Ideological Polarization in State Legislatures
Our findings in the previous chapter—that political parties play a relatively small role in financing legislative elections in most American states and that parties behave differently than other kinds of donors, tending to support moderates and challengers—raise an intriguing institutional question. If campaign finance rules allowed parties to enjoy a larger role in financing candidate campaigns, would this tend to moderate politics and diminish ideological polarization between the political parties? Our findings also raise a more fundamental question regarding political representation. Would giving party organizations more clout to shape the field of candidates lead to officeholders who more accurately reflect the preferences of voters?
Addressing these questions bears directly on future policies for campaign finance reform. To this point, contemporary debates have framed reform as promoting either equality or liberty.1 The egalitarians argue that laws should aim to reduce the ability of wealthy interests to dominate political financing, which they claim gives them an unequal voice in government. Those opposed to restrictions on political spending argue that such constraints diminish free speech. Differentiating ourselves from both egalitarians and libertarians, we suggest a third consideration. We posit that an “institutional” dimension has been overlooked in reform debates, to the detriment of representative government.
Addressing the Institutional Dimension
Our argument is that campaign finance laws shape the flow of money (what we have called the hydraulics of campaign finance) to political committees from various types of donors who have different agendas. In turn, the institutional flow of money affects the kinds of candidates who seek and win office, and also the ways in which they govern. Our main thesis is that money that flows outside party channels tends to promote ideological polarization between partisan officeholders. Conversely, money that is controlled by the party is more likely to moderate politics. Thus, the hydraulics of campaign finance have implications for political balance, stability, and the representation of majority interests.
In probing the policy question, we also raise related practical and theoretical concerns about the functioning of contemporary political parties in American democracy. From a practical perspective, our analysis in this chapter touches on whether party control over resources actually gives party organizations leverage in selecting who runs for and wins office. Since the early 20th century, the growing and widespread use of primary elections in the American states has removed the monopoly power that party officials once enjoyed in nominating candidates. The changes in the nominating process, coupled with other transformations in American society (including the rise of national-issue politics over the politics of regional interests), have given highly ideological candidates significant opportunities to pursue office with the help of like-minded partisan activists (Masket 2009). Given the importance of money in campaigns, it is entirely plausible that financially powerful factions will have considerable influence shaping the field of candidates, whether those factions are issue groups controlled by purists or formal party committees controlled by pragmatists.
From a theoretical perspective, our study helps advance our understanding about whether and how party organizations actually matter in contemporary politics. With respect to the financing of campaigns at the state level, one line of research suggests that parties have been mere conduits or empty vessels to funnel money to candidates (Krasno 2011; Rosenthal 1995). A second set of studies suggests a stronger if minimal role, arguing that party organizations are primarily technical organizations at the service of candidates who use them as addenda to campaigns (Aldrich 1995; Herrnson 1988; Kolodny 1998). Candidates, for example, benefit from the ability of party organizations to get wholesale goods and services at cheaper rates, including voter files, mailings, and even consulting services.
Finally, a third and emerging body of research claims that parties are much stronger than the conventional wisdom suggests, and that observers have missed this fact because they have been looking in the wrong places. These “network theorists” argue that the term parties actually includes not just formal party committees, but agglomerations of policy-demanding groups, consultants, and activists. And further, a significant amount of “party” campaign activity—such as the recruitment of candidates, fundraising, and mobilizing of voters—takes place by such actors outside the formal party organization (Bawn et al. 2012; Bernstein 1999; Bernstein and Dominguez 2003; Grossmann and Dominguez 2009; Herrnson 2009; Kolodny and Dwyre 1998; Kolodny and Logan 1998; Skinner 2007; Skinner, Masket, and Dulio 2012).
The network theorists might logically claim that the organizational form of political parties does not necessarily matter because contemporary parties are extended party networks comprising multiple interest groups and activists acting in concert to pursue partisan goals. The result of this dynamic is that the direction of the party reflects the aggregate of actions across dense partisan networks. Within the partisan network, the party committee can be a small or large node. But this does not matter because the “real” party combines the actions of all partisans. The implication of such theorizing is that, regardless of whether the traditional party committees are financially constrained or precluded from formally nominating candidates, the party thrives through the myriad linkages that bring partisans together for common goals.
We tend to disagree with this conclusion, even though we acknowledge the group-centered nature of partisanship. We do not object to the conceptualization of parties as consisting of networks of partisans, but we believe that organizational forms matter and that the “open systems” perspective can be pushed only so far. In our view, the party network reflects an assemblage of factions composed of various partisan interest groups, activists, consultants, and elected officials. Much of the time the factions row together. But the direction they tend to row depends critically on who controls significant resources. We argue throughout this book that money controlled by the formal party organization should give relatively more influence to an elected party leadership and the circle of pragmatists who tend to work through the party committees, rather than to the activists motivated mostly by issues and principles and operating principally from nonparty organizations.
To be sure, even the pragmatists have ideological views, which may slant their preferences toward particular kinds of candidates. But being pragmatists, they will focus primarily on winning elections by backing candidates who are not as extreme as the preferences of the ardent issue activists. We also acknowledge that not all party committees are controlled by pragmatists. Indeed, we suspect that many contemporary party leaders at the local level have been attracted to party operations precisely because they see the party as a vehicle to recruit and support candidates who favor their ideological positions. In contrast to 50 years ago, the party organization is much less a place for transactional politics than it was when traditional “bosses” controlled patronage to dispense to followers. For this reason alone the party organization will attract a more ideological leader than in the past. That said, we still maintain that local party leaders will adjust their ideological proclivities to advance electoral goals where there is two-party competition for office. Moreover, we believe that the higher the level of party, the more likely it will display moderating tendencies. State party officials will need to broaden the party appeal to ensure that the party brand does not hurt candidates statewide or in districts with moderate voters. Moreover, legislative caucus leaders want above all to win control of government; they will therefore be less inclined to impose purity tests on legislative candidates.
When laws constrain party organizations rather than affiliated partisan groups, the overall consequence is to diminish the centrality of the party organization as the venue for reconciling diverse factional interests (see chapter 1). And, more importantly for our argument, a diminished role for the party will reduce the influence of the pragmatists in party politics, who typically lead party committees precisely because they understand that it is in the interest of the organization to balance party factions. In other words, when resources for the party organization are reduced, influence moves away from pragmatists within party organizations toward the purist factions operating through policy-demanding organizations. Thus, the weakening of the formal party committee as a place to do partisan business means that collective partisan activity will be controlled more directly by financially well-off partisan factions, which get their funding from nationally based ideological constituencies. The “party” in this scenario exists as a dense network of policy demanders, but it is qualitatively different from one in which a financially strong central party organization is the major node. With weak party organizations, the ideological factions have greater leverage to recruit and support like-minded candidates at the expense of moderate party candidates (the kind favored by pragmatists) who might have a better chance of winning elections in competitive districts.
As noted in the opening chapter, our study challenges two dominant conceptions of political parties. One view sees parties primarily as controlled by “insiders”—mostly officeholders—who use the party to advance their ambitions (Aldrich 1995). The other view sees parties controlled mostly by “outsiders” who work through issue coalitions to advance policy objectives (Bawn et al. 2012). We tend to see political parties as controlled by outsiders, because our findings suggest that party organizations are not merely service providers assisting candidates, but entities that shape collective goals and electoral outcomes through the candidates they support. But we also view the “outsider” model as too basic, since it views the organizational form as of secondary importance. We demonstrate in this chapter that party organizations are of central importance because they generate substantively different outcomes when faced with more or fewer financial restrictions. In providing this empirical work, we hope to qualify the emergent school of theory expressed in the “extended party network” thesis by showing that party organizations matter. Their centrality in the network of partisan interests has the potential to moderate politics. In the last chapter of this book, we will say more about the positive outcomes that we believe will come from placing parties at the center of political networks.
Beyond parties, our analysis might also shed light on the million-dollar question: does money buy policies? Most studies look at whether donations influence votes on legislation (Ansolabehere, deFigueiredo, and Snyder 2003) or legislative effort (Hall and Wayman 1990). This approach reflects the view that the relationship between donors and candidates is essentially dyadic. Our lens is broader. We think that money has an impact at the system level. We focus on how the flow of money shapes the ideology of officeholders in the legislature. Their ideological predispositions potentially affect a whole array of specific policy choices they pursue, and shape institutional patterns that affect governing. Research shows, for example, that ideological polarization affects how parties organize themselves in legislatures, the amount of power they give leaders, and the degree of bipartisanship shown by legislators (Rohde 1991). Our study looks further back in the chain of influence. We try to show that the polarization we observe in legislatures is linked, in part, to how campaign finance regulations affect the relative power of groups engaged in elections, including party organizations, partisan factions, ideological donors, and a variety of interest groups. The electoral power these groups accrue—partly due to campaign finance laws—affects who is likely to run for office, who gets supported, and how they behave once in office.
In sum, this chapter combines our interest in understanding both political parties and campaign money using a system-level approach. Our analysis is ultimately about representation. Our goal in this chapter is to examine whether party-centered laws might create the kind of ideological moderation in legislatures that would be more reflective of the preferences of the majority of voters.
Partisan Polarization and Representation in the American States
In the opening chapter we pointed out an apparent disconnect between American voters and the people they send to office. At the federal level, representatives in Congress are much further to the right or left than voters in their districts (Bafumi and Herron 2010). One consequence is that the congressional parties are polarized ideologically, with few moderates in either party to help stitch together compromise legislation. The absence of moderates in Congress does not mean there are no districts where moderate voters prevail. Such districts are often represented by either liberal Democrats or conservative Republicans, a dynamic that Bafumi and Herron (2010) have referred to as “leapfrog” representation. This situation leads us to believe, like Fiorina, Abrams, and Pope (2005), that it is not necessarily the voters who are ideological but the choices being offered to them.
Most discussion about partisan polarization focuses on the national level. Recent research, however, indicates that polarization is also a fact at the state level (Shor and McCarty 2011). Table 4.1 uses data from Shor and McCarty (2011) to illustrate the distance between party ideological medians in the state legislatures (combined chambers). The states are arranged from most to least polarized based on voting patterns in the 2010 legislative sessions. The level of polarization of each state in 2010 is also compared with its level in 1997. The “Change” column shows the amount and direction of change that each state has experienced over those 13 years (negative numbers mean decreased polarization and positive numbers increased polarization).
First, note the wide variation in states. The California legislature stands out as representing the highest level of polarization, followed by Colorado. California is now heavily tilted toward Democrats, which enables the state to avoid some of the gridlock that the U.S. Congress faces. At the other end of the scale, some states are not very polarized because both parties are fairly liberal (like Rhode Island and Delaware) or mostly conservative (like Louisiana and Arkansas).
Table 4.1 also reveals that the distancing between the parties has increased over time in most states. Over this period, 31 legislatures have seen some increase in polarization, 14 have become less polarized, and 4 have neither increased nor decreased. (This is a total of 49 states; Nebraska is once again omitted from our analysis because of its political structure.)2 The data provided by Shor and McCarty provide an excellent resource to assess the impact of campaign finance laws on legislative party polarization. We can observe interesting variation in polarization across states and over time, which can be assessed against institutional variation—namely campaign finance laws—across and within states.
In our view, a critical problem emerging from partisan polarization is that the ideological distancing of legislators is not typically reflected in the communities they represent. If increasing polarization were mirrored by the preferences of district constituents, then it could be argued that the legislatures were merely being responsive to voters. But this is not the case. Examining data on voter ideologies from Catalist, we can see that state legislators tend to be far more polarized than the districts they represent. Figure 4.1 shows the distribution of Democratic and Republican state legislators across the ideological spectrum (the short and long dashes, respectively), compared with the distribution of the median ideologies of the districts those legislators represent (the solid black line). For legislators, the ideological scores are determined by their roll call votes in the state legislatures. For the districts, the ideological score is the ideology of the median adult in the district, as determined by Catalist.3
The figure is presented in terms of the ideology scale used by Catalist, so that 0 is the most liberal value and 100 is the most conservative. The figure clearly indicates that people in the legislative districts are not polarized in the same way that legislators are. While state legislators are distributed bimodally, with most Democrats falling clearly on the more liberal side and most Republicans falling clearly on the more conservative side, the distribution of districts is unimodal and centered in the middle of the ideological spectrum. Thus, if state legislators are polarized along party lines, the explanation does not appear to be that this is so because they are closely representing their districts.
The implications of these trends do not appear good. We observe at the national level problems associated with polarization, including partisan rancor and policy gridlock under divided government. The ongoing policy battles between the parties on immigration, health care, and other policies seem to be predicated on false policy choices that entirely lack a middle ground. The failure to act when public opinion supports a reasonable compromise raises essential questions about the ability of institutions to represent voters and function adequately. Some recent research even indicates that policy extremism appears to make voters less able to choose candidates who reflect their views. When candidates take extreme positions, they emit very strong partisan cues, which makes it harder for voters to recognize when a candidate in the opposing party might actually be closer to their own preference (Rogowski 2014).
How Party-Controlled Money Increases Moderation in the Legislature
As should be clear by this point, we maintain that strengthening party organizations by improving their finances might temper the ideological extremism of candidates. We outlined in the opening chapter the inherent pragmatism of party organizations and the mechanisms through which the moderating dynamic works. We also noted that the party organization, through its governance structures and historical label, is the broadest representation of the party and is therefore less likely to be dominated by narrow issue factions.
We described several mechanisms by which party leaders and party officials can temper extremism among the party ranks in the legislature. One direct way is by their choice of whom to support in elections. As we showed in chapter 3, pragmatists in the party organization tend to support candidates whose views are closest to the median voter and who logically have the greatest chance of electoral success in a general election. A second way party committees moderate politics is by substituting party money for purist contributions. After all, it is candidates, not parties, that receive the bulk of their donations from purist donors, and as we show in the next chapter, the candidates in the most competitive seats receive significant support through “independent” spending by issue groups. Our argument has been that a greater reliance on party support will lead to decreased reliance on ideological sources of money, which might influence how officeholders campaign and govern. If the party were a primary source of support, legislators who are fearful that straying from purist positions might hurt their fundraising prospects could take more moderate positions, knowing that the party will support them in any competitive election. This is because, as we have seen from studies of party allocation strategies, there is no relationship between the degree to which officeholders vote with the party and how much money they get from the party committees (Bianco 1999; Damore and Hansford 1999; Herrnson 1989; Leyden and Borrelli 1990; Nokken 2003, but see Jenkins 2006). In other words, the party does not necessarily use its funds to enforce party discipline, particularly for officeholders who come from moderate districts where a vote against the party might serve them well electorally. Any incentive for party leaders to use their resources to whip a particular vote will be balanced against concerns about losing seats.
A third plausible party mechanism for moderating politics comes at the prenomination stage. To the degree parties control a disproportionate share of resources, they have latent power to shape candidate recruitment and influence who gets on the party ticket. We call this the “800-pound gorilla in the room” argument. It is more speculative than the previous claims because we have not examined in depth how or whether this happens. We infer this dynamic from previous research and interviews with state party leaders. What we are arguing is that, despite the fact that political parties have legally lost control over the nomination process due to the increased reliance on primaries, partisan elites can have influence over who eventually wins a party nomination through their use of endorsements and other key electoral resources (Cohen et al. 2008; Masket 2009). While formal parties in many states cannot typically give money to primary candidates, the mass of party resources should at least allow party leaders to make credible offers of support to prospective candidates in trying to recruit them. All things being equal, party leaders will want to pledge their electoral backing to candidates who have the best chance of winning in closely contested races—usually those who are more centrist.4 Toward that end, party officials can also informally “clear the field” by discouraging others from running, backed by the weight of their campaign war chests. Ultimately, a strong party organization can choose simply to ignore an undesirable candidate in the general election. According to the director of the well-financed Florida Democratic Party, “We basically do all the recruitment . . . so we don’t have many truly contested primaries. And we stay away in the general election if a wacky candidate wins who does not reflect the district. It’s a waste of money to help them anyway.”5
Finally, a fourth way party organizations might encourage moderation is indirectly by enhancing electoral competition. Such competition tends to encourage candidates in both parties to take positions closer to those of the median voter (Burden 2004). As we noted in chapter 3, compared with other types of donors, party organizations are more generous to challengers. In supplying challenger financing, parties plausibly increase the tightness of contests, which incentivizes candidates to fight for voters in the middle.
Expected Outcomes of Party-Centered Campaign Finance Systems
Given the contrasting array of incentives offered by party organizations and factional groups, we can set forth clear hypotheses about how a party-centered campaign finance system will affect outcomes. We will then examine those hypotheses in the light of data gathered from state governments.
In states with a relatively strong party role in financing elections, we hypothesize the following:
1. Officeholder ideological preferences will be less extreme in party-centered states because well-funded parties will support moderates and challengers.
2. Legislative chambers will appear less extreme and more open to compromise under a party-centered campaign finance system because party-induced moderation will increase the number of legislators who might cross party lines.
Hypothesis 1: Legislators Are More Moderate in Party-Centered States
In the first analysis, we would like to understand whether states with party-centered (i.e., party-friendly) laws tend to elect politicians who are more centrist than states where parties lack these advantages. Using data on officeholder ideological scores from Shor and McCarty, we can make direct comparisons of legislators elected in states that allow parties to raise and contribute unlimited sums (15,026 legislators) to those running in states that have limits (35,149 legislators).6 Figure 4.2 shows the distributions of the scores of the Republican and Democratic legislators in these two sets of states. The x-axis is a legislator’s ideology score, with negative values indicating more liberal legislators and positive values representing more conservative legislators. The y-axis simply shows the relative number of legislators at each point along the ideological spectrum. The solid and long-dashed lines reflect legislator scores in states where parties have no limits on contributions to candidates and no limits on what they can raise from individuals. The shorter dashed lines represent states where parties are limited in either raising or contributing money (or both).
First, the figure highlights the importance of party labels in the American states. Most Democrats fall on the left side of the spectrum and most Republicans on the right. Officeholder preferences clearly reflect a strong bimodal distribution based on which party they belong to.
At first glance, the effect of party limits seems minimal, even though the distribution shifts toward the middle as we expect. Observing the solid and long-dashed lines, it appears that a larger proportion of legislators in party-centered states are moderate. This effect is larger on the Democratic side, which shows a peak for party-centered states that is noticeably closer to the center compared to the peak for states where parties are highly constrained. On the Republican side, the shift is less pronounced, but nonetheless shows the same moderating dynamic. Note that there is more overlap between the parties in party-centered states, suggesting that the two parties plausibly have more opportunities for policy compromise. We will address that dynamic subsequently.
Figure 4.2 includes all states in the data compiled by Shor and McCarty, but we expect to find that campaign finance laws will have their biggest impact in states where money matters a great deal. Those states tend to have professional legislatures where the value of office is quite high and competing for a seat tends to be more resource-intensive. Because money matters more, those who have control of it should have considerable leverage in the electoral process. By this reasoning, if parties are key players in financing elections in professional legislatures, we should see a more dramatic impact on the ideological moderating of officeholders. And indeed this is what we find.
Figure 4.3 shows the same kind of analysis as the previous one, but for only those 20 states that rank highest on professionalism. To determine whether a state has a professional legislature, we use the Squire Index, which classifies legislatures based on factors such as the amount legislators are paid, how many days the legislature is in session, and how many staff are employed by the average legislator (see Squire 2007). We take the 20 most professional legislatures and classify those as our “professional legislatures” in figure 4.3.7 For additional clarity, we separated the presentation into two figures, with the top showing the distribution for states with contribution or fundraising limits on parties and the bottom showing the distribution for states without either of those limits. There are significant differences, with legislatures in unlimited states showing considerably less polarization. Specifically, in states with limits, the distance between median members is 1.56, while this distance is only 1.15 in party unlimited states. The disparity between the two sets of states is like going from one competitive legislature that experienced considerable bipartisanship such as Indiana (which is less polarized than Congress) to another that is as polarized as Oregon (which is even more polarized than Congress). We also note that the shift toward the ideological middle appears similar for members of both parties in states with party-centered laws. The ideological distribution for both Republican and Democratic parties in unlimited states reflects a normal curve that skews asymmetrically toward the center, which is not true of the distributions for states with restrictive party laws, which show high peaks skewing toward the liberal or conservative poles.
Overall, these results suggest to us that a stronger role for the party in financing elections has the potential to increase the presence of moderate legislators, especially in states with professional legislatures. Of course, one concern with these results is that the states we examine may differ on some other dimensions that tend to be correlated with how partisan or polarized they are likely to be. However, recall from chapter 1 that we compared states that enacted limits on parties to those that did not and found no observable differences on a range of factors that might be associated with the probability that a legislature would polarize. Thus, we have reasonable confidence in the patterns presented in figures 4.2 and 4.3. Nevertheless, in the analysis that follows, we conduct an even stronger test of our expectations by taking advantage of the fact that the data we have collected cover a considerable passage of time.
Hypothesis 2: Party-Centered Campaign Finance Reduces Polarization in Legislatures
We next explore our second hypothesis: that a party-centered campaign finance system will make legislatures less polarized, which could improve the governing process. In Congress and many states the politics of compromise, which seems necessary in a separated system of powers, appears to be hard to practice because party members are so far apart on their preferences. But if the distance between parties could be lessened, it seems conceivable that a legislature might become more fertile ground for transactional politics and the kind of across-the-aisle bargaining that leads to bipartisan legislation.
First, we consider the prospects for finding common ground. We think this is more likely to happen when some members of both parties share ideological preferences. In times of divided government, moderates in the legislature often form the core of bargaining agreements between branches of government (Trubowitz and Mellow 2005). Moderates reflect the “pivot,” and the greater the number of legislators who occupy the pivot, the stronger their leverage to forge bipartisan agreements. We measure bipartisanship as the percentage of legislators whose ideology overlaps with at least one opposing party member’s ideology. In figure 4.4, we compare the ideological overlap between states with and without campaign finance restrictions. (Our measure simply reflects the area in figure 4.3 where the Democratic and Republican distributions overlap.) It is clear that ideological overlap is greater in states where the parties have no restrictions on making contributions and raising money from individuals. In these states, about 10.5 percent of legislators overlap in their preferences with at least one member of the opposing party, whereas in states with limits only about 8 percent do. While the difference does not appear to be large, in percentage terms the move from 8 percent to 10.5 percent reflects more than a 20 percent increase in bipartisanship.8
Finally, we seek to directly test our expectations that party-centered campaign finance laws would attenuate, and perhaps even reverse, legislative polarization in the states. While our earlier analyses used individual legislators as our units of analysis, in this case we look at the legislatures themselves. Our dependent variable is simply the distance between the median Democrat and the median Republican in each legislative chamber. When that distance is greater, it means that the parties are more polarized in the legislature; when the distance is smaller, it means that there is less polarization. The mean level of polarization during the years of our study was 1.33 (standard deviation of .47). The levels of polarization ranged from 0.42 in the Rhode Island Senate in 2003 and 2004 to 3.21 in the California House in 2007 and 2008.
Our key independent variable is, once again, an indicator of whether a state limited the extent to which parties could raise and spend money on legislative campaigns. When a state was limited in either the amount it could raise from individuals or the amount it could give to candidates, the variable was coded as a 1. When there were no such limits, the variable was set to 0. As we have in other models, we also control for whether the state limited the amount individuals or organizations could contribute to candidates.
One important note regarding these variables is that they are often highly correlated. Among all states, the prevalence of party limits is correlated with the existence of individual limits at .81 and with limits on other organizations at .46. This high correlation makes it more difficult to determine the effects of limits on political parties, especially relative to the effects of limits on individuals. However, it is important to note that the existence of different types of limits is less correlated among the 20 most professionalized legislatures. Among this group of states, the party limits variable is correlated with the individual limits variable at .66 and with the organization limits variable at just .31. In addition, of the 20 states that we identify as having the most professionalized legislatures, 11 had different laws on party versus individual limits during at least part of the time under study. While this amounts to about 11 percent of the cases, we believe that because so many different states have an experience of having different laws on the books at some point during our study, the inferences we draw from the analysis that follows are valid.
In addition to these variables indicating the type of campaign finance laws that were in place, we also include a one-year lag of the dependent variable of polarization. Including the lagged value of polarization as an independent variable in the model means that we are introducing a more stringent test for the impact of campaign finance laws. We attempted other lag structures, but these did not result in a good fit with the data and, in any event, these alternative operationalizations did not alter our main findings. We do, however, include a variable indicating the year of the observation. This information is included since polarization has been known to be increasing nationally during the period covered by our study. Accordingly, we expect the coefficient for this variable to be positive and statistically significant.
In conducting this analysis, we use cross-sectional time series regression. We include chamber fixed effects to attempt to control for any unmeasured differences across the states and chambers. Again, this choice is made to provide for the strongest possible test of whether campaign finance laws matter. Essentially, the results we show below are the effects of campaign finance laws after controlling for other differences across chambers and over time.
Table 4.2 presents the results from two models—one examining the effects for all state legislative chambers in our dataset and the other focusing on just professional legislatures. We focus first on results presented in the first column. Not surprisingly, the coefficients for both the lagged value of the dependent variable and the year are statistically significant and positive. This indicates that the amount of polarization in a chamber at a given point in time is a function of the amount of polarization during the previous year and that polarization is increasing over time. Interestingly, only the coefficient for limits on individual contributions is statistically significant. This indicates that among all chambers, placing limits on contributions by individuals tends to increase the amount of polarization there is in the legislature.9
The second model in the table limits the analysis to the 20 most professionalized legislatures. The lagged value of polarization is again significant in this model, as is the variable for year. Notably, when it comes to professionalized legislatures, only the indicator for party campaign finance limits appears to matter for polarization. States that limit what parties can raise and spend appear to experience more polarization than those that do not have such limits. The coefficients for individual and group limits are small and lack significance in this model, indicating that they do not appear to be associated with polarization in professional chambers.
table 4.2 to plot predictions for a counterfactual comparison in figure 4.5. Specifically, we imagine a state with an average amount of polarization in 1995 (which would amount to a difference of 1.48 in the party medians). We imagine that this state had no limits on parties, individuals, or organizations. We then use our regression model to set this hypothetical state on two different paths. In one path, we imagine that the state continued to set no limits on parties; in the alternative scenario, we imagine that the state enacted limits in 1996. Figure 4.5 shows what would happen to polarization in that legislature depending on which of these paths the state chose to take.
Figure 4.5 demonstrates the important implications that party finance laws have for polarization in state legislatures. In our hypothetical scenario, the state that chose to impose party limits sees polarization increase over the subsequent decade, whereas the state that continued with no limits on parties sees a decline in polarization during that same period. Within 10 years, the decision of whether to place limits on parties produces a 0.3 difference in polarization in legislatures that had identical levels of polarization 10 years earlier. That difference amounts to three-fourths of a standard deviation in levels of polarization; it is a sizable effect. In short, states with parties facing few restrictions tend to see less polarization of parties.
While these are hypothetical estimates derived from our model, simple bivariate tests reveal a similar story. We can compare the change in polarization from 1997 to 2007 in 36 professional legislative chambers that had the same laws on party limits across that entire period. We show this comparison in figure 4.6. Among this group, the increase in polarization was nearly three times as large in the 28 chambers that limited party contributions as it was in the 8 chambers that allowed for unlimited contributions. Specifically, in the states where parties were limited, the distance between the parties increased by .12, on average. However, in the states where parties could raise and spend unlimited amounts, polarization increased by just .04.
Summary So Far
In this book we have been making the argument that the institutional flow of money affects the ideological direction of political parties. In previous chapters we showed that campaign finance laws alter the flow of money, putting money in the hands of pragmatist or purist factions of the party, depending on whether the laws are party-centered or not. We then showed how these laws affect which candidates get money and from whom. In this chapter we observed the real consequences of this flow of money on representative institutions. We were able to demonstrate how the ability to raise and give money by the party organization affects state legislatures. Campaign finance laws appear to confer resource advantages (or disadvantages) on pragmatists in the party organizations, which then affect their capacity to shape the composition of preferences in legislatures.
At the individual level, we showed that officeholders in states with unrestricted party spending are decidedly more moderate than their counterparts in states that restrict party financing. This dynamic is true for members of both parties, and especially so in professional legislatures, where money matters most in campaigns. Based on this finding we were able to show that legislatures in states with “party-centered laws” achieve greater ideological overlap between members of the two parties. This overlap plausibly provides a firmer foundation for creating legislative compromise across parties.
At the institutional level, we demonstrated that there is less polarization in the legislatures of states where parties are able to raise and spend money without limits, even after controlling for other factors that might contribute to polarization. To the degree that legislatures close the distance between the parties, these bodies should become more representative of citizens in the state who, in aggregate, have preferences that fall mostly between the party medians.
The findings in this chapter have important theoretical and practical implications, which we will address in the next two chapters. For now, we simply note that our observations cast doubt on a view that party organizations are simply conduits to pass money to candidates. We observe that well-financed parties generate substantively different outcomes than weakly financed parties. This alone suggests that party organizations matter as institutional actors. Based on this study, we cannot be certain of the degree to which party actions affect the selection of candidates or the voting behavior of officeholders. But we certainly observe changes in the composition of preferences of members in legislatures, and we see how this affects polarization of a legislature over time.
Our results should refute any oversimplified theory of political parties as networks of policy-demanders, with little account for how a well-financed party organization might alter the direction of the party. If parties are merely networks of policy demanders, we should not observe differences in states where formal party organizations have a robust financing role in elections.
And to make a broader point, we have demonstrated that money matters, although not necessarily in the ways that most people think it does. The flow of money does indeed shape the preferences of officeholders, but not necessarily as a quid pro quo or as exerting unequal influence on wealthy donors. What we show is another kind of bias—one that is rooted in the incentives of actors within partisan coalitions who contend for control over the goals of the broader party coalition. For this reason we think it imperative to pay attention to the laws that affect the hydraulics of campaign money. At the very least, the findings in this chapter should dampen the momentum of a reform agenda that focuses solely on an anticorruption approach to regulating money in politics. Such a limited reform agenda would have the effect of clamping down on the financing of party organizations, with potentially toxic consequences for the party system and democratic governance.