Abstract
We analyze the effect of judge shopping in patent litigation following the appointment of a former patent litigator as the sole district judge assigned to the Waco Division of the US District Court for the Western District of Texas (WDTX). We find that patent enforcers’ ability to select, with certainty, a judge widely regarded as patentee-friendly increased the number of cases filed, especially by non-practicing entities. We show that the increase in litigation was driven by both an influx of cases that would not have been filed but for the judge’s appointment and a geographic reallocation of cases that would have been filed regardless. Overall, judge shopping in the WDTX induced over a 33-month period an increase of around 460 patent cases that otherwise would not have been filed. (JEL: K41, O34).
1. INTRODUCTION
While all US district courts are bound by the Federal Rules of Civil Procedure and United States Code, a given case may proceed rather differently depending on the court and judge to which it is assigned. Among other factors, these differences reflect district-specific “local rules,” which may impact the scope and timing of discovery and motions practice, as well as the proclivities of individual judges, who retain substantial discretion to control numerous aspects of the hearings, motions, and trials over which they preside. These formal and informal practices can have considerable impact on at least the speed, order, and cost of litigation stages and, thus, clearly may affect case outcomes.
Litigants therefore engage in “forum shopping” by strategically selecting the court in which a case will be heard, and “judge shopping” by choosing the judge who will hear it.1 In principle, neither practice is inherently harmful and, in fact, both can be beneficial. For one, forum and judge shopping tend to reduce uncertainty by making it more predictable where a given case will be filed and how it will proceed thereafter. Additionally, both phenomena can foster court specialization (Anderson 2015; Ellias 2018). Particularly in the context of complex litigation, increased judicial expertise and familiarity with specialized technologies, terminology, and the like may be beneficial to all litigants. In practice, however, forum and judge shopping are most often characterized as opportunistic behaviors that benefit plaintiffs at defendants’ expense. Because it is plaintiffs who overwhelmingly initiate litigation, selection of venue is uniquely within their control and, thus, uniquely advantageous to them. While a party that anticipates being sued may in some circumstances race to file a preemptive court action of their own, US law significantly limits courts’ jurisdiction to hear such suits.2 Moreover, while a defendant may request that the case brought against it be transferred to another eligible court, the deck is stacked against a change of venue. A defendant seeking transfer bears the burden of proof and, to prevail, must show that its preferred venue is “clearly more convenient” than the district chosen by the plaintiff.3
While the ubiquity of forum and judge shopping has long been recognized as a descriptive matter,4 the academic literature says relatively little about the resulting effects on litigation. Existing contributions include White (2006), who presents empirical evidence that forum shopping affected damages awarded in US asbestos litigation from 1987 to 2003, and Ellias (2018), who finds evidence that forum shopping in US bankruptcy cases led to more predictable outcomes in recent decades. In the context of patent litigation, which is our focus, additional contributions include Atkinson et al. (2009), who study the effect of a reduction in litigants’ incentive to forum shop following consolidation of US patent appellate jurisdiction in a single court of appeals, and Gaessler and Lefouili (2017), who present empirical evidence of forum shopping driven by plaintiffs’ preference for geographic proximity and quick proceedings in German patent litigation.
In addition to being few in number, existing studies of forum/judge shopping tend to take a narrow view of forum/judge shopping’s potential effects. In particular, the literature generally approaches strategic forum selection as something that exclusively affects the geographic allocation of some baseline level of litigation, all but ignoring the possibility that plaintiffs’ ability to select where and before whom they will litigate may affect not just the distribution, but also the size and composition, of the set of cases that are litigated. Rarely, and only in passing, have existing studies considered, for example, that greater ability to engage in forum/judge shopping may plausibly increase the overall level of litigation by inducing the filing of marginal cases that, but for the prospect of favorable treatment in a particular court, are not worth pursuing.5
Indeed, forum and judge shopping can impact litigation along two distinct margins. First, by affecting plaintiffs’ choice whether or not to litigate (i.e., whether or not to file a given potential lawsuit), the ability to engage in strategic forum selection can affect the overall level and composition of litigation. We refer to this margin as the filing decision effect of forum/judge shopping. Second, by affecting plaintiffs’ choice where (i.e., in which venue) to bring the suits that they elect to file, the ability to engage in strategic forum selection can additionally affect the allocation of litigation across courts. This margin we refer to as the venue decision effect.
In this article, we provide empirical evidence of both effects in the context of a recent expansion of the ability to judge shop in US patent litigation. Specifically, we use a difference-in-differences approach to compare the filing rates of cases that were eligible and ineligible to be heard in the Waco, Texas courtroom of Judge Alan Albright, who became a district judge in September 2018 and thereafter quickly garnered a reputation for favoring patent plaintiffs. Our results show that Albright’s elevation to the bench both redirected case filings from other courts to Waco (i.e., affected plaintiffs’ venue decisions) and induced the filing of additional cases in Waco that, but for patentees’ ability to select Albright as their judge, would not have been filed at all (i.e., affected plaintiffs’ filing decisions). In addition to documenting the existence of a filing decision effect in this context, our results suggest that this effect was quite substantial. In relative terms, we find that Albright’s overall positive effect on case filings in his district is one-third filing decision effect and two-thirds venue decision effect. In absolute terms, we estimate that over a 33-month span Albright’s appointment induced the filing of 460 patent cases that otherwise would not have been filed in any court and redirected to Waco an additional 881 cases that otherwise would have been filed in another part of the country. These totals correspond to approximately 5% and 10%, respectively, of all US patent cases filed from September 2018 through June 2021. They also exceed the total number of patent suits filed in every European nation except Germany during the same period of time.
An examination of litigant characteristics further reveals that Albright’s effect on patent litigation, and in particular his effect on patent owners’ filing decisions, is overwhelmingly driven by a disproportionate effect on “non-practicing entities” (NPEs)—that is, firms that specialize in enforcing patent rights they do not attempt to commercialize6—especially small- and medium-sized NPEs enforcing patents purchased on the secondary market. NPEs overall account for about 85% of the increase in patent litigation initiated in the Western District of Texas (WDTX) following Albright’s arrival, as well as 321 of the 460 marginal cases induced by Albright’s filing decision effect during the period of our study.
An analysis of case outcomes additionally shows that cases assigned to Albright are on average more likely to settle without reaching a substantive decision as to the infringement or invalidity of any asserted patent. Here, once again, we find that this result is driven by NPEs asserting acquired patents. While the confidentiality of settlement details forecloses further analysis, NPEs’ increased ability to achieve settlements in cases assigned to Judge Albright may, alone, help explain why so many NPEs elected to file suit in Waco, as NPEs commonly favor quick settlements to substantive rulings that might invalidate low quality patents or foreclose broad claim interpretations (Bessen and Meurer 2014; Cohen et al. 2016).
Though our data comes in the context of US patent litigation, our results advance the broader literature on strategic forum selection in a number of respects. First and foremost, our results demonstrate the importance of considering forum/judge shopping’s effect along different margins. While the existing literature focuses almost exclusively on litigants’ choice of where to litigate, our results confirm that forum and judge shopping can also influence litigants’ decisions to litigate at all. Additionally, our results show that strategic forum selection can change the overall litigation landscape and, thus, potentially also its effectiveness at achieving attendant public policy goals. As our results demonstrate, suits induced by the prospect of favorable treatment in a particular court are far from a random selection of potential disputes, and in our context the characteristics of these marginal suits leave ample room to doubt that increased ability to judge shop was welfare enhancing. NPE assertion activities have been shown to negatively affect companies targeted for suit (Tucker 2014; Cohen et al. 2017), leading many to conclude that NPEs have an overall detrimental effect on innovation (Bessen and Meurer 2014; Lemley and Feldman 2016; Cohen et al. 2016, 2019). Criticism has focused in particular on the activities of NPEs enforcing relatively small portfolios of patents acquired for assertion in court, with some empirical evidence indicating that such suits commonly (if not typically) settle for less than the cost of mounting even a simple defense (i.e., settle for “nuisance value”) (FTC 2016). At the very least, existing estimates of NPE case defense and settlements costs suggest that the 321 additional NPE suits induced during Albright’s first 33 months on the bench imposed an additional $200 million or more in aggregate costs on the companies named as defendants in those cases.7
Finally, our results underscore that in a setting where judge shopping is possible, even a seemingly mundane, low-level judicial selection can have sizeable consequences. As remarkable as the magnitude of Albright’s effect may be, there was nothing unlawful or egregiously unethical about Albright’s actions. In principle, many judges could induce similar effects under current venue and case assignment rules. In this sense, one broad implication of our analysis is that traditional case assignment norms, long applicable in many US districts, inherently introduce some risk of forum/judge shopping, including to the rather extraordinary extent that we document below.
2. VENUE AND JUDGE SELECTION
Because federal courts have exclusive jurisdiction over US patent litigation, a suit alleging infringement must be heard in the first instance by one of the more than 600 US district judges assigned to the nation’s 94 district courts. However, not every district is eligible to hear every case. Rules for establishing proper venue in patent suits dictate that an action accusing a US business entity of infringement must be filed with a district in the entity’s state of formation or another district in which the entity maintains a “regular and established place of business,” such as a headquarters or regional office.8
Shortly after a case is filed in a district meeting this requirement, it is assigned to a judge according to the case assignment rules adopted by that district. That is, each district has discretion to decide for itself how cases filed in its courthouses will be allocated across those locations and the judges who hear cases there. Most often, cases filed in a given district are assigned on a quasi-random basis; however, the subset of judges eligible for any given assignment varies by district (Botoman 2018). In particular, while some district courts allocate all cases filed in the district across all judges assigned to the district, most make case allocation decisions based on the “division” (or courthouse) in which the case was filed.9 Division-level case assignment rules typically restrict case assignments to the subset of judges with chambers located in the division, and thus allow a plaintiff to predict with higher probability which judge will be assigned to its case if it elects to file suit there. In sparsely populated divisions with relatively small courthouses and few judges, cases may be divided across a small handful of judges or, rarely, automatically assigned to a single judge.
For a number of reasons, venue and judge selection are important to, and often hotly contested by, litigants. While practical considerations concerning the location of litigants, witnesses, and documents are no doubt relevant, there is good reason to believe that choice of venue is most often driven by procedural and substantive considerations. With respect to procedure, selection of forum and judge commonly equates to selection of important rules and practices that vary across districts and courtrooms. While it is true that all district courts are bound by the Federal Rules of Civil Procedure, individual districts follow additional “local rules” and “general orders” that add greater specificity for particular types of litigation. For example, several districts—including the Northern District of California (NDCA) and District of Delaware (DDE)—have adopted “Patent Local Rules” or equivalent procedures that establish a pre-trial schedule tailored to important milestones in the patent litigation process. In addition, individual judges may establish procedural norms of their own by adopting standardized scheduling orders, discovery orders, and the like that presumptively apply in all cases assigned to them. Case assignments clearly may have substantive implications as well. It is well established in the literature that judges vary with respect to their willingness to grant certain relief, and in the specific context of patent litigation, numerous studies have documented significant heterogeneity in the rates at which motions to stay, to transfer, to dismiss, and for summary judgment have been filed and granted across districts and jurists (Leychkis 2007; Klerman and Reilly 2010; Lemley 2010; Anderson 2015; Love and Yoon 2017; Anderson and Gugliuzza 2021).
3. JUDGE ALBRIGHT AND THE WACO DIVISION OF THE WDTX
On September 18, 2018, approximately nine months following his January 2018 nomination by then-President Donald Trump, Alan Albright was sworn in as the newest member of the WDTX.10
Prior to his confirmation as a district judge, Albright was a practicing intellectual property litigator in Austin, Texas for almost two decades and, before that, served as a US magistrate judge in Austin from 1992 to 1999. Despite Albright’s long-term residence in Austin prior to (and even continuing after) his appointment, Albright surprised many by electing to locate his chambers in Waco, a much smaller city located approximately 100 miles north, where he would become the town’s (and surrounding division’s) sole district judge and thus (according to district rules in effect at the time) automatically preside over all civil cases filed there.11
Shortly after taking the bench, Judge Albright elicited surprise once more when he undertook what Anderson and Gugliuzza (2021) refer to as “a national recruitment tour” and “media blitz,” in which he spoke widely to the press and at dinners, webinars, and conferences hosted by or for patent owners and lawyers about his desire to preside over patent cases and the benefits of filing suit in his court in Waco. As reported by the Waco Tribune, during his first few weeks on the bench, Albright “let it be known in no uncertain terms that he would like his Waco courtroom to become a hub for IP cases,” and within three months’ time, multiple out-of-town law firms opened new branch offices in Waco. Figure 1 summarizes the timing of these events.
Chronology of Judge Albright’s Appointment to the WDTX.
In addition to advertising his personal interest in hearing patent cases, Albright has by most accounts followed through on his sales pitch by adopting a number of practices that tend to favor patent enforcers. While the WDTX has to date not adopted a set of local patent rules, Judge Albright has adopted a standing order that mandates many comparable requirements in patent cases assigned to him. This order also includes a default patent case schedule, according to which cases are initially slated to proceed to trial approximately 18 months after an initial case management conference.
While speed tends to benefit patent plaintiffs more generally, Albright’s default patent case schedule is additionally desirable because it interferes with defendants’ ability to challenge the validity of patents asserted against them in parallel administrative proceedings before the US Patent and Trademark Office’s Patent Trial and Appeal Board (PTAB). In recent years, PTAB has invalidated challenged patents at a consistently high rate, well above the rate of invalidation seen in court cases.12 However, a challenger must “petition” PTAB for review and PTAB panels retain substantial discretion to decline to take on a review for the sake of efficiency. Exercising this discretion, PTAB commonly elects to deny review of a patent that has been asserted in a parallel court case that is scheduled to proceed to trial before PTAB’s statutory deadline to complete its review. This rule, in combination with Albright’s default case schedule, means that patents asserted in Waco are far more likely than patents asserted elsewhere to avoid PTAB review.
Commentators wary of Albright’s popularity also commonly point to his track record in handling pre-trial motions practice. Many have observed that Albright rarely grants motions that seek to invalidate patents on subject matter eligibility grounds. Despite overseeing a disproportionate number of cases enforcing software and business method patents—patents much like those commonly invalidated in recent years by many other courts applying Alice v. CLS Bank13—Albright denied all relevant motions to dismiss or for summary judgment filed in the first three years of his tenure on the bench. In addition, it has been observed that Albright has rarely granted motions to stay cases in his court to allow PTAB review (to the extent that such review can take place given his scheduling practices) to proceed first. Both tendencies—to reflexively deny motions to dismiss and motions to stay litigation—reduce accused infringers’ odds of successfully defending the allegations against them without first producing substantial amounts of discovery at great cost.
Finally, Albright made it especially easy for patent enforcers to both select him as their judge and keep it that way. During our study period, a patent enforcer could effectively choose Judge Albright by choosing to file in Waco.14 In addition, Albright has proven to be exceptionally reluctant to relinquish control of cases filed in Waco despite his unusually high caseload. He has denied the overwhelming majority of motions to transfer filed in his cases and, moreover, has done so in the face of repeated reversals on appeal.15
4. EMPIRICAL APPROACH
To quantify the impact of forum and judge shopping on litigation, we adopt a difference-in-differences approach that exploits US patent law’s deterministic criteria for establishing venue eligibility. As explained in greater detail below, applying these criteria using publicly available information about companies sued for infringement allows us to identify which patent suits filed outside the WDTX could—and could not—have properly been filed in that district instead, and we compare filing trends among (1) the subset of patent cases filed in West Texas, (2) the subset filed elsewhere that were nonetheless eligible to be filed in West Texas, and (3) the subset filed elsewhere that were ineligible to be filed in West Texas. These comparisons allow us to estimate the overall effect of Albright’s arrival in Waco, as well as to disaggregate that impact along two margins: Albright’s effect on patentees’ choice to litigate and Albright’s effect on patentees’ choice of venue (conditional on their decision to litigate). In addition, we extend our analysis with case-level data on litigant characteristics to examine any differential effect on NPEs.
4.1 Overall effect
4.2 Venue decision and filing decision effects
While forum and judge shopping are commonly assumed to primarily affect the allocation of a given set of cases across courts, both phenomena clearly can also affect the set of cases that are filed. We define these two effects as follows:
Filing decision effect: Forum/judge shopping’s effect on the overall level and composition of litigation, by virtue of its effect on plaintiffs’ choice whether or not to litigate (i.e., whether or not to file a given potential lawsuit).
Venue decision effect: Forum/judge shopping’s effect on the allocation of litigation across courts, by virtue of its effect on plaintiffs’ choice where (i.e., in which venue) to bring the suits that they elect to file.
In our context, this means that Albright’s arrival in Waco and adoption of patentee-friendly practices may not have simply pulled cases away from other districts by making patent enforcement in his court marginally more attractive than enforcement elsewhere in the nation (his venue decision effect), but also may have increased the number of patent cases filed in the United States by making patent enforcement more attractive in absolute terms (his filing decision effect).18
To disaggregate these two effects, we rely on US patent venue rules. Applying the criteria for venue detailed above, we determine whether or not each case filed outside the WDTX could have properly been filed in West Texas instead. That is, we partition the set of cases filed in all other district courts into two sets: those that were eligible to be filed in West Texas and those that were not. Because “ineligible” cases—that is, those for which venue clearly would have been improper in the WDTX —could not have been affected by Judge Albright’s confirmation, they provide a control group to which potentially affected cases can be compared. Specifically, a comparison of eligible and ineligible litigation filed outside West Texas allows us to estimate the extent to which WDTX-eligible filings fell relative to the baseline of ineligible patent filings following Albright’s arrival. Under the assumption that this drop reflects the transition of case filings to Waco, it captures Albright’s venue decision effect. Finally, with estimates of Albright’s overall and venue decision effects in hand, we can calculate the filing decision effect by subtracting the estimated venue decision effect from the overall effect. As discussed in greater detail below, this approach assumes that Albright’s arrival did not reduce the likelihood of any subset of (eligible or ineligible) case filings. Fortunately, Albright’s unambiguously pro-patentee local rules and reputation make this monotonicity assumption reasonable in our context.
5. DATA
Our source for basic case-level data is Maxval’s Litigation Databank, which provides a comprehensive list of patent case filings. To ensure that venue rules were applied consistently to the cases used in our analysis, we restrict our data to cases filed after the release of the United States Supreme Court’s opinion in TC Heartland v. Kraft Foods in May 2017 and we include all cases filed through the end of June 2021. To ensure that our sample is restricted to original infringement cases, we exclude all declaratory judgment actions, as well as all cases generated due to the transfer of an earlier case filed in another court.19
To determine each patent enforcer’s status as an NPE or operating technology company, we cross reference data from three resources: Stanford Law School’s NPE Litigation Database, RPX Insight, and Unified Patents’ Litigation Case List. Each resource provides case-level patent enforcer classifications using its own taxonomy of patent enforcer types. Following these resources and general practice in the literature, we exclude universities and natural persons, as well as patent pool administrators, from our definition of NPE. Merging the taxonomies used by Stanford and Unified Patents, we group the NPEs in our data into the following categories: (1) entities enforcing patents acquired on the secondary market; (2) entities formed by non-practicing inventors to monetize their own patents; and (3) previously operating companies that transitioned from commercializing to monetizing their patent portfolios.20 Because NPEs enforcing acquired patents are commonly part of larger networks of related entities controlled by a parent, we additionally use RPX Insight to determine whether each NPE in our data operates independently or with an affiliation. For all NPEs controlled by a parent, we additionally note the parent’s size, measured by a count of its known subsidiaries/affiliates.
Finally, we employ a multi-step process to determine whether each entity (or business group) accused of infringement could properly be sued in the WDTX—that is, whether each entity (or at least one member of each business group) was either formed within the state of Texas or maintained at the time of the suit a physical location within any of the counties comprising the WDTX. We begin by matching accused infringers with data obtained from two resources: (1) the website of the Texas State Comptroller, which provides the mailing address and state of formation of all firms registered to pay the state’s “franchise tax,” a tax “imposed on each taxable entity formed or organized in Texas or doing business in Texas,” and (2) an additional directory of Texas companies maintained by a third party. We then check and supplement these matches in two ways. First, we manually review the pleadings filed in the overwhelming majority of cases, including all cases with an accused infringer apparently matching a company in one of the above databases,21 for addresses or other venue-related facts that may confirm or refute a match (or lack thereof). Second, in an effort to confirm the existence or absence of physical locations within the WDTX, we perform web searches using the names of all unique defendants—both alone and in combination with each of the following cities: Austin, San Antonio, El Paso, and Midland.
6. RESULTS
Figure 2 displays quarterly NPE and practicing entity (PE) case counts for the WDTX, beginning in mid-2017 and concluding in mid-2021. Across 2017 and 2018, the Western District received an average of just 13 new NPE cases and 4 new PE cases per quarter. However, in the quarters following Albright’s confirmation, case counts rose sharply, reaching a high of nearly 210 new NPE cases (and 242 total patent cases) in the second quarter of 2020 alone. On average, the Western District attracted approximately 117 new NPE case filings and almost 140 total new patent case filings per quarter in the nearly three-year period following Albright’s elevation to the bench.
Quarterly patent case filings in the WDTX, by plaintiff type. The figure shows the number of cases filed in the WDTX by quarter, distinguishing between cases filed by NPEs and PEs.
Figure 2 also indicates that while case filings began to increase almost immediately after Albright’s arrival in Waco, the rate of filings accelerated in late 2019 and early 2020.22 This is consistent with, and presumably reflects, Albright’s efforts beginning in 2019 to spread word about his desire to preside over patent suits (see Section 2). In the Supplementary Appendix, we use an event study approach to further explore the increase in WDTX case filings over time. Supplementary Appendix Figure A2 plots the coefficients on the leads and lags when we estimate the specification using quarterly data and Q4 2018 as the omitted category. Here, we see yet again that while the increase in filings begins soon after Albright’s arrival in Waco, it is particularly pronounced in the first half of 2020.
In Figure 3, we report quarterly patent case counts for the WDTX at the division level (in panel (a)) and the judge level (in panel (b)). These plots confirm that the large increase in case filings shown in Figure 2 was indeed driven by a surge of cases that were filed in the Waco Division and thereafter assigned to Albright. These plots provide further evidence that the increase in case filings shown in Figure 2 was driven by Albright’s arrival in Waco, not by confounding events generally affecting the WDTX as a whole. In the year prior to Albright’s arrival, fewer than three patent cases per quarter were filed in Waco; yet by 2020 Waco’s quarterly average topped 200. By contrast, patent enforcement activity in the district’s other divisions (and the caseloads of the district’s other judges) remained relatively flat (and largely negligible) across our sample period. The Austin Division, the Western District’s second most popular division during our study period, saw an average of 16 patent cases per quarter before Albright joined the district and an average of 11 after. All remaining divisions averaged fewer than one case per quarter. Similarly, Judges Yeakel and Pitman, who held the highest caseloads in the Western District prior to Albrght’s arrival, averaged just nine and four patent case filings per quarter, respectively, during our sample period. All other judges combined saw fewer than four new filings per quarter during the same time frame. By comparison, Albright was assigned an average of nearly 230 new cases per quarter in 2021.
Quarterly patent case filings in the WDTX, by division and judge. The figure shows the number of cases filed in the WDTX by division, judge, and quarter. (a) “Other” includes San Antonio, El Paso, and Midland. (b) “Other” includes Judges Xavier Rodriguez, Orlando Garcia, Jason Pulliam, Fred Biery, Frank Montalvo, David Counts, and David Ezra.
Figure 4 reinforces the remarkable concentration of cases assigned to Albright during the period of our study by comparing Albright’s share of the WDTX patent caseload with the share of district-wide patent case totals assigned to judges in two other patent-heavy districts (NDCA and DDE) that (in contrast to the WDTX’s practice at the time) assign new patent case filings in a quasi-random fashion across all eligible judges assigned to the district.23 For both districts, we find a roughly even share of patent cases assigned to each judge, with no judge in NDCA receiving a greater than 10% share of patent cases filed there, and all judges in DDE receiving a virtually identical share of patent cases filed in that state. By comparison, Albright was assigned well over 90% of all patent suits filed in West Texas between his confirmation and the end of Q2 2021.
Distribution of judges’ patent case assignment share in NDCA and DDE. The figure shows the share of patent cases filed in the NDCA and the DDE from September 2018 to June 2021 that were assigned to the corresponding district court judges and compares these shares with the share of WDTX patent cases assigned to Judge Albright.
Before presenting our core regression results, we begin our examination of Albright’s impact on US patent litigation with a simple visual comparison of quarterly patent case filings in the WDTX and all other districts. In Figure 5, we plot total quarterly patent case filings in West Texas alongside (1) the total quarterly count of patent cases filed in all other districts that could have been filed in West Texas under current venue rules (WDTX-eligible cases) and (2) the total quarterly count of patent cases filed in all other districts that could not have been properly filed in West Texas under current venue rules (WDTX ineligible). Here, we see that WDTX-eligible and WDTX-ineligible case filings follow an approximately similar trend in the year prior to Albright’s confirmation and diverge following Albright’s arrival, with WDTX-eligible case filings falling relative to ineligible case counts.24 Notably, it is additionally apparent in Figure 5 that the observed post-Albright drop in WDTX-eligible cases is exceeded by the post-Albright rise in the Western District’s patent caseload.
Quarterly patent case filings in the WDTX and all other district courts (by WDTX venue-eligibility). The figure shows the number of patent cases filed in the WDTX and all other district courts by quarter. Eligible: all cases filed outside the WDTX that could have been filed in the WDTX instead. Ineligible: all cases filed outside the WDTX that could not have been filed in the WDTX instead.
In Figure 6, we compare patent case filings in West Texas with quarterly averages in other district courts.25 Figure 6a plots the number of patent suits filed each quarter in the WDTX alongside the number of patent suits per district filed each quarter outside the WDTX. Here, we see that, in contrast to the marked rise in quarterly case filings in the WDTX, the average number of patent suits filed per quarter across all other districts courts remained flat throughout the sample period. Figure 6b plots the number of WDTX-eligible and WDTX-ineligible patent suits per district filed each quarter in courts other than the WDTX. Here, we observe (consistent with Figure 5) that quarterly average WDTX-ineligible case filings remained relatively stable across our study period, while quarterly average WDTX-eligible case filings decreased significantly following Albright’s confirmation. Supplementary Appendix Table A1 presents corresponding descriptive statistics.
Average quarterly patent case filings. (a) The number of cases filed in the WDTX and the average number of cases filed in all other district courts by quarter. (b) The average number of WDTX-eligible and -ineligible cases filed in all district courts (except the WDTX) by quarter.
Turning now to our regression results, Table 1 presents the coefficients obtained when we estimate the two specifications set forth above in Section 4. Columns (1) and (2) correspond, respectively, to specifications (1) and (2) and present results estimated using our full sample of cases. As expected, our analysis indicates that Judge Albright’s arrival in Waco had a significant, positive effect on the number of patent cases per quarter filed in the WDTX, with the coefficient reported in the first row of column (1) suggesting that Albright’s presence on the bench increased average case filings in the WDTX (relative to all other district courts) by 121.944 cases per quarter, with a confidence interval of [113.317–131.571]. Accordingly, our analysis suggests that Albright was responsible for the filing of approximately 1341 additional patent suits in the WDTX during the period covered by our study—that is, 121.944 additional cases per quarter multiplied by 11 post-Albright quarters (Q4 2018–Q2 2021). This figure represents Albright’s total effect on WDTX case counts, composed of his effects on patent enforcers’ filing and venue decisions.
Effect of Judge Albright on patent case filings.
| All | NPE | |||
|---|---|---|---|---|
| (1) | (2) | (3) | (4) | |
| WDTX | 121.944*** | 103.898*** | ||
| (4.047) | (3.826) | |||
| WDTX eligible | −1.146 | 0.512 | ||
| (1.301) | (0.949) | |||
| WDTX eligible | −0.899*** | −0.840*** | ||
| (0.155) | (0.176) | |||
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.908 | 0.765 | 0.864 | 0.742 |
| Observations | 1440 | 2848 | 1440 | 2848 |
| All | NPE | |||
|---|---|---|---|---|
| (1) | (2) | (3) | (4) | |
| WDTX | 121.944*** | 103.898*** | ||
| (4.047) | (3.826) | |||
| WDTX eligible | −1.146 | 0.512 | ||
| (1.301) | (0.949) | |||
| WDTX eligible | −0.899*** | −0.840*** | ||
| (0.155) | (0.176) | |||
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.908 | 0.765 | 0.864 | 0.742 |
| Observations | 1440 | 2848 | 1440 | 2848 |
Notes: Dependent variable: number of patent cases filed in a given quarter. Controls: New Judge, New Chief Judge, Change in local rules, and ln total pending cases (see Supplementary Appendix Section A). FE: fixed effects. The sample consists of 90 district courts (89 excluding WDTX in Columns (2) and (4)). Unit of observation at the case-quarter level; OLS coefficients shown; robust standard errors clustered at the case-quarter level.
significant at 10%,
significant at 5%,
significant at 1%.
Effect of Judge Albright on patent case filings.
| All | NPE | |||
|---|---|---|---|---|
| (1) | (2) | (3) | (4) | |
| WDTX | 121.944*** | 103.898*** | ||
| (4.047) | (3.826) | |||
| WDTX eligible | −1.146 | 0.512 | ||
| (1.301) | (0.949) | |||
| WDTX eligible | −0.899*** | −0.840*** | ||
| (0.155) | (0.176) | |||
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.908 | 0.765 | 0.864 | 0.742 |
| Observations | 1440 | 2848 | 1440 | 2848 |
| All | NPE | |||
|---|---|---|---|---|
| (1) | (2) | (3) | (4) | |
| WDTX | 121.944*** | 103.898*** | ||
| (4.047) | (3.826) | |||
| WDTX eligible | −1.146 | 0.512 | ||
| (1.301) | (0.949) | |||
| WDTX eligible | −0.899*** | −0.840*** | ||
| (0.155) | (0.176) | |||
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.908 | 0.765 | 0.864 | 0.742 |
| Observations | 1440 | 2848 | 1440 | 2848 |
Notes: Dependent variable: number of patent cases filed in a given quarter. Controls: New Judge, New Chief Judge, Change in local rules, and ln total pending cases (see Supplementary Appendix Section A). FE: fixed effects. The sample consists of 90 district courts (89 excluding WDTX in Columns (2) and (4)). Unit of observation at the case-quarter level; OLS coefficients shown; robust standard errors clustered at the case-quarter level.
significant at 10%,
significant at 5%,
significant at 1%.
As explained in Section 4, the results reported in column (2) of Table 1 allow us to decompose this aggregate effect to isolate Albright’s venue and filing decision effects. The coefficient reported in the third row of column (2) indicates that WDTX-eligible case filings decreased following Albright’s confirmation by an average of −0.899 cases per quarter (with a confidence interval of [−1.232, −0.567]) across the 89 other districts in which patents were enforced during the period of our study. Accordingly, our analysis suggests that Albright’s confirmation resulted in the filing of approximately 881 fewer WDTX-eligible cases in other districts during the period covered by our study—that is, 0.899 fewer eligible cases per quarter and court multiplied by 11 post-Albright quarters and 89 courts other than the WDTX in which patent suits were filed. Provided the monotonicity assumption holds in our context, this figure represents Albright’s venue decision effect, that is, the number of suits filed in Waco that but for Albright’s confirmation would have been filed instead in another district.
Because Albright’s total effect on case counts is, by definition, a sum of his venue and filing decision effects, Albright’s filing decision effect may now be computed as well, simply by subtracting his venue decision effect from his aggregate effect. Stated in terms of cases brought during the period of our study, our analysis implies a filing decision effect of 460 additional cases during Albright’s first 33 months on the bench—that is, the difference remaining when the 881 cases that Albright attracted to the WDTX from other venues are subtracted from the 1341 total additional cases filed in the WDTX in the 11 quarters following Albright’s confirmation. In relative terms, our analysis suggests that more than one-third (34.3%) of the increase in cases filed in the WDTX following Albright’s arrival is attributable to Albright’s filing decision effect. Thus, while approximately two-thirds of the WDTX’s post-Albright increase in patent case filings represents cases that would have been filed (albeit elsewhere and perhaps with differing results) regardless of Albright’s nomination, one-third of the observed increase is comprised of marginal cases that would not have been filed at all but for the ability to litigate on favorable terms before Judge Albright.
To assess the extent to which these effects are driven by the patent assertion activities of NPEs, we repeat the analysis above using only those cases filed by NPEs. Regression results for the sample of NPE cases are shown in columns (3) and (4) of Table 1. Once again, we find that Albright had a significant, positive effect on NPE case filings in the WDTX, with the coefficient reported in the first row of column (3) indicating that Albright’s aggregate effect on NPEs was to increase WDTX NPE filings (relative to NPE filings in all other districts) by almost 104 cases per quarter (103.898 cases per quarter, with a confidence interval of [95.743, 112.054]). Accordingly, our analysis suggests that Albright induced the filing of approximately 1143 additional NPE suits in the WDTX during his first 11 quarters on the bench. This result suggests, consistent with Figure 2, that the overwhelming majority (85%) of Albright’s total aggregate effect is driven by an increase in NPE filings. Using the coefficient reported in column (4) as we did above, we can also decompose Albright’s total effect on NPEs into its venue and filing decision effect components. Here, our comparison of WDTX-eligible and WDTX-ineligible NPE cases suggests that Albright’s arrival in Waco induced the transition of 822 NPE cases from other districts to the WDTX during his first 11 quarters on the bench.26 Subtracting this NPE venue decision effect from Albrights’ total aggregate NPE effect implies an NPE filing decision effect of 321 cases. Thus, our analysis suggests that Albright’s confirmation induced over the course of his first 11 quarters on the bench the filing of 321 NPE cases that otherwise would not have been filed (at all in any court). In relative terms, these results indicate that Albright’s total effect on NPEs was 72% a venue decision effect and 28% a filing decision effect. As shown in Supplementary Appendix Table A1, we find a post-Albright increase in average WDTX-ineligible NPE case filings per quarter. As a result, the venue decision effect estimated in column (4) of Table 1 reflects both an observed drop in WDTX-eligible NPE case filings and a counterfactual increase in WDTX-eligible NPE case filings that we would have expected to observe had Albright never been confirmed to the WDTX (as measured by the increase in average WDTX-ineligible case filings post-Albright).
The validity of our results disaggregating venue and filing decision effects rests on an assumption of monotonicity. In our context, this assumption would not hold if Albright’s confirmation as a member of the WDTX was by some mechanism responsible for both an increase in WDTX patent case filings and a decrease in WDTX-ineligible case filings in other districts. In principle, defendant substitutions could be one such mechanism; that is, monotonicity would not hold if Albright’s availability as a judge led some patent enforcers to substitute (eligible) litigation in West Texas for (ineligible) litigation elsewhere. However, substitutions of this kind were likely quite rare. Importantly, such a substitution would necessarily require the existence and selection of a new (WDTX venue-eligible) assertion target, not just the selection of a new court. Nonetheless, it is possible that our filing decision effect results reflect some level of ineligible-to-eligible defendant substitution that arguably should be categorized instead as a form of venue decision effect.
We also note that classifying defendants as eligible or ineligible for suit in the WDTX is not always a trivial exercise and sometimes requires a modicum of educated guesswork. As a result, it is possible that on occasion we have erroneously associated out-of-state entities with Texas entities that happen to share the same name. To the extent that we have done so, the filing decision and venue decision effects reported above are over and underestimated, respectively, as a result. To address this concern, we additionally present in the Supplementary Appendix (see Supplementary Appendix Table A2) results obtained when we estimate specifications (1) and (2) using the sample of cases filed in the 46 districts most popular with patent enforcers—that is, all districts receiving greater than or equal to the median number of new case filings during our sample period. Because patent cases filed in less popular (largely rural) districts disproportionately involve smaller accused infringers about which relatively little information is available, they present a relatively higher risk of miscategorization. Therefore, dropping these districts allows us to gauge any potential impact that misclassifying eligible and ineligible cases might have on our results. Reassuringly, we obtain very similar results in Supplementary Appendix Table A2. Using this limited sample, we find a virtually identical total aggregate effect (an increase of 122.727 versus 121.944 WDTX patent cases per quarter), and a comparable but smaller venue decision effect (of 811 versus 881 total cases over 11 quarters), which necessarily implies a comparable but larger filing decision effect (of 540 versus 460 cases over the same period). Overall, this robustness exercise suggests that, if anything, the results reported above may underestimate the size of Albright’s filing decision effect.
In the Supplementary Appendix, we present the results of two additional robustness exercises: one that uses division-level (rather than district-level) data and another that varies the treatment date in acknowledgment of the fact that Albright’s pro-patentee proclivities may not have been apparent immediately upon his confirmation.
Second, we explore in Supplementary Appendix Table A3 how our results change when we vary the treatment date. We do so in particular because media coverage of Albright’s predilection for patent cases did not begin until early 2019. As a result, some patent enforcers may not have become aware of the potential benefits of filing suit in Waco until at least three months into his tenure as a district judge. If so, our treatment date may be set too early, which could lead us to underestimate the size of Albright’s effect on patent case filings or mistakenly attribute to Albright the impact of unrelated factors taking place in late 2018. This latter scenario is unlikely at least because any confounding effects would have to specifically affect the Waco Division, and we are unaware of any plausible candidates in the relevant time period. Nevertheless, we report results in Supplementary Appendix Table A3 using alternative treatment dates in early 2019. Consistent with the gradual increase in WDTX patent case filings throughout 2019 shown in Supplementary Appendix Figures A1 and A2, we find that pushing our treatment date forward into 2019 increases our estimate of Albright’s effect on case filings. Accordingly, both additional robustness exercises suggest, if anything, that the results presented above underestimate the magnitude of Albright’s effect on patent case filings.
Next, we take a closer look at NPEs’ reaction to Albright’s confirmation. Supplementary Appendix Figure A4 plots quarterly NPE case filings by each of the three NPE types defined in Section 5. As this figure makes clear, the increase in case filings observed following Albright’s confirmation is driven not just by NPEs, but overwhelmingly by the subset of NPEs that enforce patents acquired on the secondary market. We quantify this observation in the first column of Table 2, which presents the results obtained when we run specification (1) for the subset of cases by NPEs enforcing acquired patents. These results indicate that, following Albright’s confirmation, case filings by this subset of NPEs increased by an average of almost 76 cases per quarter for a total of 833 additional WDTX case filings during the first 11 quarters of Albright’s term as a district judge. In relative terms, this result suggests that NPEs enforcing acquired patents account for nearly three-quarters of the total increase in post-Albright NPE case filings in the WDTX.29 In Table 2 column (2), we extend this analysis a step further to additionally consider the size of NPEs enforcing acquired patents. Consistent with the plot of quarterly case filings by NPE size cohorts shown in Supplementary Appendix Figure A4, our results here indicate that Albright’s effect on NPEs was driven not by members of large NPE networks, but rather by increased activity by small- and medium-sized NPEs, which account for nearly 83% of the observed increase in post-Albright activity by NPEs enforcing acquired patents—that is, about 62 of the approximately 76 additional cases per quarter or, in the aggregate, about 688 of the additional 833 cases filed across the first 11 quarters of Albright’s time on the bench.30
Effect of Judge Albright on NPE case filings.
| NPE acquired patentsa | First-time assertionsb | |||
|---|---|---|---|---|
| All | Small/mediumc | All other | WDTX-ineligible | |
| (1) | (2) | (3) | (4) | |
| WDTX | 75.861*** | 62.517*** | −0.104*** | −0.097*** |
| (2.878) | (2.485) | (0.033) | (0.030) | |
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.853 | 0.751 | 0.097 | 0.077 |
| Observations | 1440 | 1440 | 7318 | 4278 |
| NPE acquired patentsa | First-time assertionsb | |||
|---|---|---|---|---|
| All | Small/mediumc | All other | WDTX-ineligible | |
| (1) | (2) | (3) | (4) | |
| WDTX | 75.861*** | 62.517*** | −0.104*** | −0.097*** |
| (2.878) | (2.485) | (0.033) | (0.030) | |
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.853 | 0.751 | 0.097 | 0.077 |
| Observations | 1440 | 1440 | 7318 | 4278 |
Notes: Dependent variables: in columns (1) and (2), the dependent variable is the number of cases filed in a given quarter by NPEs that enforce acquired patents; in columns (3) and (4), the dependent variable is equal to one if the plaintiff asserted in the lawsuit at least one patent that had not been asserted before. Controls: New Judge, New Chief Judge, Change in local rules, and ln total pending cases (see Supplementary Appendix Section A). FE: fixed effects. Unit of observation in columns (1) and (2) at the court-quarter-level, in columns (3) and (4) at the case level; OLS coefficients shown; robust standard errors clustered at the court quarter level.
NPE enforcing patents acquired in the secondary market.
First-time assertion is equal to one if a patent has never been asserted before.
Small-/medium-sized NPE is equal to one if parent NPE has fewer than 50 subsidiaries/affiliates.
significant at 10%,
significant at 5%,
significant at 1%.
Effect of Judge Albright on NPE case filings.
| NPE acquired patentsa | First-time assertionsb | |||
|---|---|---|---|---|
| All | Small/mediumc | All other | WDTX-ineligible | |
| (1) | (2) | (3) | (4) | |
| WDTX | 75.861*** | 62.517*** | −0.104*** | −0.097*** |
| (2.878) | (2.485) | (0.033) | (0.030) | |
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.853 | 0.751 | 0.097 | 0.077 |
| Observations | 1440 | 1440 | 7318 | 4278 |
| NPE acquired patentsa | First-time assertionsb | |||
|---|---|---|---|---|
| All | Small/mediumc | All other | WDTX-ineligible | |
| (1) | (2) | (3) | (4) | |
| WDTX | 75.861*** | 62.517*** | −0.104*** | −0.097*** |
| (2.878) | (2.485) | (0.033) | (0.030) | |
| Controls | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes |
| R2 | 0.853 | 0.751 | 0.097 | 0.077 |
| Observations | 1440 | 1440 | 7318 | 4278 |
Notes: Dependent variables: in columns (1) and (2), the dependent variable is the number of cases filed in a given quarter by NPEs that enforce acquired patents; in columns (3) and (4), the dependent variable is equal to one if the plaintiff asserted in the lawsuit at least one patent that had not been asserted before. Controls: New Judge, New Chief Judge, Change in local rules, and ln total pending cases (see Supplementary Appendix Section A). FE: fixed effects. Unit of observation in columns (1) and (2) at the court-quarter-level, in columns (3) and (4) at the case level; OLS coefficients shown; robust standard errors clustered at the court quarter level.
NPE enforcing patents acquired in the secondary market.
First-time assertion is equal to one if a patent has never been asserted before.
Small-/medium-sized NPE is equal to one if parent NPE has fewer than 50 subsidiaries/affiliates.
significant at 10%,
significant at 5%,
significant at 1%.
In columns (3) and (4) of Table 2, we additionally consider whether NPE cases filed in the WDTX after Albright’s confirmation are relatively more or less likely to assert a patent that had not yet been litigated. These results shed light on whether NPEs litigating post-Albright were, in the parlance of patent practitioners, more likely to launch new enforcement “campaigns” or instead to extend existing ones. We conduct this analysis at the case—rather than district court—level; that is, we ask whether a given case enforces a patent or patents that had not yet been asserted in court at the time of the case’s filing. Column (3) presents results that compare WDTX NPE cases to all other NPE cases filed after Albright’s confirmation, while column (4) compares WDTX NPE cases to the subset of WDTX-ineligible NPE cases filed post-Albright.31 Both sets of results suggest that while Albright no doubt induced the filing of many new NPE campaigns, he induced them at a lower rate than he induced the extension of existing campaigns (i.e., the continued assertion of previously litigated patents against additional defendants). Considered together, the results reported in Table 2 suggest that Albright’s overall effect was strongest among relatively small NPEs enforcing patents acquired for the purpose of high-volume litigation—that is, NPE with characteristics most commonly associated with nuisance value patent assertion (FTC 2016)—and, at least in the near term, encouraged the marginal expansion of ongoing litigation campaigns more so than other, potentially lower-volume assertion activities.
We conclude our set of results with an examination of Judge Albright’s effect on patent case settlement, which is by far the most common outcome of US patent litigation. Table 3 presents results obtained from regressions in which the dependent variable is a dummy that is equal to one if a given case settled prior to the issuance of any substantive decision on the merits of any asserted infringement claim. Accordingly, our analysis focuses on relatively early settlements occurring before any consideration of enforced patents’ validity or alleged scope. Columns (1)–(3) present results obtained when we use as the control group all cases filed outside the WDTX, and columns (4)–(6) present results when we restrict the control group to WDTX-ineligible cases. As shown in columns (1) and (4), our results indicate that cases assigned to Albright were significantly more likely to settle early—7% and 9% more likely, respectively—regardless of the control set of cases used. In columns (2), (3), (5), and (6) we additionally consider whether Albright had a differential effect on the settlement of NPE cases. Here, the coefficients we obtain on the triple interaction term (reported in Table 3’s fifth row) indicate (in columns (2) and (5)) that Albright did not have a significant differential effect on the settlement of NPE cases generally, but nonetheless (as shown in columns (3) and (6)) did have a significant positive effect on the early settlement of cases filed by NPEs asserting acquired patents, which (as shown above) were responsible for the lion share of the post-Albright increase in WDTX patent case filings. Our results for these particularly active NPEs suggest a post-Albright increase in the likelihood of settlement of 11.6% or 12.9% depending on the control group used. While widespread confidentiality prohibits us from obtaining the terms of any settlement agreements, a higher settlement rate alone may suggest that NPEs enforcing acquired patents were able to leverage the (real or perceived) advantages of litigating in Albright’s court to induce accused infringers to settle, rather than contest the scope and validity of asserted patents.
Effect of Judge Albright on patent case settlement.
| All other | WDTX-ineligible | |||||
|---|---|---|---|---|---|---|
| All | NPE | NPE Acq.a | All | NPE | NPE Acq.a | |
| (1) | (2) | (3) | (4) | (5) | (6) | |
| WDTX | 0.070*** | 0.030 | −0.026 | 0.090*** | 0.057** | −0.008 |
| (0.020) | (0.034) | (0.030) | (0.019) | (0.024) | (0.024) | |
| NPE | 0.106*** | 0.068 | 0.100** | 0.071 | ||
| (0.031) | (0.042) | (0.038) | (0.065) | |||
| NPE WDTX | −0.034 | −0.170*** | −0.028 | −0.172*** | ||
| (0.027) | (0.009) | (0.018) | (0.017) | |||
| NPE I(t ≥ Sep 2018) | 0.096** | 0.114** | 0.122** | 0.128* | ||
| (0.039) | (0.046) | (0.044) | (0.062) | |||
| NPE WDTX I(t ≥ Sep 2018) | 0.010 | 0.129*** | −0.014 | 0.116*** | ||
| (0.025) | (0.014) | (0.017) | (0.011) | |||
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes | Yes | Yes |
| R2 | 0.036 | 0.071 | 0.065 | 0.037 | 0.079 | 0.069 |
| Observations | 12,200 | 12,200 | 12,200 | 8131 | 8131 | 8131 |
| All other | WDTX-ineligible | |||||
|---|---|---|---|---|---|---|
| All | NPE | NPE Acq.a | All | NPE | NPE Acq.a | |
| (1) | (2) | (3) | (4) | (5) | (6) | |
| WDTX | 0.070*** | 0.030 | −0.026 | 0.090*** | 0.057** | −0.008 |
| (0.020) | (0.034) | (0.030) | (0.019) | (0.024) | (0.024) | |
| NPE | 0.106*** | 0.068 | 0.100** | 0.071 | ||
| (0.031) | (0.042) | (0.038) | (0.065) | |||
| NPE WDTX | −0.034 | −0.170*** | −0.028 | −0.172*** | ||
| (0.027) | (0.009) | (0.018) | (0.017) | |||
| NPE I(t ≥ Sep 2018) | 0.096** | 0.114** | 0.122** | 0.128* | ||
| (0.039) | (0.046) | (0.044) | (0.062) | |||
| NPE WDTX I(t ≥ Sep 2018) | 0.010 | 0.129*** | −0.014 | 0.116*** | ||
| (0.025) | (0.014) | (0.017) | (0.011) | |||
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes | Yes | Yes |
| R2 | 0.036 | 0.071 | 0.065 | 0.037 | 0.079 | 0.069 |
| Observations | 12,200 | 12,200 | 12,200 | 8131 | 8131 | 8131 |
Notes: Dependent variable is equal to one if the case settled. Unit of observation at the case level; Controls: New Judge, New Chief Judge, Change in local rules, and ln total pending cases (see Supplementary Appendix Section A). FE: fixed effects. OLS coefficients shown; robust standard errors clustered at the case-quarter level.
NPE enforcing patents acquired in the secondary market.
significant at 10%,
significant at 5%,
significant at 1%.
Effect of Judge Albright on patent case settlement.
| All other | WDTX-ineligible | |||||
|---|---|---|---|---|---|---|
| All | NPE | NPE Acq.a | All | NPE | NPE Acq.a | |
| (1) | (2) | (3) | (4) | (5) | (6) | |
| WDTX | 0.070*** | 0.030 | −0.026 | 0.090*** | 0.057** | −0.008 |
| (0.020) | (0.034) | (0.030) | (0.019) | (0.024) | (0.024) | |
| NPE | 0.106*** | 0.068 | 0.100** | 0.071 | ||
| (0.031) | (0.042) | (0.038) | (0.065) | |||
| NPE WDTX | −0.034 | −0.170*** | −0.028 | −0.172*** | ||
| (0.027) | (0.009) | (0.018) | (0.017) | |||
| NPE I(t ≥ Sep 2018) | 0.096** | 0.114** | 0.122** | 0.128* | ||
| (0.039) | (0.046) | (0.044) | (0.062) | |||
| NPE WDTX I(t ≥ Sep 2018) | 0.010 | 0.129*** | −0.014 | 0.116*** | ||
| (0.025) | (0.014) | (0.017) | (0.011) | |||
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes | Yes | Yes |
| R2 | 0.036 | 0.071 | 0.065 | 0.037 | 0.079 | 0.069 |
| Observations | 12,200 | 12,200 | 12,200 | 8131 | 8131 | 8131 |
| All other | WDTX-ineligible | |||||
|---|---|---|---|---|---|---|
| All | NPE | NPE Acq.a | All | NPE | NPE Acq.a | |
| (1) | (2) | (3) | (4) | (5) | (6) | |
| WDTX | 0.070*** | 0.030 | −0.026 | 0.090*** | 0.057** | −0.008 |
| (0.020) | (0.034) | (0.030) | (0.019) | (0.024) | (0.024) | |
| NPE | 0.106*** | 0.068 | 0.100** | 0.071 | ||
| (0.031) | (0.042) | (0.038) | (0.065) | |||
| NPE WDTX | −0.034 | −0.170*** | −0.028 | −0.172*** | ||
| (0.027) | (0.009) | (0.018) | (0.017) | |||
| NPE I(t ≥ Sep 2018) | 0.096** | 0.114** | 0.122** | 0.128* | ||
| (0.039) | (0.046) | (0.044) | (0.062) | |||
| NPE WDTX I(t ≥ Sep 2018) | 0.010 | 0.129*** | −0.014 | 0.116*** | ||
| (0.025) | (0.014) | (0.017) | (0.011) | |||
| Controls | Yes | Yes | Yes | Yes | Yes | Yes |
| Court FE | Yes | Yes | Yes | Yes | Yes | Yes |
| Case filing quarter FE | Yes | Yes | Yes | Yes | Yes | Yes |
| R2 | 0.036 | 0.071 | 0.065 | 0.037 | 0.079 | 0.069 |
| Observations | 12,200 | 12,200 | 12,200 | 8131 | 8131 | 8131 |
Notes: Dependent variable is equal to one if the case settled. Unit of observation at the case level; Controls: New Judge, New Chief Judge, Change in local rules, and ln total pending cases (see Supplementary Appendix Section A). FE: fixed effects. OLS coefficients shown; robust standard errors clustered at the case-quarter level.
NPE enforcing patents acquired in the secondary market.
significant at 10%,
significant at 5%,
significant at 1%.
7. CONCLUSION
We investigate empirically the existence and effects of judge shopping in the context of US patent litigation following the confirmation of Alan Albright (a jurist widely regarded as patentee friendly) to the sole judgeship in the Waco Division of the WDTX (which provided some, but not all, patent case filers the option to ensure that their cases would be assigned to his court). We document a sharp rise in case filings in West Texas following Judge Albright’s elevation to the bench, and we confirm that this increase comprises cases that were directed to Albright’s Waco courthouse and subsequently assigned to Albright pursuant to the case assignment rules in effect at that time in the district. To quantify Judge Albright’s effect on US patent litigation, we use a diff-in-diff approach that leverages nationwide criteria for proper venue in patent suits. Specifically, we identify all US patent suits filed Q3 2017 through Q2 2021 in courts other than the WDTX, identify the state of incorporation and physical location(s) of all companies accused of patent infringement in these suits, and use this data to sort cases based on whether they could (or could not) have been filed in the WDTX.
In addition to estimating Albright’s overall aggregate effect on patent case filings, we disaggregate that effect to measure the extent to which Albright’s availability induced the filing of cases that but-for his availability would not have been filed, or instead served merely to relocate cases that would have been filed regardless of Albright. We dig deeper, as well, to examine potential differential effects resulting from Albright’s arrival in Waco. In addition to examining Albright’s effect on overall NPE patent enforcement, we consider what types of NPE patent enforcement may have been especially impacted by greater ability to judge shop. Beyond case filings, we also examine case-level data on litigation outcomes to explore Albright’s effect on case settlements.
Our results indicate that greater ability to judge shop in US patent litigation had a profound effect. We find that Judge Albright’s arrival in Waco dramatically increased patent case filings in the WDTX, with this growth primarily driven by a sharp increase in NPE patent enforcement. Importantly, our results also reveal that, while Albright’s availability did redirect a large number of cases from other districts to his, Albright also induced a substantial increase in case filings that but for his presence on the bench in Waco likely would not have been filed in any court. Our approach implies that Judge Albright’s arrival in Waco, alone, was responsible over the course of less than three years for the filing and litigation of approximately 460 US patent suits that otherwise would not have been filed. Moreover, Albright’s effect may well have been even stronger, as multiple robustness exercises suggest that our analysis likely underestimates the full magnitude of his impact. Even so, 460 cases represents 5% of all patent cases filed in the United States during the same period, and this amount alone exceeds the combined total number of patent suits filed in the UK and France—two of the world’s most prominent patent litigation venues—during the same span.
We additionally find evidence that, among various types of NPEs, Albright’s effect was strongest on the patent assertion efforts of relatively small NPEs that buy patents for the purposes of litigation. In addition, our results imply that cases filed by such NPEs were, following Albright’s confirmation, more likely to be marginal extensions of prior enforcement efforts, as well as relatively more likely to settle before the accused infringer could even partially contest the claims against it on the merits.
While we ultimately lack the data necessary to determine whether the additional litigation induced by Albright’s confirmation was welfare enhancing, our results provide several reasons to doubt that it was. Aside from the point that marginal assertions of marginal patents are less likely to vindicate valuable high-quality patent rights, we note that NPE enforcement as a whole has been heavily criticized in the literature (Bessen and Meurer 2014; Cohen et al. 2016, 2019; Lemley and Feldman 2016) and smaller NPEs in particular have been identified in the literature as disproportionately responsible for “nuisance value” patent enforcement (Bessen and Meurer 2014). Further, we note that Albright’s tendency to speed the pace of litigation, deter validity challenges, and induce settlement, plausibly may serve to foster an environment conducive to leveraging the high cost of defending even “low-quality” patents (Cohen et al. 2019). At the very least, it is reasonable to assume that the 321 marginal NPE suits induced by Albright’s filing effect alone generated well in excess of $200 million in litigation and settlement costs, little of which plausibly inured to the benefit of anyone directly engaged in innovative activity (Scott Morton and Shapiro 2014; Lemley and Feldman 2016).
Finally, when viewed more broadly, our results underscore the importance of considering the effect of forum and judge shopping along different margins; that is, its effect not just on litigants’ decision where to litigate, but also on their decision to litigate at all. Forum and judge shopping remain largely unaddressed in the economics literature, and when these topics have been studied (primarily in the legal literature), they have been presumed to primarily (if not exclusively) affect choice of venue for a static set of suits. Our results show that, to the contrary, judge shopping can have a large effect on the set of cases that are filed.
Footnotes
To be clear, the concept of “judge shopping” does not typically refer to a litigant’s ability to select a judge that is favorable to a specific person or company, for example, due to a specific personal or financial conflict of interest. Rather, judge shopping refers to a litigant’s ability to select a judge that is perceived to be generally advantageous because, for example, the judge typically makes procedural choices (such as setting early trial dates) that tend to give one party leverage over the other.
See, for example, Aetna Life Ins. Co. v. Haworth, 300 U.S. 227, 240–41 (1937).
See, for example, In re Volkswagen of America, Inc., 545 F.3d 304, 315 (5th Cir. 2008) (en banc).
In the specific context of US patent litigation, for example, a large body of descriptive research documents substantial differences in cases, litigants, and outcomes across district courts (e.g., Lemley 2010; Chien and Risch 2017).
White (2006: 396), for example, mentions that “higher damage awards and higher settlement probabilities make it more profitable for plaintiffs’ lawyers to […] file additional claims,” but does not attempt to isolate or quantify this effect.
Our definition of NPE includes: entities enforcing acquired patents, individual-inventor-controlled entities, and failed operating companies that transitioned to patent monetization. We exclude from the definition: universities, natural persons, and patent pool administrators.
According to the AIPLA’s 2021 Report of the Economic Survey, defense costs for a simple NPE case (one with less than $1 million at stake) average approximately $70,000 through initial case management, almost $500,000 through discovery and motions practice, and over $1 million through trial and appeal. According to FTC (2016), suits filed by “litigation PAEs” settled for an average of approximately $370,000 between 2009 and 2014.
See 28 U.S.C. §1400(b); TC Heartland v. Kraft Foods, 137 S. Ct. 1514 (2017); In re Cray, 871 F.3d 1355, 1362–64 (2017). Foreign entities may be sued in any district; however, strict limits on the extraterritorial reach of US courts’ jurisdiction make such cases relatively rare. Accordingly, suits against foreign entities most often target those entities’ domestic subsidiaries.
Almost all federal districts are subdivided into multiple geographic “divisions,” each of which frequently corresponds to the location of a federal courthouse located within the district. Litigants who file suit in a divided district choose the division to which their case will be assigned.
Between his nomination and swearing in, Albright responded to the Senate’s Questionnaire for Judicial Nominees and testified at a Senate Judiciary Committee hearing held on April 25, 2018. A month later, on May 24, the Committee voted to advance Albright’s nomination to the full Senate, which voted to confirm his nomination on September 6.
The WDTX is divided into seven divisions and eight courthouses that collectively house the chambers of nineteen district judges, including Albright.
To be clear, PTAB’s greater rate of invalidation is by design and derives, at least in part, from the fact that PTAB follows a different and significantly more compact procedure.
573 U.S. 208 (2014).
Likely in response to controversy surrounding Judge Albright, in July 2022 the WDTX adopted new case assignment rules ending this practice. All cases in our sample were filed before this announcement.
Motions to transfer, unlike most substantive rulings in a US patent litigation, may be appealed on an interlocutory basis—that is, while the case remains pending. Appeals of substantive rulings, such as claim constructions, occur much less often because they are allowed only when a case has been litigated to a final judgment, something that typically requires multiple years of litigation and the expenditure of several hundred thousand to several million dollars in legal expenses.
We conduct our analysis at the district court level because litigation in the Waco division accounted for nearly all litigation in the WDTX in the 11 quarters following Albright’s confirmation (see Figure 3).
Note that we rely on quarterly data to account for the fact that many district courts receive relatively few court filings. However, as discussed in greater detail below, our results do not depend on that decision and hold when we use monthly data instead.
In principle, it is possible that forum and judge shopping could also affect the overall level of litigation by affecting pre-litigation settlements. For example, accused infringers who expect suit in an unfavorable jurisdiction may be more willing to settle before a lawsuit is filed. However, in practice, pre-litigation settlement of NPE patent infringement claims, which (as shown below) constitute the overwhelming majority of suits filed in Waco, is rare. Instead, patentees who favor a particular jurisdiction overwhelmingly elect to sue without prior warning to ensure that accused infringers have no opportunity to file a proactive declaratory judgment suit of their own.
When a case is transferred between districts, it is assigned a new case number and filing date by the destination district. We drop these subsequent “cases” from our sample.
We additionally grouped a small number of NPEs into a fourth category: NPEs that enforce patents generated by their own technical staff. However, no NPE in this category filed suit in the WDTX during the period of our study. Accordingly, this category plays no role in the analysis that follows.
Our manual review additionally covers all remaining cases filed in any of Texas’ four district courts, as well as all remaining cases naming (1) more than one accused infringer, (2) an accused infringer that was sued more than once during the period of our study, or (3) an accused infringer with an entity suffix other than “Inc” or “LLC.” This latter check ensures that we identify (and treat as eligible for filing in the WDTX) the small minority of cases that exclusively target a foreign company—that is, accuse a foreign company that lacks US-based subsidiaries or accuse a foreign parent company without additionally suing its US-based subsidiary.
Monthly WDTX case filings, which show this increase a bit more clearly, are provided in Supplementary Appendix Figure A1.
We exclude from this analysis: magistrate, senior, and visiting judges. Our set of Northern California judges consists of Judges Alsup, Chen, Chhabria, Davila, Donato, Gilliam, Hamilton, Illston, Koh, Orrick, Rogers, Seeborg, Tigar, and White; our set of Delaware judges consists of Judges Andrews, Connolly, Noreika, and Stark.
The spike in eligible case filings observed in Q2 2020 is driven in substantial part by an unusual spike in case filings by IP Edge-controlled NPEs during the final two days of June 2020. RPX additionally links the increase in Q2 2020 case filings to a series of contemporaneous patent sales by Intellectual Ventures. See Q2 in Review: Second-Quarter NPE Filings Spike as Prior Patent Transfers Spark New Litigation (July 7, 2020).
The following district courts were excluded from our analysis due to a lack of patent litigation activity: District of Guam, District of the Virgin Islands, Eastern District of Oklahoma, and District of the Northern Mariana Islands.
That is, 0.840 “missing” eligible NPE cases per quarter and district multiplied by 11 quarters and 89 districts produces a product of 822 total NPE cases.
This analysis requires us to drop from our sample all cases filed in single-division districts.
This is due to the fact that, before Albright’s confirmation, patent cases filed in West Texas were predominantly filed in Austin, rather than Waco.
Unreported specifications reveal that inventor-controlled NPEs account for nearly 20% of the total increase in post-Albright NPE case filings in the WDTX. NPE successors to failed operating companies account for the remaining approximately 7% of the increase.
Note that the overall contribution of relatively small NPEs is even larger in the complete sample of NPE cases because few large inventor-controlled NPEs or failed operating companies exist, and none filed a case in the WDTX during the period covered by our study.
Arguably, WDTX-eligible cases are not an appropriate control group in this context because that subset of cases is affected by Albright’s confirmation. The analysis in column (4) avoids this concern.
Acknowledgments
We thank the editor, Jeffrey Staton, and two anonymous referees for helpful comments that significantly improved this article. We also thank seminar and conference participants at Santa Clara University, KU Leuven, IPSC 2022, EPIP 2022, and the Royal Economic Society Annual Conference 2023. We acknowledge financial support from the Computer and Communications Industry Association.
Supplementary material
Supplementary material is available at Journal of Law, Economics, & Organization online.
Conflict of interest statement. None declared.
Data availability
The data underlying this article are available in its online supplementary material.
