A. Magistrate Judge Selection and Retention

Federal trial judges—Article I and Article III—arrive on the district court through dissimilar processes and under disparate conditions. District judges are chosen through a politicized and public process, are evaluated based on their political bona fides and personal lives, and are nearly free from control once appointed. Magistrate judges, by contrast, are chosen through a formal and confidential merit process, are evaluated based on their ability to get work done efficiently and effectively, and are supervised by those who select them.

The Article I magistrate judge appointment process has little in common with the more familiar Article III method. Congress originally delegated unconditional discretion over magistrate selection to district judges. After litigators and repeat parties voiced concern over the uneven quality of magistrate appointees (Smith Reference Smith1987; McCabe Reference McCabe2016), Congress responded by passing the Federal Magistrate Act of 1979, creating explicit steps in the selection process (Smith Reference Smith1987; Judicial Conference of the United States 2014). The overhauled process now channels and constrains the selection through its formality, confidentiality, and focus on substance: the district court issues a call for applications when a vacancy is created, candidates submit a long and detailed application, a district-level merit panel of local attorneys and nonattorneys conducts interviews and a substantive evaluation of applicants and makes a written recommendation of the five best candidates, district judges vote on the recommended candidates (and only those candidates), and the new magistrate judge is announced (28 U.S.C. § 631; Foschio Reference Foschio1999). The only visible steps in this process are the first and the last—the announcement of an open position and the announcement of who will fill it—allowing district judges greater latitude in their selection among the finalists.

The goals of individuals who select district judges versus magistrate judges sharply contrast (George and Yoon Reference George2016). Whereas elected officials picking district judges are seeking to maximize reelection chances, district judges choosing magistrate judges are trying to maximize their workload relief. Presidents and senators want district judges who satisfy their supporters, their political party, and the attentive public (Epstein and Segal Reference Epstein2005). District judges likely care about magistrate judges’ perspectives, including ideological views, when those viewpoints affect the magistrate judge’s work. District judges’ most pressing considerations are their caseloads.

District judges retain authority over whomever they select to be a magistrate judge, whereas presidents and senators know that they have only a few weak levers to control district judges after their appointment (e.g., refuse to dispense promotions or pay increases or pursue costly discipline such as impeachment). District judges, in contrast, have numerous mechanisms of control over their judicial appointees. District judges direct assignments to magistrate judges and dictate the quantity and nature of their workload. District judges can also refuse to renew a magistrate judge’s contract when it expires.

B. Magistrate Judge Tenure and Scope of Duties

Both magistrate judges and district judges are federal judicial branch employees. Their salaries and retirement benefits are roughly comparable: full-time magistrate judges earn 92% of the salary of district judges.Footnote ² Both become eligible for a pension equal to their final salary if they serve for 14 or 15 years and reach the age of 65 (28 U.S.C. §§ 371, 377(a)).

Their job tenure, however, is a key difference between the positions. District judges enjoy lifetime tenure. Magistrate judges are appointed for an 8-year, renewable term. District judges decide whether to renew a magistrate judge at the end of the 8-year term.Footnote ³ A majority of district judges may remove a magistrate judge during that term for incompetence, misconduct, or neglect of duty (28 U.S.C. § 631(i)). By contrast, Congress may remove a district judge only for “high crimes and misdemeanors” through the Article I impeachment process.

Another notable distinction between the two judicial offices lies in the scope of judicial duties. While Congress, within constitutional limits, defines the scope of authority of district judges, district judges generally set the boundaries of magistrate judge authority. In short, district judges can take it upon themselves to handle all aspects of a case, whereas magistrate judges can do only a subset of these tasks. District judges have wide discretion in determining what tasks to delegate to magistrate judges.

These meaningful differences between district and magistrate judges—selection, job security, autonomy, and scope of work—only matter if magistrate judges play an important role in the federal trial courts. The next two sections explore that question in distinct ways.

A. Pretrial and Preliminary Matters

We begin with the judicial work most frequently associated with magistrate judges: pretrial and preliminary matters. The most common civil pretrial matters handled by magistrate judges are attorney conferences (e.g., settlement efforts or discovery disputes) and motions. A magistrate judge may hold a motion hearing with or without ruling on the issue presented. If the motion is nondispositive, the district judge may assign the task of hearing and deciding the motion entirely to the magistrate judge (28 U.S.C. § 636(b)(1)(A)). If the motion is dispositive, then the magistrate judge hears the motion and makes findings and a recommendation to the district judge.

To better understand the variation across districts in how magistrate judges are used to manage the civil pretrial process, we develop a measure of the Combined Civil Pretrial Matters Rate. To create this measure, we add the number of magistrate judge matters handled, per district court, for settlement conferences, other pretrial conferences, motion hearings without a decision, and motion decisions from 2000 to 2016. We then divide that total number for each district court by the total number of filed civil cases in the district during that same time frame.

Figure 1 depicts the resulting Combined Civil Pretrial Matters Rate for three groups of district courts: those with median rates, those with rates in the top 10% of all districts, and those with rates in the bottom 10%. As the figure reveals, the median combined magistrate judge activity rate in civil filed matters is 1.35 per filed case. The top 10% of district courts use magistrate judges for civil pretrial matters much more frequently (up to nearly five times total per filed case), while the bottom 10% of districts have magistrate judges handle civil pretrial matters in, on average, less than one-third of their filed civil cases.

Figure 1.

Distribution across districts of the combined number of civil pretrial matters handled by magistrate judges per filed civil case (2000–2016). Data source: Administrative Office of the US Courts, annual judicial business tables. Data exclude territory district courts.

We take a similar approach to our analysis of magistrate judge participation rates in the most common criminal pretrial matters: pretrial conferences, motion decisions, guilty plea proceedings, motion hearings, and writs. Consistent with our civil pretrial findings, we observe sizable levels of magistrate judge activity in key criminal pretrial matters in many districts. Figure 2 depicts a Combined Criminal Pretrial Matters Rate for the median district courts, along with those in the top and bottom 10% (computed in the same way as the combined civil pretrial rates described above). The median combined magistrate judge activity rate in criminal pretrial matters is 1.82 per filed criminal defendant. The top 10% of district courts use magistrate judges for criminal pretrial matters much more frequently (up to nearly 7.5 total per defendant), while the bottom 10% of districts have magistrate judges handle criminal pretrial matters for, on average, less than a third of their defendants.

Figure 2.

Distribution across districts of the combined number of criminal pretrial matters handled by magistrate judges per filed civil case (2000–2016). Data source: Administrative Office of the US Courts, annual judicial business tables. Data exclude territory district courts.

Magistrate judges frequently handle a range of felony preliminary proceedings including search warrants, arrest warrants, initial appearances, arraignments, preliminary examinations, detention hearings, and bail review. Figure 3 plots the Combined Felony Preliminary Proceedings Rate by which district courts use magistrate judges. We can see that every district is active in some notable way when it comes to using magistrate judges to handle felony preliminary proceedings, with the bottom 10% of districts having rates ranging from 1.43 to 2.5 per defendant. The top 10% of districts lean very heavily on magistrate judges for these proceedings, with rates from 4.8 to 5.7 per defendant.

Figure 3.

Distribution across districts of the combined number of felony preliminary proceedings handled by magistrate judges per filed civil case (2000–2016). Data source: Administrative Office of the US Courts, annual judicial business tables. Data exclude territory district courts.

Our data on pretrial and preliminary matters, taken together, provide strong evidence of district courts’ reliance on magistrate judges. Certain districts use magistrate judges relatively frequently to handle civil and criminal pretrial and preliminary matters while others opt to use them infrequently. Three district courts—the Eastern District of Virginia, the Western District of Oklahoma, and the Western District of Washington—rank among the 10% lowest districts for magistrate judge use in two categories. On the other side, six districts rank in the top 10% for two pretrial/preliminary magistrate judge use categories: the District of Colorado, the District of Massachusetts, the District of Minnesota, the Southern District of Georgia, the Western District of Louisiana, and the Western District of Missouri. Looking at the cumulative weight of these magistrate judge use data also confirms the zero-sum nature of these judges’ time. We see no districts where magistrate judge use ranks highly across all three categories, and, indeed, we see examples in which high use in one category is balanced by very low use in another (such as the District of Wyoming, the Southern District of Georgia, and the Western District of Missouri).

We find that magistrate judges are key decision makers in the range of pretrial matters that dominate federal litigation. District judges, to varying degrees, are relying on magistrate judges to direct the early crucial stages of litigation, the point at which many cases end. Congress, however, authorized district courts to go further and allow magistrate judges to fulfill the judicial role at the merits stage. We next look at whether district judges haven ceded greater authority to magistrate judges.

B. Substantive and Merits Decisions

Magistrate judges also affect merits outcomes either directly, with party consent, or indirectly, through highly influential recommendations to district judges. In a civil action, both parties may consent to all proceedings, including trial (with or without a jury), being conducted by a full-time magistrate judge (28 U.S.C. § 636(c)(1)). Thus, the magistrate judge has power over the entry of final judgment in the dispute, acting as a district judge is otherwise authorized to act.

Party consent may operate through an opt-out or an opt-in system, with the former system likely to result in higher rates of magistrate judge use. In a number of districts, magistrate judges are on the random case-assignment wheel for all or some civil cases, and magistrate judges are randomly assigned a case from the outset (Alexander, Dahlberg, and Tucker Reference Alexander2019). Any party may choose to opt out of the assignment to a magistrate judge, and then the case is reassigned to a district judge. Other districts take a less active approach, choosing instead to assign cases to district judges but then immediately refer them to a magistrate judge and simultaneously encourage the parties to consent to that magistrate judge’s jurisdiction. The remaining districts continue to allow party consent to occur later in the process. But even there, the timing of the official entry of the consent may happen almost immediately before the magistrate judge’s opinion and order resolving the case.Footnote ⁴

Despite a persistent belief that magistrate judges are limited to pretrial matters, party consent to having their case assigned to a magistrate judge rather than a district judge is not rare and, in some districts, is common. From 2000 to 2016, the median district in the data has a magistrate judge serving as the presiding judge in about 6% of its civil cases. Again, substantial interdistrict variation is the norm. In some districts, there is effectively no chance that the parties in a case will consent to a magistrate judge acting in lieu of the district judge for the entirety of their case. In other districts, however, 25%, 33%, or even nearly 50% of the civil cases are formally resolved on the merits by magistrate judges filling the role of the district judges. Figure 4 presents our findings.

Figure 4.

Distribution across districts and years of the number of civil consent cases (Section 636(c)) terminated by magistrate judges per filed civil case (2000–2016). Data source: Administrative Office of the US Courts, annual judicial business tables.

District judges may also assign magistrate judges the task of issuing Reports and Recommendations (R&Rs) on the resolution of civil case evidentiary and dispositive motions (such as motions to dismiss and motions for summary judgment). While district judges do not have the power to formally assign final decision-making authority directly to the magistrate judges without the consent of the parties, magistrate judges’ R&Rs are afforded significant deference by district judges. Indeed, prior empirical evidence indicates that district judges rarely fail to adopt magistrate judge R&Rs (Eisenberg Reference Eisenberg1981; Boyd Reference Boyd2016; Alexander et al. Reference Alexander2019). Thus, when a district judge asks a magistrate judge to issue an R&R, the district judge is signaling a preference to rely on the magistrate judge’s judgment to resolve the case.

Magistrate judges in many district courts are advising district judges on the resolution of civil motions. The median R&R rate across 2000–2016 falls between 6% and 10%. A few districts have few or no recorded R&Rs in civil cases in multiple years (including Wyoming and both districts in Wisconsin). Other districts are regularly on the upper end of the distribution, with a third to one-half of their civil cases including an R&R from a magistrate judge, as reflected in figure 5.

Figure 5.

Distribution across districts and years of the number of reports and recommendations issued by magistrate judges per filed civil case (2000–2016). Data source: Administrative Office of the US Courts, annual judicial business tables. The figure excludes one extreme outlier.

Viewed together, figures 1–5, and our underlying discussion, reveal notable variation in the use of magistrate judges across districts and time. What explains the ways and degree to which districts lean on magistrate judges? Magistrate judges are a “flexible corps” of judges for individual district courts, allowing them to address their specific workload and other needs (Hanks Reference Hanks2015). As former district and magistrate judge Dennis Cavanaugh (Reference Cavanaugh2008, 7–9) has argued, “District courts utilize magistrate judges differently for many reasons, including the varying caseload demands and the mixes of cases, differences in court culture, different views of the most effective role of magistrate judges, and sometimes to benefit from the particular skills of incumbent magistrate judges.”

A. SSA Project Data

We draw upon two data sources to examine the role of magistrate judges in SSA cases. First, we continue to take advantage of the AO’s comprehensive annual data. Second, we use the new SSA Project Data collected by Boyd, Ainsworth, and Lynch.Footnote ⁵ The SSA Project Data include a stratified random sample of up to 40 SSA district court cases per district with termination years from 1997 to 2014. The data cover 90 district courts, excluding only the four territory district courts. The SSA Project Data are coded from docket sheets and case documents. As such, they enable us to systematically examine the frequency with which magistrate judges serve as the final decision maker in SSA cases.

B. Rationale for Magistrate Judge Involvement in SSA Cases

There are good reasons to expect heavy reliance on magistrate judges in SSA cases. First, most district court dockets have a high volume of SSA cases. For this reason, districts are looking for a way to process these cases in a timely manner without overburdening their district judges or getting in the way of the other docketed cases. The high, consistent volume also means that district courts have an incentive to develop a standardized routine for SSA cases. Allowing magistrate judges to bear the bulk of the burden in these cases and develop specialization offers a potentially desirable solution.Footnote ⁶

Second, SSA cases are often less complex than many civil lawsuits in district courts. Because they involve appellate review of an administrative law judge’s denial of benefits, the district court has the narrow task of reviewing whether that denial aligns with the law. There is no discovery, development of the record, extensive motions practice, settlement conferences, pretrial hearings, or trials. As such, district courts may be less hesitant than in other types of cases to turn SSA cases over to magistrate judges.

Finally, there are incentives for SSA case parties to approve of and foster vast magistrate judge participation to help speed their case determinations, including via consent to magistrate judge jurisdiction. Claimants want their cases resolved, preferably positively, as soon as possible. The faster they win, the faster the disability benefits begin. Claimant attorneys, when present, often specialize in disability law or personal injury law. Their work is high volume and low pay per case, meaning that their incentive structure mandates minimal effort per case. The SSA and its counsel, the US Attorney’s Office in the district and general counsel within the SSA, are repeat players in the district court. With that comes numerous advantages, including vast comfort from a highly standardized and speedy resolution of these cases.

C. Patterns of Magistrate Judge Use in SSA Cases

From the data and documents used in the SSA Project Data and the AO’s annual data, we can observe common patterns in how magistrate judges are used by district courts to process SSA cases.

1. Patterns related to magistrate judge consent practices in SSA cases

Party consent to magistrate judge jurisdiction is a common occurrence in SSA cases in many districts. We observe some districts where magistrate judges are on the SSA case random-assignment wheel (opt-out consent) and others that assign cases to magistrate judges well after filing. For districts that prioritize the use of magistrate judges to resolve SSA disability cases, it is not uncommon to see explicit court efforts to obtain the parties’ consent to magistrate judge jurisdiction.

In figure 6, we use the SSA Project Data to plot, by district, the percentage of SSA cases ultimately decided by a magistrate judge.Footnote ⁷ Figure 6 focuses on the top and bottom 10% of districts in the data, along with the median districts. The median district court value is just over 18% of outcomes decided by magistrate judges. Districts on the low end have very close to 0% of their cases formally concluded by magistrate judges. Magistrate judges often still play a notable referral role in the SSA cases in these districts, but the districts do not actively encourage consent and likely have alternative preferred ways to handle SSA cases. This stands in stark contrast to the top 10% of districts in the data, where the percentage of SSA cases in the data ultimately decided by magistrate judges is at or above 70%.

Figure 6.

Percentage of Social Security case outcomes decided by magistrate judges (MJs). Data source: SSA Project Data. Data exclude Reports and Recommendations.

2. Patterns related to magistrate judge referral and R&Rs in SSA cases

While districts vary in the frequency and timing when referring SSA matters to magistrate judges, notable patterns emerge. Most (but not all) districts refer at least some SSA matters to magistrate judges, whether those matters are preliminary, nondispositive activities like case scheduling, motion timing, or attorney withdrawals or are issuing R&Rs. Some districts have standing orders that refer SSA cases, as a matter of practice, immediately to magistrate judges for these preliminary matters.

Turning to a more systematic examination of the data, figure 7 reveals the significant role played by magistrate judges in issuing R&Rs in SSA cases. Once again using the AO’s reported annual data, the figure plots the distribution of SSA R&Rs issued by magistrate judges across districts and years, from 2000 to 2016. The figure allows us to draw a number of conclusions. First, R&Rs are quite common in SSA cases. The median annual R&R rate per SSA case ranges from 0.20 to around 0.40. There are still a few districts with SSA R&R rates at or very near 0, but even more notable are the large interquartile ranges (with sizes of up to 0.65) and high reaches of the full boxplot quartiles/whiskers (nearly always above a rate of 1).

Figure 7.

Distribution across districts and years of the number of reports and recommendations issued by magistrate judges per filed social security case (2000–2016). Data source: Administrative Office of the US Courts, annual judicial business tables. Data exclude the four territory districts. The figure excludes one extreme outlier.

As we expect, figure 7’s R&R rates stand in sharp contrast to those present for all civil cases. In figure 5, median R&R rate values for all civil cases never rose above 0.15. The distribution of figure 5’s civil R&Rs also has a much tighter spread than we observe in figure 7.

D. Overall District-Level Reliance on Magistrate Judges in SSA Cases

Combining the SSA Project Data’s insights on magistrate judge outcomes with the AO’s data about R&Rs in SSA cases (fig. 7) allows us to observe the significant role that magistrate judges perform on SSA cases, either through R&Rs or primary oversight during the pretrial stage. Of note, it appears that four districts—Colorado, the Eastern District of New York, the Northern District of Alabama, and the Southern District of Iowa—rely on magistrate judges for R&R and consent jurisdiction in under 5% of their SSA cases. Another 14 districts have notable magistrate judge R&R and consent roles in under 25% of their SSA cases. Interestingly, many of these low-magistrate-judge SSA districts rank in the top 10% in figure 1 for their magistrate judge rates of handling civil pretrial matters. While some of these districts rely on magistrate judges to fulfill similar-to-civil pretrial duties in SSA cases in these districts, given the nature of SSA cases and their low complexity relative to other civil lawsuits, some districts likely opt to handle their SSA cases with other judges, including senior status district judges, while allocating their magistrate judge resources to other district needs.

E. Magistrate Judges’ Involvement in SSA Cases: Does It Matter?

There are logical reasons to explain why district judges heavily rely on magistrate judges in many SSA cases. However, this frequent practice also raises at least a few potential concerns about significant magistrate judge use that are worth addressing in more detail. SSA disability cases, despite being handled efficiently as routine cases by many district courts, are policy-rich disputes that attract substantial political attention and ire from the public and interest groups. Each year, the federal government expends approximately $200 billion in disability payments to about 20 million disabled Americans through the SSA (SSA 2015). Additionally, SSA disability cases in district courts are appellate cases rather than matters of first instance. It is hard to imagine litigants consenting to have non–Article III judges decide their appeals in the federal circuit courts or the courts approving a policy that would allow law clerks or non–Article III judges to recommend case outcomes, on the record, that are nearly always adopted by the circuit judges. If such practices would not be acceptable (or even constitutional) for appellate cases resolved in the federal circuit courts, why would they be permissible in the federal district courts?

Another potential concern with the heavy reliance on magistrate judges in resolving SSA cases is that they may decide the cases differently than district judges. Existing scholarship on magistrate judges provides scant evidence that magistrate judges behave inconsistently with district judges. Most relevant to the current research is Krent and Morris’s (Reference Krent2015) study of SSA disability appeals decided from 2010 to 2012 in district courts finding no differences between remand rates for district and magistrate judges.Footnote ⁸ Similarly, Boyd’s (2016) study of magistrate judges in employment discrimination litigation finds virtually no difference in whether a plaintiff or defendant wins litigation based on whether the case’s assigned judge is a district or magistrate judge. After viewing the relationship between district and magistrate judges through the lens of principal-agency theory, Boyd and Sievert (Reference Boyd2013) find evidence of consistency between magistrate judges and district judges: magistrate judges assigned to cases by the consent of the parties are more likely to rule in favor of liberal positions (e.g., in support of civil rights and personal injury plaintiffs) when their districts’ district judges are themselves liberal. In his study of discovery, Zambrano (Reference Zambrano2017) finds a small difference in district and magistrate judge rulings in the discovery setting.

Despite the limited prior evidence on decision-making differences, we posit that SSA cases may result in magistrate judges diverging from district court judges. District judges, on average, are busy and have packed and inflexible caseloads (Taha Reference Taha2004; Morriss, Heise, and Sisk Reference Morriss2005; Boyd Reference Boyd2015). By contrast, magistrate judges, while also busy, may have more workload flexibility (Cavanaugh Reference Cavanaugh2008) and thus can more easily provide “careful consideration” of the cases before them. As the district courts have argued, magistrate judge case assignment, by the parties’ consent, can “promote timely and careful consideration” of the case.Footnote ⁹ Surveys of magistrate judges also indicate that these judges relish opportunities to take on more-demanding and time-consuming judicial roles (Hanks Reference Hanks2015).

Magistrate judge–assigned SSA disability cases may resolve differently than district judge–assigned ones because of the greater time, energy, and flexibility that magistrate judges necessarily devote to their decision making. Recall that in SSA cases, district judges serve as appellate judges, reviewing the decisions of SSA administrative law judges (ALJs) to deny benefits to disability claimants. There are no appeals in which the SSA is appealing an ALJ’s award of benefits to a claimant. SSA ALJs make their denial decisions in a difficult climate that includes nonadversarial hearings and high caseloads that can reach over 1,000 cases per year per ALJ (Fahrenthold Reference Fahrenthold2014). While all district judges examine SSA disability denials under a deferential standard of review, the thoroughness of this review likely depends on a judge’s circumstances. Busy district judges may be much more likely to defer to the ALJs in this review and, thus, more mechanically affirm the ALJs, even when the quality of the ALJs’ decisions is questionable. Magistrate judges, by contrast, may be more likely to have the time and flexibility to assess the case record, including the claimant’s arguments, the evidence, and the ALJ’s evaluation of the case and handling of the original hearing. This opens the door for magistrate judges to be more likely, on average, than district judges to remand the case to the ALJ for additional proceedings. Our hypothesis captures our expectation:

To empirically assess whether there is support for our hypothesis, we use SSA Project Data for nine full federal district courts: the Central District of California, the District of Arizona, the District of Idaho, the District of Maryland, the District of South Dakota, the Eastern District of Kentucky, the Eastern District of Pennsylvania, the Middle District of Alabama, and the Southern District of California.Footnote ¹⁰ Our dependent variable is Pro-Claimant, a dichotomous variable coded as 1 if the claimant receives a favorable outcome in the district court, including all reversals and remands, and coded as 0 for all outcomes that affirm the ALJ’s denial of benefits to the claimant. We drop from our analysis all transferred cases and filing and administrative errors. After excluding these cases, our nine-district-court data set was left with approximately 5,700 cases.

Our independent variable of interest is Magistrate Judge. This is a dichotomous variable that captures whether the case’s outcome judge is a magistrate judge (coded as 1) or an Article III district judge (coded as 0). Within our unweighted data, district judges decided 62% of the outcomes, while magistrate judges decided 38% of the outcomes.

Beyond the status of the case’s judge, we control for other key factors that may affect whether there is a pro-claimant outcome in the SSA case. We account for the type of SSA disability action that the claimant has brought with the variable SSI Title XVI. SSA disability benefit cases are brought two ways under federal law: through the Title XVI Supplemental Security Income (SSI) program (coded as 1) and through the Social Security Disability Insurance (SSDI) program (coded as 0). Coded based on the claimant’s selected nature of suit code, SSDI-brought claims are made by individuals who have paid sufficient Social Security taxes to be vested, while SSI-brought claims are made by low-income disabled, blind, or aged claimants and are funded from general tax revenues. Because of the nature of the weaker-positioned claimants in SSI cases, we expect these cases to be less likely to result in a pro-claimant outcome than SSDI cases.

Additionally, we control for Judge Ideology. Our measure of judicial ideology uses the senatorial courtesy-based score from Giles et al. (Reference Giles2001) and Epstein et al. (Reference Epstein2007) that ranges from −1 (most liberal) to +1 (most conservative). Since magistrate judges were not selected through Article III’s political appointment process, they do not have a comparable ideology score similar to district judges. Following prior studies, we create a proxy ideology score for magistrate judges that accounts for their distinct selection process in their district by using the median (active) district judge ideology score for the year that the magistrate judge was first appointed to his district court. We expect that, all else equal, more-conservative judges will be less likely to issue pro-claimant rulings in SSA cases.

Our model includes the variable Pro Se Claimant to account for SSA claimants representing themselves without counsel throughout their SSA appeal. This variable is coded as 1 for pro se claimants and 0 for those with counsel during some or all of their district court appeal. Consistent with what is well documented in other legal settings (e.g., Galanter Reference Galanter1974; Haire, Lindquist, and Hartley Reference Haire1999; Smith Reference Smith1999), we anticipate that pro se claimants will face significant disadvantages and be much less likely than represented parties to receive favorable outcomes in their SSA cases. Finally, we include district and year fixed effects in our modeling to account for differences that may emerge in decisions in given districts and years of the data.

Table 1 provides the results of our logistic regression for the likelihood a district court outcome favors the claimant.Footnote ¹¹ As expected by our hypothesis, the presence of a magistrate judge as the assigned case judge increases the likelihood that the claimant prevails in her SSA disability case. Figure 8 further explores the substantive effect of the type of judge for a case’s outcome: for magistrate judges, the average predicted probability of a pro-claimant outcome is .59, meaning that 59% of the SSA disability cases in the nine districts examined here do not end with an affirmance of the SSA’s denial of benefits when a magistrate judge serves as the case’s outcome judge. This number falls to just .46 for district judges’ cases. The right side of the figure plots the difference between these two—.13—and the 95% confidence interval around this difference. Because the difference’s confidence interval does not contain 0, we know that the difference between magistrate judge and district judge case outcomes is statistically meaningful.

Table 1.

Logistic Regression Results for Whether an SSA Case’s Outcome is Pro-Claimant

Note.—SEs are clustered on the individual judge. The model’s percentage reduction in error is 16.8%. Baseline year is 2007; baseline district is the Southern District of California. Modeling includes sampling weights. $N=\mathrm{5,680}$ .

* p < .10.

** p < .05 (two-tailed).

Figure 8.

Predicted probability of a pro-claimant case outcome based on whether the outcome judge is a magistrate judge or a district judge and the average difference between the values for magistrate and district judges

The control variables Judge Ideology and Pro Se Claimant perform as expected. For the former, the results provide marginal evidence that as the case’s assigned judge becomes more conservative, the more likely the SSA’s denial of benefits will be upheld (and the less likely there will be a pro-claimant decision). Similarly, when the SSA claimant is pro se, he is much less likely to receive a favorable district court outcome than those claimants represented by counsel.

This analysis of magistrate judges’ effects on outcomes provides just a first step, given its focus on a single, unique type of district court case and its examination of only nine federal district courts. Future data-driven efforts can better untangle cases in which magistrate judges work in tandem with district judges from those in which district judges or magistrate judges operate alone. Additionally, more information and statistical analysis is needed to understand whether nonrandom selection effects in the cases that magistrate and district judges hear or distinctions between types of SSA case remands drive some of the differences we observe. Nonetheless, these empirical findings provide evidence that the variability in how district courts use and rely on magistrate judges matters for parties involved in cases and affects outcomes, at least in SSA disability cases. Given the large number of SSA disability cases heard in district courts each year, this alone is a meaningful finding and is likely to provide encouragement to claimants to eagerly welcome magistrate judge participation in resolving their cases. Additionally, our findings raise the question whether magistrate judges decide cases differently than district judges in other fact-intensive types of cases. This divergence might be most likely in cases in which magistrate judges’ time and flexibility relative to district judges position them to evaluate the case more thoroughly. For similar reasons, magistrate judges may also manage their cases differently than district judges in ways that may indirectly, but importantly, affect case outcomes and the timing of settlement.