Abstract
Despite a long history of nonpartisanship, the Nebraska state legislature has polarized rapidly within the past decade. Using interviews and campaign finance records, we examine politics in the modern Unicam to investigate nonpartisan polarization. We find that newly instituted term limits created opportunities for the state’s political parties to recruit and finance candidates in an increasingly partisan fashion. Social network analysis suggests that there is a growing level of structure to campaign donations, with political elites increasingly less likely to contribute across party lines. The results offer a compelling example of parties overcoming institutions designed to eliminate them.
Keywords
Introduction
Nebraska’s legislature is polarizing rapidly—more rapidly than any other state or federal legislative chamber in the nation. A claim of rapid party polarization raises few eyebrows in modern American politics, of course, as such findings have become commonplace in studies of the Congress and state legislatures. Such a claim warrants particular scrutiny, however, in Nebraska, officially the only nonpartisan state legislature in the land. How does partisanship emerge in a legislature where parties do not officially exist?
In this article, we examine the recent political history of the Nebraska state legislature through a series of interviews and analyses of campaign donations. We find that the imposition of term limits in the middle of the last decade helped to instigate greater partisan polarization within the chamber. With at least a quarter of the chamber now regularly being turned out of office, a handful of partisan actors, including the Republican governor and Democratic political operatives, have become intensely involved with recruiting partisan candidates to run for office and keeping them faithful to a partisan agenda once in office. We compare campaign donation patterns in Nebraska with those of other states with similar financial rules and find Nebraska to have experienced a relatively sharp rise in donor partisanship during this time period. The evidence suggests that Nebraska’s legislative political system—once disorganized and unstructured—has developed a more rigid structure in recent years, and this structure is evident both within the chamber and in the elections that send people there. The findings reinforce the notion that nominations are the key to party control.
Who Runs the Party?
An influential strain in the literature on political parties suggests that parties are the creatures of politicians (Aldrich 1995, Cox and McCubbins 1993, 2005, Schwartz 1989). Ambitious legislators build parties because they recognize their value in developing a coherent public reputation, turning out voters, passing legislation, and fulfilling their campaign promises. Politicians understand political life to be rife with collective action dilemmas that threaten their own careers; parties are a solution to these problems.
Such arguments are certainly persuasive but run into difficulty outside the Congressional context in which they were developed. For example, under a system known as cross-filing (1913–59), California politicians were free to run in as many party primaries as they wished without their party label appearing next to their name on the ballot, which resulted in a dramatic drop in legislative partisanship. Interestingly, oral histories and roll call records suggest that not only did legislators tolerate this weakly partisan system quite well, but they also actually protected it. They had a choice between programmatic government and an easy legislative life largely lacking acrimony and chose the latter (Masket 2009). As we explain below, Nebraska has largely followed that same path since the advent of its nonpartisan legislature in 1937, at least until recently.
Other studies have found that parties are pursuing their own agendas and not those of their incumbents (Galvin 2012). Recent history finds examples of parties seeking to enhance their ability to punish errant lawmakers (Van Oot 2011) or to winnow the field of candidates in competitive races (Dominguez 2005; Masket 2011). Increasingly, it seems that politicians are the creatures of parties, rather than the other way around (Bawn et al. 2012). Such cases are probably better explained by a theory of political parties that focuses on nominations as the source of party control (Schattschneider 1942, 64). Recent work by Cohen et al. (2008), Bawn et al. (2012), and Masket and Noel (2012) expands on this theory, arguing that when a group of ideologues determines party nominations—whether by recruiting a candidate, manipulating a convention, or strategically deploying campaign resources—the resulting officeholders will behave in a partisan manner. They will do so both because they have been selected for such behavior and because it is in their career interests to do so; incumbents who disappoint their nominators credibly fear facing a well-funded, well-endorsed challenger in the next primary. Conversely, when partisan actors are prevented by institutional rules from influencing party nominations, legislators will tend to sideline partisan concerns and increasingly focus on their general election constituency.
If legislators are truly most responsive to whomever can most easily remove them from office, then we should consider the ideological preferences of these constituencies if we want to understand legislative behavior. The typical general election voter follows politics only sporadically and very rarely can recall much information about congressional candidates, much less state legislative candidates. Conversely, those who are pivotal in making nominations tend to be highly aware of politics and consistently ideological in their preferences. It follows that when legislators are left under the supervision of the general electorate only, legislative politics will be relatively unstructured, with members voting their constituencies and adhering to whatever temporary coalitions may suit their reelection needs. However, when partisan actors control nominations, legislative politics will tend to be organized in a more partisan fashion.
This certainly does not mean that any change in institutional rules that disadvantages party elites will result in depolarized politics. Party elites have proven themselves to be highly adaptable to changing institutional conditions. As a result, shifts in rules about who may participate in primaries have had an impact on the partisan behavior of incumbents that, while expected to be dramatic, has instead ranged from modest (Gerber and Morton 1998; Kanthak and Morton 2001) to imperceptible (McGhee et al. 2014). But an institutional rule change that deprives or provides voters with a fundamental piece of information about the candidates on their ballots (such as their party affiliation) is expected to have a more dramatic effect.
The state of Nebraska provides a useful context for a test of theories about the role of partisan election rules. The state shifted from a partisan to a nonpartisan legislature in 1937, leading to a decline in legislative partisanship. In recent years, however, the adoption of term limits has spurred new electoral involvement by the state’s political parties, increasing the role of partisanship in state politics. We explain these changes below.
A History of Nonpartisanship
Only two American states have experimented with nonpartisan legislatures: Minnesota from 1913 to 1973 and Nebraska from 1937 to the present day. Evidence from Minnesota suggests that while the adoption of nonpartisanship did initially have a profound effect on legislative voting behavior, legislators nonetheless began to form “Liberal” and “Conservative” caucuses by the 1950s, voting in blocs that closely corresponded with the state’s two major political parties. Minnesota’s legislature adopted partisanship in 1973 at least in part because the nonpartisan tradition had become a fiction (Manning 2004, Mitau 1960).
Nebraska’s history was different, however; voters instead of legislators established the nonpartisan legislature there. U.S. Senator George Norris (R-NE) and others had made the case that the state’s bicameral, partisan legislature was needlessly expensive and incompetent—unsuited for dealing with the state’s difficulties during the Great Depression. Despite the opposition of state party leaders, over 90% of Nebraska’s counties and precincts gave the initiative majority support in 1934 (Sittig 1986). The use of direct democracy in Nebraska’s case made it impossible for legislators to reinstate partisanship without the voters’ approval. Unlike in Minnesota, parties remained virtually nonexistent for decades. Welch and Carlson (1973) found that the establishment of nonpartisanship in 1937 led to an enduring decline in legislative partisanship.
Under state laws, the nonpartisan system affects every aspect of legislative business. There is no official majority or minority caucus in the legislature (known popularly as the Unicam), nor are there whips or party leaders. The Speaker is elected by his colleagues through a secret ballot, as are the chairs of all the standing committees. The 49 senators are elected to staggered four-year terms via nonpartisan ballot. The primary election is an open, blanket contest, with all eligible candidates for a legislative seat appearing on every ballot regardless of the voters’ or candidates’ party registrations. The ballots are constitutionally forbidden from containing information about the state legislative candidates’ affiliations with political parties. The top two winners of the primary go onto a similarly nonpartisan November runoff election, even if they are registered with the same political party.
Just because the ballots lack evidence of party affiliation does not mean that it is impossible to determine the partisan leanings of legislative candidates in Nebraska. Newspaper and online coverage of campaigns often refers explicitly to the party affiliations of the candidates, and researchers and political activists can readily learn this information from the media, state legislative voter records, and a roster published by the state government. Still, the typical voter is less likely to pay such information collection costs when making a choice between state legislative candidates, and it seems reasonable to say that many voters, who see party labels on neither the primary nor the general election ballot, have little idea which candidates belong to which parties. Such institutional rules certainly do not favor partisan behavior among legislators.
However, recent roll call evidence suggests that the Unicam is polarizing. In their innovative study of state legislative roll call votes over the past two decades, Shor and McCarty (2014) place state legislator ideal points in a common space so that we can compare them across states and years. They find that despite Nebraska’s nonpartisan charter, its chamber does not stand out as unusually depolarized. Quite the opposite, actually: according to Shor and McCarty’s measures, Nebraska’s Unicam has polarized more quickly—just about doubling the distance between unofficial party medians—over the past decade and a half than any of the other 98 state legislative chambers or even the U.S. House and Senate.
Figure 1 shows a trace line of legislative polarization (measured as the inter-median distance between the legislative parties) for every state legislature since 1993. The Nebraska Unicam is highlighted in the figure. As can be seen, Nebraska actually saw the lowest level of polarization in the nation back in the mid-1990s, but it has quickly surpassed many other states, with a dramatic increase in polarization coming in the second half of the last decade. In 2010, it was about as polarized as Kansas, the state against which it fared so poorly in terms of legislative partisanship just a decade ago (Wright and Schaffner 2002).
Polarization of state legislatures over time.
How is this possible? Party leaders and activists have made no secret of their desire to reinstate partisanship, but roughly a score of formal attempts to do so over the past 75 years have failed to secure sufficient votes from legislators or signatures from voters (Sittig 1986). Even if party activists want to see more ideologically extreme candidates winning their party’s nomination, it is no small thing to get those candidates past voters in a nonpartisan blanket primary election. Primary voters, after all, must choose among a handful of candidates, often of multiple parties, without the aid of party labels. What would allow them to coordinate on the most ideologically extreme—but still electable—candidate?
One possible explanation is the rise of term limits in Nebraska. Nebraskans passed term limits back in 1992, but that initiative was invalidated by the state’s Supreme Court. After several other attempts, an initiative limiting state legislators to two consecutive four-year terms was passed by the voters in 2000 and started turning out incumbents in 2006. We would expect to find the most profound impact of term limits on state politics occurring in 2006, at the moment it began affecting the composition of the legislature (and this is consistent with the trend seen in Figure 1), but given that its enactment seemed nearly inevitable since the early 1990s, we might expect to see political actors adapting to it prior to ratification.
Term limits have been associated with heightened partisanship in state legislatures (Bailey and Sinclair 2009), although other studies have found no such link (Cain and Kousser 2004). Beer (2001) finds evidence suggesting that term limits encourage party-line voting in Mexico’s national and state legislatures. Kurtz, Cain, and Niemi (2007), meanwhile, note that term limits may have inadvertently empowered parties and interest groups by creating less experienced candidates who are more dependent on them. That is, while term limits may not directly create a more partisan legislature, they do unmistakably increase legislative turnover; seats held by incumbents for three or four decades will come open just as frequently as battleground districts. This creates an opportunity for partisan actors to play a more direct role in candidate recruitment, and the legislators produced by such a system may behave in systematically different ways from ones protected by decades of incumbency. Term limits will also result in a pool of officeholders with attenuated institutional memory and legislative expertise, effectively transferring some of their political power to parties, lobbyists, and the executive branch (Cain and Kousser 2004). Finally, term limits introduce a large number of freshmen with no commitments to long-standing chamber norms, such as bipartisanship. Such an influx of new members can cause these norms to collapse (Diermeier 1995).
Have term limits had such an impact on Nebraska politics? To address this topic, we have conducted a series of interviews with political actors across the ideological spectrum in Nebraska politics. These actors include officeholders, party officials, journalists, interest groups leaders, and prominent donors. In the next section, we provide some of the evidence offered by these qualitative data. The section that follows seeks to match the interview responses up with a quantitative examination of sources of polarization in Nebraska.
(Non-)Partisanship in Nebraska: An Assessment from the Inside
In meeting with incumbents in the Nebraska Unicam, one of the most consistently expressed opinions is a clear and nearly unanimous love for the nonpartisan system. “We really are a family in this legislature,” boasts Senator Amanda McGill, a registered Democrat
1
from Lincoln. “I am just as good of friends with the Republicans in the body as I am the Democrats.” Unicam Speaker Mike Flood, a registered Republican from Norfolk, echoes, If you come down to Lincoln and expect to see people vote with what you think the core principles of the party are every time, you’re going to be sorely disappointed. . . . There’s no caucusing. . . . There’s no minority, majority whip. Every issue requires the hustle of every vote.
Instead, relatively consistent coalitions form around particular issues, including abortion and urban/rural issues, but those coalitions regularly transcend party lines.
Despite this celebration of the nonpartisan tradition, there is an acknowledgment that the chamber has become somewhat more partisan in recent years. “There are still some moderates,” says United Transportation Union lobbyist Ray Lineweber, “but not as many as years past.” One might look to the state parties themselves as the cause of this—party leaders have sought for years to strengthen their role relative to the legislature and even to formally reintroduce partisanship, only to be rebuffed by legislators’ content with the nonpartisan system. According to former state Republican chair David Kramer, The legislature was not particularly receptive to having the party involved in much, once the elections got over. You know, they all wanted our help at election time . . . They did not often appreciate that the party took positions on key legislative issues, did not appreciate that we would communicate publicly with them what our positions were on those issues, and encourage them to vote one way or another . . . We were trying to partisanize the legislature, and that was unbecoming of the body, and you know, I’m not sure when or where being partisan became a bad word.
Nebraska’s term limits initiative, which began terming out members in 2006, had a dramatic effect on the chamber’s composition. The percentage of freshmen in the Unicam, usually hovering between 15% and 20% (Niemi and Powell 2004), shot up to nearly 50% after the 2006 elections. When the 101st legislature was sworn in on January 7, 2009, only 12 of the chamber’s 49 senators had been in the chamber four years earlier. Term limits had removed many long-standing incumbents, requiring someone to recruit new candidates to run. That job largely fell to Nebraska’s formal parties.
Barry Rubin, who had just moved from Maryland to become the chairman of Nebraska’s Democratic Party in 2004, saw term limits as a unique opportunity for aggressive recruitment. “From a tactical standpoint,” Rubin says, “[term limits were] great for the Democratic Party, because it opened up seats where . . . we haven’t been able to compete more for in years.” Rubin and others then went about finding candidates: “We would literally scour the earth looking for good, moderate, forward-thinking candidates that we could find to run for office.” One such candidate was Amanda McGill, who in 2006 had been a communications director for the state Democratic Party: “Before I knew it, my own colleagues had recruited me to run for the state legislature. So, it wasn’t something I was planning on doing.” According to Rubin, such recruitment could get very detailed if a solid candidate was not initially obvious: I’ve literally spent days and days along with some of my coworkers going through registered voter files data and scouring the data for “five of five” Democrats—“five of five” meaning they’ve voted in five of the last five elections—who, in looking at occupational codes and looking at income and things of that nature and figuring out, “Okay, here’s a lawyer who votes in every election,” and we’ll sit around with a group of five or ten people from different walks of life that are all like-minded organizations or party people or whatever and said, “Yeah, I know that guy. He’s a lawyer down the street; he’s real active in politics . . . ” That was how we would target them and pick up the phone and say, “Hey, you ever thought about running for office?”
Recruitment efforts did not end at just identifying good candidates. Both parties, according to Kramer, “work like crazy in these non-partisan races to support candidates of their particular persuasion. . . . We probably targeted four to six legislative races a cycle, and put thousands of dollars into those races.” Rubin says that the Democratic Party would not just hand money to candidates, but would still help them raise their own funds: “We provided them with leads . . . We had a finance staff that helped them do call sheets and call time with them and things of that nature or hire their own people to set up those operations.” With fundraising in place, Rubin says, the party would also train their candidates in best campaign practices at a campaigning “school.”
To the extent they could, the party leaders coordinated their recruiting efforts with allied interest groups. According to former state Republican Party chair Kramer, We would collaborate with other organizations at that, you know, shared similar kinds of interests. The Chamber [of Commerce] being one of them, the National Federation of Independent Business, the realtors, other organizations that, while not overtly political themselves, their philosophy of governance generally meshed closer to ours than it did to others.
Terry Moore, the president of the Omaha Federation of Labor, boasts, “There’s a multitude [of legislators] that are in there right now that we recruited—probably about four or five of them that I personally recruited and brought in front of our board through our system.” Other interest group leaders claimed an active role in candidate recruitment, as well. Julie Schmit-Albin of Nebraska Right to Life, for examples, says, “We network with certain like-minded people who are also looking for good candidates, as long as they understand that anyone they send our way has to pass our pro-life endorsement criteria.”
Another actor who has been extremely active in recruiting candidates is Nebraska’s Republican governor, Dave Heineman. Heineman ascended to the governor’s office in 2005—roughly the same time that term limits kicked in and that Rubin began running the Democratic Party’s efforts. Heineman’s experience within his party is somewhat atypical for Nebraska governors: a former executive director of the state’s Republican Party, Heineman is seen as an exceptionally partisan creature with a hands-on approach to politicking. According to Senator Heath Mello, a registered Democrat from Omaha, “We’ve seen a level of involvement from the executive branch that we haven’t seen before, in regards to recruiting candidates . . . hand-picking candidates, raising funds and expending funds on legislative candidates’ behalf.” Senator McGill found herself going toe-to-toe with the governor in her 2010 reelection campaign: “The governor was the one who appeared on the mail pieces against me . . . supporting my opponent. So, I may trust most of my Republican colleagues in the legislature, but I have no love for the governor.”
Having so many members of the Unicam personally indebted to him has given the governor considerable power over the legislative branch. Beyond that, term limits has created a larger contingent of junior legislators with little experience or institutional knowledge who are more dependent on the governor’s office for information and resources. “Our governor is very active in the processes here,” says Senator Beau McCoy, a registered Republican from Omaha. “His office is obviously in the building here in our Capitol. He makes the rounds . . . He’ll poke his head in our office from time to time just to say hi.” Lobbyist Walt Radcliffe adds, “We . . . seem to have more of a unicameral parliament than anything else. The governor would have to really come up with something pretty wacky to not get what he wants.”
The Most Important Actors
To get a more representative sense of the structure of Nebraska politics, we asked respondents who they believed would be the three most important people to contact if one were interested in running for the state legislature. The answers suggest that formal party sources are overwhelmingly seen as essential for those interested in mounting a campaign for office. Among Democrats, party officials, including the chairman and the executive director, were the dominant recommended actors, along with current or former state senators from the area. This was generally suggested for the purpose of giving the prospective candidate a sense of what the job would be like, although several respondents suggested this as a specific source of recruitment. A middle tier of actors includes the mayor of Lincoln, the staff of U.S. Senator Ben Nelson, and prominent donors. A few lobbying organizations and consultants rounded out the bottom of the list. Even if pooled together, the interest groups do not nearly rise to the level of the party sources in citations.
On the Republican side, the lesson was roughly the same. Judging from the interviews, respondents of both parties broadly saw the governor as the de facto head of the state Republican Party and its chief strategist. Thus, it is not surprising to find that he was the number one choice among Republicans as a crucial actor in legislative campaigns. As on the Democratic side, respondents advised speaking to incumbent senators, party staff, and the staff of Speaker Mike Flood. Donors received somewhat higher billing here than on the Democratic side.
One should not overinterpret these results as the survey was not randomly selected. However, they do offer some insight into the mind-set of Nebraska’s political elites, who clearly believe that the formal parties and the governor are pivotal actors in legislative campaigns. These results suggest a good deal of formal, but often barely visible, structure to Nebraska’s current political system. They also suggest that while interest groups and donors may be important, they are subservient to the formal parties and the governor in the matter of recruitment.
An Empirical Test of Structure
The interviews made clear that this structure did not always exist. Term limits, it was noted, gave the parties the incentive and the opportunity to organize state politics in a way impossible previously. As Senator Mello remarked, “Term limits has ushered in, I think, a level of partisanship that most people would agree that they haven’t seen before.” Has their increased role in recruitment given the parties the influence over legislators they have long sought?
It is certainly possible that some of the reported change in the attitudes and beliefs of Nebraska’s political class is post hoc rationalization. However, we might expect to see the behavior of political actors change in observable ways with the advent of term limits. We know from the Shor–McCarty (2014) common space roll call data that legislative polarization was increasing at an extremely high rate in the state throughout this time. The session at which we would expect polarization to increase the most as a result of term limits would be 2007–08, which is exactly when the measured distance between party medians sees its most rapid rise. In 2006, the distance was around 0.717, while in 2008, it was 0.835 and 1.12 in 2010. It continued rising to 1.4 in 2011-12, although dropped a bit to 1.25 in 2013. In comparison, in 1996, the level was 0.467.
Another expected change in elite behavior would be a shift in the patterns of campaign donations. If a partisan structure emerged in the middle of the decade as a result of term limits, then we would expect a reflection in donation patterns. In a relatively weak party system, we would not expect to see elite donors giving exclusively to one party. Rather, they would donate for any number of idiosyncratic reasons—ideology, friendships, reputation, and so on. Conversely, if parties are strong, it makes sense to donate exclusively to one party.
The donation patterns of the Nebraska Republican and Democratic Parties (NRP and NDP, respectively) reflect this expectation to some extent. In the three election cycles prior to the imposition of term limits in 2006, the NRP donated $20,657 to candidates, but $27,304 in the three subsequent elections. The shift was even starker for Democrats; the NDP donated $6,551 in the three cycles prior to term limits and $36,906 in the three cycles thereafter. But these are fairly paltry sums, only donated to a handful of legislative candidates. Party activity, to the extent it occurred, would likely best be captured by the activities of a broader range of donors loosely affiliated with, but loyal to, the party (Hassell 2013).
We would not necessarily expect to detect much of a shifting contribution pattern among all donors—low-level donors are known to be inconsistent in their preferences and attached to just one or two candidates. But elite donors—those who make large donations and set donation patterns followed by others—should indicate such a tendency (Campaign Finance Institute, and Institute for Politics Democracy and the Internet 2006).
To examine the behavior of elite donors in Nebraska, we have collected campaign finance records for state legislative races from 2000 to 2010 from the National Institute on Money in State Politics. 2 For our first analysis, we have set a lower limit of $2,000 in year 2000 dollars 3 for all state legislative contributions by a donor. This eliminates most of the donors from analysis. In 2000, for example, a total of 674 unique donors made 2,463 contributions to state legislative campaigns in Nebraska. There were only 45 donors that year giving at least $2,000 to state legislative candidates, however, making 106 such contributions. Reducing the number of cases this way allows us to make the pattern clearer graphically without changing the underlying results.
A quick comparison between two different years offers some stark contrasts. Elite donors in 2000 often contributed to candidates across party lines. The Nebraska State Education Association distributed $8,500 across three legislative candidates who were registered Republicans, while giving another $21,732 to five Democratic candidates. The Nebraska Realtors’ Association, meanwhile, donated $9,000 to four Republican candidates and $5,000 to two Democrats. Most of the major donors were quite bipartisan then in their contribution patterns, in stark contrast to 2008, when major donors tended to stick within their party.
We have depicted these patterns more systematically in Figure 2 using social network analysis techniques. We have used the campaign finance data to create a one-mode network of candidates for the 2000 Unicam. As the figure shows, nearly every Republican candidate is connected to a Democrat, and nearly every Democratic candidate is connected to a Republican, via elite donors. The donor space is depolarized. In other words, if a major donor contributed $2,000 to a Democrat, she almost certainly contributed that much to a Republican, as well. While Democrats are somewhat clustered to the left of the diagram, it is difficult to identify a faction of either Republicans or Democrats in the network. Compare that network with the one from 2008, depicted in Figure 3. Here, we see that the network of candidates connected by donors has changed dramatically from a few years earlier. Instead of one large community of candidates, we see two distinct factions, divided largely along party lines.
One-mode network of candidates for Unicam, 2000.
One-mode network of candidates for Unicam, 2008.
Figures 2 and 3 suggest that some change occurred over the last decade. This is backed up with an analysis of the two-faction modularity (Zhang et al. 2008) of the networks: the modularity equals 0.013 in 2000 and 0.393 in 2008. The implication is that a two-faction structure has emerged in the network, with elite donors donating far less across party lines. To the extent that candidates and officeholders in Nebraska are obligated to their donors, those donors are compelling them to behave in an increasingly partisan fashion.
We can further examine the rising role of partisanship in Nebraska campaign donations by looking at contribution patterns. It may be, that is, that the people who make major donations to Nebraska state legislative races today are more politically polarized than the donors who contributed a few years ago. A simple analysis along these lines can be seen in Figure 4, which shows the partisan contribution patterns of repeat donors, focusing on those who gave at least $500 (in year 2000 dollars) to at least two state legislative candidates in at least two years. We have categorized the donors based on their donation patterns. If they made at least 80% of their donations to Democratic state legislative candidates, we have categorized them as “Dem-leaning.” Conversely, those labeled “GOP-leaning” contributed at least 80% of their donations to Republican legislative candidates. The rest are categorized as “bipartisan.” The figure charts the percentage of each donor group in each year.
Trends among repeat donors.
As can be seen, Dem-leaning donors have generally comprised a low percentage of contributors, but that number increased slowly during the decade and jumped substantially between 2004 and 2006. The percentage of donors categorized as GOP-leaning also took a sharp upward turn after 2004. For both groups of party-leaning donors, their proportion of the donor pool each year under term limits has been higher than it was in any year prior to term limits. Bipartisan donors, meanwhile, have been on the decline, most sharply after 2004. This suggests that donors in Nebraska legislative elections are polarizing. State legislators are being funded by an increasingly partisan group of elite donors.
Another take on changing fundraising patterns can be seen in Figure 5. Here, we have divided up Nebraska’s legislators by their ideal points (as calculated by Shor and McCarty 2014). Those whose ideal point’s absolute value is greater than the absolute value of their party’s median ideal point are classified as more ideologically extreme; the rest are considered less ideologically extreme. As the figure shows, there was little difference in the fundraising prowess of legislators across the ideological spectrum in the early part of the decade. Regardless of ideological extremism, legislative candidates prior to term limits raised an average of $29,000. Under term limits, however, we have seen a substantial divergence between these candidates. While candidates across the board are raising more money than they used to, the more ideologically extreme legislative candidates are each raising approximately $10,000 more than their moderate counterparts. This is suggestive of a recent change in fundraising patterns, and it is consistent with other evidence suggesting that the electoral environment has become more favorable to a more polarized set of candidates.
Campaign fundraising patterns by ideological extremism.
We perform a more sophisticated analysis on these donation patterns to include appropriate statistical controls. Studying campaign contribution patterns, however, can present some methodological difficulties. One large problem is the possibility that donors influence each other, that is, one donation to a candidate may encourage friends or colleagues of the donor to make a similar donation. Such behavior would mean that the act of contributing is not an independent one, making analyses of such patterns produce results that are potentially biased and misleading.
Studying donation patterns through the use of an exponential-family random graph model, or ERGM, can help resolve this difficulty. ERGM functions like a logistic regression equation in networks research. In essence, an ERGM investigates the differences between an observed network and a randomly generated network with an equal number of nodes (Cranmer, Desmarais, and Menninga 2012). It allows us to examine the impact of different variables on the likelihood of any two given nodes sharing a link, without concerns about the independence of cases. 4 In this particular case, we can assess the likelihood that any two nodes (candidates) in a network share a common donor. We can use a variety of independent variables to predict this likelihood.
For this equation, we have used six independent variables:
Links—measuring the probability of observing a connection (also known as a “link” or an “edge”) between any two candidates.
Party—comparing whether two candidates in a dyad are of the same party or not. This is based on the notion that two members of the same party would likely share a partisan donor.
Ideal point—measuring the absolute difference of the candidates’ ideal points in a dyad, as calculated by Shor and McCarty. This is based on the idea that a donor would be likely to contribute to candidates of similar ideological persuasions, even if party itself is relatively meaningless.
Cohort—comparing whether two candidates in a dyad were elected the same year or not.
Lincoln/Omaha—comparing whether two candidates in a dyad are both from the state’s major cities (Lincoln and Omaha) or both from rural areas, or if one is from an urban district and one is from a rural one.
Presidential year—comparing whether two candidates in a dyad were elected in a presidential year or not.
We add the network term geometrically weight edgewise shared partner distribution (GWESP) to this model, with an alpha value of zero. This helps control for the clustering and transitivity that tends to occur in networks of legislators. We expand the analysis by lowering the donation minimum to $500 in year 2000 dollars. This has the virtue of substantially increasing the number of cases—288 unique donors making 893 donations in 2000, rather than 45 unique donors making 106 donations—although the results from the analysis are substantively identical to those at the $2,000 level. The analysis is limited to officeholders only. The models were estimated using the R packages Statnet and SNA. The results of the models appear in Table 1.
ERG Model Predicting the Likelihood of Officeholders Sharing an Elite Donor, by Year.
Note. Cell entries are exponential-family random graph model coefficients predicting the likelihood that any two given officeholders in the Nebraska legislature share an elite donor (contributing at least $500 in year 2000 dollars). Standard errors appear in parentheses beneath coefficients. ERG = exponential random graph; GWESP = geometrically weight edgewise shared partner distribution; BIC = Bayesian Information Criterion.
p ≤ .05. **p ≤ .01. ***p ≤ .001.
The coefficients are log-likelihoods, not lending themselves to easy interpretation, although their signs and levels of statistical significance are of interest. The main finding of note is that the party coefficient is statistically indistinguishable from zero in the years prior to term limits, and is actually in the wrong direction in 2004, suggesting that candidates of different parties were more likely to share a donor than those of the same party. In 2006 and 2008—after term limits have kicked in—the party coefficient is suddenly positive and statistically significant. (The coefficient is positive but not statistically significant in 2010.) For the sake of interpretation, the predicted probability of candidates sharing a donor if they are of the same party in 2000 is 49%, but 46% if they are of different parties. In 2008, conversely, that probability is 38% for those not sharing the same party and 53% for co-partisans. These results suggest that party has become more important to structuring elite donations since the advent of term limits.
Ideology is in the expected direction, with ideologically similar candidates more likely to share a donor, but statistically significant in only one year—2010. We can understand this as only once the parties have polarized does it become possible for donors to distinguish between candidates in terms of ideology. The cohort variable is statistically significant from 2004 on, suggesting that the high turnover rates during this period produced increased attention by donors to new candidates. The region and presidential year coefficients vary a great deal from year to year with little apparent meaning.
These coefficients have been translated into predicted probabilities for two years, 2000 and 2008, and displayed graphically in Figure 6. This figure charts the probability of any two given candidates sharing a common donor based on the similarity of their ideal points and whether or not they share a political party. In 2000 (at left), we can see a modest effect for ideology—the more ideologically distant the candidates, the less likely they are to have a donor in common. Party, however, seems not to matter at all. Co-partisans are only 3% more likely to share a donor than cross-partisans, a result that is not statistically significant.
Predicted probabilities of sharing a donor by party and ideology.
The chart at right shows the probabilities for 2008, during the term limits era. While the figure appears to indicate that candidates are more likely to share a donor the more ideologically distant they are from each other, that coefficient is actually statistically insignificant. Party, however, suddenly matters a great deal; co-partisans are 15% more likely to share a donor than cross-partisans. 5 These results suggest that party has become more important to structuring elite donations since the advent of term limits.
Conducting the above analysis separately for each year may ignore some important time dependencies inherent in the network model. For example, the interaction of donors and candidates in one year may affect their interactions in future years. To account for these temporal dependencies, we have conducted a TERGM (temporal ERGM) analysis on the same data, combining all six observations from 2000 to 2010 into one network list. We added two new variables to the analysis here that were not in the previous one: a term limits variable (equaling 0 from 2000 to 2004 and 1 from 2006 to 2010) and an interaction term for term limits and party. A positive coefficient on the interaction term would be consistent with the finding that party became a better predictor of sharing donors in the term limits era. The results for this TERGM can be found in Table 2. The results are consistent with expectations. The interaction term is positive and statistically significant at the p ≤ .05 level. 6
TERG Model Predicting Likelihood of Officeholders Sharing an Elite Donor, 2000–10.
Note. Cell entries are TERGM coefficients predicting the likelihood that any two given officeholders in the Nebraska legislature share an elite donor (contributing at least $500 in year 2000 dollars). 95% confidence intervals appear in brackets beneath coefficients. TERGM = temporal exponential-family random graph model; GWESP = geometrically weight edgewise shared partner distribution.
p ≤ .05.
We can increase our confidence in these findings by exploring counterfactuals suggested by other states. That is, while we have made efforts to control for confounding factors, it is possible that Nebraska simply experienced a secular polarizing trend in campaign donation patterns that happened beyond the state’s borders, as well. To examine this possibility, we collected identical state legislative campaign contributions records from nine other control states over the same time period. These nine states—Alabama, Iowa, Indiana, Missouri, North Dakota, Oregon, Pennsylvania, Texas, and Utah—were selected because they, like Nebraska, have no upper limit on contributions to state legislative candidates. And while partisanship may have increased in those states, none of the states adopted term limits during this time period. We gathered campaign finance records from these states from the National Institute on Money in State Politics. 7
To compare the states, we look at all donor ties between state legislative candidates between 2000 and 2010 and measure the percentage of those ties occurring between legislators of the same party. We estimate this once for 2000–04 (prior to term limits in Nebraska) and once for 2006–10 (after the adoption of term limits in Nebraska). We expect, due to polarizing trends, that all the states will see increases in the percentage of donor ties occurring within the same party. However, if the increase is atypically large in Nebraska relative to the other states, that would suggest that Nebraska’s rise in polarization is attributable to specific events within that state rather than from larger national trends.
The results of the analysis, which are displayed in Table 3, suggest that Nebraska has experienced a low rate of donor polarization relative to the other states, but saw a substantial rise in donor polarization under term limits. The greatest rate of polarization was observed in Utah, with an astounding 25-point increase between the two time periods. Pennsylvania comes in second with an 8.7-point increase, followed closely by Nebraska at 6.2 points. This is well above the median (2.5) or mean (3.5) increases between the two eras. Although the small number of cases prevents this difference from meeting conventional tests of statistical significance, it does suggest that Nebraska has experienced a greater than average polarization of its elite donors base, that it has come to more closely resemble other states (those with parties) in the way it funds state legislative races, and that this shift was roughly concurrent with the adoption of term limits.
Donor Polarization in 10 States with No Upper Contribution Limit for Legislative Races.
Note. Donor polarization is measured as the percentage of candidate network ties, via shared donors, that are between candidates of the same party.
Discussion
Assessing a state’s partisanship is no simple task. Just as there are multiple facets of parties (legislative caucuses, voter allegiances, formal and informal organizations, etc.), so are there multiple ways of describing just how important parties are in a given state and how that role may have changed over time.
Nonetheless, using a variety of methods, we have sought to do just this for the state of Nebraska. The results suggest that while early twentieth century reformers were largely successful at reducing the impact of parties on the state government, partisanship has returned to Nebraska with a vengeance in the past few years. To be sure, the Nebraska Unicam is hardly the California Assembly, no less the British House of Commons; there is considerable ideological overlap between the parties, and legislators deeply value their nonpartisan tradition. Voting coalitions, several senators suggested, do not regularly follow party divisions, but rather fall along splits determined by particular issues stances, and bill authors must sometimes build winning coalitions from the ground up.
And yet the preponderance of evidence suggests that this familiar system is giving way to a more partisan one. The forced retirement of a large segment of the legislature in 2006 due to term limits spurred the parties and the governor into action, recruiting, training, and funding candidates at levels not previously seen in modern Nebraska. The state’s new legislators are increasingly being chosen for their expected adherence to party agendas, as determined by party leaders and the governor. And the donation patterns of elite campaign contributors are increasingly following a partisan and ideological pattern, suggesting that to the extent that legislators want to keep their donors happy, they will do so by voting more with their party.
This is not to suggest that the trends we report here from 2000 to 2010 will continue unabated. Indeed, roll call voting records from 2013 suggest that polarization in the Nebraska legislature has plateaued recently (see Figure 1). The trends of donor polarization and ideological candidate recruitment do not appear to have reversed. However, we did see evidence of a bipartisan pushback against the termed-out Governor Heineman as his tenure drew to a close in 2014. In a follow-up interview in June of 2014, Senator Heath Mello remarked, “I think over the last few years, we’ve seen a move towards a more independent voting pattern away from what the governor has been advocating in the legislature.” Indeed, as Mello noted, majorities in the chamber, including large numbers of Republicans, have overridden several of the governor’s recent vetoes on education issues and health care funding (Schulte 2014). This may well represent the flip side of term limits—even as it may open the door to greater partisanship in recruitment, those compelling partisan behavior lose much of their power as they become termed out themselves.
Nonetheless, Nebraska’s political system experienced a significant shock in the middle of the last decade and appears unlikely to return to its earlier nonpartisan style. As the parties continue to recruit candidates in an ideological fashion and as the influence of colleagues from an earlier, less partisan era fades away, we are likely to see increased legislative polarization there in the coming years.
Why then has Nebraska’s version of antiparty reform—the nonpartisan legislature—apparently worked for so long, while parties in other states seem to adapt and overcome antiparty reforms relatively quickly? A good parallel for Nebraska’s story is that of California under cross-filing (1913–59), under which candidates could run in as many primaries as they desired without their party label appearing on the ballot, which substantially depressed legislative partisanship for four decades. Other states have attempted to limit partisanship and empower moderate candidates through the creation of more open primary elections, with little apparent success (McGhee et al. 2014). Why might parties be able to adapt to some of these reforms but not to a ballot without party labels? The latter has the feature of depriving voters of one of the most essential pieces of information the ballot has to offer—arguably, more important than the identity of the candidates themselves. Yes, high-information voters and political activists can discern the partisans on the ballots, but the vast majority of voters will not conduct research on these matters and are at the mercy of whatever information actually appears on the ballot. The fact that party labels still do not appear next to the names of Nebraska’s Unicam candidates will certainly continue to provide a ceiling on legislative partisanship there as many voters still lack the necessary information to vote their party. Nonetheless, it appears that the parties themselves are developing ways to structure elections in a more partisan manner and give advantages to their preferred candidates. Ironically, it appears that voters themselves have given parties an institutional vehicle to accomplish this end.
Footnotes
Authors’ Note
The authors are grateful for valuable assistance and suggestions from Skyler Cranmer, Daniel Galvin, Peter Hanson, Ari Kohen, Georgia Kernell, Philip Leifeld, Hans Noel, Betsy Sinclair, Jennifer Nicoll Victor, and Michael Wagner. Most of the interviews conducted for this research occurred in 2011. Since that time, Nebraska Speaker Mike Flood left the chamber in 2013. Lobbyist Ray Lineweber has retired.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors received funding from NSF Grants SES-1059716 and SES-1060092 and from the Robert Wood Johnson Scholar in Health Policy program for this project.