature rooted in the study of the modern presidency transfer to the study of presidents who served in earlier years and who were likely to have quite different incentive structures. Enhancing our knowledge about how earlier presidents affected the nation's policy making is consequential not simply because our understanding of the historical presidency is not as richly detailed as it could be. Historical studies provide important opportunities for revealing changes in the office over time and exploring how these changes affected the relationships between the presidency and the other branches of government, and answering these questions has important implications for research programs on political institutions more generally. Important contributions from existing scholarship (e.g., Cohen 1988, 2012; Ellis and Wildavsky 1989; McCarty 2009; Skowronek 1993) underscore the value of probing presidential history to better understand the development of the office and its consequences for American government.

In this article I explore how late- nineteenth-century presidents' incentives as party leaders affected the operation of the administrative state. I contend that Republican-era presidents were driven largely by their desire to build and support the party apparatus and that their control over appointments and patronage well positioned them to do so. Using data on the county-level distribution of post offices between 1876 and 1896, I report evidence that the president's strongest geographic bases of support were disproportionate recipients of federal post offices. In a period of decentralized party structures, these findings suggest that presidents used federal resources such as post offices--and the patronage positions they created--to reward their core constituencies and further their party's goals. Importantly, the results also reveal that presidents appeared to be successful in using the bureaucracy to meet these goals. Consistent with research that calls into question periodization schemes that distinguish modern presidents from their predecessors (Adler 2013; Ellis and Walker 2007; Galvin and Shogan 2004; Klinghard 2005, 2010; Skowronek 1993, 2002; Young 2011), the findings presented in this article suggest that nineteenth-century presidents were more central in directing government activity than is generally recognized, though their efforts may have been in service of different goals.

The Presidency in the Republican Era

The advent of the modern presidency is widely believed to have marked a turning point in American politics. The Progressive vision of the president whose power was derived directly from the people began to take shape in the late nineteenth century, gained steam with Wilson's (1905) articulation in Constitutional Government, and, according to proponents of the modern presidency thesis, had been fully realized by the presidency of Franklin Roosevelt. Lowi (1986, vii) argued that this transformation of the office was tantamount to the founding of a "Second Republic" in which the country entered "an entirely new constitutional epoch." A large body of scholarship details how the public demands of, and authority provided to, modern presidents paved the way for increasing the president's advantages relative to other political actors through, for instance, the institutionalization of the office of the presidency (Moe and Wilson 1994) and the increased use of unilateral prerogatives (Moe and Howell 1999).

The modern presidency paradigm thus rests on the claim that earlier presidents enjoyed no such advantages. Prior to what Skowronek (2002, 748) termed the "big-bang transformation of the presidency," presidents are believed to have been relatively powerless and ineffectual at guiding government policies. The vigor and leadership displayed by Andrew Jackson, James K. Polk, and Abraham Lincoln are viewed as the exceptions rather than the norms. Though scholars disagree about when exactly the modern presidency was forged (for a range of opinions, see, e.g., Klinghard 2005; Lowi 1986; Tubs 1987), this development is widely believed to have represented a sharp break with the past.

Existing accounts do not allow much room for nineteenth-century presidents to have affected the operations of the administrative state. During this period, "national political power was vested chiefly in Congress and not in the presidency" (De Santis 1963, 556). After all, the presidents of this era--Ulysses S. Grant, Rutherford B. Hayes, James Garfield, Chester Arthur, Grover Cleveland, Benjamin Harrison, and Cleveland again--functioned mostly as clerks of the administrative state and stand in contrast to contemporary characterizations of a strong, individualistic presidency. Skowronek (1993, 54) classified the period from 1832 to 1900 as a Partisan era in which presidents were "unabashed representatives of their party organization." The clerical nature of the presidency left Lord James Bryce ([1888) 1995) mostly unimpressed and prompted him to entitle a chapter of his book "Why Great Men Are Not Chosen President." Studies of media coverage in the nineteenth century further suggest that presidents were not especially salient news figures. For instance, Kernell and Jacobson (1987) show that newspaper coverage of national politics during this time period focused far more on Congress than it did on presidents; to the extent that presidents were the focus of news stories, the media coverage usually emphasized the importance of presidential elections.

Presidents, Patronage, and Party Building in the Republican Era

Understanding the extent to which presidents directed governmental behavior in earlier periods, however, requires specifying exactly what presidents hoped to accomplish. Modern presidents are likely motivated by different incentive structures compared with their predecessors. According to Moe and Wilson (1994), presidents in the modern era are fundamentally concerned with being regarded as strong leaders. This incentive leads presidents to seek autonomy in political decision making wherever it is possible. But presidents in earlier eras were not likely driven by the same motivations. Instead, nineteenth-century presidents were charged chiefly with brokering a national coalition of local party organizations, and patronage appointments were the currency of presidential power (Skowronek 1993; see also Galvin 2014). Above all else, presidents were expected to further their party's goals and contribute to party development (James 2000, 2005, 2006). Indeed, nineteenth-century parties complained loudly when they believed the president did not provide sufficient support for the party program. For instance, just several months into his term, party insiders accused Hayes of "a lack of devotion to the interests of his party" (Stoddard 1889, 77), possibly as a means to secure greater compliance from Hayes with party goals.

The incentives to serve the party likely led presidents to pursue strategies in their administration that would reward and support their party strongholds. Unquestionably, the most important tool at the president's disposal was patronage. Indeed, the dispensing of patronage accords quite well with what presidents engaged in party-building activities would be expected to do. In his study of presidential party building, Galvin (2010, 5) defines presidential party building as efforts "to enhance the party's capacity to provide campaign services; develop human capital; recruit candidates; mobilize voters; finance party operations; and support internal activities. By creating positions that state and local party organizations could use to recruit party members and reward the party faithful, the distribution of patronage would have easily accomplished most if not all of these goals.

I examine how a president's partisan incentives affected the establishment of local post offices and the valuable patronage positions that accompanied them. The Postmaster General's Office was formally responsible for designating new post offices, which created new postmaster positions over which presidents had sole appointment authority. According to Johnson and Libecap (1994, 104), "the President was the key figure in the allocation of patronage, and the party that controlled the White House essentially controlled the allocation of patronage." If executive patronage was the currency of presidential power in this period, no arm of the national government afforded the president with greater power than the Post Office Department, as local postmaster positions were prized by both local parties and prospective officeholders and the post office was the largest federal employer. For instance, Kernell and McDonald (1999) report that members of Congress spent vast amounts of time meeting with prospective local postmasters, on whose behalf legislators often lobbied the president and the postmaster general.

However, though the president appointed the postmaster general (along with other cabinet officials) with the consent of the Senate, existing research leaves open the question of whether nineteenth-century bureaucracies were responsive to the president. In fact, Woodrow Wilson (1885, 177) pointedly declared that bureaucratic officials were subservient to Congress: "[M]embers of the Cabinet, being confined to executive functions, are altogether the servants of Congress. Biographical accounts of postmasters general during this time period, however, suggested that presidents granted broad autonomy to them to run the department as they saw fit (e.g., Calhoun 1988). Recent work by Mashaw (2012) argued that while appointment and removal powers provided presidents with broad control over executive branch agencies, this authority did not enable presidents to direct all administrative activities. Thus, presidential success in achieving their partisan goals depended upon the degree to which they influenced administrative behavior.

The Nineteenth-Century Post Office

The post office was one of the earliest indicators of the American state, and its role in American society has been widely celebrated by both contemporaries and historians. As early as 1833, Supreme Court Justice Joseph Story (1833, 3:306) proclaimed the importance of the post office:

   The post-office establishment has already become one of the most
   beneficent, and useful establishments under the national
   government. It circulates intelligence of a commercial, political,
   intellectual, and private nature, with incredible speed and
   regularity. It thus administers, in a very high degree, to the
   comfort, the interests, and the necessities of persons, in every
   rank and station of life. It brings the most distant places and
   persons, as it were, in contact with each other; and thus softens
   the anxieties, increases the enjoyments, and cheers the solitude of
   millions of hearts. It imparts a new influence and impulse to
   private intercourse; and, by a wider diffusion of knowledge,
   enables political rights and duties to be performed with more
   uniformity and sound judgment ... Thus, its influences have become,
   in a public, as well as private view, of incalculable value to the
   permanent interests of the Union.

Historians and scholars of American political development have emphasized the importance of local post offices for connecting communities and contributing to nationalization (e.g., Carpenter 2001; Fuller 2003; John 1995; Mashaw 2012). In the era of state building, the post office stood out as one of the only signs of the federal government in local communities.

Political parties were likely to have placed great value on post offices for reasons that extended beyond patronage. As Fowler (1943, 146) wrote, postal workers devoted much of their time to party work, rather than to postal activities. For instance, upon assuming the presidency, Cleveland complained that Republicans used their vast network of local postmasters for electioneering purposes. Perhaps in validation of Cleveland s complaint, Kernell and McDonald (1999, 796) observed that local postmasters often inserted campaign literature into local residents' mail. Thus, by using post offices as informal bases of local operations, post offices would have been valuable hubs for partisan activity.

Based on the discussion above, I hypothesize that presidents would have had the strongest incentives to distribute post offices to their core partisan constituencies. In doing so, presidents could largely ensure that the post office would be controlled and staffed by their co-partisans and thus supply the local party with patronage appointments over which they could wield influence. This expectation contrasts with what we might expect from presidents who were chiefly concerned with electoral matters in which presidents may have perceived incentives to shore up political support by distributing post offices to swing or marginal constituencies. For instance, Kriner and Reeves (2015) show that federal grants in the contemporary era are disproportionately awarded both to a president s strongest partisan supporters as well as electorally important swing states. Moreover, this expectation also suggests that presidential incentives generated distinct patterns of nineteenth-century state building in which resources such as post offices were disproportionately awarded based on explicitly political concerns and in ways that reflected the president's partisan incentives.

[FIGURE 1 OMITTED]

Data and Methods

My exploration of presidential influence over the establishment of post offices builds upon recent research on the influence of the president over the distribution of federal resources (Berry, Burden, and Howell 2010; Dynes and Huber 2015; Kriner and Reeves 2015). I test the hypotheses outlined above using data I collected on the distribution of federal post offices from 1876 to 1896. These data were collected from the United States Official Postal Guide (the Guide), published annually (with monthly supplements) from 1874 to 1954. The Guide included a complete listing of post office locations by county and state or territory. These data were collected for even years, with two exceptions: because I could not locate Guides for 1884 or 1888,1 used information from Guides published in January 1885 and January 1889, respectively, in their place. (1) Given the availability of the Guide, the post-Reconstruction era is a nearly ideal time period to study the distribution of post offices. Free mail delivery to larger cities began in 1863, and by the time rural free delivery was initiated in 1896, additional post offices were no longer needed to continue the expansion of postal operations. (2)

Figure 1 shows the county-level distribution of post offices in 1876, 1886, and 1896.3 The median county had 16 post offices, and the average was just over 20. However, the maps also indicate that the variance was quite high, as post offices were not distributed evenly across the country. Every county had at least one post office, but the distribution was heavily skewed. As one might expect, the concentration of post offices was greatest in the Northeast, and the number of post offices increased substantially over time across the South and West. The distribution of post offices also varied widely within states. For instance, Douglas County, Wisconsin, in the northwest corner of the state, had considerably more post offices than the more densely populated counties around Madison, the state capital, and Milwaukee. Indeed, the distribution of post offices correlated only weakly with population. In 1896, for example, Allegheny County, Pennsylvania (home of Pittsburgh) had a population of just under 700,000, and 181 post offices. In contrast, the city of St. Louis, Missouri, had a population of 525,000 in the same year, but only 20 post offices. Thus, the high variation in the number of post offices across time, states, and counties suggest that political factors could have been an important determinant in their distribution.

The hypotheses outlined above concern how electoral and partisan context influenced the distribution of post offices. The party-building hypothesis predicts that post offices were distributed to the states that provided the president with the greatest support in the most recent election. The electoral hypothesis predicts that post offices were disproportionately targeted toward counties in swing states, as support from these states would have been most critical for the president's reelection chances and for ensuring his party's success in the next national election. If post offices were distributed universally, however, a state's support for the president in the previous election should not be related to the number of post offices in that state's counties. I assessed these hypotheses by coding states on the basis of their support for the president in the most recent election. Following Kriner and Reeves (2015), states won by the president by more than 10 percentage points were labeled core states, and states where the margin of victory was 10 percentage points or less were coded as swing states. Thus, the omitted category comprises states where the president lost by more than 10 percentage points and represents hostile states.

Figure 2 displays a preliminary assessment of these hypotheses. For each year, the height of the bars displays the average number of post offices in counties located in core, swing, and hostile states. Several key patterns are apparent in the figure. First, consistent with the data presented in Figure 1, the average number of county-level post offices increased steadily over this time period. More interestingly, however, the distribution of post offices appeared to differ considerably based on electoral and partisan context. In each year, counties located in swing states had considerably greater numbers of post offices than counties in hostile states. These differences were not insubstantial. For instance, in 1876, the average county in states that strongly opposed Grant in the 1872 elections had 8.5 post offices, compared with an average of 16.4 post offices in counties located in states where Grant narrowly won or lost. Similarly, in 1892, counties located in states that had strongly backed Cleveland in the 1888 election rather than the victor, Harrison, had an average of 16.3 post offices, while states that had been closely contested in 1888 had an average of 30 post offices per county.

Figure 2 also indicates that states that strongly supported the president were generally advantaged relative to states that strongly opposed the president. In 9 of the 11 years, counties located in core states had greater numbers of post offices than counties in hostile states. Finally, in every year except 1876, counties in swing states also had considerably larger numbers of post offices than counties in core states. Thus, the descriptive statistics suggest that a state's partisan and electoral context was associated with the post offices distributed among its counties.

[FIGURE 2 OMITTED]

However, the raw data, while informative, cannot definitively test the hypotheses outlined above because of the many other county-, state-, and year-specific factors that may also influence the distribution of post offices. Thus, following other research on the distribution of federal resources (Berry, Burden, and Howell 2010; Dynes and Huber 2015; Kriner and Reeves 2015), 1 regress the county-level number of post offices on the indicators for the state partisan and electoral context. Because the distribution of post offices is highly skewed, I use the logged number of post offices as the dependent variable. I also include both county and year fixed effects in these models. County fixed effects control for all time-invariant county characteristics that may also be associated with the distribution of post offices, while the year fixed effects adjust for differences in the allocation of post offices across years. I also control for county population density (measured in hundreds per square mile), as increased density (as a result of population growth) is also likely to affect the allocation of post offices. (4) All standard errors were clustered by county.

I also estimated models that include a variety of control variables that may also affect the distribution of post offices. In particular, previous research suggests that certain members of Congress are better positioned than others to secure distributive resources. (5) Because the distribution of post offices could have been influenced by a legislator's partisan alignment with key government actors, I included indicators for whether a county's House member was a member of the House majority party (e.g., Cox and McCubbins 2005) or was a co-partisan of the president (e.g., Berry, Burden, and Howell 2010). Other research emphasizes the importance of membership on key congressional committees (Deering and Smith 1997; Ferejohn 1974; Shepsle and Weingast 1987). Thus, because both the Appropriations and Post Office and Post Roads Committees played key roles in affecting post office policies and priorities, I included indicators for whether a county's congressional representative served on these committees. (6) At the outset, however, I point out that the inclusion of these additional covariates reduces the number of observations that are included in the regression models, as counties that are split between multiple congressional districts or located in states served exclusively by at-large congressional districts cannot be directly mapped to a specific legislator. Fortunately, however, these counties constituted a relatively small fraction of the data set and were located mostly in urban areas.

Results

The results of these regression models are shown in Table 1. Column (1) displays the coefficient estimates when the logged number of post offices is regressed on the indicators for core and swing states and population density. The coefficient for core state is positive and statistically significant and indicates that counties located in core states received about 8.6% more post offices than counties located in hostile states. Given that the average number of post offices per county was approximately 20, this translates into an average difference of nearly 2 post offices. Aggregated to the state level, this result suggests that providing political support to the president was a significant boon to state parties who sought to use post offices and the attendant patronage positions to reward faithful party operatives and local supporters.

At the same time, the results in column (1) provide no support for the hypothesis that counties in swing states were disproportionate recipients of post offices. In fact, perhaps surprisingly, the coefficient for swing states is negative and statistically significant (- 0.025), indicating that counties in swing states received fewer post offices than counties in both hostile and core states. While I am somewhat reluctant to make too much of the negative coefficient, the data clearly provide no evidence that post offices were systematically directed toward counties in the most competitive states. Finally, the coefficient for population density also indicates that increases in density were associated with receiving greater numbers of post offices.

The results in column (2) largely corroborate these findings when the additional control variables are included. The coefficient for core state is again positive and statistically significant, indicating that counties in states that strongly supported the president received about 6.4% more post offices than counties in hostile states. The coefficient for swing states, however, again provides no evidence that post offices were targeted to states in an effort to shore up their support in the next election.

The results from these regression models provide strong support for the party-building hypothesis. Counties in states that supported the president at high levels, and thus were likely to have strong party organizations, received substantially more post offices than counties located in states that granted strong support for the president's opponent. The data provide no support for the electoral hypothesis, however. While post offices were disproportionately targeted to counties in states that had previously provided strong support for the president, counties in the most competitive states received no such advantage. Thus, the results presented here suggest that presidents used their positions to both reward and strengthen the party organizations that had been the core of their electoral coalitions.

Recipients of Presidential Targeting

Presidents interested in advantaging their party operations were unlikely to distribute post offices uniformly within states. Instead, presidents committed to party building were likely to distribute post offices to the areas that provided the strongest bases of support. Thus, post offices were likely to be targeted in ways that reflected the provision of electoral support within states. In particular, within core states, counties that provided the highest levels of support were likely to be prime targets for post offices.

To assess this hypothesis, I identified a core county as one that provided the president with a margin of victory greater than 10 percentage points in the most recent election. I then interacted this variable with the core state indicator. The interaction term thus allows the relationship between core county and the provision of post offices to vary based on whether the core county was located in a core state. If presidents targeted post offices to help build their party, I expect that they directed post offices not only to counties in states that strongly supported them, but also to the counties that provided the greatest support within core states. To assess this hypothesis, I estimated the models shown in Table 1 and included these additional terms.

The results are shown in Table 2. Consistent with the results in Table 1, the coefficients in both columns (1) and (2) show that counties in core states receive larger numbers of post offices than counties in other states. The results also suggest that core counties received more post offices than other counties, though the coefficient falls short of standard levels of statistical significance in column (2). More importantly, however, the coefficient for the interaction between core state and core county is positive and statistically significant, indicating that within states that were strong supporters of the president, counties that provided the strongest bases of support were disproportionate recipients of federal post offices.

Figure 3 graphically displays these differences. Using the estimates from column (2) of Table 2, the plotted points show the percentage point difference in the number of post offices distributed to counties in core states, compared with counties in swing states and hostile states. Positive values along the x-axis indicate that greater numbers of post offices were distributed to counties in core states. The horizontal lines are the 95% confidence intervals associated with these differences. The dashed vertical line at zero indicates where these plotted points would fall if there were no differences in the distribution of post offices based on whether a county was located in a core, swing, or hostile state.

The solid circles display the percentage point advantage in the number of post offices received by noncore counties in core states. Noncore counties in core states received 7.7% additional post offices compared with counties in swing states and 5.9% additional post offices compared with counties in hostile states. These differences are themselves statistically distinguishable at p < .04. The advantages were even larger for core counties in core states, which received 10.1% more post offices relative to counties in swing states and 8.3% more post offices than counties in hostile states. Thus, within core states, core counties received an average of 2.4% more post offices than noncore counties (p < .01).

[FIGURE 3 OMITTED]

The results presented here suggest that federal resources, such as post offices, were concentrated within nineteenth-century presidents' strongest partisan constituencies. Rather than use federal resources to curry favor with electorally valuable states, as contemporary presidents are found to do (Kriner and Reeves 2013), the evidence suggests that nineteenth-century presidents saw incentives to disproportionately distribute post offices within states that provided the presidents' largest bases of support. These findings indicate that presidents have long occupied valuable positions as head of the executive branch and that they used this position to achieve key objectives.

Discussion and Conclusion

Understanding how political actors wield influence in the separation of powers system requires examining their incentives and capacities for acting upon them. In studying historical phenomena, however, we must be attentive to whether incentive structures and institutional capacities transfer seamlessly across time. In the case of the presidency, this would seem to be a particular concern. In this article, I have attempted to sketch some potential motivations for presidential behavior prior to the modern era and present some preliminary evidence to assess whether the data are consistent with what those incentives may have led presidents to do. In so doing, the article paints a portrait of a nineteenth-century presidency largely motivated to dispense the perquisites of office to core partisan supporters, both as rewards for loyal support and as contributions toward party-building efforts.

The results presented here suggest that presidential influence over government activity is not confined to the modern presidency and that presidents in earlier eras were likely motivated by different factors than contemporary presidents. While incentives to display leadership (Moe and Wilson 1994) and consolidate power (Howell 2013) may indeed undergird the behavior of modern presidents, presidents in earlier eras were largely motivated by partisan goals. These different sets of incentives complicate efforts to compare presidential influence over time. Presidents who served in earlier eras may not have been as influential as later presidents, simply because they were not expected to be, but instead sought to use the office in the service of building and maintaining their party s organization and coalition. As the findings in this article suggest, moreover, presidents appeared to be influential in directing bureaucratic activity well before the modern era. Thus, periodization schemes that conflate changing incentive structures with changes in the president's institutional capacities risk missing an important part of the story.

While the findings shown in this article are of some interest on their own, the argument and evidence presented here need to be extended. In canvassing the breadth of U.S. history, when were presidents relatively more successful in accomplishing their goals, and when did they fail? How did these patterns of success and failure vary with other features of the institutional environment such as majority party control of Congress, bureaucratic design, and changing electoral institutions? How did institutional changes in the executive branch, such as civil service protections and the creation of units such as the Executive Office of the President, affect the president's ability to direct government affairs?

These questions--all of which concern core components of research programs on political institutions--can be studied by more thoroughly integrating political history with approaches used to study contemporary institutions. Scholars such as Adler (2013), Ellis and Walker (2007), Galvin and Shogan (2004), and Klinghard (2005, 2010) have provided rich and nuanced accounts of presidential behavior in earlier points in history, while the evidence in support of theories about presidential action in the modern era tends to be largely quantitative in nature (e.g., Berry, Burden, and Howell 2010; Cameron 2000; Canes-Wrone 2006; Howell 2003; Howell, Jackman, and Rogowski 2013; Kriner and Reeves 2015; Lowande 2014). Following the approaches described by Wawro and Katznelson (2014), scholarship on the presidency could be meaningfully advanced by combining detailed, historical knowledge with quantitative approaches designed specifically to explore how institutional changes affected political outcomes of interest. Recent scholarship by McCarty (2009), who shows how changing presidential incentives in the earlier nineteenth century affected the issuance of vetoes, and Cohen (2012), who shows how presidential agenda setting evolved over the course of U.S. history, illustrate the advantages of such an approach. In so doing, the field promises to make great progress in understanding not only how presidents act in service of their goals within our political system, but also, following Skowronek's (1993) call, how presidents attempt to change the system itself and transform our politics.

ACKNOWLEDGMENTS: The data used in this project relied on the hard work of Chris Gibson, Alex Bluestone, Michael Byrne, Madelyn Josel, Sophie Schuit, and Joe Sutherland. The Department of Political Science, the College of Arts and Sciences, and the Office of Undergraduate Research at Washington University in St. Louis provided generous funding. I thank Richard Ellis for helpful suggestions on an earlier version of the manuscript and Cameron Blevins, Will Howell, Richard John, Doug Kriner, Terry Moe, John Patty, and Andrew Reeves for stimulating discussions about the ideas presented in this paper.

JON C. ROGOWSKI

Washington University in St. Louis

References

Adler, William D. 2013. '"Generalissimo of the Nation': War Making and the Presidency in the Early Republic." Presidential Studies Quarterly 43 (2): 412-26.

Berry, Christopher R., Barry Burden, and William G. Howell. 2010. "The President and the Distribution of Federal Spending." American Political Science Review 104 (4): 783-99.

Bryce, James. [1888] 1995. The American Commonwealth. Indianapolis, IN: Liberty Fund.

Calhoun, Charles W. 1988. Gilded Age Cato: The Life of Walter Q. Gresham. Lexington: University Press of Kentucky.

Cameron, Charles M. 2000. Veto Bargaining: Presidents and the Politics of Negative Power. New York: Cambridge University Press.

Canes-Wrone, Brandice. 2006. Who Leads Whom? Presidents, Policy, and the Public. Chicago: University of Chicago Press.

Carpenter, Daniel P. 2000. "State Building through Reputation Building: Policy Innovation and Coalitions of Esteem at the Post Office, 1883-1912." Studies in American Political Development 14 (2): 121-55.

--. 2001. The Forging of Bureaucratic Autonomy: Networks, Reputations and Policy Innovation in Executive Agencies, 1862-1928. Princeton, NJ: Princeton University Press.

Cohen, Jeffrey E. 1988. The Politics of the U.S. Cabinet: Representation in the Executive Branch, 1789-1984. Pittsburgh, PA: University of Pittsburgh Press.

--. 2012. The President's Legislative Policy Agenda, 1789-2002. New York: Cambridge University Press.

Cox, Gary W., and Mathew D. McCubbins. 2005. Setting the Agenda. Cambridge, UK: Cambridge University Press.

De Santis, Vincent. 1963. "American Politics in the Gilded Age." Review of Politics 25 (4): 551-61.

Deering, Christopher J., and Steven S. Smith. 1997. Committees in Congress. Washington, DC: Congressional Quarterly Press.

Dynes, Adam, and Gregory A. Huber. 2015. "Partisanship and the Allocation of Federal Spending: Do Same-Party Legislators or Voters Benefit from Shared Party Affiliation?" American Political Science Review 109 (1): 172-86.

Ellis, Richard J., and Aaron Wildavsky. 1989- Dilemmas of Presidential Leadership from Washington through Lincoln. New Brunswick, NJ: Transaction Publishers.

Ellis, Richard J., and Alexis Walker. 2007. "Policy Speech in the Nineteenth Century Rhetorical Presidency: The Case of Zachary Taylor's 1849 Tour." Presidential Studies Quarterly 37 (2): 248-69

Ferejohn, John A. 1974. Pork Barrel Politics: Rivers and Harbors Legislation, 1947-1968. Palo Alto, CA: Stanford University Press.

Fowler, Dorothy G. 1943. The Cabinet Politician: The Postmasters General, 1829-1909. New York: Columbia University Press.

Fuller, Wayne E. 2003. Morality and the Mail in Nineteenth-Century America. Champaign: University of Illinois Press.

Galvin, Daniel J. 2010. Presidential Party Building: Dwight D. Eisenhower to George W. Bush. Princeton, NJ: Princeton University Press.

--. 2014. "Presidents as Agents of Change." Presidential Studies Quarterly 44 (1): 95-119.

Galvin, Daniel J., and Colleen Shogan. 2004. "Presidential Politicization and Centralization across the Modern-Traditional Divide." Polity 36 (3): 477-504.

Greenstein, Fred I. 1982. The Hidden-Hand Presidency: Eisenhower as Leader. New York: Basic Books.

Howell, William G. 2003. Power without Persuasion: The Politics of Direct Presidential Action. Princeton, NJ: Princeton University Press.

--. 2013. Thinking about the Presidency. Princeton, NJ: Princeton University Press.

Howell, William G., Saul P. Jackman, and Jon C. Rogowski. 2013. The Wartime President. Chicago: University of Chicago Press.

James, Scott C. 2000. Presidents, Parties, and the State: A Party System Perspective on Democratic Regulatory Choice, 1884-1936. New York: Cambridge University Press.

--. 2005. "The Evolution of the Presidency: Between the Promise and the Fear." In Institutions of American Democracy: The Executive Branch, eds. Joel D. Aberbach and Mark A. Peterson. New York: Oxford University Press, 3-40.

--. 2006. "Patronage Regimes and American Party Development from 'The Age of Jackson' to the Progressive Era." British Journal of Political Science 36 (1): 39-60.

John, Richard R. 1995. Spreading the News: The American Postal System from Franklin to Morse. Cambridge, MA: Harvard University Press.

Johnson, Ronald N., and Gary D. Libecap. 1994. "Patronage to Merit and Control of the Federal Government Labor Force." Explorations in Economic History 31 (1): 91-119.

Kernell, Samuel, and Gary C. Jacobson. 1987. "Congress and the Presidency as News in the Nineteenth Century." The Journal of Politics 49 (4): 1016-35.

Kernell, Samuel, and Michael P. McDonald. 1999. "Congress and America's Political Development: The Transformation of the Post Office from Patronage to Service." American Journal of Political Science 43 (3): 792-811.

Klinghard, Daniel P. 2005. Grover Cleveland, William McKinley, and the Emergence of the President as Party Leader." Presidential Studies Quarterly 35 (4): 736-60.

--. 2010. The Nationalization of American Political Parties, 1880-1896. New York: Cambridge University Press.

Kriner, Douglas L., and Andrew Reeves. 2015. The Particularistic President: Executive Branch Politics and Political Inequality. New York: Cambridge University Press.

Lowande, Kenneth S. 2014. "After the Orders: Presidential Memoranda and Unilateral Action." Presidential Studies Quarterly 44 (4): 724-41.

Lowi, Theodore J. 1986. The Personal President: Power Invested, Promise Unfulfilled. Ithaca, NY: Cornell University Press.

Mashaw, Jerry L. 2012. Creating the Administrative Constitution: The Lost One Hundred Years of American Administrative Law. New Haven, CT: Yale University Press.

Moe, Terry. 1985. "The Politicized Presidency." In The New Direction in American Politics, eds. John Chubb and Paul G. Peterson. Washington, DC: Brookings Institution Press, 235-72.

Moe, Terry, and William Howell. 1999- "Unilateral Action and Presidential Power: A Primer." Presidential Studies Quarterly 29 (4): 850-72.

Moe, Terry M., and Scott A. Wilson. 1994. "Presidents and the Politics of Structure." Law and Contemporary Problems 57 (2): 1-44.

Neustadt, Richard E. 1960. Presidential Power: The Politics of Leadership. New York: Wiley.

Rossiter, Clinton. 1960. The American Presidency. New York: New American Library.

Rudalevige, Andrew. 2002. Managing the Presidents Program: Presidential Leadership and Legislative Policy Formation. Princeton, NJ: Princeton University Press.

Schickler, Eric, and John Sides. 2000. "Intergenerational Warfare: The Senate Decentralizes Appropriations." Legislative Studies Quarterly 25 (4): 551-75.

Shepsle, Kenneth A., and Barry R. Weingast. 1987. "The Institutional Foundations of Committee Power." American Political Science Review 81 (1): 85-104.

Skowronek, Stephen. 1993. The Politics Presidents Make: Leadership from John Adams to George Bush. Cambridge, MA: Harvard University Press.

Skowronek, Stephen. 2002. "Presidency and American Political Development: A Third Look." Presidential Studies Quarterly 32 (4): 743-52.

Story, Joseph. 183.3. Commentaries on the Constitution of the United States. Boston: Hilliard, Grey.

Stewart, Charles H. 1989/ Budget Reform Politics: The Design of the Appropriations Process. New York: Cambridge University Press.

Stoddard, William O. 1889. Lives of the Presidents: Hayes, Garfield, and Arthur. New York: F. A. Stokes and Brothers.

Tubs, Jeffrey A. 1987. The Rhetorical Presidency. Princeton, NJ: Princeton University Press.

Wawro, Gregory J., and Ira Katznelson. 2014. "Designing Historical Social Scientific Inquiry: How Parameter Heterogeneity Can Bridge the Methodological Divide between Quantitative and Qualitative Approaches." American Journal of Political Science 58 (2): 526-46.

Wilson, Woodrow. 1885. Congressional Government: A Study in American Politics. Boston: Houghton Mifflin.

--. 1905. Constitutional Government in the United States. New York: Columbia University Press.

Young, Christopher J. 2011. "Connecting the President and the People: Washington's Neutrality, Genet's Challenge, and Hamilton's Fight for Public Support." Journal of the Early Republic 31 (3): 435-66.

(1.) In addition, the only Guide published in 1876 that could be located was missing the first page of entries, and thus data for about half the counties in Alabama are missing for that year.

(2.) For instance, Carpenter (2000, 140) notes that the number of post office locations peaked at 76,945 in 1901, and then reduced considerably to 59,580 by 1910, as the expansion of delivery routes led to the closing or consolidation of existing post office locations (Kernell and McDonald 1999).

(3.) Some caution is warranted in overinterpreting these maps, which were constructed using contemporary county borders. County names and borders in the West, and particularly in then-territories, changed during the time period under investigation and may have also changed since that time. Though data on the distribution of post offices were collected for all counties in all states and territories, some counties do not have clear contemporary analogs, particularly in territories that were settled over this time period such as Oklahoma, New Mexico, and Arizona.

(4.) Data on county population density were obtained from Inter-university Consortium for Political and Social Research (ICPSR) study #2896, "Historical, Demographic, Economic, and Social Data: The United States, 1790-2002."

(5.) Sources for these data include David Canon, Garrison Nelson, and Charles Stewart, "Historical Congressional Standing Committees, 1st to 79th Congresses, 1789-1947"; ICPSR study #3371, "Database of [United States] Congressional Historical Statistics, 1789-1989"; and ICPSR study #8611, "Electoral Data for Counties in the United States: Presidential and Congressional Races, 1840-1972."

(6.) The Appropriations Committee handled appropriations for the Post Office Department until 1885, at which point jurisdiction was transferred to the Post Office and Post Roads Committee (see Schickler and Sides 2000; Stewart 1989, chap. 3). The Post Office and Post Roads Committee was solely responsible for legislation related to the designation of post roads.

Jon C. Rogowski is an assistant professor of political science at Washington University in St. Louis. His research interests include political institutions, policy making, and public opinion, and he is coauthor of The Wartime President: Executive Influence and the Nationalizing Politics of Threat, which won the 2014 William H. Riker Award for the best book in political economy from the Political Economy Section of the American Political Science Association.

TABLE 1
State Electoral Context and the Targeting of Post Offices, 1876-96

                                  (1)                (2)
Core state                  0.086 ** (0.005)   0.064 ** (0.006)
Swing state                 -0.025 ** (0.007)  -0.040 ** (0.007)
Population density          0.030 ** (0.007)   0.108 ** (0.033)
  (100s per square mile)
Intercept                   2.422 ** (0.008)   2.477 ** (0.013)
N (observations)                 26,477             22,334
N (units)                        2,539              2,388
Controls                           No                Yes

** indicates p < .05 and * indicates p < .10, two-tailed
tests. The dependent variable is the logged number of post
offices in even years, 1876-96. Entries are linear
regression coefficients with standard errors clustered on
county shown in parentheses. County and year fixed effects
were also included. Control variables include indicators for
whether a county's House member was a member of the House
majority party, a co-partisan of the president, served on
the House Appropriations Committee, or served on the House
Post Office and Post Roads Committee.

TABLE 2

Electoral Context and the Targeting Post Offices within
States, 1876-96

                                  (1)                 (2)

Core state                  0.075 ** (0.008)   0.059 ** (0.009)
Swing state                  -0.002 (0.008)    -0.018 ** (0.008)
Core county                 0.008 * (0.005)      0.005 (0.005)
Core state x Core county    0.027 ** (0.011)    0.020 * (0.010)
Population density          0.033 * (0.009)    0.134 ** (0.032)
  (100s per square mile)
Intercept                   2.536 * (0.007)    2.549 ** (0.013)
N (observations)                 23,222             20,205
N (units)                        2,536               2,386
Controls                           No                 Yes

** indicates p < .05 and * indicates p < .10, two-tailed
tests. The dependent variable is the logged number of post
offices in even years, 1876-96. Entries are linear
regression coefficients with standard errors clustered on
county shown in parentheses. County and year fixed effects
were also included.
Control variables include indicators for whether a county's
House member was a member of the House majority party, a
co-partisan of the president, served on the House
Appropriations Committee, or served on the House Post Office
and Post Roads Committee.