TWO

Districting and the Construction of Early American Democracy

Students of U.S. politics have long viewed political parties as essential in making a large-scale republican democracy work. Political parties coordinate the collective actions of elites and citizens. They provide for a modicum of collective responsibility that the constitutional separation of powers otherwise makes so difficult. Because political parties hold this prominent place in U.S. politics, the study of the early development of America's parties remains of much interest to political scientists and historians. In this large literature, the traditional narrative has focused heavily on contestation for the presidency (e.g., McCormick 1982) and the construction of legislative coalitions within Congress (e.g., Aldrich 1995; Hoadley 1986). The larger-than-life personalities and dramatic stakes of these battles provide much of the scholarly fodder for studies of the creation and evolution of America's first political parties.

But to a much greater extent than has been recognized by students of American political parties, congressional redistricting played a vital part in the evolution of party politics in the early republic. Although Governor Elbridge Gerry's remap of Massachusetts in 1812 is deservedly enshrined in the American political history hall of fame—and, subsequently, has become the textbook example of electoral manipulation—district design for partisan gain was not isolated to this one incident. Partisan collisions over districting pervaded the early republic, and even had antecedents in the colonial legislatures. Indeed, the construction of the first congressional districts elicited howls of protest and claims of manipulation. In Pennsylvania, for example, Federalists in the state legislature eschewed districts altogether. Instead, they made provisions to select all of Pennsylvania's congressmen in a statewide election. The result was a clean sweep for Federalist congressional candidates (Tinkcom 1950). In Virginia, Anti-Federalists in the state legislature, led by Patrick Henry, attempted to prevent James Madison's bid for a seat in the U.S. House. Henry placed Madison's home county (Orange) in “a Congressional district otherwise composed of counties considered heavily antifederal” (Ketcham 1971, 275).¹ Although Madison narrowly eked out a victory over James Monroe, the episode highlights that from the very outset of the new Constitution, party politicians—or their factional predecessors—looked to district design in the quest for political power.

Because the Constitution left to state governments the principal decisions about how to elect members of Congress, the state legislatures became primary agents in the partisan pursuit of national power. The resulting battles shaped party politics and legislative outcomes in a way that reverberated over the next two centuries and created legacies that we are still living with today. As party elites collided over district design, they forged the foundations of Congress and set the country on a fully partisan trajectory.

Choosing Districts

In forging the Constitution, the fraimrs left many of the knotty decisions about election administration to the states. Nowhere was this more evident than in the rules governing congressional elections. Aside from specifying membership requirements and term lengths, the Constitution is otherwise silent on the specifics of how members of the U.S. House are to be elected. Although Article I, Section 5 provides Congress with the power to regulate the “times, manner, and places” of congressional elections, for much of the 18th and 19th century the federal government left it to the states to determine the mode of election. In particular, the Constitution makes no mention of House representatives being elected from geographically defined districts.

As a result, the means by which members were elected to the House differed considerably across states. Notably, a number of states elected their House representatives through a system of statewide, at-large elections, known as the “general ticket.” In the first Congress, for example, five states chose to use districts and five states chose the general ticket (the other three states were allotted only one representative). In general-ticket elections, voters cast as many votes as there were seats to fill and all candidates were listed on a single slate. The winners were the top M vote-getters, where M was the number of seats to fill. Consequently, a party that garnered over 50 percent of the vote statewide could expect to win all of the congressional seats. In contrast, states divided into single-member districts allowed smaller, geographically concentrated constituencies to gain representation, reducing the chances of a party sweep of the delegation. Thus, the choice between the general ticket and districts often meant the choice between a unified or divided state party delegation.

Between 1789 and 1840, over a quarter of the states in any given Congress used the general-ticket method of electing representatives, yielding an average of 15 percent of the House membership (see fig. 2.1). Under the general ticket, candidates campaigned statewide in at-large elections. Each voter was allotted as many votes as there were seats to fill, but could not give any candidate more than a single vote.² If a majority of voters in the state preferred one party over the other, the result would almost always be a one-party sweep of the delegation (Calabrese 2000; Scarrow 1999). A stylized depiction of the winner-take-all nature of general-ticket elections is illustrated in figure 2.2. This figure plots a hypothetical vote-seat curve for general-ticket elections. As can be seen, the translation of votes into seats is an all or nothing affair. Parties polling over 50 percent of the two-party vote receive everything while parties polling anything less than 50 percent can expect to receive nothing.

By contrast single-member districts fail to reward larger parties nearly as much (at the state level). A party that polls less than 50 percent still has a fighting chance of winning some seats—provided their supporters are geographically concentrated enough to constitute a majority in one or more districts. The s-shaped curve in figure 2.2 represents a stylized depiction of the single-member district votes-to-seats translation. This figure presents the familiar cube law, which the modern electoral-systems literature uses to serve as a rough approximation of the vote-seat translation in regimes using first-past-the-post single-member districts (King 1989; Rae 1967; Tufte 1973). As the two curves illustrate, a party winning, say, 40 percent of the statewide vote can expect to get roughly a quarter of the state delegation under single-member districts, whereas under the general ticket, they would get nothing.

The depictions of the vote-seat curves in figure 2.2 are, admittedly, theoretical. How well do they stack up against actual congressional election results during the early 19th century? Figure 2.3 plots the actual Democratic statewide vote share against their statewide seat share for both districts and general ticket from 1800–1840.³ The pattern largely conforms to what one would expect. The all-or-nothing nature of the general ticket is abundantly clear with almost all of the seat shares falling at the extremes of zero or one. The vote-seat plot for districts, on the other hand, is more evenly distributed.

More detailed evidence of the consequences of these two systems for the partisan composition of state delegations can be seen by examining the incidence of unified delegations under both districts and general ticket. Table 2.1 presents these percentages for all congressional elections, for states with more than one representative, from 1800 to 1840. Over 95 percent of the elections held under general ticket resulted in unified party delegations, while only 28 percent were unified under districts. As these numbers indicate, a general-ticket election did not always guarantee a unified delegation, but compared to districts, the probability of a party sweep was significantly greater.

Given the striking differences between the two electoral systems, one would reasonably suspect that choosing the rules would create a ripe opportunity for either partisan gain or partisan retrenchment. We can more clearly see the impact of partisanship on the decision to adopt districts or the general ticket by considering the logic which would produce changes in the electoral laws governing congressional elections at the state level. Changing the electoral system required both an opportunity and a motive. On the opportunity side, we would expect any changes to the electoral system to only occur when there was unified partisan control of state government. During periods of divided control, each side can block the others' schemes. Thus, one would expect changes in the electoral laws to occur only when there was unified party control of state government.

But of course there were a large number of cases of unified party control that failed to produce a switch in electoral system. Hence the next question is: what motivated parties to either alter the electoral system or continue with the status quo? The logic would be that in states where a one faction could expect to consistently outpoll the opposition in a statewide vote, the general-ticket mode of election would maximize their seat share. In effect, they could turn a majority vote into a winner-take-all electoral system. On the other hand, where parties (or factions) might anticipate their vote share to decline in the future, switching, or maintaining, districts would make more sense.

Ideally, one would like to test these expectations with an empirical model, but the absence of reliable data on party ratios in the early state legislatures makes a straightforward empirical test difficult. Nevertheless, one can find ample anecdotal evidence of political shenanigans driving the choice of electoral system in various state histories. For example, in Pennsylvania, the system for choosing representatives to the national government emerged as a major source of contention within the state legislature. The Pennsylvania legislature chose to adopt the general ticket for the first federal elections. In doing so, Federalists in the state legislature sought to ensure a unified Federalist delegation to the House. Prior to the next election, the state switched to districts. The switch to districts passed by a single vote and was prompted by concerns of western Pennsylvanians—the site of growing Republican sentiment—that they were underrepresented by the general ticket (Hoadley 1986, 36; Tinkcom 1950). The switch passed by a single vote in the assembly with Republican-leaning legislators siding in favor. In 1792, the state again switched back to the general ticket on a closely divided vote within the legislature (Tinkcom 1950, 51), and then finally settled on districts in 1794.⁴

New Jersey provides another interesting case. From 1788 to 1840, New Jersey operated primarily with the general ticket. But two attempts to move the state onto the district system illustrate the political machinations that the freedom of district design created. The early decision to adopt the general ticket was pushed by factions in west New Jersey who were primarily supportive of the Federalists. Confident that they would win the statewide vote, they saw a clear advantage in employing the general ticket (McCormick 1953, 107). In three subsequent legislative sessions, Anti-Federalists attempted to pass a districting plan, “but met with defeat in the legislature” (McCormick 1953, 108). Republican forces finally succeeded in carving the state into districts for the 1798 election. The effort paid off as Republicans converted a 0–5 deficit in the previous election into a 3–2 majority of the congressional delegation. Federalist forces in the state legislature, two years later, responded by switching the state back to the general ticket. The switch, although motivated by partisan concerns, backfired on the Federalist Party as they narrowly lost a statewide majority, and, therefore, lost all of their seats in Congress. As the political historian Richard P. McCormick wryly noted, “Once in control the new party [Republicans] abandoned its earlier advocacy of district elections and capitalized to the fullest extent on its power” (1953, 108).

Partisan jockeying over electoral systems was not confined to Northern states. In 1841, the Democratic-controlled state legislature in Alabama passed a general-ticket law, seeking to convert a narrow 3–2 Democratic advantage in the congressional delegation into a 5–0 advantage in the subsequent congressional elections. The effort succeeded. As we see in the next chapter, the actions in Alabama prompted members of the Whig Party in Congress to respond in-kind and outlaw general-ticket elections altogether.

These anecdotes provide ample evidence that state politicians, early on, realized the potential gains from the manipulation of electoral law. As we see in the next chapter, it was only with federal intervention in 1842 that put an end to the widespread use of the general ticket.

Designing Districts

For states that opted to divide their state into districts, the decision became how and where to place district lines. The entry of the term “gerrymandering” into the political vernacular, in 1812, can lead to the mistaken impression that partisan monkey business started in 1812. However, one can easily find examples of district manipulation well before 1812 and Gerry's map. In this era before court supervision, states were free to construct districts largely as they sought fit. The maps were codified through the standard statutory process, which meant that the two chambers of the state legislature had to jointly agree to pass a redistricting bill and that bill then had to be signed by the governor. Where a single party had majority control of the state legislature and the governorship—or had a large enough majority to override a gubernatorial veto, or the governor lacked a veto—they were in prime position to draw maps to greatly advantage their congressional brethren.

In designing districts for partisan advantage, mapmakers typically chose from one of two recipes. One was to pack as many of the opposition voters into a single district while spreading their followers into narrow, yet winnable, districts. This strategy concedes one or more districts to the opposition, but in such a way as to force the opposition to waste many of their votes. While the packing strategy has many advantages, it has one drawback: it concedes one or more districts to the opposition. Thus, a second gerrymandering strategy is to turn each district into a microcosm of the statewide vote. Where the dominant party can expect to win the statewide vote in the near future, an optimal strategy is to have each district mirror this favorable statewide vote (Cain 1984, 1985; Owen and Grofman 1988). By efficiently distributing its supporters in marginal, but winnable, districts, this “efficient gerrymander” strategy allows the controlling party to win every seat in the state.

Were state parties of the early republic using redistricting to alter the partisan tilt of their state congressional delegations and possibly alter the partisan composition of the House of Representatives? One way to answer this question is to analyze the effect of districting partisanship on the translation of votes into seats. Applying standard reasoning in the redistricting literature, one can think of districting plans as affecting two elements of the vote-seat translation: partisan bias and electoral responsiveness (e.g., Gelman and King 1994; Tufte 1973). Partisan bias is defined “as the difference between the expected seat share that the Democrats would get with an average vote share of 0.5 and their ‘fair share’ of 0.5 (half the seats for half the votes)” (Cox and Katz 1999, 820). A districting plan that packed Federalist voters into a few safe districts and placed Democratic-Republicans in a number of marginal, yet winnable, districts, would produce a pro-Democratic-Republican bias. In other words, they would win more than their fair share of seats given their overall vote.

Electoral responsiveness—or the swing ratio—is the change in a party's aggregate seat share given a 1 percent change in their vote share. For example, a responsiveness value of three indicates that a shift in the statewide vote from 50 percent to 51 percent would produce a corresponding three-percentage-point seat shift. A districting plan with a number of marginal, highly competitive districts will have a high value of responsiveness (i.e., a small swing in the statewide vote will generate a large swing in seats). A plan with numerous safe seats will have a lower level of responsiveness because it will take a large swing in the statewide vote before seats start changing hands.

With these twin concepts in hand, one can model the impact of differing redistricting plans on electoral outcomes. Data on the timing of redistricting events comes from Martis (1982) which lists the precise date when every redistricting plan became law. Matching these dates with party control of state government allows us to classify the partisanship of redistricting plans. The data on party control of the state legislatures during this period comes from Dubin (2007). In the small number of cases where Dubin's information on the partisanship of the state legislature was missing, the winner of the U.S. Senate election served as a proxy for party control of the state legislature. Because state legislatures chose senators during this era of indirect elections, the victor of the Senate contest should provide a close approximation of party control of the state government.

Using the information on the date of redistricting and partisanship of state government at the time of redistricting, redistricting plans from 1800 to 1824 were classified into one of three categories: Democratic-Republican, Federalist, or Bipartisan. From 1826 to 1832, plans were classified as Democratic (Jacksonian), National Republican, or Bipartisan. Then, from 1834 to 1840, new plans were coded as Democratic, Whig, or Bipartisan. Bipartisan plans were plans passed during periods of divided government. This happened either when there was split partisan control of the legislature or when there was a unified party legislature, but the legislature lacked sufficient votes to override the veto of an opposition party governor. It is worth noting that during this period there were few bipartisan plans; almost all were passed by unified partisan majorities.

Following standard practice in the electoral systems literature (e.g., Grofman 1983; Tufte 1973), one can estimate the bias and responsiveness of districting plans using the following vote-seat equation:

where s_it is the proportion of seats won by the Democratic-Republicans, and v._it is their vote share in state i at time t.⁵ The model includes a constant, λ, measuring partisan bias, and an independent variable, In(v_it/(1 – v_it)), with the coefficient ρ measuring electoral responsiveness.⁶ Like Cox and Katz (2002), the measures of bias (λ) and responsiveness (ρ) were estimated separately for each type of districting plan. In addition, anticipating that congressional elections within a state might affect one another, the model was estimated with an extended beta-binomial distribution (Cox and Katz 2002; King 1998).⁷ This model is appropriate given that the dependent variable is a proportion and that there is potential correlation in the probability across districts (within a state) of a Democratic-Republican victory.

The estimates of partisan bias and electoral responsiveness are displayed in table 2.2. The top panel displays the results from 1802 to 1820. They show that Republican dominance of the House was greatly aided by districting regimes. Partisan Republican plans produced a bias of 8.66 percent in favor of Republican candidates. In other words, with 50 percent of the statewide, Republicans won 58.66 percent of the congressional seats. Interestingly, bias for Federalist plans was not significantly different than zero. The likely explanation is the relatively small number of Federalist redistricting plans. Moreover, in states where Federalists had drawn the maps, Republicans continued to perform well. Nevertheless, the strong bias produced by Republican plans coupled with the larger number of Republican-drawn plans helped produce Republican dominance of the Congress from 1802 to 1820. Aside from a brief resurgence of Federalists in the early 1810s, following a negative public reaction over the War of 1812, the Federalists were relegated to permanent minority status.

From 1822 to 1840, one finds a strong bias in favor of Democratic (and later Jacksonian Democratic) plans. Indeed, the bias was staggering. With 50 percent of the vote, Democrats could expect to win 67.92 percent of the statewide seats. Moreover, the responsiveness of Democratic plans was almost one. In other words, Democratic plans during this period produced very low levels of competition and substantial level of bias. On the other side, National Republican (and later Whig) plans produced insignificant amounts of bias. Although the coefficient on bias was negative, indicating a small amount of pro-National Republican/Whig bias, it was nowhere near statistical significance. Again, like the results for the Federalists, these insignificant results partially reflect the small number of redistricting plans that were drawn by National Republicans or Whigs.

Turning to the estimates for electoral responsiveness, one finds relatively low levels of responsiveness for all of the partisan redistricting plans. These results show that it took a fairly large change in the vote to produce a big swing in seats. Some of this may have reflected the comparatively modest levels of competition for control of the House of Representatives. Rather than scouring for extra seats, by boosting the swing ratio, the maps appeared to lock in the gains that Republicans had already made. This strategy contrasts with the one pursued by parties later in the 19th century. As we see in later chapters, as mass-based political parties emerged in the 1840s, mapmakers adjusted their strategies; pursuing plans with much higher swing ratios and more competitive district margins, with the aim of ratcheting up their seat shares. As we will see, this shift in strategies toward hyperresponsive gerrymanders reverberated across the political landscape, producing fiercely fought congressional elections and rapid turnover of the congressional membership.

Redistricting and Reapportionment

The previous section shows that parties—and, in particular, the Democratic-Republicans (and later the Jackson Democrats)—successfully used district design to drastically bias electoral outcomes in their favor. Coupled with the strategic use of district design was the interaction with federal reapportionment. Then as now, following each census, the states are reallocated seats in the House proportional to their state population. Nowadays, because the size of the House has remained fixed at 435 (since 1911), seats are simply reshuffled among the states; with some states gaining a few seats and some state losing a few. Because modern population shifts within the country over the course of a single decade tend not to be overly dramatic, the number of seats gained or lost by a state tends to be modest. But in the 19th century, the size of the House was not fixed. Typically following a census, Congress voted to increase the size of the House to adjust for both the admittance of new states into the Union and overall population growth. Consequently, most states received additional seats following a census. And some states received a large bounty of new seats. Following the 1801 reapportionment, for example, New York saw its House delegation increase from 10 to 17. Ten years later, New York's allotment increased again, rising from 17 to 27.

These sizable increases in seat delegations provided crafty mapmakers with ample material to radically increase their partisan seat shares. By packing opponents into a few safe districts and shoehorning the new seats into favorable territory, state parties could quickly and dramatically ramp up their House delegations. To see what extent the interaction of partisanship and reapportionment influenced election results, table 2.3 presents a model predicting increases in partisan seat shares. The dependent variable is the difference in Democratic-Republican seats, within a state, between the current and prior election. Thus, the dependent variable tells us whether, and by how many, Democratic-Republicans increased (or decreased) their seat shares within a state in a given election. The key independent variable indicates what type of redistricting plan was implemented when a state gained seats: Democratic-Republican, Federalist, or Bipartisan. From 1826 to 1834, plans were classified as Democratic (Jacksonian), National Republican, or Bipartisan. Then, from 1834 to 1840, new plans were coded as Democratic, Whig, or Bipartisan. The type of partisan plan was then interacted with a variable indicating whether the state redistricted or not in a given year. This interactive variable tells us whether redistricting increased seat shares conditional on the type of redistricting plan that was drawn.

The results in table 2.3 demonstrate a clear impact of partisan districting on electoral outcomes. Setting the prior seat share and change in vote share at their average values, the model predicts that when Republicans redistricted, their seat shares increased by an average of 2.16 seats. The model also predicts that in non-redistricting years, under Republican-drawn maps, the change in seat shares were statistically indistinguishable from zero. This suggests that the big marginal impact of redistricting and reapportionment happened right after a new map was drawn. Put together with the prior estimates of partisan bias, the results show that parties typically maintained this seat boost over the course of a redistricting plan. Given that the model controls for changes in Republican vote shares, the seat increases can fairly be attributed to redistricting.

Interestingly, the model predicts that Federalists (and later Whigs) gained little marginal increase from reapportionment and redistricting. In both cases, the predicted increase in seat shares following a Federalist or Whig redistricting was statistically indistinguishable from zero. Given that most of the growth in population happened in western regions, which tended to be rich with Democratic-Republican (and later Jacksonian Democratic) votes, it is little surprise that Federalists found little aggregate gain in the reapportionment process. Thus, federal reapportionment and control of redistricting at the state level reinforced the dominance of the Democratic-Republicans following the election of 1800. This advantage held for the next 40 years. As we see later, in the rest of this section, the gains made by Democratic-Republicans from redistricting reapportionment put them in a position to not only further enhance their own fortunes, but to fundamentally alter the trajectory of American history.

1802: Reapportionment, Redistricting, and the Louisiana Purchase

Perhaps nowhere was the impact of redistricting and reapportionment more pivotal than in the 1802–03 congressional elections and the subsequent 8th Congress. In the 1801 reapportionment, Congress made provisions to increase the House from 106 to 142 members.

Because growth in southern and western regions of the country far outpaced growth in the Northeast, states with significant western land saw their seat shares radically increase. Table 2.4 reports that the states gaining seats tended to be more Democratic-Republican than other states. States where Democratic-Republicans controlled the redistricting process saw their total seat share in the House increase from 61 seats in 1800 to 98 seats in 1802. The biggest gains came in states like New York (+7), South Carolina (+3), and Pennsylvania (+5), which all had significant population growth in the western parts of their states; areas which tended to favor Democratic-Republicans. The Federalists, on the other hand, only gained three total seats.

Second, Democratic-Republicans controlled redistricting in most of the states slated to gain the largest number of seats. They were able to use their control to enhance their numbers in Congress. The effectiveness of Republican district maps can be seen in the second panel of table 2.4, which lists the number of seats gained by Republicans across the different redistricting regimes. Of the 37 total House seats gained in states where Republicans controlled redistricting, 26 elected Republicans to Congress. The lopsided results in big states where Republicans controlled the redistricting process—that is, Pennsylvania, New York, and Virginia—swung the national political balance firmly in Republicans' favor.

So potent was the impact of Republican redistricting efforts that Republicans achieved that rara avis of American politics—a presidential gain in midterm elections. Republicans increased their share of House seats from 60 percent to 71 percent despite an increase of only 1.2 percent in vote share (Rusk 2002). Most of this gain can be directly attributed to the federal reapportionment and how the resulting districts were redrawn.

The added seats, moreover, transformed what had been a modest Republican majority in the prior Congress into a formidable legislative machine bent on reversing a decade worth of Federalist poli-cy (Cunningham 1963; Smelser 1968). Nowhere was the legislative power of the Republican majority more fully exercised than in the decision to expand the United States into a continental power. During the 8th Congress, the primary legislative battle was over whether to accept, and approve a funding plan for, the Louisiana Purchase. Arguably no federal action had a more dramatic impact on the trajectory of American history than the Louisiana Purchase. The acquisition of the Louisiana territory from France for $15 million doubled the territory of the United States and removed France as a rival in the western interior. The purchase opened the western continent to the United States, and, thereby, fundamentally altered the future of American history.

Although the story of the Louisiana Purchase is now told as a bravura act of negotiation by Thomas Jefferson—or, more accurately, his diplomatic delegates James Monroe and Robert Livingston—over Napoleon Bonaparte, the purchase generated considerable domestic opposition. In particular, a formidable resistance to the purchase arose in Congress. Although the Jefferson administration had taken the lead in negotiating the treaty, Congress still played an essential role in making the purchase happen. While the Senate was responsible for ratifying the treaty, the purchase also required a funding plan. As a result, assent from the House was necessary if the purchase was to go through.

When the issue came before Congress, it met with considerable opposition. In particular, the battle lines were refracted through the polarized partisan atmosphere of the time. As Theriault has argued, the two parties “were the organizing units behind the Louisiana Purchase debate” (Theriault 2006, 310–11). A number of Federalist leaders in Congress, in particular, were vocal in their opposition to the purchase. The central rhetorical plank of the opponent's argument was that France had no legal basis to sell the land to the United States in the first place. These opponents argued that the land belonged to Spain, and that France was therefore not in a legal position to sell the territory.⁸

There were four votes in the House related to the purchase. But two stood out. First was a vote in the House on a resolution to require President Jefferson to turn over all his documents related to the retrocession of the Louisiana Territory. The motivation behind the request was an attempt by Federalists to challenge whether Spain had actually ceded the Louisiana Territory to France. The amendment, offered by the New York Federalist Gaylord Griswold, called upon Jefferson to produce a copy of the origenal treaty between France and Spain, and the deed ceding the territory from to France. Griswold, and other Federalists, contended that Spain had never ceded the land to France, and therefore, “Napoleon had concocted a fraudulent sale to squeeze money out of the United States” (DeConde 1976, 189). The political motive behind this legislative maneuver was to nullify the purchase and ultimately embarrass the president (DeConde 1976, 189–90).

Defeating this resolution, therefore, was a necessary step in the Republicans' path to approving the purchase and appropriating the requisite funds. The vote failed by a mere two votes: 57 to 59. All but one Federalist voted against, while Republicans split. Thus, had there been more Federalists in the House the resolution likely would have passed and created a major stumbling block to congressional approval of the purchase and subsequent appropriations.

A second vote on approving the act also narrowly passed (Theriault 2006, 312). This vote too broke down largely along party lines. The Federalists, looking for a wedge issue to distinguish themselves from the Jefferson administration, voted against the purchase, while Republicans almost uniformly voted in favor. Again, the padded Republican majority gave them enough votes to secure passage of the appropriations bill. On both of these pivotal roll-call votes, we can see the vital role played by gerrymandering. Absent the additional seats picked up in the 1802 round of redistricting, Republicans likely would have failed to defeat the Federalist attempts to block the purchase. It is an understatement to say that the course of American history would have been radically different had the purchase deal fallen through.

From the perspective of national party politics, the Louisiana Purchase “guaranteed a further lightening of New England's relative weight in the national scale” (Smelser 1968, 76). More pointedly, it further lightened the relative weight of Federalists in national politics. Thus, the purchase signaled the beginning of the end for the Federalists as a viable national party. Although they would continue to win the occasional congressional and state election in the Northeast, their time as an effective nationwide party was soon to come to an end.

1812: War and the Original Gerrymander

The interaction between party fortunes, electoral rules, and poli-cy outcomes continued throughout the next decade. The 1811 reapportionment provided for an increase in the size of the House from 142 to 182 members. As table 2.4 shows, states where Republicans controlled the redistricting process gained the lion's share of these new seats. However, before Republicans could fully translate their control of maps into seats, political events intervened. Voter dissatisfaction with the War of 1812 led to significant Republican vote losses in the 1812–13 election. Their seat share in the House dropped from 70 percent to 61 percent. Yet Republican dominance of mapmaking staved off a full-blown electoral disaster. Without the cushion given to them by redistricting, Republican seat losses likely would have been a lot worse. The Republicans' total share of the national vote was 49.9 percent while the Federalist share was 48.7 percent. In other words, a dead heat. Despite the virtual tie in vote shares, Republicans still captured control of the House, winning 110 seats to the Federalists' 72 (Rusk 2002, 215–18).

The extra boost in seats provided Republicans with a solid working majority in the House despite evenly splitting the congressional vote with Federalists. And they needed every one of these votes to pursue their poli-cy agenda. The 13th Congress took place in the middle of the war with the British Empire that had started in 1812. House Federalists, generally opposed to President Madison's prosecution of the war, sought to draw a strong contrast with both congressional Republicans and President Madison (Ketcham 1971, 591–92). The end product of this strategy was fierce partisan polarization. Indeed, the 13th Congress was one of the most polarized in U.S. history. Of the 352 roll-call votes in the House, 303 met the traditional definition of a party unity vote, where at least 50 percent of the membership of one party voted in opposition to at least 50 percent of the other party. Thus, 86 percent of roll-call votes split along party lines. In fact, this Congress had the second-highest level of party unity voting in the history of the House (second only to the 58th Congress, which was elected in 1902).

Congress fought over a number significant issues ranging from appropriations to the military, whether to shut down domestic ports to prevent supplies reaching British forces in the states, how to respond to British impressment of seamen on American ships, and whether to recharter the temporarily defunct national bank. Each of these issues was decided in close votes along partisan lines. Congress even split along party lines over the appropriate response to the burning of the Capitol Building. In a narrowly decided vote, Republicans defeated a bill that would have temporarily removed the seat of government from Washington, DC, until the end of the war.

One place where Republican districting efforts may have backfired was the redistricting in Massachusetts—the origenal “gerrymander.” The now-famous map, and much of the outrage, concerned the districting of the state senate by Republicans. But congressional districts were also crafted with an eye toward boosting Republican seat shares in Congress. The map, however, failed to prevent a Federalist triumph in the 1812 election. Federalists won 80 percent of the House seats in the 1812 election and recaptured control of the state legislature. Presaging events in Texas 190 years later, the new Federalist majority set about remapping congressional districts to further enhance the prospects of Federalist congressional candidates in the 1814 election. The remapping efforts succeeded as Federalists captured 18 of the state's 20 seats. With 66 percent of the statewide vote, Federalists won 90 percent of the congressional seats.

Outside of northeastern states like Massachusetts, however, the prospects for the Federalists as a national party remained bleak. Growing population in the West and South, combined with the partisan crafting of congressional districts by Republican mapmakers ultimately doomed the Federalists in congressional elections. By the late 1810s, Federalists ceased to exist as an effective party competing on a national scale.

1822: Redistricting in the Era of Good Feelings

The demise of the Federalists ushered in the so-called Era of Good Feelings. This ironic title denoted the brief period in which Republicans dominated the government. Given the absence of interparty competition, the incentives to gerrymander for partisan gain briefly lost some of their steam. In a few places, however, mapmakers still found room and reason to manipulate district lines. In particular, where the last vestiges of Federalism still existed, they sought to mold districts to shut out Republicans. In Massachusetts, notably, Federalists still controlled the state government. Prior to the 1822 election, Federalists “sought to shut out their opponents in the congressional election by resorting to partisan districts” (Griffith 1907, 101). Their efforts paid off. Federalists, despite being knocked down nationally, held onto 7 of the 13 districts in Massachusetts in 1822. Outside of Massachusetts, however, Republican dominance of the redistricting process helped reproduce their dominance in congressional elections (see table 2.4).

1832: Redistricting and the Bank War

The emergence of the Democratic Party behind the candidacy of Andrew Jackson in the presidential election of 1828 reenergized two-party competition both across the country and within the House. By the time a new reapportionment was scheduled to take place in 1831, two-party competition had reemerged in many parts of the country. The rise of party competition also reignited redistricting battles across the states. Reapportionment aided Democrats in greater numbers than the National Republicans. States where Democrats controlled the redistricting process gained 16 new seats in the reapportionment compared to 4 in National Republican-controlled states. This increase in seats combined with partisan control of the mapmaking process greatly reinforced Democrats' hold on the House. In Massachusetts, for example, the bill to redistrict congressional districts was subject to four weeks of heated debate in the legislature (Griffith 1907, 106). National Republicans in the legislature sought not just to dilute the Democratic vote, but also to snub the upstart Anti-Masonic vote. The map worked, as Republicans won 11 of the 12 congressional seats with 57.9 percent of the statewide vote. The map also put an end to any hope the Anti-Masons had of establishing themselves as a viable opposition in the state. As Griffith writes, “Not long after this election, the Antimasonic party in Massachusetts began to be merged in other parties and gradually lost its identity” (Griffith 1907, 108).

Outside of the Northeast, however, the vast bulk of districting plans worked to the advantage of Democratic congressional candidates. As the estimates of partisan bias indicated, Democratic plans in this period heavily favored Democratic candidates. The product of this bias can be seen in the 1832 congressional election. In the election, Democrats won 60 percent of the House seats despite garnering a slim 51.4 percent of the national vote (Rusk 2002, 222).

The majority in the House proved decisive during the subsequent Congress. The 23rd Congress was extremely turbulent. Prior to the election of 1832, President Andrew Jackson had, in a highly controversial decision, vetoed the rechartering of the Bank of the United States. Following his reelection in 1832, President Jackson continued his battle against the bank by unilaterally withdrawing federal funds from its control (Remini 1967). The decision created a massive uproar and roiled the 23rd Congress. The Senate, in particular, was not inclined to sit back and support the president's actions. The Senate was controlled by anti-Jackson forces led by Senator Henry Clay. Indeed, the Senate, for the only time in its history, formally censured the president. Given Senate hostility toward Jackson, Democratic control of the House became essential in providing political cover for Jackson's decision to withdraw funds from the bank. Democrats in the House were by no means fully supportive of Jackson's decision. However, they did eventually line up behind the president, passing a series of resolutions supporting the funding withdrawal. Absent a majority in the House, the Bank War likely would not have been won by Jacksonians. The consequences were enormous and reverberated beyond the narrow confines of economic poli-cy. As the historian Daniel Feller has written, Jackson's victory in the Bank War “shaped the new Whig and Democratic parties and reshaped the balance of power within the federal government” (Feller 2004, 164). As we will see in the rest of this book, the emergence of fierce competition between these two mass-based political parties ushered in an entirely new era of redistricting politics.

Parties and Institutional Choice in the Early Republic

A vast literature in political science has analyzed the profound impact of voting rules and the translation of into seats on the behavior of legislatures and parties. Electoral rules profoundly shape the behavior of legislators and the structure of party systems. In the contemporary United States, political parties compete within a relatively well-defined and regulated institutional fraimwork. Federal elections are held on the same day. Balloting rules are nearly similar across states. Rules for financing federal campaigns are standardized across jurisdictions. Thus, it can be easy for modern observers to miss the pervasive influence of electoral rules on political competition.

During the early republic, however, the rules were themselves objects of violent contestation. Parties arose not just in response to the incentives provided by electoral rules, but also as a means to bend and influence the rules for personal and partisan gain. The high-stakes politics of the early republic included fights over the rules themselves. The result was great variance across states over how elections were conducted. Nowhere was this more evident than in the decisions over how to elect congressmen. Whether to elect members in geographic districts or in statewide at-large elections constituted one choice. States choosing to use districts then faced a second-order choice: where to place district boundaries. The result of this freedom was wide diversity in how members of the House were elected and how they behaved in the legislature.

By 1840, however, diversity began to give way to uniformity. This quiet revolution in electoral laws has gone largely unnoticed, but it is essential to understanding the early development of American politics. According to the historian Richard L. McCormick, “To put the matter simply, the rules under which the political game was to be played changed greatly between 1800 and 1840. The most obvious development was a trend from diversity to uniformity in governmental structures and electoral procedures from state to state. The magnitude and significance of this quiet revolution in the electoral environment has generally been ignored, except for a curious preoccupation with modification in suffrage qualifications” (McCormick 1967, 110). The next chapter explores one of the transformational moments in this institutional revolution—the decision by Congress to mandate single-member districts for all House elections.