Recusal as Remedy: Disincentivizing Donors

Abstract As judicial elections become increasingly expensive, recusal has emerged as a way to address concerns about the impartiality of judges who receive contributions from lawyers or potential litigants. While it is unclear if strict recusal rules are the best remedy for conflicts of interest created by contributions, they may disincentivize potential donors from investing in judicial campaigns by negating their potential goal of influencing decisions. We consider whether donor behavior in judicial campaigns – especially for those donors most likely to be interested in specifically currying favor with judges – responds to differences in recusal standards. Using data from 219 state supreme court races in 22 states from 2010 to 2020, we find that states with stricter recusal rules attract fewer campaign donations to judicial races, and states with more lax rules attract more overall and, most especially, for attorney donors.


Introduction
In a speech to Fordham Law School, former Justice Sandra Day O'Connor expressed her characteristic scorn for electing judges."No other nation in the world" elects its judges, she said, "because they realize you're not going to get fair and impartial judges that way" (O'Connor 2008).Of most concern to most observers: Money.
Judicial elections have been part of the American legal landscape since the midnineteenth century (see Goldschmidt 1994).While this method of selecting judges is popular, the process creates a "fundamental tension between the ideal character of the judicial office and the real world of electoral politics" (Chisom v. Roemer. 1991, 501 US 380, at 400) and for courts reliant on legitimacy, money in elections becomes worrisome (Gibson and Caldeira 2012).The concern has only been amplified with judicial elections becoming more contentious and costly every year (Bonica and Woodruff 2012;Bonneau and Hall 2017;Geyh and Thrapp 2017;Gibson et al. 2011;Gibson and Nelson 2017).
Still, most states favor elections, and while some scholars demonstrate that campaigns for such elections can cause trouble for legitimacy, especially via fundraising (Gibson 2008;Gibson et al. 2011;Goodwin-Gill 2006;Peters 2018;Streb 2007), there is evidence that electing judges also enhances legitimacy through accountability (Gibson 2012;Gibson et al. 2011;Woodson 2017) and serves an important informationprovision function (Bonneau and Hall 2017).However, mitigating any appearance of bias (or actual bias) that may occur due to fundraising pressures continues to be necessary, and there is ample reason to believe that when judges participate in cases where the litigants are campaign donors, at the very least, an appearance of bias results. 1  In many cases, state courts have turned to recusal as a remedy.When judges have before them a litigant who has contributed to their campaign, they can limit the appearance of bias that might result through recusal.Indeed, several states counsel recusal when donations reach problematic levels, making recusal rules a way to help ensure impartiality (or at least its appearance). 21 A 2011 survey found that 94% of North Carolina voters believed that campaign contributions have some influence on judicial decision making.Moreover, 43% of respondents thought that contributions greatly influenced decisions (Bonneau 2017).In 2009, a USA/Gallup survey showed that almost 90% of respondents considered the effect of campaign contributions on judges' rulings to be a problem (Biskupic 2009).Justice at Stake reports that 81% of citizens were concerned by the fact that, in some states, almost half of all supreme court cases involve individuals that contributed to the campaign of one of the judges (2002).Even comedians are riffing on the problem of campaign contributions to judges, with John Oliver quipping, "Giving money to judges wouldn't be acceptable in a state fair squash growing competition" (Oliver 2015)!And regardless of whether the research can demonstrate that campaign contributions actually drive judicial decision making (and the findings are decidedly mixed; see, e.g., Gibson and Nelson 2017, Bonneau and Cann 2009, Cann 2007, Champagne 1988, Hazelton, Montgomery, and Nyhan 2016, Ware 1999, Waltenburg and Lopeman 2000, Kang and Shepherd 2011, Cann 2002, Cann, Bonneau, and Boyea 2012), an appearance of bias may well be enough to damage the legitimacy of the state courts (Rottman 2009;Benesh 2006).In addition, these potential conflicts of interest are avidly reported upon and are fodder for judicial campaigns.In the recent Wisconsin Supreme Court race, a widely-televised Protasiewicz ad accused her opponent, Kelly, of ruling in favor of the Wisconsin Institute for Law and Liberty six times while taking money from its board members (Petrovic 2023).The media then reports on and amplifies such ads, especially when they are negative (Hughes 2022;Ridout and Smith 2008).
Judges themselves have joined the chorus.Approximately 60% of New York State judges reported that they believe campaign contributions raise reasonable questions about judges' impartiality (Commission to Promote Public Confidence in Judicial Elections 2006).Sonia Sotomayor, writing in Woodward v. Alabama (2013), argued that Alabama's partisan elections have "cast … a cloud of illegitimacy over the criminal justice system" (571 U.S. 1045).In a poll of 2,428 state court judges, 46% said that they thought campaign contributions influenced their decisions in some way (Justice at Stake 2002).And then there's Ohio Supreme Court Justice Paul E. Pfeifer's lament: "I never felt so much like a hooker down by the bus station in any race I've ever been in as I did in a judicial race… [Attorneys] mean to be buying a vote, whether they succeed or not, it's hard to say" (quoted in Grimes 2009, 863).Attorneys seem to agree with Judge Phifer's conclusion: A survey of Texas attorneys revealed that over 79% think judicial campaign contributions significantly affect decision-making (Spanhel 1999).
2 There is little research on whether recusal actually counters perceptions of bias created by, among other things, campaign contributions.There are some theoretical considerations in the law reviews.Bam (2011) argues that current processes likely harm legitimacy by calling attention to bias, especially given the media's propensity to cover only those cases where recusal is requested and the judge denies it.Marbes (2017) notes that our presumption of judges' impartiality is the real problem and suggests procedural reforms including use of per se rules for recusal along with ending self-recusal.Phillips and Poll (2007) argue for auto-strike provisions (like some states use), so that recusal can be done without discussion or debate.
The only attempt to empirically ascertain whether recusals move public opinion is Gibson and Caldeira (2012).Employing an experimental vignette modeled after the dispute in Caperton, Gibson and Caldeira find

Recusal and campaign contributions
The US Supreme Court decision in Caperton v. A.T. Massey Coal Company (2009, 556 US 868) involved a judge who had received significant campaign contributions from one of the parties involved and refused to recuse himself.The Court held that the judge's failure to recuse himself violated the due process rights of the opposing party, as it created a significant risk of bias.The Court explained, "[t]here is a serious risk of actual bias when a person with a personal stake in a particular case had a significant and disproportionate influence in placing the judge on the case by raising funds or directing the judge's election campaign when the case was pending or imminent" (Caperton, at 871).The decision clarified that the Due Process Clause requires recusal in instances when litigants demonstrate actual bias and when certain "extreme facts" create an unacceptable "probability of bias" (Caperton,(886)(887).
The Caperton decision prompted the American Bar Association to renew its call to states to include recusal rules relating to campaign spending in their Judicial Codes. 3 The ABA had already included this logic in the 2007 Model Code, which called for recusal when "[t]he judge knows or learns by means of a timely motion that a party, a party's lawyer, or the law firm of a party's lawyer has … [within a time period determined by state] made aggregate contributions to the judge's campaign that is greater than [amount to be determined by state]" (Rule 2.11(A)(4)).However, at the time of this writing, only a handful of states have adopted a version of these rules, and two states, Nevada and Wisconsin, expressly rejected such reforms (Corriher and Paiva 2014).
It may be, however, that removing the perception of bias (and thereby, perhaps, enhancing legitimacy) is not the most intriguing effect of these recusal rules.Indeed, many policies have unintended consequences, which are also important to consider.In our case, we suggest that these recusal rules, regardless of whether they boost perceptions of the courts as impartial, may influence donor decisions to contribute (or not) to judicial candidates.When recusal rules, which are designed to negate an appearance of bias, are adopted by states, counsel recusal whenever donations exceed that recusal can moderately improve the public's perception of a court or judge but that "the effect of recusal is far from the complete restoration of the institution's impartiality and legitimacy" (Gibson and Caldeira 2012:18).And, one might expect that criticism of experimental effectsthat they are nearly always likely to be the maximum potential effectis especially operative here where is it unlikely the members of the public know about recusal or very often notice when a recusal has taken place.Not true, potentially, of those engaged enough to contribute to judicial campaigns.See the next section.
3 Concern for the partiality of candidates is unique to judicial elections.States that utilize elections to staff their judiciaries have promulgated codes of judicial conduct comprising a set of electoral rules "which in the interest of promoting the independence, impartiality, and integrity of the courts, constrain what judicial candidates can say while campaigning, how they may interact with political organizations, and how they may participate in fund-raising" (Peters 2018, 1).These rules, of course, would be shocking and likely unconstitutional if promulgated for other elected offices.
Judicial elections are addressed in Canon 4 of the American Bar Association's Model Code: "A judge or candidate for judicial office shall not engage in political or campaign activity that is inconsistent with the integrity and impartiality of the Judiciary" (American Bar Association 2011), which derives historically from the American Bar Association's Model Canons of Judicial Ethics (Peters 2018).Three campaign activities believed to harm judicial independence were targeted in the 2011 Code: candidate speech, political party involvement, and soliciting campaign donations.The federal courts have weighed in on many of these provisions, generally (though not universally) upholding limits on campaign promises, party engagement, and personal solicitation of campaign funds.
State Politics & Policy Quarterly some threshold, donors experience a new roadblock to influence.In effect, races in states with recusal rules become less attractive to donors looking to gain influence.As a result, moneythe root of most of the criticism of judicial electionsis less prevalent there.

Donor motivations in judicial elections
However, is it reasonable to assume that donors to judicial campaigns are primarily seeking influence?Ansolabehere et al. (2003) suggest that, in Congressional campaigns, individuals are unlikely to be motivated solely by the desire to influence the individual decisions of those to whom they contribute.In other words, it is unlikely that individuals calculate returns on investment and much more likely that individuals are ideologically motivated in their donation decisions.Canes-Wrone and Gibson (2019) and Barber, Butler, and Preece (2016) agree, noting that individual motivations to donate appear to be policy-related or ideological at their core, given their lack of long-term relationships or desire for future access to legislators (Barber, Butler, and Preece 2016).Other scholars make the case that self-interest (material interest) is at least part of the donation calculation (see, e.g., Grier and Munger 1991;Francia et al. 2003;Brown, Powell and Wilcox 1995).
The focus of all of these articles, though, is on Gubernatorial or legislative (often Congressional) races, which draw far more donations and feature candidates who do not have specific professional groups that may be individually interested in the decisions they make.Judges may well be different.First, research shows that donors receive favorable treatment when they go before a judge they support (e.g., Cann  2007; Cann, Bonneau, and Boyea 2012; Waltenburg and Lopeman 2000; Ware  1999) 4 , making it reasonable to expect donors to consider factors that aid or impede their ability to "cash in" on their contribution.Second, this should be especially evident among attorneys, and attorneys typically account for a sizeable proportion of contributions to judicial candidates (e.g., Goldberg et al. 2005).As repeat players (e.g., McGuire 1995;Sample and Pozen 2007), attorneys have more opportunities to benefit from the election of their preferred judge (Williams and Distlear 2007, 137).Miller and Curry (2013) link attorney donors to the concept of "investor" donors conceived by Francia et al. (2003: 43), who provide the following description of this class of donors: These "investors" typically desire broad policies to benefit their industry, narrow policies to benefit their company, or even narrower policies to benefit themselves.Investors often view a contribution as part of a larger government relations strategy.They give contributions to make it easier for a lobbyist to get the attention of a member of Congress the first time, to thank a member for supporting a specific piece of legislation, or to help maintain an ongoing relationship that affords company or industry lobbyists continued political access.
Furthermore, as legal professionals, attorneys should be especially familiar with the judicial candidates and a state's recusal rules (Miller and Curry 2013: 131).
Therefore, we expect a relationship between contributions and recusal overall, particularly for attorneys and other groups who bring litigation, including business and labor groups.(After all, the Caperton matter did not involve an attorney but the CEO of a large coal company.)Other groups, say, single-issue ideological groups, might be more like contributors to Congressional campaigns, driven not by their own potential successes in court but rather by their interest in the over-time decisionmaking cumulating into preferred policy.We test for these possible differential influences by testing a model of all donations as well as models of contributions from subgroups most and least likely to be interested in access to specific judges.In the aggregate, we find that recusal rules matter to donation decisions.But they matter most to attorneys and least to single-issue ideological groups.These results hold with controls for other aspects of the race that likely influence donations and other characteristics of the state's rules that are likely related both to recusal rules and donation decisions.

Modeling campaign contributions
Because we do not have the means to test the relationship between recusal rules and donations using inference-friendly techniques, we must settle for an observationbased strategy.We pay particular attention to the specification of our model in terms of other independent effects on campaign contributions and other aspects of the state context that may be related to their decisions to have strict or lax recusal rules.
The current literature on campaign fundraising suggests that a number of factors influence fundraising.Personal characteristics of the candidate, like race and gender, may be influential with some extant research showing that women and minorities are less successful fundraisers (see, e.g., Thompson, Moncrief, and Hamm 1998;Barber, Butler, and Preece 2016; but see Hogan 2007).Incumbency is also likely to matter in donation decisions (Bonneau 2007, Jacobson 1997.)Researchers have identified a set of electoral characteristics influential on the decision to donate to a particular race as well.According to Bonneau, "the context of the race can either assist or hinder the amount of money a candidate can raise, independent of anything else" (Bonneau 2007: 71).Candidates in partisan elections raise more than candidates in nonpartisan elections (Bonneau 2007).And, when more than one seat is up for election, candidate fundraising decreases (Bonneau 2007).
Also important, though, are potential influences that might explain both our independent variable of interestrecusal -AND the levels of campaign contributions we see in a particular race.To avoid omitted variable bias, we also consider the following characteristics of state judicial elections that suggest a state particularly concerned with the potential ill-effects of judicial campaigns: term length (longer terms are expected promote independence); odd-year elections (designed to decrease attention and partisan influence on judicial elections); nonpartisan election (in attempts to remove party influence); state-wide elections (limiting accountability); professionalism (enhancing the likelihood of quality judges who focus on law); campaign finance regulation (in attempts to limit influence); and whether or not the state employs a Judicial Campaign Conduct Committee (used to safeguard judicial independence).It is quite likely that states who employ these "protections" of their courts would also be likely to promote recusal for campaign contributions (and that these protections will result in fewer campaign contributions).Including them allows us to treat the influence of recusal more carefully, making our claim that recusal affects donors more robust.(Table 1 provides all variable descriptions.)

The influence of recusal rules
In addition to all of the above influences on levels of campaign donations, we posit our main hypothesis: Recusal rules that mention campaign contributions as a basis for recusal or allow recusal on that basis will decrease election fundraising as compared with states with rules that forbid recusal for campaign donations.Specifically, we argue that a state's recusal rules will affect the total amount of fundraising in an election.If a state's code of judicial conduct calls for disqualification in cases involving campaign donors, then any expected payoff from a campaign contribution disappears.Therefore, total election fundraising in these states should decrease.On the other hand, if a state's code of judicial conduct expressly says that contributions CANNOT be a basis for recusal, contributors may be more attracted to investing in its judicial campaigns.In these states, individuals and businesses can contribute without worrying about repercussions and gain the access or advantage they presumably seek (or at least maximize that probability).Therefore, total fundraising in states with more relaxed recusal procedures will be greater than in states with more stringent recusal rules.
While recusal may not directly fix any legitimacy problem that electoral contributions might create (see, e.g., Gibson and Caldeira 2012), recusal has the potential to get to the root of the actual problem.Strict recusal rules, while surely designed with legitimacy and due process in mind, may serve to disincentivize electoral investment in the first place.In other words, regardless of Gibson and Caldeira's findings, recusal could decrease the very thing that worries everyone about judicial elections: their cost. 5 We model donor decisionsthe logged total amount of spending on a given state supreme court raceto ascertain whether or not recusal rules influence them.We consider many races over time and across states and a fully specified model to determine whether recusal rules serve to disincentivize donors from investing in judicial elections.Evidence that states that employ the most stringent recusal rules attract fewer campaign contributions would provide evidence that such rules are one way by which states can maintain the benefits of judicial elections while decreasing their legitimacy costs, even indirectly. 6

Measurement and model
We argue that total contributions, rather than the number of large contributions, is the most relevant dependent variable here for a number of reasons.First, if donors are reacting to recusal rules, we suspect they will respond regardless of how large a donation they had planned on making.Regardless of size, donors will experience a disincentive to donate to judicial elections in states with rigid recusal rules.Second, focusing only on "large" donations would require determining the level at which such donations portend legitimacy problems, which is subjective, at least in part a function of overall spending in the race, and less 5 Of course, this comes at the expense of information, which, Bonneau argues in much of his work, decreases the very democratic accountability we seek when we elect our judges.6 Miller and Curry (2013) also use this strategy of evaluating recusal rules effectiveness.They study campaign donations to Supreme Court candidates in Alabama finding that "the existence of a per se recusal statute significantly decreases the likelihood of observing large donations from several categories of donors" (Miller and Curry 2013: 125).Their focus only on one state without considering the variety of recusal rules employed by the states limits their generalizability.Their attention to Alabama in particular further limits generalizability and also, potentially, validity given that AL's rule is statutory and unenforced.(Most recusal rules are adopted by the judiciary.)Additionally, they do not offer a fully-specified model of campaign contributions.Our study builds on their initial intuition.likely to be noticed by the public.In addition, the overall amount of money in judicial elections is what motivates legislators and judiciaries to make rules requiring recusal in response to public opinion about the effects of increasing amounts of money in judicial elections.Finally, the most frequent contributors to judicial campaigns are attorneys, though they tend not to be among the largest donors.Therefore, a measure that counted only the largest donations could potentially overlook the group most likely to benefit from the election of a particular judgethose who actually appear before them.Examining total campaign fundraising provides a comprehensive view of the donor landscape, ensuring that all forms of participation are acknowledged.
In order to measure the influence of recusal rules on donor behavior, then, we focus on total contributions per race (logged), and analyze multiple states over several elections in contested and uncontested elections. 7We collected data on judicial fundraising for 219 contested and uncontested partisan and nonpartisan elections to the state's highest appellate court, along with the recusal rules in 22 states for the years 2010-2020. 8Data on fundraising for state supreme court races was obtained from the National Institute of Money in State Politics.Due to differences in state campaign financial disclosure requirements, contribution data are only available for these 22 states.Fundraising data for each candidate were aggregated by election to capture the total fundraising for each race.The contributions were adjusted to constant dollars (2020) using the Consumer Price Index to make the estimates comparable over time.Thus, the dependent variable in this analysis is the log of the total amount of money raised in each election, in 2020 dollars.9Data on candidate, institutional, and electoral characteristics for 2010-2012 came from Brent Boyea's dataverse (Boyea 2016).For elections after 2012, we gathered information from state court websites, state secretary of state websites, and from the New Politics of Judicial Elections reports published by Justice at Stake and the Brennan Center for Justice after each election cycle.10

Recusal rules
To measure each state's recusal rules regarding campaign contributions, we examined the recusal requirements dictated by each state supreme court's code of judicial conduct and coded them for their content.State codes of judicial conduct for the year 2010 were obtained from Peters (2018) via his dataset of 37 state codes of judicial conduct over a 21-year period. 11We gathered data on state codes of conduct for the years 2011-2020 through state supreme court websites.Using the ABA's Model Code as a reference, we coded state rules into one of three categories that range from statements that contributions may be grounds for recusal if they implicate bias, to those who do not mention campaign contributions at all, to specific requirements that campaign contributions not be considered grounds for recusal.(Table 2 shows our categorization of each state in our study; Table 3 shows median fundraising by category; Appendix B provides the language on which we based our coding.)The relevant portion of the ABA's Model Code says the following: (A) judge shall disqualify himself or herself in any proceeding in which the judge's impartiality* might reasonably be questioned, including but not limited to the following circumstances: (4) The judge knows or learns by means of a timely motion that a party, a party's lawyer, or the law firm of a party's lawyer has within the previous [insert number] year[s] made aggregate* contributions* to the judge's campaign in an amount that [is greater than $[insert amount] for an individual or $[insert amount] for an entity] [is reasonable and appropriate for an individual or an entity].(American Bar Association 2011).
Our measure ranges from 0 to 2, measuring how stringent state recusal regulations are as they pertain to campaign contributions.As mentioned earlier, following Caperton in 2009, a handful of states put new recusal rules into place, which vary in scope and language used.If a change occurred during our study timeframe, we updated that state/year accordingly. 12 Some states follow the Model Code and have adopted rules counseling disqualification for contributions, often worded with reference to the Caperton decision.These include Arkansas, Georgia (2012-2020), Michigan, Mississippi, North Dakota, New Mexico, Pennsylvania, and Washington. 13These states, which explicitly mention that large campaign contributions should be a reason a judge should recuse, are coded 2 (see Appendix B for the actual wording of the rules).
States with disqualification rules that make no mention of campaign contributions, neither making recusal necessary nor saying recusal is not necessary, were scored as 1.These states -Alabama, Georgia (2010), Idaho, Illinois, Kentucky, Louisiana, Minnesota, Montana, North Carolina, Oregon, Texas, and West Virginia have recusal rules that simply provide that a judge ought to disqualify whenever his or her impartiality might be reasonably questioned without detailing why it might be.
Finally, some states adopted rules that explicitly state that contributions CAN-NOT BE the sole basis for recusal; they are coded 0 on our measure.Ohio, Nevada, and Wisconsin have these rules.We use fundraising in these states as our baseline, given that these rules are the most encouraging to interested donors.

Other influences on contributions and recusal
In order to ascertain whether recusal policies influence donors, we control, as mentioned above, for a variety of other potential influences on their behavior, including race-and candidate-specific characteristics shown in the literature as influential on campaign contributions.We also think carefully about the sorts of states that adopt strict recusal rules and consider omitted variables that might influence both that adoption and the levels of fundraising in the state, connoting as they do the comfort level of state legislatures with the election of judges.
We rely on Bonneau (2007) and Boyea (2017) to identify specific influences on judicial elections and on other, general elections research for the rest.In the interest of space, we present Table 1, which provides a description of each of the variables and how they are measured.As shown there, we control for characteristics of the race (including incumbency, contestation, position, minority candidate, and female candidate), the electoral context (competitiveness, potential shift in balance of the court), institutional characteristics (number of seats, tort docket), and state characteristics (multi-seat election, voting-age population, and attorneys per voter) in our model of campaign contributions. 14 In addition, we control for influences that may be related both to our primary independent variable of interest (recusal) and our dependent variable (total logged spending), thereby considering the potential impact of omitted variable bias, as mentioned above.We call these "State Acceptance of Elections" variables, and they include term length, odd year, partisan election, district election, professionalism, campaign finance regulation, and whether or not the state employs a Judicial Conduct Committee, as detailed above.Their coding information is also in Table 1.

Data and analysis
Appendix A provides summary statistics.As demonstrated in Table 3, median contributions to state supreme court races appear to vary across the categories of 14 All of these variables derive from the literature and some have had only mixed success in explaining campaign contributions.In addition to Bonneau (2007) and Boyea (2017), see Berch (1996), Brace and Hall (2001), Bonneau (2004Bonneau ( , 2005)), Bonneau and Cann (2011), Bonneau andHall (2009, 2017), Frederick and Streb (2008), Hall (2007Hall ( , 2015)), Hall andBonneau (2008, 2013), Herrick (1995), Hogan (2007), McMahon (2007), Ruckman (1993), Shipan and Shannon (2003), Sorauf (1988), Squire (2008), Wilhite andThielemann (1986), andWitko (2015).A few coding notes are necessary.First, there are six multi-seat elections in our dataset.For those six, we treat them as open if more than half the seats are available to non-incumbents.This means that PA 2015 and WV 2012 are open seat races, while MI 2010MI , 2012MI , 2014MI , 2018 are not open seat races.In multi-seat elections, if any of the seats is for the Chief Justiceship, we code the election to be for Chief.In six elections in our dataset, judges are chosen in elections with multiple seats and multiple candidates.All six, then, are labeled competitive under this measure.Given a lack of agreement over how best to measure competition in multidistrict legislative races, let alone judicial races that are conducted in a nonpartisan manner (see, e.g., Niemi, Jackman, and Winsky 1991, Cox, Fiva, and Smith 2020, Weber, Tucker, and Brace 1991, Niemi, Jackman, and Winsky 1991) and given the vote distributions in those six races, we are comfortable that treating them as competitive taps the underlying phenomenon.Media coverage of each election determined whether the ideological balance of any given court was at stake in any given election.We include Michigan and Ohio as partisan election states.Although party information is not displayed on the ballot, candidates for the Michigan Supreme Court are nominated at Democratic, Republican, and minor party conventions (see: Michigan Election Law Section 168.392).Moreover, parties are deeply involved in election campaigns."[I]n terms of campaign style [and] the behavior of judicial candidates … Michigan More closely resembles a partisan state" (Cann, Bonneau, and Boyea 2012: 42).Likewise, candidates for the Ohio Supreme Court are elected through nonpartisan general elections, but they first must participate in partisan primary elections.Moreover, political parties endorse judicial candidates and are very involved in campaigns.Some studies treat these states as "quasi-partisan" (see Nelson, Caufield, and Martin 2013).We estimated a model treating Ohio and Michigan as "quasi-partisan" and the results did not change.Finally, because Louisiana districts 1, 6, and 7 all include portions of both Jefferson and Orleans Parishes, we use the total population of both parishes as our measure for those districts.recusal rules, especially at the extremes.To wit, states that forbid recusal for contributions have far higher medians, over time, than those who counsel recusal in light of large contributions.This provides initial support for our assertion that recusal rules might affect donor decisions.
To isolate that effect from all the other variables that likely influence donor decisions, we estimate a multivariate model of (logged) campaign contributions.Since the dependent variable is continuous, we use ordinary least-squares regression (OLS) to estimate the model, using robust standard errors clustered by state. 15Table 4 displays the results for the full model, including all influences on campaign contributions noted to date in the literature.
The model's explanatory power appears to be strong (adjusted R 2 = 0.694).In addition, the statistical results are largely consistent with the findings of previous studies.We find evidence for race-specific influences (a race with an open seat, a contested election, or for the Chief Justice's seat are all more expensive), for competitiveness (more competitive races generate more money), and for state acceptance of election characteristics (partisan elections generate more money; more campaign finance regulation results in less money).While not all of the influences we test exert statistically significant effects on spending in state supreme court races, many of the categories of effects have at least some impact.Given that many of the effects for which we test have been deemed unimportant in previous literature as well, including the number of female or minority candidates, previous competitive election, oddyear election, size of the voting-age population, number of seats at stake, and tort docket, we feel comfortable that the model performs well.
Most important for our analysis, though, is the effect of recusal rules on donor incentives.As shown in Table 4, when contributions are mentioned as potential reasons for recusal, fewer campaign dollars flow into the races as compared with the baseline (the recusal rules that explicitly state that contributions alone cannot be a basis for disqualification).Even a recusal rule that says nothing about campaign contributions results in fewer campaign dollars than elections in states that ensure recusal will not result from campaign contributions.This suggests that both the presence of a rule that contributions matter and, even more strongly, the presence of a rule that contributions cannot, influence donor decisions. 16n states with rules that do not mention contributions as a basis for recusal, the total contributions decrease by 75% (p < 0.01) over states for which contributions cannot be the basis for recusal.State supreme court races in states with rules mentioning campaign contributions as a basis for recusal see a 68% decrease in total campaign contributions (p < 0.05).Further analysis shows that much of the work of the recusal effect is a distinction between states that FORBID recusals and all othersentering the variable as a single dummy that notes whether recusal is forbidden for campaign contributions or not, results in a statistically significant (p < 0.01) and positive value, increasing fundraising by 276% (not shown).It appears to be the case that recusal rules are signals to potential investors of their potential influence.
To further test our theory, we estimate two additional models, isolating the types of donors that, according to our theory, should be most and least likely to be influenced by recusal rules (e.g., those most and least likely interested in garnering favor) to ascertain whether they, in fact, are more or less influenced by a given state's recusal rule.We start with the group most likely to fit our donor profile of specific influence: attorneys.17As shown in Table 5, recusal rules have an even stronger effect on the donation decisions of attorneys than they do in the aggregate, even while some of the acceptance of elections variables exert stronger influences than the full set of donors.Attorney contributions decrease by over 78% (p < 0.01) in states with rules that do not mention contributions as a basis for recusal over states that forbid contributions to be the basis for recusal.Elections in states that mention contributions see fundraising decrease by more than 75% (p < 0.01).This suggests that attorneysa group particularly well-positioned to benefit from the election of their preferred judicial candidateare especially sensitive to recusal rules when donating to a state supreme court election campaign.
But, as expected, ideological groups are not at all influenced by recusal rules.In fact, few of the influences found in the literature matter to those groups, focused as they likely are on long-term ideological goals. 18iscussion Judicial elections are not just like every other election, though many of the same forces apply to the choices voters make over candidates (Hall 2001).Only for judicial elections are there rules governing behavior during elections that range from prohibitions on making statements that may appear to commit a judicial candidate to a particular position to regulations about how campaign contributions can legally be solicited to requirements about recusal in cases involving donors.We worry about the legitimacy of courts, and the potential effect elections have on that legitimacy, in ways we decidedly do not concern ourselves with when it comes to other elections.That concern has led to regulations around campaigning and judicial speech that the Supreme Court has upheld as necessary to the compelling state interest of maintaining confidence in the judiciary.
Those rules, though, likely have some unintended consequences, and those are what we explore here.While recusal rules are borne of a desire to protect the legitimacy and perceived impartiality of judges and courts, they have, as we have shown, also affected donors' decisions.If Bonneau and Hall (2009) are correct and judicial campaigns serve as important information provisions to voters, then the very thing that judges do in promulgating these recusal requirements may make the environment surrounding judicial elections even more information-poor than it already is, especially in nonpartisan elections.
The increasing role that money plays in elections has been a considerable source of controversy and debate and ought to continue to be so.Critics argue that elected officials are beholden to the interests of campaign contributors.This is especially salient when it comes to judicial elections, the one branch of government idealized for being above politics.And, in the last few decades, judicial elections have become increasingly partisan and expensive.The possibility that the public might view campaign contributions as driving decisions is of valid concern.Of concern, too, though, is that voters in judicial elections have enough information to make an informed choice.These recusal rules produce a conundrum: While they likely help to protect the legitimacy of the judiciary (especially indirectly, through fewer donations), they do so by dissuading donors from participating, which may translate to a less information-rich environment.That may be desirable as that disincentive stems from the understanding that dollars cannot affect judicial outcomes when judges recuse from cases in which donors are interested/participants, which will be good for legitimacy.But it may also harm democratic accountability, which is why Americans elect their judges in the first place.

Table 2 .
Recusal rules by state 11Via correspondence with Peters.

Table 3 .
Median total contributions by recusal rule

Table 4 .
Total campaign contributions in state supreme court elections 2010-2020 a Note: Statistics reported are ordinary least-squares regression with robust standard errors clustered on state.a Dependent variable: log of total contributions (2020 dollars).b p < 0.001.c p < 0.01.d p < 0.05.

Table 5 .
Donor-specific models of campaign contributions in state supreme court elections 2010-2020 Statistics reported are ordinary least-squares regression with robust standard errors clustered on state.a Dependent variable: log of contributions from attorney/lobbyist donors (2020 dollars).