Do Interest Group Endorsements Cue Individual Contributions to House Candidates?

Abstract

Non-incumbents face an uphill battle in their quest to raise sufficient funds to compete effectively against seated House incumbents and, alternatively, in competitive open-seat House contests. Interest group endorsements are thought to contribute to their electoral success, but whether endorsements help non-incumbent House candidates raise contributions from individuals, as a component of this success, remains unknown. Their heavy reliance upon individual contributions to finance their campaigns as well as the prominence of the groups making a large number of endorsements justifies an explicit test of this relationship. Using contribution data from the U.S. Federal Election Commission (FEC) between 2006 and 2012 paired with endorsement tabulations in both a set of regression and matching analyses, I uncover evidence that endorsements increase individual contributions to the candidate over the course of the election cycle. The results underpin the value of endorsements to non-incumbent candidates as well as their utility for the endorsing groups wishing to influence House elections.

Keywords

endorsements elections individual contributions fundraising

Congressional non-incumbent candidates’ difficult quest for campaign funds is well documented in numerous political science studies and in the mainstream media’s coverage of the money race during every election cycle. There are multiple factors that make building a fundraising base challenging. First, there is the need to generate fundraising momentum early in the campaign to pay for start-up costs and to build confidence in the competitiveness of the candidate’s bid for office (Biersack, Herrnson, & Wilcox, 1993; Herrnson, 2008, 2012; Krasno, Green, & Cowden, 1994). Second, gaining funding from political action committees (PACs) can be challenging as many PACs follow access contribution strategies by giving money to incumbent members of Congress in hopes of extracting policy outputs (Brunell, 2005; Grier & Munger, 1993; Rudolph, 1999; Stratmann, 2002; Wright, 1985)—resources that non-incumbent candidates can only access if they win. As a consequence, non-incumbent candidates rely primarily upon individual contributions to finance their campaigns (Herrnson, 2008, 2012; Jacobson, 2000). Third, incumbency advantage can also make it difficult for non-incumbents to attract sufficient funds. Relatively unknown candidates must make appeals to reveal their issue positions, ideology, and personal credentials because, unlike incumbents, they cannot point to their record in office (Druckman, Kifer, & Parkin, 2009; Herrnson, 2008, 2012; Jacobson, 2000). In other words, they need to establish their credibility as a worthwhile investment opportunity. One way in which challengers and open-seat candidates make such appeals is by publicly advertising the endorsements they receive from interest groups (Druckman et al., 2009; Herrnson, 2008, 2012). These endorsements are an understudied resource in the non-incumbent candidate’s toolbox, which could affect the fundraising success of the campaign. Nonetheless, whether interest group endorsements help non-incumbent candidates recruit funds from individual donors remains unknown.

This study seeks to address this lacuna in the literature by directly testing whether interest groups’ endorsements of non-incumbent candidates in the general election lead to increases in individual contributions these candidates are able to fundraise. Using U.S Federal Election Commission (FEC) contribution data from the itemized individual and committee files for the 2006, 2008, 2010, and 2012 election cycles in both a set of regressions and matching analyses, I find both the number of endorsements and having at least one quality endorsement from a national advocacy organization significantly increases both the total amount and number of individual contributions received by the candidate over the full course of a single election cycle. In addition, the value of any endorsement rises during presidential election years, when excitement about the election is higher and the flow of money to all political groups and candidates tends to increase (Interviews with Party Officials, May 2010, January 2011, October 2011; Interview with EMILY’s List, June 13, 2014; Interview with Conservative Education 501c, July 21, 2014; Interview with Gun Owners of America, September 3, 2014).

The Impact of Endorsements

Despite an extensive number of studies devoted to the influence of interest groups’ campaign financing role in elections and the subsequent effects of their contributions on election outcomes and congressional members’ behavior in office (see Stratmann, 2005, for a review of the literature), few studies investigate the impact of interest group endorsements in elections and none analyzes the relationship between the receipt of endorsements and candidate fundraising. Rather, the literature on interest group endorsements tends to focus on the influence of endorsements on the voters’ choice of candidate and the candidates’ overall vote share or likelihood of winning (Arceneaux & Kolodny, 2009; Hannagan, Pimlott, & Littvay, 2010; Karpowitz, Monson, Patterson, & Pope, 2011; Kenny, McBurnett, & Bordua, 2004; Lupia, 1994), on interest group members’ support for particular candidates or parties as cued by their group’s endorsements (e.g., Clark & Masters, 2001; Delaney, Masters, & Schwochau, 1990; McDermott, 2006; Rapoport, Stone, & Abramowitz, 1991), or on game theoretic explanations for these relationships (e.g., Baron, 1994; Grofman & Norrander, 1990; Grossman & Helpman, 1999; McKelvey & Ordeshook, 1985).

Where voters are concerned, studies test whether interest group endorsements operate as information shortcuts to help voters identify either the candidate’s ideology or specific issue positions (e.g., Karpowitz et al., 2011; McDermott, 2006) and/or they conceptualize endorsements as lending credibility to the candidate’s bid for office by providing voters with a sense of the quality of the endorsed candidate, particularly for candidates without a track record in office (e.g., Dominguez, 2011). This conceptual framework can reasonably be extended from voters and interest group constituencies to donors. Although studies investigate the motivations of donors (see Brown, Powell, & Wilcox, 1995; Crespin & Deitz, 2010; Francia, Green, Herrnson, Powell, & Wilcox, 2003, 2005; Gimpel, Lee, & Pearson-Merkowitz, 2008; Wilcox et al., 2003), the investment cues and solicitations they receive remain understudied. It seems reasonable to presume, in addition to direct solicitations by parties, candidates, and interest group bundlers (Herrnson, 1992, 2008, 2012; Marshall, 1997; Sorauf, 1992), donors are attentive to other cues indicating the candidate is a worthwhile investment whether because of electoral competitiveness, ideology, issue positions, or other possible competencies that might prove useful in office, such as the candidate’s professional background or education.

Endorsements may operate as one cue that the candidate is worth the donor’s investment. Not only might endorsements provide the donor with clues about the specific qualifications of the candidate, but also, more generally, any endorsement is an indicator that the candidate is connected to one or more pre-established social networks and in many cases, corresponding donor-pools. In general, network ties are thought to benefit candidate fundraising (Bramlett, Gimpel, & Lee, 2011; Crespin & Deitz, 2010; Magleby, 2011). Although variation in the impact of different types of endorsements is not well understood (see Karpowitz et al., 2011), these potential network ties provide good reason to believe a group endorsement might hold a different significance than an endorsement by a single celebrity or politician.

Phone interviews with senior staff at one union, one professional association, two federal PACs, and four 501c groups confirm the value of group endorsements lies not only in their importance as a information-laden heuristics but also in connected efforts by the endorsing group to mobilize supporters and donors on the endorsed candidate’s behalf (Interviews, May 21, 2014a and b; June 13, 2014a and b; June 23, 2014; July 21, 2014; August 22, 2014; September 3, 2014).¹ Although the degree to which groups engage in funds-focused mobilization varies, senior staff reported that most interest groups with which they were familiar, including their own, underwrote the costs of fundraising for the candidate at least in part. Groups offering the low end of support provide information about the candidate on the group’s website and in newsletters (either emailed and/or hardcopy) to members. Groups offering a high level of support go beyond simple publicity by making phone calls soliciting funds and reminding members to vote, assigning staff or volunteer fundraisers to bundle funds from the groups’ membership to fund the campaign, and by issuing tailored appeals for contributions through micro-targeted direct mail to members. While many groups do provide direct contributions to most, if not all, of their endorsed candidates, not every group has the capacity to do this in addition to this range of activities. Nonetheless, the potential value of a group endorsement to a congressional candidate is manifold and for this reason, group endorsements will be the focus of this study—specifically, advocacy organizations’ endorsements including those from unions, professional associations, connected and non-connected PACs, Super PACs, and 501c social welfare organizations, are counted. Candidate endorsements by corporations, political institutions (e.g., state legislatures), newspapers, famous individuals, and political parties are not included in the analysis.

Endorsements of candidates in the general election are also the primary focus of this study. Senior interest group staff reported that the majority of prominent organizations wait to endorse House candidates until either clear front-runners emerge at the end of the primary election or, more usually, at the beginning of the general election cycle when a winner exists (Interviews, May 21, 2014a and b; June 13, 2014a and b; June 23, 2014; July 21, 2014). The process of collecting endorsement data helped confirm this because for many endorsements it was possible to see the dates on the press releases advertising the receipt of the endorsement. In addition, staff explained there are costs to endorsing candidates in the primary election, which is why many groups chose not to do so. If the group backs a losing non-incumbent, the endorsement might lead to a poor relationship with either the winning non-incumbent or with the current incumbent. As the goal of supporting non-incumbent candidates is to gain new allies in Congress, any choice that risks this relationship is considered costly (Interviews, May 21, 2014a and b; June 13, 2014a and b). While groups such as the American Nurses Association and Gun Owners of America (Interviews, August 22, 2014; September 3, 2014) do endorse candidates in the primary and general elections, several groups reported that they only endorse candidates in the general election to prevent any chance of alienating prospective legislative allies and other supporters of the group (Interviews, May 21, 2014a and b; June 13, 2014a and b). In light of these findings, the models discussed below analyze the change in the candidate’s total of individual receipts between the end of the primary election cycle and the end of the general election cycle as the dependent variable, thereby treating the primary cycle receipts as a baseline measure of fundraising capacity in the absence of the full slate of available endorsements.

Unfortunately, for the purposes of social science research, the candidate’s receipt of an endorsement is not a random process. Endorsing organizations decide whom to support, based upon the factors I include as controls in the models discussed below—such as not only the candidate’s previous experience in elected office but also for unknown or even unquantifiable reasons. As a consequence, improvements in fundraising might be misattributed to the endorsement of the advocacy group, when in reality, other factors are to be credited. As the direction of causality is not entirely clear for all these reasons and a full experimental design including all House candidates is not feasible, the remaining option is to conduct a matching analysis to simulate an experimental design in which similar candidates randomly receive endorsements whereas others do not (Hannagan et al., 2010, also recognize this difficulty in their study of EMILY’s List endorsements). Along with the regression results, the matching analyses help confirm the fundraising worth of endorsements and suggest they are a valuable resource for both the candidates hoping to gain a greater number of individual contributions and the interest groups hoping to influence the outcome of the election.

Data and Method

The two explanatory variables of primary interest are the number of endorsements received by the non-incumbent House candidate and a dummy variable indicating whether the non-incumbent House candidate received at least one endorsement from a major national advocacy group, such as the National Rifle Association, National Right to Life, the American Federation of Labor and Congress of Industrial Organizations (AFL-CIO), or the Sierra Club. Congressional candidate endorsements from the 2010 and 2012 general elections were tabulated primarily using candidate profiles on the Center for Responsive Politics’ (CPR) and the League of Women Voters’ (LWV) websites. Most endorsement information for the 2006 and 2008 elections was gleaned from the Library of Congress’ online election archives of campaign websites.² Prominent organizations’ lists of endorsed candidates, the parties’ congressional campaign committees’ websites, and Google searches were also conducted to confirm these findings for each candidate in the sample (2006-2012)—often in these searches, the endorsements were mentioned in newspaper articles featuring the particular contest or in press releases (see Tables A and B in Online Appendix A). A list of the most frequently endorsing organizations designated as “quality” endorsers between 2006 and 2012 is available in Online Appendix A, Table A. Where possible, membership totals as well as the number of endorsements issued were used to help classify these groups as prominent organizations. On average, quality endorsing organizations boasted 2.7 million members and made at least nine endorsements. Examples of lesser-known endorsing organizations can be found in Table B in Online Appendix A. Between 2006 and 2012, there were endorsements by 511 distinct advocacy organizations, not counting individual state chapters of major advocacy groups. Only endorsements for the general election are included because the timing of the endorsement during the fundraising cycle of the campaign is an important factor in terms of uncovering evidence of causation.

Candidate summary files, the itemized individual contribution files, and itemized committee files generated by the U.S. FEC between 2006 and 2012 serve as the primary source for all the variables concerning House candidate campaign finances for the series of models to follow. Contribution data are used to determine whether general election endorsements result in changes in total individual contributions to non-incumbent House candidates between the end of the primary cycle and the end of the general election cycle. The logic behind this dependent variable is that the total amount of individual receipts raised during the primary cycle represents a baseline measure of the fundraising capacity of the campaign prior to the receipt of most endorsements. Thus, the dependent variable gauges the increase or decrease in individual receipts attributable to endorsements received by the non-incumbent candidate during the course of the general election cycle starting from this baseline. This is achieved by calculating the total individual receipts raised by the campaign during the primary election and subtracting this figure from the total receipts the campaign raised from individual donors during the general election cycle. It is important to note that the length of the primary cycle and consequently the general election cycle varies by state and by election year. Thus, it was necessary to calculate each individual non-incumbent candidate’s total primary receipts and total general election receipts from individual donors according to these parameters.³ The FEC records the date each individual’s contribution was made to the candidate’s campaign so it is possible to accurately calculate these totals. For example, in 2006, Ohio held its congressional primary in mid-May whereas New York held its primary in mid-September. Thus, for candidates running in Ohio in 2006, primary receipts from individual donors are added together from January 1 to the date of the primary in May, and general election individual receipts are totaled from mid-May to December whereas conversely the totals for New York candidates reflect the January to mid-September primary period and mid-September to December general election period. Candidates with successful fundraising strategies should report an increase in funds raised between the primary and general election cycles rather than a decrease after controlling for the length of the general election cycle. Thus, the primary dependent variable ranges from −US$703,519 to US$1,309,944 in the regression models.

As a second dependent variable, using the same FEC files, I tabulate the number of individual contributions the non-incumbent candidate received over the course of both election cycles to determine whether increases in the amount raised reflect the mobilization of many or just a few individual donors. Interviews with senior staff at a variety of interest groups suggest endorsements have just such a mobilizing effect. Because every candidate in the sample received at least one contribution from an individual donor, I utilize a tobit regression model left censored at one.

In the same way as the first dependent, party and interest group contributions to the candidate are also totaled separately for the primary and general elections while accounting for variations by state and election year, and the differences are taken. Party contributions include direct, coordinated, in-kind, and independent expenditures for the candidate and against the candidate’s opponent (the latter were matched by using a race identification number as the U.S. FEC associates the totals spent against the candidate with that candidate’s identification number, rather than the identification number of the candidate those expenditures are helping). Interest group contributions include direct contributions to the candidate by federal PACs, as well as independent expenditures and electioneering communication expenditures made by all manner of interest groups that report to the U.S. FEC, including Super PACs. For purposes of comparison between midterm and presidential election cycles across years, all monies are placed in constant 2006 dollars. These variables are included in all the models.

In addition to these financial variables, I control for a range of factors that previous research suggests affect the fundraising and general electoral success of candidates, particularly for non-incumbent candidates. A major determinant of success lies with the candidate’s previous experience in elected office prior to running for Congress (Abramowitz, 1991; Damore & Hansford, 1999; Gerber, 1998; Jacobson, 2000; Jacobson & Kernell, 1981; Lublin, 1994; Squire, 1989; Squire & Smith, 1986)—Such experience is connected to fundraising success because experienced candidates already have an established donor pool from which to recruit an initial set of funds. Consequently, I control for the quality of the challenger using a dummy indicating whether the non-incumbent candidate held previous elected office (Quality Candidate). In addition, the change in the competitiveness of the race certainly affects the amount of funds the campaign can raise as Election Day approaches. To account for this, I use the Rothenberg Political Report’s ratings of race competiveness, which classifies races by increasing competiveness as safe, favored, leaning, titling, and toss-up. Following this classification scheme, I placed races on a five-point scale ranging from safe to toss-up and took the difference between the ratings either at the end of May or June (available reports vary by election year) and the rating at the beginning of November. As a consequence, increases in competition are represented by positive values, no changes in competition are reflected by zero, and decreases in competition are represented by negative values (Change in Competition). The variable ranges from −4 to 4. The degree of support for the candidate’s party heading into the election also makes a considerable difference in the likelihood of success and the willingness of partisan donors to support the candidate’s bid for office. I control for district ideological preferences using the percentage of the two-party presidential vote received by the Democratic candidate (Dem Pres Vote Share). I also include dummy variables representing the party of the candidate (Republican) and whether the contest is for an open seat (Open). Finally, as mentioned previously, a dummy variable reflecting states with primaries in June or earlier (according to each election year’s primary calendar; see www.FEC.gov) is included in all the models to reflect the fact that a longer general election cycle might increase the ability of the candidate to raise a greater amount of funds in the general election than the primary election (Early Primary).⁴

These variables are utilized in the three sets of regression models as well as an analysis using coarsened exact matching (CEM; Blackwell, Iacus, King, & Porro, 2009). Matching procedures mimic random assignment to treatment and control groups in experimental studies. The goal of this analysis is to discover the value of a single type of endorsement on fundraising while controlling more effectively for candidate characteristics and contextual factors that might cloud the direct effects of the endorsement on fundraising success. The dependent variable continues to be increases or decreases in total individual contributions between the end of the primary cycle and the end of the general election cycle. The treatment variable is a dummy variable reflecting a type of endorsement received by the candidate (see Table 3). In Model 1, the treatment variable equals 1 with the receipt of any endorsement and 0 otherwise. In Model 2, the treatment variable represents a quality endorsement, coded 1, or the lack of a quality endorsement, coded 0.

The first step is to construct the matched sample from the full sample of candidates. In CEM, the data are coarsened into substantively meaningful groups that are used to make matches. As this technique is dependent upon good matches, larger samples yield better and more accurate results. The L1 statistic (ranges from 0 to 1, with lower values representing better balance and reliability in the matching procedure) as well as the match rates are used to determine the extent to which the matching analysis produced a reliable set of results. These statistics are used to compare the “fit” of the user-specified or coarsened model with the un-coarsened automated model produced by the CEM algorithm. The idea is to improve upon the fit of the automated model. The specifics of this procedure as well as more information about these goodness-of-fit statistics can be found in the online appendix. After the cases are matched, a simple regression of the treatment on the change in individual contributions is run with the weights generated by the CEM algorithm during the matching process added to it.

Endorsements Support Fundraising Success

The results of the ordinary least squares (OLS) regression models can be found in Tables 1 and 2, which display the results for the number and the quality of endorsement models, respectively, for the 2006, 2008, 2010, and 2012 elections. The first pattern that is immediately apparent is the difference in the observed monetary values of endorsements in midterm versus presidential election years. Regardless of the identity of the endorsing group, endorsements boost fundraising from individual donors more during presidential elections than they do in midterm elections. In the midterm elections of 2006 and 2010, for each additional endorsement from any advocacy group major or minor, the candidate’s predicted individual contribution totals increase, all else equal, by US$3,433 and by US$4,372, respectively (see Table 1). Given that the mean number of endorsements per candidate in the sample is three, most candidates can expect to gain between US$10,299 and US$13,116, all else equal, in a midterm election year. Conversely, in 2008, each additional endorsement is worth US$17,404, and in 2012, each endorsement is worth US$22,210, all else equal. Thus, each additional endorsement in a presidential election year is worth more than three endorsements in a midterm election year on average.⁵

Table 1.

Effect of the Absolute Number of Endorsements on Individual Contributions Totals.

Variables	2006	2008	2010	2012
Number of endorsements	3,433.9*** (942.04)	17,404.02*** (2,395.7)	4,372.6** (1,350.5)	22,210.6*** (2,319.7)
Change interest group contributions	0.108*** (0.015)	0.121*** (0.032)	0.154*** (0.040)	0.060** (0.020)
Change party contributions	0.059*** (0.014)	0.017 (0.053)	0.029 (0.017)	0.097*** (0.014)
Change in race competitiveness	6,115.2 (6,748.5)	27,705.3** (8,304.1)	-4,349.4 (5,403.6)	21,281.9 (11,402.4)
Candidate quality	26,471.5 (14,773.4)	16,134.2 (19,253.1)	26,433.9* (12,482.4)	3,072.4 (16,265.9)
Open	73,774.8*** (19,196.5)	31,067.3 (19,457.6)	−13,919.2 (13,764.1)	41,144.3* (18,914.9)
Republican	−3,780.9 (15,087.9)	60,040.5*** (15,352.3)	22,543.9* (11,121.3)	78,412.04*** (15,811.4)
Dem presidential vote (%)	−504.24 (527.2)	−805.1 (576.6)	−125.3 (176.3)	−983.6*** (273.3)
Early primary state	61,629.5*** (12,617.9)	90,289.6*** (14,973.2)	8,645.3 (10,095.95)	91,169.8*** (14,687.8)
Constant	10,462.9 (26,476.7)	−22,432.4 (31,988.98)	−7,568.96 (10,073.7)	−71,931.4*** (15,092.8)
Number of observations	422	373	473	460
R ²	.49	.46	.19	.45

Note. Standard errors in parentheses. DV = difference between general election and primary election individual contribution totals; all monies in 2006 constant dollars.

p < .05. **p < .01. ***p < .001.

Table 2.

Effect of Quality Endorsements on Individual Contributions Totals.

Variables	2006	2008	2010	2012
Quality endorsement (1/0)	30,234.8* (13,750.3)	89,048.1*** (16,184.8)	22,794.02* (10,632.9)	104,675.9*** (16,385.9)
Change interest group contributions	0.111*** (0.015)	0.156*** (0.033)	0.187*** (0.041)	0.053* (0.021)
Change party contributions	0.06*** (0.014)	0.026 (0.055)	0.025 (0.017)	0.123*** (0.015)
Change in race competitiveness	8,180.3 (6,765.6)	31,662.2*** (8,517.04)	−7,902.4 (5,462.5)	16,838.6 (11,959.3)
Candidate quality	25920.4 (14779.1)	15844.3 (19854.1)	29457.3* (12455.5)	6,993.7 (17410.3)
Open	90,112.8*** (18,801.7)	47,133.3* (19,786.5)	−4,906.7 (13,908.9)	57,050.9** (19,734.7)
Republican	−9,479.9 (15,047.8)	39,022.8* (15,630.01)	23,504.4* (11,397.3)	65,937.4*** (16,515.8)
Dem presidential vote (%)	−335.8 (537.35)	−324.3 (601.9)	−31.33 (179.6)	−524.4 (290.1)
Early primary state	61,679*** (12,765.8)	87,488.3*** (15,410.0)	14,879.01 (10,269.7)	87,226.6*** (15449.3)
Constant	2,939.5 (28,833.7)	−49,642.3 (35,313.1)	−10,884.3 (10,362.9)	−68,152.4*** (16,394.9)
Number of observations	425	372	494	460
R ²	.48	.42	.19	.39

Note. Standard errors in parentheses. DV = difference between general election and primary election individual contribution totals; all monies in 2006 constant dollars.

p < .05. **p < .01. ***p < .001.

Similarly, quality endorsements result in greater returns to the candidate during presidential election years than in midterm election years (see Table 2). The difference in the predicted value of a quality endorsement issued in a midterm versus presidential election cycle is substantial. House candidates receiving an endorsement from the National Rifle Association or the League of Conservation Voters in 2008 as opposed to 2006 can anticipate gaining an additional US$58,814 from the endorsement. Those running in 2012 rather than 2010 are predicted to enjoy a US$81,881 increase in individual receipts, all else equal. This result suggests not only the origin of the endorsement matters but also the electoral context in which the group makes the endorsement can strongly affect fundraising outcomes. Moreover, regardless of the timing, endorsements from prominent organizations are worth more than an endorsement from any organization (see Table 2). A quality endorsement from a major national advocacy group increases the candidate’s individual contributions between the primary election and the general election by an estimated US$30,234 in 2006, US$89,048 in 2008, US$22,794 in 2010, and US$104,675 in 2012, all else equal. This suggests one quality endorsement, such as the U.S. Chamber of Commerce or the AFL-CIO, holds roughly the same fundraising value as approximately five endorsements from less prominent organizations. However, quality endorsements in the 2006 election cycle proved to be an exception. In that year, a major endorsement is worth approximately nine individual endorsements from a diverse array of organizations. In 2006, there were contests for 32 open House seats—a large number for a midterm election. As the most competitive races, open seats typically generate more donor interest and investment (Herrnson, 2008, 2012; Jacobson, 2000). Both of these factors may account for the higher value associated with quality endorsements in that election year. The highly significant coefficient for the open-seat dummy in the 2006 quality endorsement model (see Table 2) certainly indicates this possibility.

Although candidates can expect to receive more in total dollars as a consequence of endorsements, how many individual donors are actually motivated to give in response to the endorsement? The results of the tobit regression analyses are presented in Tables 3 and 4. In the midterm elections of 2006 and 2010, a single quality endorsement increases the predicted number of individual contributions made to the campaign by 69 and 45 checks, respectively. In the presidential cycles, the number of predicted contributions per quality endorsement climbs to 173 and 238, all else equal (see Table 3). Conversely, a single endorsement from any interest group results in an estimated 8 contributions in 2006 and 2010, 39 contributions in 2008, and 46 contributions in 2012 (see Table 4). The maximum number of individual contributions raised by any non-incumbent in this sample was 2,031 with a mean of 209 contributions. Thus, endorsements appear to have a mobilizing effect on individual donors over the course of the election cycle that represents a truly meaningful expansion of the average candidate’s fundraising base, particularly in presidential election years.

Table 3.

Effect of Quality Endorsements on the Number of Individual Contributions.

Variables	2006	2008	2010	2012
Quality endorsement (1/0)	69.06* (29.02)	173.18*** (38.03)	44.87* (20.71)	238.3*** (28.58)
Change interest group contributions	0.0003*** (0.00003)	0.0002** (0.0001)	0.0004*** (0.0001)	0.00002 (0.00003)
Change party contributions	0.0001* (0.00003)	0.0002 (0.0001)	0.00002 (0.00003)	0.0002*** (0.00002)
Change in race competitiveness	10.49 (12.88)	54.09** (17.20)	2.761 (10.61)	7.118 (19.73)
Candidate quality	101.27** (29.98)	73.48 (40.40)	102.2*** (23.31)	9.353 (29.33)
Open	158.25*** (36.56)	88.58* (43.84)	−31.91 (26.74)	89.09* (33.30)
Republican	−89.76** (30.36)	12.99 (34.28)	62.73* (21.97)	24.70 (28.08)
Dem presidential vote (%)	1.575 (1.234)	−0.213 (1.33)	−0.564 (0.340)	0.398 (0.513)
Early primary state	63.28* (26.86)	85.27* (33.45)	8.343 (19.36)	58.15* (26.77)
Constant	−26.02 (64.83)	−20.16 (79.97)	−5.134 (20.06)	−26.24 (29.75)
-2 Log Likelihood	−2,201.94	−2,074.89	−2,062.46	−2,874.15
Number of observations	332	310	337	428
Pseudo R²	.062	.032	.027	.034
Sigma	232.24 (9.232)	275.24 (11.42)	165.70 (6.654)	255.26 (8.922)

Note. Standard errors in parentheses. Left-censored tobit regression at 1. DV = count of individual contributions to campaign; all monies in 2006 constant dollars.

p < .05. **p < .01. ***p < .001.

Table 4.

Effect of Number of Endorsements on the Number of Individual Contributions.

Variables	2006	2008	2010	2012
Number of endorsements	7.857*** (1.784)	39.45*** (5.115)	7.61* (2.701)	46.38*** (3.800)
Change interest group contributions	0.0003*** (0.00003)	0.0001 (0.0001)	0.0003** (0.0001)	0.00003 (0.00003)
Change party contributions	0.0001* (0.00003)	0.0001 (0.0001)	0.00003 (0.00003)	0.0002*** (0.00002)
Change in race competitiveness	4.863 (12.70)	44.19* (16.23)	7.372 (10.58)	15.66 (18.28)
Candidate quality	99.47** (29.27)	69.56 (38.05)	100.17*** (23.42)	4.545 (26.70)
Open	121.04** (36.87)	53.38 (41.67)	−44.56 (27.01)	58.62 (31.02)
Republican	−74.87* (30.02)	66.26* (33.32)	56.98* (21.87)	49.59 (26.15)
Dem presidential vote (%)	0.970 (1.197)	−1.441 (1.241)	−0.622 (0.334)	−0.747 (0.467)
Early primary state	61.09* (26.19)	91.02** (31.48)	−0.659 (19.24)	66.93** (24.73)
Constant	2.133 (58.12)	27.29 (67.91)	−2.139 (19.71)	−23.67 (26.19)
−2 Log Likelihood	−2,195.48	−2,064.55	−1,976.19	−2,842.44
Number of observations	332	310	337	428
Pseudo R²	.064	.040	.027	.045
Sigma	227.84 (0.052)	259.27 (10.74)	161.91 (6.629)	236.09 (8.255)

Note. Standard errors in parentheses. Left-censored tobit regression at 1. DV = count of individual contributions to campaign; All monies in 2006 constant dollars.

p < .05. **p < .01. ***p < .001.

While the evidence thus far suggests endorsements positively affect individual fundraising outcomes, are endorsements and interest group contributions one and the same tactic, or, as senior staff indicated in interviews, are they distinct tactics? Turning to the results in both sets of OLS regression models (see Tables 1 and 2), interest group contributions have a positive impact on campaign fundraising from individual donors without exception. As expected, interest group contributions and expenditures are highly correlated with both the number of endorsements and the candidate’s receipt of a quality endorsement (r = .18 and r = .27, respectively, p < .000). However, tolerance scores for each explanatory variable and variance inflation factors with values close to 1 in each instance suggest multicollinearity is not a threat. Importantly, these statistics indicate while money and endorsements are related, they likely have separate direct effects on fundraising from individual donors.

The regression results above represent the effects for the pre-matched sample. The next step is to examine a set of candidates who share similar characteristics and are competing in similar electoral contexts and, therefore, have a similar likelihood of receiving an endorsement. To estimate the causal effect of the treatment in the matching analysis requires calculation of the sample average treatment effect on the treated (SATT). To calculate SATT, I regress the outcome variable—in this case, the change in individual contributions raised over the course of the election cycle—on the treatment dummy. In Model 1, the treatment is simply whether the candidate received any endorsement or not. In Model 2, the treatment dummy reflects whether the candidate received a quality endorsement or not (see Table 5). I add the strata-based weights generated by the respective CEM logarithms for each matched sample to the two different regression models. In this set of models, the treatment coefficient is the SATT. The components of these analyses as well as the results are displayed in Table 5. In Model 1, between the user-coarsened and auto-coarsened models, there is an improvement in the L1 statistic of 0.21, indicating an improvement in fit (see Table 5). Although the match rate does not improve, the number of additional unmatched cases is not excessive—25 additional cases in the control group and 71 cases in the treatment group are unmatched in the coarsened model (see Table 5). In Model 2, there is significant improvement in the L1 statistic with only a slight decrease in the match rate—18 additional cases went unmatched in the control group and 47 in the treatment group (see Table 5). A discussion of how each of these statistics is used to balance the composition of the treatment and control groups can be found in Online Appendix B.

Table 5.

Coarsened Exact Matching of House Non-Incumbent Candidates.

	Model 1 (any endorsement)				Model 2 (quality endorsement)
	Auto baseline model		User coarsened model		Auto baseline model		User coarsened model
Balance criteria	C	T	C	T	C	T	C	T
Number of cases	588	1,211	588	1,211	772	997	772	997
Number of matched cases	544	779	519	706	707	618	689	571
Number of unmatched cases	44	432	69	503	65	379	83	426
Multivariate L1 distance	0.493		0.280		0.42		0.284
Number of strata	437		455		433		443
Number of matched strata	71		98		78		95
Treatment effect (1/0)	—		16,073.68** (5,461.186)		—		17,410.84** (6,049.3)
Constant	—		27,862.05*** (4,145.92)		—		35,711.86*** (4,072.3)
Number of cases			1,225				1,260
R ²			.01				.01

Note. Standard errors in parentheses; C = control group; T = treatment group.

p < .05. **p < .01. ***p < .001.

In Model 1, it is evident the candidates who received the treatment of a single endorsement from any group can expect to raise US$16,073 more in individual contributions between the primary and general election cycles, all else equal (see Table 5). Conversely, in Model 2, in which the candidate receives a quality endorsement, the predicted individual contribution increase is US$17,410. The first major difference between this analysis and those above lies in the effect sizes. In the matching analyses, the value of a single endorsement from any organization is greater than a single endorsement from any organization in a midterm election year in the regression analyses, but it is worth about the same as a single endorsement from any organization in a presidential election year in the regression analyses. Moreover, even when the regression models are pooled (and clustered by state), the predicted increase in receipts from a single endorsement from any organization is just US$7,422 (results not shown) whereas in the matched sample the predicted increase is US$16,073. Conversely, in a pooled regression model, the fundraising boost associated with a quality endorsement is predicted to be US$68,563, all else equal, whereas the matched sample predicts a US$17,410 return associated with a quality endorsement (results of pooled model not shown). The second major difference is that in the regression models, there exists a great disparity in the value of receiving a single endorsement from any group in comparison with an endorsement from a prominent organization regardless of whether the endorsement is issued in a presidential or midterm election year. In contrast, in the matching analysis, the predicted value of a quality endorsement is only slightly greater than the value of an endorsement from any group—a prominent organization’s endorsement is worth US$1,337 more to the candidate, all else equal. As the matching analysis is restricted to candidates who have a similar chance of success, these estimates likely reflect the true effects of different endorsements better than the regression analysis. The advantage of the regression analysis is that the explanatory variables are placed in competition with one another as a way of determining whether other variables drown out the direct effect of endorsements on fundraising—they do not.

Endorsements as a Fundraising Tool

The results suggest endorsements contribute to the fundraising success of non-incumbent House candidates. In all the models, endorsements increased the amount of individual contributions the candidate raised between the end of the primary election and the conclusion of general election cycle. In the matching analysis, the gains associated with an endorsement from a prominent organization versus a single endorsement from any organization were similar in value—In the first instance, the candidate is predicted to receive US$16,073 more in individual contributions and in the second, US$17,410 more. In the regression models, the worth of a quality endorsement ranged from a low of US$22,794 in 2010 to a high of US$104,675 in 2012. To understand just how meaningful these gains are, it is useful to consider the wide variation in the fundraising success of the non-incumbent candidates who competed in for seats in the House between 2006 and 2012. An examination of total receipts in constant 2006 dollars reveals candidates in the 25th percentile raised US$28,808, those in the 50th percentile raised a total of US$168,061, and those in the 75th percentile raised US$1,183,688, respectively. The maximum raised by any non-incumbent House candidate was a little over US$9 million. Clearly, several endorsements from prominent organizations or a large number of endorsements from lesser-known organizations could greatly improve the candidate’s fundraising portfolio by shifting that candidate into a different fundraising quartile and, as the count data suggest, increase the size of their individual fundraising base. An endorsement from a prominent advocacy organization could generate more than a 100 additional contributions in a presidential election cycle. This is significant not only because money follows money but also because the media track fundraising success as part of their coverage of elections (Herrnson, 2008, 2012) and the parties’ congressional campaign committees also consider the candidate’s ability to independently fundraise prior to making any allocations (Interviews with Party Officials May 2010, January 2011, October 2011). As non-incumbents rely upon individual contributions more than PAC contributions (Herrnson, 2008, 2012; Jacobson, 2000), their overall fundraising profile could be strongly affected by interest group endorsements.

Although many endorsing organizations will also make direct contributions to their endorsed candidates, both the quantitative and qualitative results suggest these two tactics are distinct. When organizations formally and publicly endorse a candidate, the reputation and brand of the organizations become associated with the candidate. And, as discussed, the endorsement usually comes with a significant investment of organizational resources and time (Interviews, May 21, 2014a and b; June 13, 2014a and b; June 23, 2014; July 21, 2014; August 22, 2014, September 3, 2014). Arguably writing a check does not associate the organization with the candidate as closely either in a symbolic sense or in terms of the number of interactions the organization has with the campaign. As discussed below, the interview data suggest endorsements are issued in connection with a larger number of goals than a direct contribution to the campaign or independent expenditure that is made to support the candidate.

When an interest group endorses a candidate, there are two possible cuing effects, one extending to members and close affiliates of the endorsing group and one encompassing a broader pool of partisan donors. As discussed earlier, very few groups publicly state their endorsement of a candidate in a press release and leave it at that. Senior staff at a variety of interest groups, based upon their experiences with their own groups as well as others during the course of their careers, described a range of outreach efforts associated with the group’s formal endorsement (Interviews, May 21, 2014a and b; June 13, 2014a and b; June 23, 2014; July 21, 2014; August 22, 2014; September 3, 2014). Interviews revealed three different approaches with varying levels of donor activation. At the high end of the outreach spectrum lies EMILY’s List (Interview, June 13, 2014). As a non-connected federal PAC, EMILY’s List is devoted to bundling contributions to benefit its endorsed candidates and to raising general awareness of its pro-choice candidates. EMILY’s List provides information on endorsed candidates to its members through electronic communications (such as newsletters), events, and social media as well as in direct mail pieces containing solicitations for money. Their endorsements are also publicized in press releases, on the organization’s website, and through social media accounts. In sum, members are asked to contribute by bundlers using multiple approaches. The Service Employees International Union (SEIU) and the American Nurses Association lie in the middle of the outreach spectrum (SEIU Interview, May 21, 2014, and EMILY’s List, June 13, 2014). Like EMILY’s List, they publicize candidates to members via electronic newsletters, on their website, through press releases, and by direct mail. SEIU also calls their members to mobilize them around Election Day. However, their goal is to activate a wide range of support among members, which includes financial support of endorsed candidates. Members are primarily encouraged to provide contributions to each organization’s connected PAC to benefit the organization’s endorsed candidates. However, a senior staff member at SEIU explained the solicitation is broad enough that habitual donors are likely motivated to give directly to the endorsed candidates as well. At the low end of the outreach spectrum lies a conservative education issue 501c group with a connected PAC (Interview, July 21, 2014). Their endorsed candidates are publicized in electronic newsletters and in emails to members, in press releases, and on their website, but direct solicitation for money is not part of these communications. While a senior staff member acknowledged that their endorsements might encourage members to make contributions to the group’s endorsed candidates, the staffer believes the group’s endorsements helped encourage members to vote for the candidate more than anything else. While there is variation in the degree to which groups directly activate donors among their membership, there is good reason to believe outreach efforts generate both direct and indirect fundraising effects, which increase the candidate’s likelihood of receiving a contribution from a member of the group.

What remains less clear is whether an endorsement might spur contributions from a broader pool of partisan donors, although several elements of the results, in particular, the contribution count data, suggest this as a possibility. The fact that the prominence of the endorsing organization rather than simply the quantity of endorsing organizations increased the returns to the candidate in the regression models, provided a slight gain to the candidate’s predicted fundraising in the matching analysis, and increased the number of contributions to the campaign overall indicates a broader cuing effect. In addition, these findings indicate that during presidential elections, a larger pool of donors is contributing to congressional candidates when excitement surrounding a particular election is higher. There was consensus among the senior interest group staff I interviewed that the amounts of money flowing to congressional races tends to increase during presidential elections making fundraising for these races less labor intensive. There are a number of mechanisms likely triggering this broader cuing effect. Candidates publicize interest group endorsements in press releases, on their websites, and in their own solicitation letters to donors (Interviews, May 21, 2014; June 13, 2014; June 15, 2014; June 21, 2014a and b; June 23, 2014). The Democratic Congressional Campaign Committee and the National Republican Congressional Committee occasionally discuss endorsements received by their candidates on their websites’ blogs (Interviews with Party Official, January 2011). Party fundraisers also use endorsement information to convince donors that the candidate’s issue stances align with the particular policy interests of the donor (Interviews with Party Officials May 2010, January 2011, October 2011). In short, the fundraising benefits may still be attributable to the interest groups’ endorsements, but interest groups are not the only organizations publicizing those endorsements.

In addition, for the group of donors outside of the membership network of the organization, both the number of the endorsements and the quality of those endorsements likely matter because each is an indication of the strength of candidate’s fundraising profile—that is, they indicate whether the candidate represents a worthwhile investment or a lost cause. Organizations and donors, similarly, do not want to waste their limited resources on non-incumbents who passionately represent their issue perspectives but never have a chance of gaining a seat in Congress. Organizations also want their brand name associated with successful candidates more than half the time (Interviews, June 13, 2014; June 15, 2014). For these reasons, senior staff stressed that while candidates must adhere to the organization’s preferred issue stances as a non-negotiable pre-condition for endorsement, this was by no means the only criterion in their decision-making calculus (Interviews, May 21, 2014a and b; June 13, 2014a and b; June 23, 2014; July 21, 2014; August 22, 2014; September 3, 2014). As previous research suggests, the competiveness of the race and whether candidate had any prior experience serving in elected office were both important considerations for all the groups in the interviewed sample (Interviews, May 21, 2014a and b; June 13, 2014a and b; June 23, 2014; July 21, 2014; August 22, 2014; September 3, 2014). The controls in the regression and matching analyses reflect these considerations. In this way, each new endorsement and prominent endorsement helps other investors gauge the electability of the candidate as well as his or her loyalty to particular issue perspectives. In other words, the risk of investment diminishes with more information and less uncertainty.

The counts of individual contributions cannot reliably reveal whether interest group members or a broader pool of partisan donors lie behind this mobilization effect because the U.S. FEC does not maintain profiles of individual donors but instead relies on individual campaigns to report each time a donor makes a contribution. However, the count data does support the presumption held by senior staff in the interviewed sample of interest groups that endorsements play a role in activating donors and thereby significantly expand the candidate’s donor base as well as the candidate’s campaign receipts. Endorsements underwrite the costs of fundraising for non-incumbent House candidates primarily by lowering the costs donors face in identifying candidates for support. Endorsements serve as a heuristic, providing likely donors with an understanding of whether the candidate’s issue stances align with those of the donor. Endorsements also operate as signal to donors and the rest of the electorate identifying the candidates who represent worthwhile investments rather than lost causes. Finally, they trigger mobilization and fundraising efforts on behalf of the candidate, which further facilitate the connection of donors to like-minded candidates. As a consequence, endorsements are a simple but effective fundraising tool.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Anne E. Baker is an assistant professor at Santa Clara University. Her research focuses upon political parties and party polarization in Congress but also extends to congressional elections and the influence of money in politics.

References

Abramowitz

(1991). Incumbency, campaign spending, and the decline of competition in the U.S. House elections. Journal of Politics, 53, 34-56.

Arceneaux

Kolodny

(2009). Educating the least informed: Group endorsements in a grassroots campaign. American Journal of Political Science, 53, 755-770.

Baron

(1994). Electoral competition with informed and uninformed voters. American Political Science Review, 88, 33-47.

Biersack

Herrnson

Wilcox

(1993). Seeds for success: Early money in congressional elections. Legislative Studies Quarterly, 18, 535-551.

Blackwell

Iacus

King

Porro

(2009). CEM: Coarsened exact matching in Stata. The Stata Journal, 9, 524-546.

Bramlett

Gimpel

Lee

(2011). The political ecology of opinion in big-donor neighborhoods. Political Behavior, 33, 565-600.

Brown

Powell

Wilcox

(1995). Serious money: Fundraising and contributing in presidential nomination campaigns. New York, NY: Cambridge University Press.

Brunell

(2005). The relationship between political parties and interest groups: Explaining patterns of PAC contributions to candidates in Congress. Political Research Quarterly, 58, 681-688.

Clark

Masters

(2001). Competing interest groups and union members’ voting. Social Science Quarterly, 82, 105-117.

10.

Crespin

Deitz

(2010). If you can’t join ’em, beat ’em: The gender gap in individual donations to congressional candidates. Political Research Quarterly, 63, 581-593.

11.

Damore

Hansford

(1999). The allocation of party controlled campaign resources in the House of Representatives, 1989-1996. Political Research Quarterly, 52, 371-385.

12.

Delaney

Masters

Schwochau

(1990). Union membership and voting for COPE endorsed candidates. Industrial and Labor Relations Review, 43, 621-635.

13.

Dominguez

(2011). Does the party matter? Endorsements in congressional primaries. Political Research Quarterly, 64, 535-544.

14.

Druckman

Kifer

Parkin

(2009). Campaign communications in U.S. congressional elections. American Political Science Review, 103, 343-366.

15.

Francia

Green

J. C.

Herrnson

P. S.

Powell

L. W.

Wilcox

(2003). The financiers of congressional elections: Investors, ideologues, and intimates. New York, NY: Columbia University Press.

16.

Francia

Green

J. C.

Herrnson

P. S.

Powell

L. W.

Wilcox

(2005). Limousine liberals and corporate conservatives: The financial constituencies of the democratic and republican parties. Social Science Quarterly, 86, 761-778.

17.

Gerber

(1998). Estimating the effect of campaign spending on Senate election outcomes using instrumental variables. American Political Science Review, 92, 401-411.

18.

Gimpel

Lee

Pearson-Merkowitz

(2008). The check is in the mail: Interdistrict funding flows in congressional elections. Journal of Politics, 52, 373-394.

19.

Grier

Munger

(1993). Comparing interest group PAC contributions to House and Senate incumbents, 1980-1986. Journal of Politics, 55, 615-643.

20.

Grofman

Norrander

(1990). Efficient use of reference group cues in a single dimension. Public Choice, 64, 213-227.

21.

Grossman

Helpman

(1999). Competing for endorsements. American Economic Review, 89, 501-524.

22.

Hannagan

Pimlott

Littvay

(2010). Does an EMILY’s List endorsement predict electoral success, or does EMILY pick the winners? PS: Political Science & Politics, 43, 503-508.

23.

Herrnson

(1992). Campaign professionalism and fundraising in congressional elections. Journal of Politics, 54, 859-870.

24.

Herrnson

(2008). Congressional elections: Campaigning at home and in Washington (5th ed.). Washington, DC: CQ Press.

25.

Herrnson

(2012). Congressional elections: Campaigning at home and in Washington (6th ed.). Washington, DC: CQ Press.

26.

Jacobson

(2000). The politics of congressional elections (4th ed.). New York, NY: Longman.

27.

Jacobson

Kernell

(1981). Strategy and choice in congressional elections. New Haven, CT: Yale University Press.

28.

Karpowitz

Monson

Patterson

Pope

(2011). Tea time in America? The impact of the Tea Party movement on the 2010 midterm elections. PS: Political Science & Politics, 44, 303-309.

29.

Kenny

McBurnett

Bordua

(2004). The impact of political interests in the 1994 and 1996 congressional elections: The role of the National Rifle Association. British Journal of Political Science, 34, 331-344.

30.

Krasno

Green

Cowden

(1994). The dynamics of campaign fundraising in the House. Journal of Politics, 56, 459-474.

31.

Lublin

(1994). Quality, not quantity: Strategic politicians in U.S. Senate elections, 1952-1990. Journal of Politics, 56, 228-241.

32.

Lupia

Arthur

. 1994. “Shortcuts versus encyclopedias: Information and voting behavior in California insurance reform elections. American Political Science Review, 88, 63-76.

33.

McDermott

(2006). Not for members only: Group endorsements as electoral information cues. Political Research Quarterly, 59, 249-257.

34.

McKelvey

Ordeshook

(1985). Elections with limited information: A fulfilled expectations model using contemporaneous poll and endorsement data as information sources. Journal of Economic Theory, 36, 55-85.

35.

Magleby

(2011). Electoral politics as team sport: Advantage to the democrats. In Green

Coffey

(Eds.), The state of the parties: The changing role of contemporary American parties (pp. 81-100). New York, NY: Rowman & Littlefield.

36.

Marshall

(1997). Bundling the cash: Why do interest groups bundle donations? American Review of Politics, 18, 291-308.

37.

Rapoport

Stone

Abramowitz

(1991). Do endorsements matter? Group influence in the 1984 democratic caucuses. American Political Science Review, 85, 193-203.

38.

Rudolph

(1999). Corporate and labor PAC contributions in the House elections: Measuring the effects of majority party status. Journal of Politics, 61, 195-206.

39.

Sorauf

(1992). Inside campaign finance: Myths and realities. New Haven, CT: Yale University Press.

40.

Squire

(1989). Challengers in U.S. Senate elections. Legislative Studies Quarterly, 14, 531-547.

41.

Squire

Smith

(1986). A further examination of challenger quality in Senate elections. Legislative Studies Quarterly, 21, 235-248.

42.

Stratmann

(2002). Can special interests buy votes? Evidence from financial services legislation. Journal of Law & Economics, 45, 345-373.

43.

Stratmann

(2005). Some talk: Money in politics. A (partial) review of the literature. Public Choice, 124, 135-156.

44.

Wilcox

Cooper

Francia

Green

Herrnson

Powell

. . . Webster

(2003). With limits raised, who will give more? The impact of BCRA on individual donors. In Malbin

(Ed.), Life after reform: When the Bipartisan Campaign Reform Act meets politics (pp. 61-79). New York, NY: Rowman & Littlefield.

45.

Wright

(1985). PACs, contributions, and roll calls: An organizational perspective. American Political Science Review, 79, 400-414.