Chapter 1 Why what Happened in Montana won't Stay in Montana
Montana witnessed a Senate campaign for the ages in 2012. That race itself was a singular event and the state certainly is unique. But the record of that campaign, and the insights it provides, are applicable to other competitive races, whether for House or Senate seats and irrespective of location. Through careful study of this campaign, we gain a better understanding of the politics of the American West and of the substantial political clout it has gained. And, we can see more clearly how members of Congress understand the process of representation and its electoral consequences.
Incumbent Senator Jon Tester was elected by a razor-thin margin in 2006 and was considered to be one of the most vulnerable Democrats running for reelection. Although characterizing Montana as a Republican state is overly simplistic, voters frequently cast their ballots for Republican presidential candidates. In 2008, Barack Obama may have had the best showing of a Democratic presidential nominee since Bill Clinton, who eked out a win in 1992, but the newly elected president quickly became unpopular among Montanans.1 The Republican nominee was expected to win the state by double digits in 2012. The Tea Party-fueled midterms in 2010 that brought Republicans a stunning 60-seat gain and a majority in the House of Representatives coincided with similar gains for Montana Republicans. Democrats lost 18 seats in the lower house of the state legislature, becoming a tiny minority in the chamber.
Two years before Election Day 2012, internal Tester polls showed the senator losing to Congressman Denny Rehbergâthe widely anticipated Republican nomineeâby more than eight points.2 Tester was in troubleâand if Tester fell, the Democratic majority in the U.S. Senate likely would follow. In February 2011, Rehberg, who had served as the state's lone congressman since 2001, announced his candidacy. This prompted respected national political prognosticators Charlie Cook, Larry Sabato, and Stu Rothenberg to rate Tester as âextremely vulnerable.â All considered the race a toss-up.3 Many months later, in his final election prediction for the New York Times, election forecasting guru Nate Silver gave a 66 percent chance of a Republican Senate victory in Montana.
But Silver would be wrong. Jon Tester defied nearly everyone's expectations and beat Republican Denny Rehberg by nearly four percentage points while Republican presidential candidate Mitt Romney handily carried the state by thirteen. Out of the thirty-three Senate races in the 2012 cycle, Montana was only one of two that Silver predicted incorrectly. Why were Silver and so many other prognosticators wrong? How did Tester survive? Finding an answer to that puzzle is one of the reasons I wrote this book.
It is not the only reason. Through an in-depth analysis of the Montana Senate race, we learn a lot about congressional campaigns, elections, and the process of representation. One might ask whether we can learn anything generalizable from one race unfolding in such a distinctive place. Admittedly, Montana is unusual, as any Montanan will proudly tell you. While it is the fourth-largest state geographically, it is sparsely populated (and only one of seven states with just one member serving in the House of Representatives). Montana also is one of the poorest states in the country and has little ethnic or racial diversity. The two largest minority groups in the countryâLatinos and African-Americansâtogether make up a mere fraction of the state's population. In many ways, Montana appears to be an outlier from which one could not make broad, sweeping conclusions about political phenomena.
But as unique as Montana is, the stateâand this particular campaignâis exactly the place a student of politics should study to look for widely applicable conclusions about electioneering and representation at the dawn of the 21st century for four reasons. First, because the race was competitive from start to finish, the 2012 Montana Senate campaign sheds light on whether and under what conditions campaigns affect electoral outcomes. Second, the rise of the West politically throughout the 20th century has made the region critical to the balance of political power nationally while simultaneously altering the content of the national political debate. Third, Montana is no different when it comes to how members of Congress craft and communicate representational styles to constituentsârepresentational styles which are central in the campaign narratives of all congressional campaigns (and especially meaningful in competitive campaign environments). Finally, because I enjoyed unprecedented access to both candidates during the campaigns, I saw campaigns in a way few political scientists do. I travelled with the candidates, watched them interact with constituents, interviewed their staff members after the campaign, and had access to some of their internal data. As a result, we get an up-close-and-personal look at the campaign and its issues.
Do Campaigns Matter? Montana Votes âYesâ
Let's begin by acknowledging a larger debate among political scientists that has beset the discipline since its very beginnings: Do campaigns matter? Some of the earliest research suggested âno.â These studies of voter behavior indicated that casting a ballot was mostly a function of a person's partisan identification and socioeconomic status.4 Given that these characteristics are immutable, campaigns really didn't matter much because voting was a reflexive act. Add more recent scholarship indicating that partisanship actually acts as an information filterâindividuals seek information conforming to their preexisting views and discard information contracting those beliefs5âand some political scientists quickly conclude that campaigns are full of sound and fury, signifying nothing.
Alternatively, the authors of The American Voter, who identified partisanship as a critical determinant of voter behavior, addressed a puzzle contradicting this theory. In 1952 and 1956, the electorate was overwhelmingly Democratic, and yet Republican Dwight Eisenhower won two presidential races decisively. Something about Eisenhower caused Democrats to abandon their partyâat least in the short termâfor someone carrying the other party's banner. The authors attribute Eisenhower's appeal to his âprojecting a strong personal image and stressing the foreign policy concerns of voters. This allowed less attached and less interested Democratic adherents to overcome their predispositions and cast a vote for Eisenhower.â6 In short, the Republican campaign for Eisenhower mattered. Other scholars have found important campaign effects on the ability of voters to identify and recall candidates, the ways in which candidates and parties are evaluated, increasing voter turnout, and how accurately voters recall the ideology and issue positions of a candidate.7 This book sheds light on this larger debate among political scientists about campaign effects, concluding that campaigns matter.
A related reason to look at one Senate race in one stateâeven if that state may not be representativeâis that if we believe that campaigns can matter, they should matter the most in a competitive race. The fact is that campaigns often do not matter in many congressional elections because most congressional races are no longer competitive. A singular feature of the modern electoral landscape is the disappearance of so-called marginal congressional districts, a phenomenon first noted by David Mayhew in the early 1970s. Mayhew wrote that between 1956 and 1972, the number of contested House seats with incumbents running for reelectionâwhere the Democratic Party vote share is between 45 percent and 55 percentâdeclined by roughly half.8 He argues that one factor for the decline is the increasing value of incumbency; political scientists spent the next two decades measuring incumbency's precise electoral value.9 When a House incumbent chooses to run for reelection, they win 95 percent of the time throughout the postwar period. According The Cook Political Report, the number of swing congressional districts declined 45 percent between 1998 and 2013, from 164 to just 90.10 A substantial minority of House members run uncontested each cycle, while another sizeable percentage never face a quality, well-funded challenger (defined as a candidate who has successfully run for and served in an elected office before running for Congress).
But Senate incumbents are not as well-protected as House members. They represent states, not districts, so they are not subject to the redistricting that increases electoral security. Senators are also more likely to draw a quality challenger to run against them. Even so, they enjoy a substantial incumbency advantage. From 1946 through 2012, Senate incumbents choosing to run for reelection who reached the general election won reelection 82 percent of the time. On average, those incumbents won with 63 percent of the vote; only 37 percent of Senate incumbents received less than 55 percent of the vote in the previous election. Forty-five percent of Senate incumbents in the postwar era do not even draw a quality challenger.11
In noncompetitive congressional races, voter partisanship and the incumbent's greater name recognition yield the almost universal outcome of the incumbent winning reelection regardless of the campaign effort the incumbent undertakes. But these situations are not a true test of the hypothesis that campaigns matter, because the incumbent dominates the information environment. To better test the proposition that campaigns affect election outcomes, we must look at races where the information environment is up for grabs, such as open seat races or races where the incumbent faces a competitive challenge from a quality candidate.
Congressional incumbents actively discourage the emergence of competitive challengers, so races where they are at an informational disadvantage are rare. Incumbents establish a veneer of invulnerability by constructing massive financial war chests and winning elections by substantial margins. These two factors signal to would-be candidates that most challenges would be costly and unsuccessful. The best candidatesâthose who have run and won a campaign for elected officeâoften chose not to run against incumbents because the costs of entry are high, and the return on investment low. Open seats represent the obverse scenario: a lower cost of entry with a higher potential payoff. As a result, the best quality candidates often choose not to challenge incumbents, waiting instead for the incumbent to retire. The end result is that most incumbents breeze easily to reelection while open seats are often expensive, hard-fought affairs where either party's nominee could win. Campaigns can and often do matter in open seat races.12
But incumbents can and do lose, and they are much more likely to do so when they present a target of opportunity for a strategically minded challenger. Incumbents embroiled in scandal, as well as those who are perceived as out of touch with constituents, have health problems, or simply have not created a large reelection constituency shatter the cloak of invulnerability. These situations draw quality challengers to run against weakened incumbents. One way to systematically identify incumbent vulnerability is to classify incumbent campaigns as competitive or not by using three factors: the incumbent's previous vote share, whether the challenger they face in the general election is âquality,â and whether the incumbent represents a state that was won by their party's presidential candidate in the last presidential election.
I examined each Senate election featuring an incumbent between 2000 and 2012 and coded a race as âcompetitiveâ if the incumbent won the previous election with less than 55 percent of the vote and faced a quality challenger. I also coded a race as competitive if the incumbent's presidential party did not win the state in the most recent presidential election prior to the incumbent's reelection bid. This yielded 79 competitive Senate races with an incumbent candidate running for reelection out of a total of 186. Of the 107 races defined as noncompetitive, the incumbent only lost three timesâa reelection rate of 97 percent. In competitive races, 23 incumbents lost for a reelection rate of 71 percentâ11 percentage points less than the postwar average of 82 percent.13
The Montana Senate race between Jon Tester and Denny Rehberg in 2012 was, by this definition, a competitive race. Rehberg, as the state's lone congressman, was certainly a âqualityâ challenger. Democrat Tester won his initial campaign in 2006 with only 49 percent of the voteâand, in fact, beat incumbent Senator Conrad Burns by less than 3,600 votes. Finally, Republican presidential nominee John McCain won the popular vote in Montana by a little more than two percentage points in 2008. The most important question for our purposes, however, is whether the Montana Senate race as a case is representative of the larger population of competitive Senate races. If so, it is much more likely that we can apply the lessons learned from this single case to competitive Senate races (and to the bigger question of whether campaigns matter). If it is not, then we have an interesting story to share but not much else.
In three important respects, the 2012 Montana Senate race is a typical example of a competitive Senate campaign. The average Senate incumbent spent $9.2 million dollars in competitive races between 2000 and 2012. The average challenger? $5.4 million. In Montana, incumbent Tester spent a little more than $13...