Abstract
The Office of the Solicitor General wins the vast majority of Supreme Court cases in which it participates. Does it enjoy a built-in advantage, or does it win because it employs experienced litigators, enjoys resource advantages, or carries the executive’s sword? To answer these questions, we employ cutting edge matching methods. After matching OSG lawyers with nearly identical non-OSG lawyers in nearly identical cases, we find that OSG attorneys still are more likely to win their cases. We discuss possible reasons for this built-in advantage along with some practical implications of our findings.
The United States Supreme Court’s decision in District of Columbia v. Heller (2008) marked the first time in the nation’s history where the Court determined the Second Amendment to the United States Constitution supports an individual right to possess a firearm. Despite its significance, one aspect of the ruling was ordinary: as per usual, the Court adopted the position advocated by the United States Solicitor General, the lawyer for the federal government in the Supreme Court. In the post-war era, the Court has sided with lawyers in the Office of the Solicitor General (OSG) more than any other attorneys who routinely appeared before it.
Given this astonishing success, it is reasonable to ask whether OSG lawyers enjoy a built-in advantage that favors them. Is there something unique to the OSG that leads its lawyers to succeed? Asked otherwise, if a nearly identical non-OSG lawyer tried a nearly identical case in a nearly identical context, would the outcome be different? If so, the OSG clearly enjoys an advantage before the Court. If not, OSG success can be attributed to some other characteristics that other lawyers might possess.
In what follows, we employ a novel matching approach to answer these important questions. More specifically, we match Supreme Court cases so that they are as similar to each other as possible, save for our treatment—the presence of an attorney from the OSG. Our goal is to determine how the Court would rule in the absence of an attorney from the OSG and, thereby, to determine whether the OSG enjoys a unique advantage before the Court. Along the way, we reexamine a number of contemporary theories of OSG success, and provide a strong rebuttal to one.
Our results show that OSG attorneys wield significant influence over the Court and enjoy a built-in advantage. When we compare an OSG lawyer to a lawyer who never worked in the Office—but who otherwise is nearly identical to the OSG lawyer and who argues a nearly identical case—the OSG lawyer nevertheless enjoys a higher probability of winning. Additionally, when we compare an OSG lawyer to an otherwise identical lawyer who formerly worked in the OSG, we find that the OSG lawyer still enjoys this affiliation-based advantage. That is, once the lawyer leaves the OSG, her advantage disappears. Finally, the data show that attorneys who formerly worked in the OSG are just as likely to win as nearly identical attorneys (in nearly identical cases) who never worked there. In short, there is something uniquely powerful about presenting an argument from the Office of the Solicitor General, and whatever caché an attorney enjoys while working there vanishes on her departure.
We believe these findings make three distinct contributions. First, and most obviously, they show that the Office of the Solicitor General enjoys a significant built-in advantage before the Court (see also Black and Owens 2011, 2012). This finding is important for normative and empirical reasons. On a normative level, the finding suggests that attorneys who square off against the SG—even when they are nearly identical in all relevant aspects—must work extra hard just to achieve parity with the OSG in the eyes of the Court. The finding questions whether parties observe “equal justice under law” when they square off against the OSG, an institution that seems to have the Court’s ear. On an empirical level, our findings illuminate whether the OSG influences justices, as opposed to whether it simply succeeds before the Court.
Second, the results bring us closer to understanding why the OSG is so influential, as the research design is able to control for—and challenge—existing theories of OSG success. We discover that some long-held arguments for SG success do not stand up to empirical rigor. More specifically, we revisit the argument made by McGuire (1998), which holds that SG experience drives OSG success. Our data suggest that OSG experience does not singularly lead to its influence. We find that OSG attorneys are more likely to win than non-OSG attorneys with the same amount of experience. In short, these results strongly challenge a dominant theory of SG success.
Third, the findings suggest researchers should examine more fully the Office of the Solicitor General. By analyzing attorneys within the office, and not just the SG himself or herself, we hope to push further research toward the internal institutional dynamics of the office. Indeed, as we point out in the conclusion, our sense is that professionals within the office drive its success.
To examine OSG influence over the Court, we begin by providing a brief background on the Office and the existing theories that seek to explain SG success. We then discuss matching methods and the reasons scholars use them. Having explained matching, we then discuss our data and methods, present our results, and offer a discussion for what leads to OSG influence.
The Solicitor General and the Court
Today’s Solicitor General serves two important roles. His first role is to represent the United States before the Supreme Court. Lawyers in the SG’s office “do the bulk of the legal work in Supreme Court cases in which the federal government participates, including petitions for hearing, the writing of briefs, and oral argument” (Baum 2007, 88). In this capacity, OSG lawyers advocate for the United States in the Supreme Court, relying on precedent, policy, and other methods to persuade the Court to side with the federal government. The SG’s second role is to coordinate U.S. appellate strategy (Pacelle 2003). The Solicitor General determines whether the United States will seek Supreme Court review of lower court rulings. 1 The SG also decides whether the United States will file an amicus curiae brief in a case where the United States is not a party but has an interest in the case’s outcome. Simply put, not only does the SG’s office argue cases before the Court, it also must determine which cases will go there.
When fulfilling these dual functions, the Solicitor General and the lawyers in the OSG are tremendously successful (Deen, Ignagni, and Meernik 2003; Segal 1990; Scigliano 1971). For example, during the Rehnquist Court era, the OSG won more than 62% of all cases in which the United States was a direct party and more than 66% of cases in which it participated as amicus curiae (Epstein et al. 2007). During the Burger Court era, the OSG won more than 67% of cases in which the United States was a party and, during the Warren Court era, more than 59% of such cases (Id).
Despite this impressive success rate, scholars do not agree on why the SG succeeds so often. Students of the Court (and the executive) have proffered three general theories for SG success. One theory holds that the OSG is a unique group of actors who consider broader institutional features regarding the Court, and that their concern for the Court leads to heightened success (Caplan 1988). A second theory holds that there is nothing unique about the office, and that OSG success simply turns on attorney experience, extensive resources, and other attributes that non-OSG lawyers can (but often do not) possess (McGuire 1998). Finally, a third theory posits that the SG succeeds for separation of powers reasons, with justices deferring to the executive—and, thus, the OSG—to avoid political rebuke (Johnson 2003; Epstein and Knight 1998). We explain these theories below.
The OSG as a Unique “Actor” before the Court
According to the first theory, the Solicitor General’s Office possesses a group of characteristics, unknown to private law firms, that drive its winning ways. Proponents of this theory claim that the Solicitor General is like a “Tenth Justice” (Caplan 1988). The SG plays a special role for the Court, screening out cases and advocating positions that advance the goals of the Court as an institution (Salokar 1992; Scigliano 1971). As one former clerk told us, the SG is expected to “play as an honest broker of the facts” when communicating with the Court. Indeed, Perry’s (1991) seminal text is filled with comments from justices and clerks who assert that the SG serves as an arm of the Court. One justice stated: “Every solicitor general who has been here my . . . years has taken this job very seriously . . . not to get us to take things that don’t require our attention relative to other things that do” (Perry 1991, 132). In short, the SG thinks and acts likes a justice, with the OSG largely behaving like an extension of the Court itself.
These characteristics translate into OSG success because, much like teammates, the OSG and Court seek the same goals. They both seek “correct” legal outcomes and seek to protect the institution of the Court. What is more, even if they seek different policy outcomes, SGs know they cannot stray too far from legitimate legal posturing (Wohlfarth 2009). As a result, the SG observes substantial success, success that flows from the OSG’s inherent tendencies to think about what is best for the Court as an institution. And because private litigants do not have the same motivation toward the Court, they do not enjoy the same deference from it, and therefore do not generate the same success rates.
The OSG as Experienced Litigators with Superior Resources
A second theory of Solicitor General success holds that OSG attorneys win so often because they simply are more experienced than other attorneys and because they have more resources (McGuire 1998). Why should this matter? Experienced attorneys, the argument goes, are more successful than their inexperienced counterparts because through frequent contact with the Court they acquire relevant information, develop personal relationships with justices, and enjoy a built-up stock of goodwill on which they draw to win cases. Experienced lawyers are more informed as to the questions the Court recently reviewed, legal issues of particular interest to particular justices, and pending cases on the Court’s docket that may affect the disposition of the current case (Ennis 1984, 605). As such, an experienced lawyer knows how to compile effective written and oral presentations (Scigliano 1971, 183). That is, understanding that their reputations for credibility are important, lawyers who want long-term, successful practices at the Supreme Court bar must provide the Court with credible information. Justices can therefore trust the information provided by experienced lawyers, which, in turn, leads to a higher probability of success for them (McGuire 1998).
Existing empirical evidence supports this theory. McGuire (1993) finds that Washington, D.C.–based attorneys are much more likely than non-Washington attorneys to succeed before the Court. In similar work, McGuire (1995) finds that attorneys who are more experienced than their opposing counsel are more likely to win their case. Because the more experienced attorneys provide credible information to justices, he argues, they observe more victories. As he puts it, “the Solicitor General is merely one of many successful lawyers who appear before the Court . . . there is no evidence for the literature’s frequent assertion that the [SG’s] success is derived from an uncommon reputation as the Supreme Court’s leading practitioner” (McGuire 1998, 506).
A number of studies have moved in a similar direction, arguing that parties with more resources fare better before the Court (Collins 2004, 2007; Grossman, Kritzer, and Macaulay 1999). Galanter (1974) found that resource-advantaged litigants fared better than “one-shotters” because they use their resources to stack the deck in their favor (Songer, Kuersten, and Kaheny 2000, 540). They tend to hire better lawyers who can conduct more extensive research. They engage the services of better expert witnesses who thereby create more influential trial court records. They can anticipate legal challenges to their actions and inoculate themselves against those challenges by creating “comprehensive litigation strategies” (Songer, Sheehan, and Haire 1999, 812). Stacking the deck provides these resource-rich players a formidable advantage.
Experience and resource advantages translate into OSG success because these attorneys learn the information justices want and provide it. Justices, in turn, have reason to rely on this information to overcome their information constraints. Lawyers in the OSG nearly always have more of such information than the attorneys against whom they square off. In short, when the OSG lawyer has more experience or resources than her opposing counsel (which she usually does), the United States wins.
The OSG and the Separation of Powers
A third theory of OSG success turns on the separation of powers and holds that justices defer to the OSG out of political necessity. Strategic separation of powers theory argues that justices must make decisions that satisfy not only their own preferences but also those of other actors with whom they must interact, such as the president and Congress. These political actors have the power to undo Court decisions and even take retributive action against the Court. As such, justices may not be able to rule sincerely in cases but, rather, may need to temper what they want with what they are politically able to achieve (but see Owens 2010).
Within this paradigm, the Solicitor General is often perceived to be the mouthpiece of the executive in Court. The SG is nominated by the president and, while confirmed by the Senate, serves at the pleasure of the president. Presidents can be expected to nominate loyalists to the post. Indeed, most observers recognize that Solicitors General have a strong incentive to advance the president’s agenda in the Court. Thus, when the SG (or an attorney within the office) speaks to the Court, justices may believe they are actually listening to the president. 2 As such, justices will do what they can to stay within the bounds of permissible behavior, which means listening to—and following—what the SG recommends. Since the Solicitor General is the Court’s only direct line to the executive, justices have a real incentive to defer to him or her. 3
Of course, justices need not defer to the SG in every case. Rather, justices will defer to the SG when the Court is constrained by political context. That is, when the Court is ideologically “in between” the relevant legislative pivots (the president, House, and Senate), justices have little need to worry about the political response to their ruling since there will be none—there will be no political response that is Pareto superior to the Court’s decision (Owens 2010). On the other hand, when the Court is more liberal (conservative) than all the relevant legislative pivots, justices will be more likely to defer to the OSG. Accordingly, if the separation of powers theory is correct, OSG attorneys would be more successful than non-OSG attorneys when the Court is constrained by the separation of powers. When, however, the Court is unconstrained by the separation of powers, OSG attorneys will be less successful.
Each of these three theories is plausible and makes intuitive sense. Our goal is to examine all of them with the hopes of answering the broader question whether the OSG enjoys some type of built-in advantage before the Court or whether its success is tied to a characteristic that any lawyer could have. It is to this task that we now turn.
Matching and Data Preprocessing
To examine whether the OSG enjoys a unique built-in advantage or whether its success is tied to other features, we examine what legal scholars refer to as a “but for” question—but for the fact that a particular attorney hails from the OSG, would that attorney still win (or lose) her case? Of course, in a perfect world, to analyze this question we would execute an experiment. We would randomly select a sample of attorneys and randomly assign them to be affiliated either with the OSG or not. In so doing, we would randomize treatment status across numerous confounding variables. If we observed OSG lawyers winning at a higher rate than non-OSG lawyers, we would infer that the OSG possesses unique characteristics that lead its attorneys to success (Boyd, Epstein, and Martin 2010).
Unfortunately, such experiments are often impossible for social scientists, who must rely on observational data (as we do here). To overcome this limitation, researchers have relied on a number of tools like parametric models (e.g., OLS or logistic regression) that include potential confounding variables as controls. While generally useful, such models can be limited. As Boyd, Epstein, and Martin (2010) point out, “we cannot depend on random assignment to ensure that our covariates are systematically unrelated to our treatment variable” (394). And since “performing causal inference requires researchers to limit their analyses to the range of values for which they have data in the treatment and the control groups” one must pursue additional avenues to make inferences from the data, avenues that observe more balance between the treatment and control groups (Boyd, Epstein, and Martin 2010, 395).
To mitigate against the weaknesses inherent in parametric models—while aspiring to the purpose behind randomized experiments—researchers have increasingly turned to matching methods. At least two recent studies have applied matching methods to analyze important judicial phenomena. Boyd, Epstein, and Martin (2010), for example, investigate whether the presence of a female judge on a federal circuit court panel causes the males to vote differently than if they sat on an all-male panel. By matching cases on similar relevant features—but for the presence of a female judge—the authors were able to determine whether the female judge caused her male counterparts to vote differently than they otherwise would. Similarly, Epstein et al. (2005) employ matching methods to examine whether the presence of war causes justices to restrict civil liberties.
As these and other studies suggest, the logic of matching is simple. The analyst takes the data she has collected on the topic of interest and then matches observations such that the values of covariates in the control group and treatment group are as close as possible to each other. Observations that do not match across groups are discarded. (This is called balancing the data.) The goal is to retain data such that the control group is identical to the treatment group, with the only difference between the two being the presence of the treatment. Matching attempts to replicate the essence of an experiment.
The most intuitive matching technique is “exact matching,” whereby the analyst “matches a treated unit to all of the control units with the same covariate values” (Blackwell et al. 2010, 2). That is, the analyst identifies the treatment and then seeks out exact matches among the control group. An example illustrates. If we wanted to examine OSG influence over Court decisions while controlling (only) for the effect of attorney experience, we could simply locate OSG lawyers with x previous oral arguments before the Court and then compare their successes in the Court with non-OSG lawyers who made (only) x previous oral arguments before the Court. If we observed differences between the two groups we could infer that those differences were the result of the attorney’s status as an OSG attorney. 4 The approach is simple enough, until one must match on multiple covariates. At that point, multidimensionality makes exact matching intractable. To return to our example, if we wanted to examine the effect of attorney OSG status while simultaneously controlling for attorney experience, party resources, and organized interest involvement, we would quickly reduce the number of exact matches between the treatment and control groups. Indeed, it is unlikely that we would observe any exact matches.
To circumvent this limitation of exact matching, we capitalize on recent scholarship and utilized coarsened exact matching (CEM) (Iacus, King, and Porro, 2012). 5 The benefits of CEM appear on the front and back ends of the approach. At the front end, the approach is useful because it allows the analyst to define the level of imbalance between the treatment and control groups that is acceptable. The software then evaluates how many matched units remain, given that tolerable level of imbalance. In other words, the CEM approach allows the analyst to define acceptable balance not just in some overall sense but for each individual pretreatment variable of interest. This approach is important because exact matching is often unnecessary for continuous variables. Return to our example above, where we sought to examine the effect of OSG status while controlling for attorney experience and other covariates. The problem with exact matching was that we would be unlikely to retain any matches between the treatment and control groups. To be sure, there are certain instances where we might want one-to-one matches, particularly among the inexperienced attorneys. But once our attorneys gain experience, the marginal value of each additional appearance surely diminishes. To wit, the difference between attorneys who argued, say, 11 cases versus 15 cases will be negligible. And therein lies the value of the CEM method. Under the CEM approach, one can specify the level of coarsening that is acceptable. We can match on clusters or groupings of covariates to achieve balance while still maintaining enough observations to execute the study. Thus, we can treat attorneys with 11 and 15 previous cases as being equivalent (or, as exact matches, if you will) while still forcing exact matches on attorneys with, say, 0 or 1 previous case under their belts. 6
At the back end, once the data are balanced (or less imbalanced), the analyst can fit a standard parametric model to the data. Such approximately balanced data lead to stronger statistical estimates and less model dependence. 7 “The only inferences necessary [after approximately balancing the data] are those relatively close to the data, leading to less model dependence and reduced statistical bias than without matching” (Iacus, King and, Porro, 2012, 1).
To be sure, matching forces the researcher to discard portions of the data, sometimes large. Nevertheless, the loss in data is a small price to pay to make stronger causal inferences. As Boyd, Epstein, and Martin (2010) state: “While it may seem counterintuitive, balanced data that are comparable—even if smaller in number—are preferable to a complete sample for the purpose of estimating causal effects” (398, fn. 25).
Data and Measures
To examine whether the OSG enjoys a built-in advantage before the Court, we began by collecting all orally argued cases decided between the Court’s 1979 and 2007 terms (N = 2839). We then followed the logic of earlier studies and matched cases on a host of characteristics that likely influence whether a litigant wins or loses in the Court but that are nonrandomly distributed between members of the OSG and other attorneys. Our goal, again, is to compare cases such that they are identical in all relevant respects, but for the presence of our treatment—whether the attorney in question hailed from the OSG.
Attorney Experience
We first matched on the prior experience each attorney enjoyed when arguing before the Court. We do so to address head on the argument that attorney experience drives OSG success. For each case, we downloaded its oral argument transcript from LexisNexis and identified each attorney who appeared at oral argument. Across the 2837 cases we examined were 6704 attorney appearances distributed over approximately 3900 unique attorneys. Since our unit of analysis is each attorney in each case, we started with 6704 potential observations and, from there, matched the data. We calculated the number of total prior cases each attorney in our sample orally argued before the Supreme Court before the case at issue. We do so for two primary reasons. First, oral argument experience has already been established as a reasonable measure of attorney experience (Johnson, Wahlbeck, and Spriggs 2006). 8 Second, as McGuire (1998) points out, the United States Reports do not consistently record which attorneys were on the briefs in a case, but do consistently record the identity of the attorney who orally argued the case.
As we note above, while CEM allows the user to specify a set of values to pool, in working with these data we have generally found that the automated binning executed by the software results in a great reduction of imbalance. This also allows the coarsening to be sensitive to unique characteristics of a specific treatment/control data set. For example, the initial level of imbalance between an OSG attorney and a non-OSG attorney is significantly greater than when we compare an OSG attorney versus a former OSG attorney. Accordingly, because our overriding goal is to maximize balance across the treatment and control group, we allow the CEM software to construct the bins for this and all other variables described below. In any event, our substantive effects are consistent if we manually coarsen the data (but with the caveat that we retain greater imbalance in the data).
We measure two different variations of experience. First, we start by including the amount of experience for the attorney who is arguing before the Court. To this value, we also add the experience of any additional attorneys who appear at oral argument as an amicus participant. We take this step to control for the possibility that while a specific attorney might be an oral argument novice, the presence of a veteran advocate as amici who also supports the novice’s side could help tip the scales toward victory in a case. 9 Second, because the argument process is an adversarial one, we also include the cumulative amount of experience possessed by attorneys on the opposing side. The basic intuition here is that the likelihood of a seasoned veteran winning in a case should be higher when he is squaring off against a side represented by first-timers as opposed to frequent Supreme Court litigators.
Net Resource Advantage
To avoid inferring an OSG advantage that might simply be a function of its resource advantages, we also match litigants on their resource advantages. To do so, we follow the trend among scholars and rank order litigants along a sliding continuum (Black and Boyd 2012; Collins 2004, 2007; Songer, Sheehan, and Haire 1999; Sheehan, Mishler, and Songer 1992; Songer and Sheehan [add space] 1992; Wheeler et al. 1987). We follow the approach of Collins (2004, 2008) and assign each petitioner and respondent to one of ten potential categories, which we present in ascending order of resources: poor individuals, minorities, nonminority individuals, unions or interest groups, small businesses, businesses, corporations, local governments, state governments, and the U.S. government. The weakest category—poor individuals—is coded as 1 whereas the strongest category—the U.S. government—is coded as 10. We then subtract the ranking for the opposing side from the ranking for the side being supported by the specific attorney we examine. This approach identifies the relative differential between the two sides. Positive (negative) scores reflect cases in which the attorney’s side (opposing side) was advantaged. 10 Thus, if the OSG squared off against a state government, the resource differential would be 1. If a state government squared off against a local government, the resource differential would be 1. We would then match on that resource differential of 1. (We would not match either of those two instances with a case where the resource differential was anything other than 1.)
Separation of Powers Context
As we stated above, some scholars argue that the Court defers to the OSG out of separation of powers concerns. That is, justices accept the OSG’s position for fear of the political ramifications of ruling otherwise. If that theory is true, we should observe heightened OSG success under one key condition. When the Court is constrained by the political branches and falls outside the legislative equilibrium (i.e., the median justice is more liberal or conservative than the president, House, and Senate), it will be more likely to defer to the OSG. Conversely, when the Court is unconstrained by the political branches and falls within the legislative equilibrium in a case (i.e., the median justice is ideologically between the key pivots in the House and Senate, and the president) it will be less likely to defer to the OSG. Accordingly, following standard spatial separation of powers models (Owens 2010), we devised a binary variable for whether the Court was politically constrained (1) or unconstrained (0). 11 We match constrained Courts with constrained Courts and unconstrained Courts with other unconstrained Courts. If, after matching these types of Courts with one another, we still observe a heightened rate of success for the OSG, we can dispel the separation of powers theory.
Net Number of Supporting Amicus Briefs
A host of studies suggest that amicus curiae briefs influence the choices justices make (Collins 2004, 2008) and that amici participate more often in cases where the OSG is present (Hansford 2004). As such, we believed it necessary to control for the presence of amicus briefs. Using data provided by Collins—which we updated for the 2002-2007 terms—we created a variable that measured the net number of amicus briefs supporting a particular attorney’s side, which we then used in the CEM algorithm. More specifically, we matched cases with other cases when the attorney in question enjoyed the same net number of supporting amicus briefs.
Petitioner Status
We also matched cases based on whether the attorney—including attorneys that appeared as amicus curiae—represented (or supported) the petitioner. We did so to control for the Court’s well-known trend to reverse cases it reviews (Perry 1991).
Ideological Congruence with the Court
Finally, a large portion of judicial decision making turns on the ideological preferences of justices (Segal and Spaeth 2002). As such, for each attorney in our sample, we determined his or her ideological distance from the Court median, following the same procedure as Johnson, Wahlbeck, and Spriggs (2006). We determined, first, the ideological direction of the lower court decision as reported in the Supreme Court Database. If the lower court decision was liberal (conservative), we coded the petitioner as making a conservative (liberal) argument. If the petitioner’s argument was conservative, we coded our variable as the Court median’s ideal point, as estimated by Martin and Quinn (2002). If the argument was liberal, we coded our variable by multiplying the Court median’s Martin-Quinn score by –1.
Before we move to the Results section, we reiterate the data patterns we would observe if each of the three theories we discussed above were true. If the OSG possesses some unique characteristics that lead it to success, we should observe OSG attorneys winning more cases than non-OSG attorneys after matching on all our relevant covariates. If OSG attorneys win because of their experience advantages (McGuire 1998), we would observe no difference in success rates between OSG and non-OSG attorneys in cases where they are equally matched in terms of experience. Likewise for the resources argument. If the OSG wins because of separation of powers constraints, we would observe differences in success rate when comparing constrained Courts with unconstrained Courts, but when we match constrained Courts with other constrained Courts, there would be no difference in success rates between OSG and non-OSG attorneys.
Methods and Results
After we used CEM to match on the pretreatment covariates described above, 12 we estimated a series of logistic regression models for each treatment variable of interest. In particular, we considered three treatment effects:
OSG attorneys (the treatment) against non-OSG attorneys (attorneys who never served in the OSG);
OSG attorneys (the treatment) against former OSG attorneys; and
former OSG attorneys (the treatment) against non-OSG attorneys.
Our dependent variable was whether the attorney won (1) or lost (0). In addition to a dummy variable for whether an attorney was in the treatment group (1 = yes; 0 = no), we also include each of the pretreatment variables identified above: attorney experience, net resource advantage for the attorney’s side, net number of supporting amicus briefs for the attorney’s side, petitioner status, and the ideological distance from the Court median. In so doing, we follow the advice of Ho et al. (2007), who advocate including pretreatment variables in parametric models to neutralize any imbalance that remains in the data.
The two panels of Figure 1 present our results. The top panel of the figure illustrates the benefits obtained in terms of data balance from preprocessing our data using CEM. We summarize balance using L1, a measure that provides an easy-to-interpret index of the degree of imbalance across all multivariate combinations of pretreatment variables in a data set (Iacus, King, and Porro 2011). 13 Zero, which is the minimum possible value taken by L1, corresponds to perfect balance between the treatment and control groups. L1’s theoretical maximum, 1, indicates that no overlap exists between the two groups. As the panel makes clear, we observe a sizable reduction in imbalance for each of the three treatments considered in our analysis. The postmatching value of L1 for our OSG versus non-OSG treatment (0.07) is roughly 86% smaller than its unmatched value (0.49). The relative changes for our other treatments are also notable (65% and 70%), indicating that matching has greatly enhanced our data’s balance and, thus, the value of our statistical models.
Summary dot plot of data balance (top panel) and treatment effect (bottom panel) for our three treatment variables of interest. CEM matching was performed using the
The bottom panel of Figure 1 illustrates the substantive size of the treatment effect for each of our three treatment variables. These values come from stochastic simulations (similar to Clarify) performed after estimating our regression models (see the supporting information for tables with parameter estimates). Consider, first, the top line on this panel, which compares OSG lawyers against otherwise identical lawyers in otherwise identical cases but who never served in the OSG. It shows, in other words, the treatment effect that an attorney from the OSG has versus someone who is not—and has never been—affiliated with the OSG. As the figure makes clear, holding everything else equal, attorneys from the OSG can expect a 0.13 increase in the probability that their side will win the case. This 13% increase is attributable exclusively to the OSG’s participation in the case. That is, this probability boost occurs even after we have eliminated a host of competing explanations for why the OSG is so successful. That we recover any effect while accounting for these other plausible drivers of success speaks to the OSG’s unique advantage.
Consider, next, the second line of the panel, which compares OSG attorneys against former OSG attorneys. Here we compare an OSG attorney against an otherwise identical lawyer in an otherwise identical case who previously served in the OSG. Once again, our results show that the OSG attorney enjoys a built-in advantage. Attorneys from the OSG can expect a 0.10 increase in the probability of winning—an increase, again, that is attributable simply to arguing from the OSG. Stated otherwise, holding all else equal, save for the fact that in one case the petitioner attorney hails from the OSG and in the other a former OSG attorney represents the petitioner, the OSG attorney still has a 10% higher chance of winning.
Finally, consider the third line in the panel, which compares former OSG attorneys against attorneys (of equal experience in equal cases) who never worked in the OSG. The treatment effect of having previously been employed by the OSG is not statistically significant. There is no difference in the probability of winning between attorneys who previously worked for the OSG and similar attorneys who never worked there.
These results provide a host of information—which we will get to momentarily—but combined, they compel one conclusion: the OSG is uniquely powerful and leads its attorneys to be more victorious than they would be if they worked in any other law firm. After matching OSG lawyers against nearly identical lawyers in nearly identical cases, we observe that the OSG still succeeds more often. 14 Simply put, there is something inherently unique about arguing from the OSG that leads its lawyers to heightened success. We can safely argue that the OSG enjoys a built-in advantage.
Along the way, these results cast doubt on existing theories of SG success. Take, first, the theory that attorney experience drives OSG victory. By matching attorneys based on levels of experience we remove the possibility that OSG victory arises simply from an experience advantage. If experience explained OSG success, we would have observed no statistically significant difference between the probability of an OSG victory and a non-OSG victory while matching on experience. We do, however, observe a statistical advantage for the OSG, showing that attorney experience does not explain the OSG’s high success rate.
Consider next, a related argument that the OSG simply hires the best and the brightest lawyers. That is, attorney quality may have less to do with experience than with some type of innate abilities. While this, too, is a plausible hypothesis, our results cast significant doubt upon it. Legal prowess and skills would travel with the attorney once she leaves the OSG. Yet, the comparison between OSG attorneys and former OSG attorneys (as well as the comparison between former OSG attorneys and non-OSG attorneys) suggests otherwise. If the OSG simply hired the best and brightest lawyers, we would observe no statistically significant difference between the probability of victory between a current OSG attorney and a former OSG attorney. At the same time, we would observe a statistically significant difference in the probability of victory between a former OSG attorney and an attorney who never worked in the OSG. Instead, we observe the opposite. Indeed, the latter finding suggests that the influence an attorney enjoys while working in the OSG vanishes once she leaves the office. Of course, beyond the raw informational value of this finding, there is a monetary value to future litigants. To the extent that hiring former OSG attorneys costs them more money than retaining equally experienced attorneys who never worked in the OSG, our results suggest that taking on the additional expense to retain the former OSG attorney might be misguided.
Next, consider the resource advantage argument. Does the OSG win simply because the United States enjoys more resources than opposing counsel? Likely not, though we emphasize the word likely. We matched on each party’s relative resource advantage so that our models compared OSG attorneys with equally advantaged non-OSG attorneys. If OSG success comes from having more resources than other litigants, we would have observed no statistically significant difference between the probability of an OSG victory and a non-OSG victory. When we matched on resource advantage, though, we observe that the OSG’s advantage remains. Nevertheless, we are somewhat hesitant to discard the resource advantage theory altogether. The scale we employed here may not fully capture the resources enjoyed by OSG attorneys. OSG attorneys have a number of advantages that non-OSG attorneys lack, such as using the relevant executive agencies as sounding boards, and not having to worry about direct financial incentives in cases. These features may escape a simple scale of resources. Still, though, our approach is in line with existing studies and, we believe, at least calls into question the resource advantage argument. 15
Additionally, the data fail to support the separation of powers theory. Some scholars have argued that the Court follows OSG recommendations because of political context. That is, they argue that justices will defer to the OSG in a politically constrained environment. We matched cases on such political regimes (constrained Courts with similar constrained Courts and unconstrained Courts with other similar unconstrained Courts). Yet, the results still showed a heightened probability of OSG success. Put plainly, separation of power dynamics do not appear to explain SG success.
We are mindful of other considerations that might come into play when examining Solicitor General influence over the Court. One possible alternative argument is that the OSG wins more often because it largely participates in cases its attorneys believe they will win. While we are sympathetic to this selection argument, a recent study by Nicholson and Collins (2008) casts doubt on it. The authors “reject the hypothesis that the SG selects cases based on their perceived winnability” (404). Our own data suggest that strategic case selection does not account for SG success. To check whether strategic selection was at work, we compared the success of the OSG as petitioner versus non-OSG attorneys as petitioners, as well as the OSG as respondent versus non-OSG attorneys are respondents. We find that an OSG attorney representing the petitioning party receives a 0.18 [0.10, 0.25] boost in the probability of winning over a non-OSG petitioner. We likewise observe that an OSG attorney representing the respondent receives a 0.20 [0.14, 0.26] boost in the probability of winning over a non-OSG respondent (both confidence intervals are 95%, two-tailed). These results do not suggest that the OSG wins because it selects the right cases to appeal. 16
These results bode poorly for any party squaring off against the United States. Even when an attorney is nearly identical to the OSG attorney, the OSG still is more likely to win. This means, then, that non-OSG attorneys must work harder than the OSG to win the case. And even hiring high-priced Washington, D.C., attorneys and former OSG attorneys may not be enough to secure victory. Rather, the private litigant must sink vast resources into a case just to achieve parity with the OSG. Perhaps it is no wonder, then, that organized interests combine to file amicus curiae briefs, often on varying legal issues within the same case, to influence the Court. And perhaps it is no wonder that these groups raise and donate significant sums of money to win Supreme Court cases. They must in order to offset the OSG’s advantage before the Court.
Conclusion
We asked whether the Office of the Solicitor General wins because of something unique to the office or because of shared qualities that its counterparts typically lack. We found that the OSG enjoys a built-in advantage before the United States Supreme Court. We compared the success of attorneys working for the OSG against the success of otherwise identical attorneys who never worked in the OSG or who previously worked for the OSG. The results showed that after matching on characteristics most likely to lead to success, attorneys in the OSG remain significantly more likely to succeed than similarly situated lawyers outside the OSG. These numbers attest to the importance of the Solicitor General’s Office. We therefore come a step closer to explaining OSG success.
So what does explain OSG success? While our data do not provide a clear answer, we believe that OSG success likely stems from the office’s longstanding relationship with the Court and with the professionalism its attorneys display. Like Fisher (1990), we think it dubious to label the SG (and the OSG) as something akin to a “Tenth Justice.” The label “Justice” must reside exclusively with those who survive the Supreme Court nomination and confirmation process. Still, our data suggest that the OSG observes something unique that leads to its built-in advantage. The answer, we speculate, might be found in the OSG’s professionals.
As many before us have argued, placing professionals in key positions is one way to ensure overall efficiency. Professionals may serve as a “credible commitment” (North and Weingast 1989) to justice, as they have a continuing drive to advance their profession and strengthen its public image while, at the same time, they also focus on specific issues, making them (largely neutral) experts (Mosher 1982). 17 Indeed, in a number of other contexts, the use of professionals has led to increased levels of stability and efficiency in policymaking. Consider the (near) eradication of boss politics. Machine politics in many cities encouraged personal ties between economic groups and “the boss.” With such extreme rent-seeking occurring rampantly, business could not advance. Once professionals, rather than political hacks, filled key positions business could rely on a set of stable rules (Miller 2000). More recently, Riccucci (1995) discussed how professionals in the SEC blew the whistle on the Savings and Loan scandal that involved a number of members of Congress, including Speaker Wright and powerful senators. Professionals, in short, often disregard short-term gains for longer term benefits that reflect on the institutions within which they act. If that is the case, then, the Court has ample reason to trust the information provided to it by OSG lawyers, as their professional goals will be more in line with the Court’s institutional focus than the goals of private attorneys who more closely follow the directions of their principals (see also Bailey, Kamoie, and Maltzman 2005). Put plainly, we believe that this type of professionalism may generate heightened success for the OSG, and thus, future research should derive theoretical arguments from scholars who discuss the virtues of professionals.
Social science is about rigorously testing theories with observable data. We tested a number of theories that seek to explain Solicitor General success. And we did so using cutting edge methods that have just recently become empirically tractable. We urge future scholars to examine more fully the office of the Solicitor General and the professionals within it. It is there, we believe, that the answer to OSG success may be found. The answer is important, since “our understanding of the Supreme Court will not be complete until the role of the solicitor general is better understood” (Segal 1988, 142-43).
Footnotes
Authors’ Note
A previous version of this paper was presented at the 2010 meetings of the Midwest Political Science Association (April 22-25, Chicago, IL). For providing helpful comments and suggestions, we thank Carlos Estevez, Scott Hendrickson, Tim Johnson, Al Leong, Rich Pacelle, Rachel Schutte, Justin Wedeking, and Patrick Wohlfarth.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.