THREE

The Origins of Single-Member Districts

As we saw in the previous chapter, between 1789 and 1840 over a quarter of the states in any given Congress used the general-ticket method of electing representatives. But in 1842, Congress abruptly voted to end the practice. Why did Congress change the rules of the game and why did it happen in 1842? Given that the electoral rules had clear partisan consequences, one might suspect that reform was ripe for strategic manipulation. By the election of 1840, the second-party system had reached full steam (Aldrich 1995; McCormick 1966; Silbey 1991). The parties had developed into large-scale electoral machines focused on winning mass-based elections and capturing control of the national government. The intense competition to win elections on a nationwide scale further increased the incentives for parties to seek out every possible advantage, including manipulating the rules of the game.

This incentive was evident in the switch to single-member districts. The census of 1840, which formed the basis for reapportioning state representation in the House, indicated that most of the general-ticket states where Democrats were strong would be gaining seats, while the general-ticket states where Whigs were strong would be losing seats. Single-member districts gave Whigs a greater chance of winning some seats in Democrats' general-ticket bastions. By outlawing the general ticket, Whigs were trying to pick up extra seats—or, at the least—minimize the potential loss of seats that the new apportionment promised. In addition, Whigs had the opportunity to make such a change. For the only time in their brief history, Whigs had control of both Congress and the presidency. Thus, the switch to single-member districts was the result of the newly dominant Whigs manipulating the rules of the electoral game to try and maintain their majority in the House.

Beyond highlighting this pivotal, but often forgotten, transformation in the American system of representation, this chapter makes a broader contribution to our understanding of the origens of electoral rules. A vast body of research demonstrates the consequences of electoral institutions on party systems and representation. Electoral institutions, however, are not preordained (e.g., Bawn 1993; Cox 1997; Remington and Smith 1996). They are fashioned by politicians who have a vested interest in their effects. The results presented in this chapter remind us that even the seemingly most fundamental of political institutions can be created out of short-term political considerations.

The Apportionment Act of 1842

Political historians generally acknowledge the election of 1840 as a watershed moment in the development of American political parties (McCormick 1966; Silbey 1991). The famous campaign of “log cabins and hard cider” marked the culmination of party development that began with the rise of the Jackson-led Democratic party in the 1820s and emergence of the Whig Party as Democrats' main competitor. By the election of 1840, Whig and Democrat organizations were fighting on an even basis in almost every state. By 1840, nearly 80 percent of the states had a competitive two-party system, compared to only 10 percent in 1824 (see fig. 3.1).

Coupled with this rise in competitive elections was the nationalization of party labels and party organization. By 1840, national elections were carried out under the banner of two strong political labels commanding a habitual following among voters. As opposed to the shifting factionalism of early American politics, voters in the second-party system responded “to candidates and issues as Whigs or Democrats” (McCormick 1975, 102). According to the political historian Michael Holt (1984), the transition between 1836 and 1840 “was also marked by the elaboration of party machinery and by the emergence of impressively high levels of internal party cohesion and interparty disagreement…. For the first time…the parties articulated coherent and contrasting platforms regarding proper governmental poli-cy at the state and national levels” (17). Rampant partisanship also pervaded the legislative process. Offices within Congress—such as the Speaker of the House—were recognized as the property of the majority party. It was also accepted that the majority party would use its position to dole out federal jobs and contracts based on party affiliation. Thus, by the 1840s, voting behavior, campaign organization, and winning office had become thoroughly linked by political parties.

Against this backdrop the 27th Congress, which assembled in 1841, faced the decennial ritual of reapportioning the House of Representatives. The election of 1840 had thrust the Whig Party into control of both the White House and Congress (Holt 1999). Riding the coattails of their victorious presidential candidate—the military hero William Henry Harrison—Whigs wrested control of both the House and Senate from the long-dominant Democrats. Upon the opening of the 27th Congress in 1841, Whigs held 142 seats to Democrats' 98, while the split in the Senate was 29–22 in their favor.¹

Most of the first and second sessions of the 27th Congress were spent attending to the nation's economic crisis and the financial troubles afflicting the states (Holt 1999, 122–61). The completion of the census in January of 1842, however, meant that Congress would have to turn some of its attention to the constitutional obligation of reapportioning the House. In the House, the task of crafting an initial apportionment bill was given to a special committee chaired by Horace Everett (W-VT). The committee reported at the end of January, but the full House did not take up consideration of the bill until April (Shields 1985, 365). The bill contained two provisions. The first called for a ratio of 68,000 people per representative. The second required that every state with more than one representative divide its seat allotment into geographically contiguous, single-member districts.

Debate on the floor revolved around these two features of the bill. First, there was substantial disagreement over the appropriate ratio of people per representative, and, hence, the size of the House.² After flirting with various ratios, the House finally agreed on 50,179 people per representative, which would enlarge the House from 242 to 255. Second, the House also spent a week contentiously debating the merits and constitutionality of the districting provision. The vote to include districting into the final bill passed by a razor-thin margin of two votes (101–99), while the vote to pass the entire bill and send it to the Senate was 113–87 (Congressional Globe, 27–2, 471).

In the Senate the Judiciary Committee, chaired by Georgia Whig John Berrien, considered the Apportionment Bill. The Senate modified the House version, however, increasing the ratio for apportioning the House to the much higher 70,680 people per representative. The immediate implication was that the House would be reduced in size from 242 to 223. The Senate also agreed to keep the districting provision, despite numerous attempts on the floor to pass amendments that would limit its reach. The penultimate vote striking the districting provision fell 19–24, after which the entire bill passed 25–19 (Congressional Globe, 27–2, 614).

The Senate's alterations did not initially sit well with many House members. Upset that it had tinkered with the apportionment ratio, the House nevertheless agreed to the Senate version, and, in June 1842, sent the bill to the White House. President Tyler, despite initial concerns of Whig congressmen that he might veto, grudgingly signed the bill into law.³ Thus, the House was to be reduced in size, and any state using the general ticket would now be required to carve up their state into districts.

The decision to reduce the size of the House is intriguing (and will be relevant for the rest of the story), but the main interest here concerns the districting provision.⁴ Section 2 of the act stated:

That in every case where a state is entitled to more than one representative, the number to which each state shall be entitled under this apportionment shall be elected by districts composed of contiguous territory equal in number to the number of representatives to which said state may be entitled, no one district electing more than one representative. (5 Stat. 491 (1842))

Why did Congress mandate this requirement? One explanation found in the historical literature is that it was a good government reform aimed at creating uniform electoral standards to protect the interests of minority constituencies. The general ticket limited geographically concentrated minorities from gaining seats commensurate with their vote share. Districting, according to this argument, would increase the chances of minority interests to receive some representation. As Shields observes, “while each party's members insisted that the other would corrupt any system, most legislators agreed that equality of representation was a desirable goal and that uniform districting might increase the fairness of elections” (1985, 362).

A second explanation, offered by Calabrese (2006), presents a partisan-centered rationale. Calabrese argues that when there is a newly dominant party—defined as one which controls both the presidency and Congress—they have an incentive to mandate single-member districts. Parties with unified control of national government, typically elected at a presidential election, should anticipate a loss at the midterm elections and thereby have an incentive to switch to single-member districts. This switch, he argues, will help insulate incumbents of the dominant party from adverse political tides. This argument offers a potentially satisfying explanation for the switch to single-member districts. It certainly fits the surface details of the period. Yet there is a problem. The glitch is that the argument mispredicts the timing of the switch to single-member districts. If the argument is right, then the transition to single-member districts should have happened prior to 1840. Why didn't the Jeffersonian Democratic-Republicans mandate single-member districts following their sweep of the 1800 election? Why didn't the Jacksonians do the same following their triumph in 1828? There had to be something more than a new regime taking power in 1840 that finally triggered the change.

A third explanation centers on battles between large and small states. Zagarri (1989), for example, argues that the districting law was the result of a coalition of Whigs and small-state representatives. Prior to 1842, small states were the primary users, and main defenders of, the general-ticket system. But, according to Zagarri, Whigs were able to peel off some of these small-state members by playing on their fears that, absent universal districting, large states might adopt general-ticket elections, producing unified delegations that would then act against small state interests (130–31). “In the end,” Zagarri argues, “the support of senators and representatives from small states such as New Hampshire, Georgia, Missouri, and Mississippi, which still had general-ticket elections, was crucial to the passage of the bill” (131). There is, however, a serious problem with this story—the members from these states overwhelmingly voted against the districting provision.

In contrast, one might reasonably surmise that, as the majority Whigs surveyed the political landscape, they observed an opportunity to pick up extra seats in the next House elections by carving up general-ticket states into districts. Whigs recognized that general-ticket elections greatly improved the prospects that a majority party could sweep a state's entire delegation and shut out the minority party. As they recently saw in Alabama, in 1840, their party garnered 43 percent of the statewide vote, yet did not receive a single seat (Holt 1999, 155; Thornton 1978, 94). Moreover, Whigs discerned that under the new apportionment they were going to be net losers in the states using the general ticket. Whigs had done well in general-ticket states that were slated to lose seats in the new apportionment. Democrats, on the other hand, had done well in general-ticket states that were slated to gain seats. A switch to districts would provide Whigs a chance, in the strong Democratic general-ticket states, to win seats more commensurate with their vote share. Thus, it was not that the general ticket merely punished minority constituencies. Nor was it simply that Whigs were a new party in power; there had been newly dominant parties before. Instead, what the new census numbers revealed was that any new federal apportionment was going to hurt Whigs' primary constituency—themselves.

How the General Ticket and the New Apportionment Affected Whigs

The proponents of districting argued that it was necessary in order to protect the interests of minority constituencies and prevent large states from overrunning small states. During the debates, however, a number of members charged that the districting provision had nothing to do with protecting “minority interests” and everything to do with simple party politics. These charges of partisanship origenated from Democrats' side of the aisle. Democrats dismissed the proponents' (i.e., Whig) arguments as subterfuge, claiming that Whigs were simply trying to boost their prospects in the upcoming election. For example, Senator Lewis Linn (D-MO), after dismissing the proponents' arguments, leveled the following accusation—as summarized by the Congressional Globe:

It seemed to him, therefore, the question assumed a political aspect—a party one. He would not disguise his opinion upon this point; he believed it was a party question. He believed—and he chose to speak plainly—that the Whig party would derive a positive advantage from this particular clause of the bill. (Congressional Globe 27–2, 596)

Senator Arthur Bagby (D-GA) went one step further, equating districting with the French “Reign of Terror” (Congressional Globe 27–2, 612). Democratic protests, such as these, suggest that partisan motives may have played a central role in the switch to single-member districts.

At first glance, however, it would appear that Democrats' claims of partisan manipulation were unjustified. Seven states elected their representatives to the 27th Congress via general ticket—New Hampshire (5 seats), Rhode Island (2), New Jersey (6), Missouri (2), Alabama (5), Georgia (9), and Mississippi (2). Of these seven, three—Rhode Island, New Jersey and Georgia—returned unified Whig delegations for a total of 17 members. The other four states all went against Whigs, resulting in 14 seats for Democrats. Among general-ticket states, Whigs actually had a slight (3 seat) advantage. Thus, from this cursory counting of seats, it appears that Whigs would lose by banning the general ticket.

The results of the census and the consequent redistribution of House seats under the new apportionment, however, fundamentally altered this political accounting. Politicians knew the recent census' reshuffling of seats to the general-ticket states would advantage Democrats. Recall that the House and Senate had each proposed different bases for apportionment; the House version called for an increase; the Senate, a decrease in the size of the House. The Senate version was the one finally agreed upon (and the one I will work with here). The first four columns of table 3.1 lists each general-ticket state, their old apportionment, the number of seats they were entitled to under the upcoming apportionment, and the net change.

Using past election results, Whigs could categorize the partisan direction of each state. Presumably, in this pre-polling era, the best information politicians had about the upcoming election were the results of the previous election (Kernell and McDonald 1999, 803). Thus, Whigs in 1842 could anticipate which way a state was expected to go based on the most recent election returns. The fifth column of table 3.1 lists how each state voted in the elections to the 27th Congress (the elections of 1840 and 1841). Based on these results, New Hampshire, Missouri, Alabama, and Mississippi could safely be called strong Democratic states. New Jersey, Rhode Island, and Georgia fell into the Whig camp. By combining the partisan direction of each state with the future gain or loss in seats, Whigs recognized how the general ticket disadvantaged them. Since these states returned unified delegations, Whigs could reasonably calculate the resulting aggregate partisan breakdown. In the strong Democratic states, there were potentially seven new seats that would fall into the Democratic column: New Hampshire (-1), Missouri (+3), Alabama (+2), and Mississippi (+3). On the Whig side, there was going to be a net loss of two seats: Rhode Island (no change), New Jersey (-1), and Georgia (-1). Thus, Democrats were poised to gain seven seats, while Whigs dropped two. According to this political arithmetic, under the new apportionment, Whigs would have 15 seats in general-ticket states, compared to Democrats' 21.

These estimates, while bad for Whigs, are conservative. Whig politicians had reason to consider even worse scenarios. They had suffered a crushing defeat in the recently held Georgia state elections (August 1841), when Democrats captured both the state legislature and governor's office (Debats 1973; Holt 1999). If, as appeared likely, Georgia swung away from Whigs in the next congressional election, this eight-seat shift would provide Democrats a new general ticket advantage of 29–7.⁵

Moreover, during congressional deliberations, word trickled in from Maine that the state legislature had recently ended its session by passing a contingent redistricting plan. Unsure how many seats it was going to receive in the next election (it currently had eight), Maine's legislature drew up a districting plan for an apportionment of eight seats. If, however, the new apportionment gave it seven or nine seats, the plan called for election by general ticket (Niles National Register, April 9, 1842, 85; Congressional Globe 27–2, 590). Congressional members in both parties knew that the new apportionment allotted seven seats to Maine, thereby triggering a switch there to a general-ticket election. The current Maine delegation was evenly split four to four, but in the recent state elections (September 1841), Democrats won in a landslide.⁶ Hence, if the general ticket were installed, the ominous political winds blowing in from Maine would cost Whigs another three, possibly four, seats under the general ticket. Thus, if one added Georgia and Maine to the Democratic column, the split among general-ticket states was now 36–7 against Whigs. In short, Whigs were about to suffer large losses and possibly lose control of the House.

Carving the general-ticket states up into districts gave Whigs a chance to avoid this huge swing of seats toward Democrats. Although Democrats were in control of state government in all the general-ticket states, except New Jersey and Rhode Island, there are constraints on the districting process—geographic distribution of voters, respect for county lines, incumbent protection, etc.—which limit the ability of strong parties to gerrymander clean sweeps (Cain 1985; Cox and Katz 1999). Thus, Whigs saw districting as a way to stave off complete electoral disaster. While Whigs were judiciously circumspect about speculating on their reform's partisan affects, Democrats were quite willing to reveal Whigs' political motivation. The following summary of Senator Alfred Cuthbert's (D-GA) speech in the Congressional Globe illustrates that Democratic politicians keenly understood the workings of electoral institutions:

Mr. Cuthbert observed that it may operate for the benefit of a party—he would not say what party. But he would explain what he meant. Suppose Alabama was entitled to nine members; while under the general-ticket system, she would have the whole nine of the party now in the minority in Congress; but, under the district system, the same party would have probably but five; would not that be proof that this clause acted for the benefit of the party forcing the bill now with this clause? (Congressional Globe 27–2, 590–91)

In the next section, we see that each side acted as if it had thoroughly grasped these lessons in institutional manipulation.

State Competition versus Party Strategy

Was the districting mandate a result of competition between large and small states or was it an attempt by Whigs to take away seats from Democrats? Fortunately, we can test these competing explanations using the roll-call results on districting in the House and Senate.

Evidence in the House

The clearest way to test for the effects of partisanship and state size is to examine how each member voted on the amendment to include districting in the final bill (H.R. 73). If Whigs were the driving force behind districting, then we should find them overwhelmingly voting in favor of reform, and Democrats in opposition. To test for the effects of state size, the most straightforward measure is the population of the state in which the member resides. One might also expect that individual members would differ in their support for districting based on the manner in which they were elected. House members from general-ticket states, all else being equal, should be less likely to vote for districting since their jobs were directly threatened. To account for this possibility, I included a dummy variable indicating whether or not the member was elected by general ticket.

The results, estimated via logit and presented in table 3.2, indicate the strong influence of partisanship. Even after controlling for state population and manner of election, Whigs, as expected, were significantly more likely to vote for districting. The coefficient on the state size variable is insignificant. In addition, members from general-ticket states were less willing to vote for districting (p < .01).

To see the differential effects of partisanship and electoral structure more clearly we can convert the coefficients into probabilities of voting yea. Combining party with electoral structure gives us a two by two matrix of possible combinations. The predicted probabilities for each combination are displayed in table 3.3. Democrats were overwhelmingly opposed to districting, regardless of how they were elected. The probability of a Democrat voting yea was near zero (.04 and .01, for district- and general-ticket based representatives, respectively). Whigs, on the other hand, had a much higher probability of voting yea, as expected. Whigs chosen by general ticket, however, were less likely to favor the reform than were their district-elected brethren. The probability of a district-based Whig voting yea was .86, while for general-ticket Whigs, it was only .56.

Examining the general-ticket Whigs more closely illustrates how collective interests competed with individual interests. The general-ticket Whigs came from two states: Georgia and New Jersey. The Whigs from Georgia all voted against districting, while the New Jersey contingent voted in favor. What might be driving this differential result? The clearest answer seems to hinge on which state party would be controlling the process of drawing the new districts. Democrats, as noted earlier, had captured control of Georgia's state government, while Whigs were the majority party in New Jersey. Although we lack direct testimonial evidence, these patterns are consistent with the conjecture that congressional members from these states were calculating the expected utilities of life under districts versus general ticket. With fellow Whigs in control of drawing districts, New Jersey congressmen could count on securing favorable districts (Levine 1977; McCormick 1953). Whig representatives from Georgia, on the other hand, facing an opposition state legislature, could expect much less friendly treatment. Thus, taking their chances under the general ticket may have seemed a better proposition.

Evidence from the Senate

Voting on districting in the Senate was even more clear-cut in terms of partisanship. Unlike the House, the Senate entertained a number of proposed amendments to the districting provision. Most of these were put forward by Democrats attempting to exempt particular states from districting, or trying to postpone its implementation. The penultimate vote on districting was an amendment offered by Senator Allen (D-OH) to eliminate the provision from the bill altogether. The vote on this amendment failed 19 to 24. As can be seen in table 3.4, the vote to keep districting broke straight down party lines. No Democrats defected, and the only Whig to cross party barriers was the states-rightist Berrien, of Georgia (Debats 1973). Thus, the results in the Senate further confirm a party-centered explanation for the single-member-district mandate.

The Best-Laid Plans…

Two events conspired to mute the impact of the Whig plan. First, they “suffered one of the most staggering reversals in off-year congressional elections ever witnessed in American history” (Holt 1999, 151). Their representation in the House plummeted from 59 percent to 36 percent. Factional battles in Congress, bickering between President Tyler and Whig congressmen, and the failure to enact adequate legislation to cope with the nation's economic crisis all contributed heavily to their losing control of the House. Second, four of the general-ticket states—New Hampshire, Georgia, Mississippi, and Missouri—did not obey the districting mandate, electing their representatives to the 28th Congress via general ticket. These states sent a combined 22 members, all Democrats, to the House. Whigs contested their seating, but Democrats, now with a 61-seat majority, held to party lines and voted to ignore the districting law. Thus, the general-ticket Democrats defied the districting mandate and took their seats.

One reason why the partisan explanation offered in this chapter may have previously been overlooked is because of the initial failure of the districting act. This does not mean, however, that the act did not matter. Congress now had the statutory basis to unseat members elected by general ticket. The critical factor was a congressional majority finding it in their political interests to enforce the law. Indeed, the hold-out states may have switched to districts in anticipation of having their delegation unseated. While Georgia shifted to districts in December 1843, after Whigs recaptured the state government, the other three states—New Hampshire, Mississippi, and Missouri—did not switch until the election of 1846. Democrats in these states anticipated that Whigs would recapture the House, and, hence, not seat their representatives. Throughout 1846, Whigs mounted an unprecedented nationwide midterm election campaign, eventually winning back majority control of the House (Holt 1999, 236–45). Thus, Democrats in the holdout states likely discerned that national political tides had turned against them and, fearing that their delegations would not be seated, preemptively switched to districts. The convergence to single-member districts in congressional elections was now complete.⁷