Introduction
There exists a large literature on the determinants of court decisions (see, for example, Segal and Spaeth Reference Segal and Spaeth1993, Reference Segal2002; Sunstein et al. Reference Sunstein, Schkade, Ellman and Sawicki2006; Boyd, Epstein, and Martin Reference Boyd, Epstein and Martin2010; Epstein, Landes, and Posner Reference Epstein, Landes and Posner2012; Harris and Sen Reference Harris and Sen2019; Bourreau-Dubois et al. Reference Bourreau-Dubois, Doriat-Duban, Jeandidier and Ray2020; Choi, Harris, and Shen-Bayh Reference Choi, Andrew Harris and Shen-Bayh2022). Much less is known, however, about the determinants of administrative appeals decisions (notable exceptions include Taratoot and Howard Reference Taratoot and Howard2011; Nakosteen and Zimmer Reference Nakosteen and Zimmer2014; Taratoot Reference Taratoot2014b) and even less about them in countries other than the United States (exceptions include Chang Reference Chang2014; Hüschelrath and Smuda Reference Hüschelrath and Smuda2016; Hirata Reference Hirata and Hoshakaigakkai2023). But as administrative law scholars have shown, most appeals against agency decisions are brought to administrative review before they go to court and, indeed, never make it to court at all (Cane Reference Cane2009; Asimow Reference Asimow2015; Bradley Reference Bradley, Peter Cane, Hofmann and Lindseth2020). For instance, in the United Kingdom, the courts received 2,835 “judicial review” cases in 2020 (Ministry of Justice 2023a), whereas administrative tribunals received 304,566 cases of appeals against administrative decisions (Ministry of Justice, n.d.). Of course, some countries, such as Germany or Taiwan, require citizens to file appeals to administrative agencies prior to bringing a lawsuit, but in others, such as France, Japan, and South Korea, citizens may bring either an administrative appeal or an administrative case to court (Chang Reference Chang2014).Footnote 1 In countries where there is choice, citizens can typically file an appeal at much lower cost than a lawsuit, and decisions are generally reached more expeditiously. Administrative appeals thus offer the first, and, for most citizens, the only, avenue via which they can seek redress from administrative decisions.
Rulings over administrative appeals lack the authority of court rulings, but they do at times induce real policy change. According to Cole Taratoot (Reference Taratoot2013), decisions by administrative law judges (ALJ) were the most important factor shaping US National Labor Relations Board decisions between 1991 and 2006. In Japan as well, a series of rulings by the national administrative tribunal led the Ministry of Health, Labor, and Welfare to revise its policies to support workers returning to work after being affected by an occupational hazard (Oe Reference Oe2021, 15). Existing studies find strong effects of partisanship on ALJ decisions in the United States (Taratoot and Howard Reference Taratoot and Howard2011; Taratoot Reference Taratoot2014a), but the determinants of administrative appeals decisions may be different in less polarized contexts. Japan is a useful case for exploring this issue because its political context is much less polarized (Kobayashi and Yokoyama Reference Kobayashi and Yokoyama2018). Moreover, given Japan’s major reform of its administrative appeals system in 2014, which led to the creation of a third-party tribunal to review administrative appeals—namely, the Administrative Complaint Review Board (ACRB) (Gyosei Fufuku Shinsakai)—an assessment of the determinants of appeals decisions there is especially important.
In this study, we ask two questions: first, whether, and to what extent, ACRBs with more lawyers are more or less likely to uphold agency decisions and, second, why governors appoint greater or fewer numbers of lawyers on ACRBs in the first place. Prior to the 2014 reform, the review largely consisted of internal review by the same agency, and often by the same agency official, that had made the initial decision.Footnote 2 The 2014 reform required agencies to assign a different official to review the decision than the one who had made the initial decision, and it introduced a new third-party tribunal system to review the internal review decision, unless requested otherwise by the claimant. The ACRBs are still housed inside the administrative branch, but the new review system offers greater independence from the administrative apparatus than the prior system.Footnote 3
The 2014 reform stipulated the establishment of new ACRBs, both at the national and local levels. Each of the forty-seven prefectural governments in Japan established an ACRB to review citizens’ challenges against their decisions. In all prefectures, ACRB members are appointed by the governor. As with administrative appeals panels of many developed democracies, ACRB members are not required to have professional legal backgrounds. But lawyers represent the largest professional group serving on prefectural-level ACRBs, although as will be shown later, the proportion of lawyers varies across different ACRBs.Footnote 4 The inclusion of legal professionals from outside of the government represents a major break from the existing Japanese administrative appeals system, in which administrative officials, the vast majority of whom were lacking professional legal experience, had dominated the review process.
Existing studies of judicial decision making offer competing insights as to whether having more lawyers on ACRBs might lead to more or fewer administrative review decisions that uphold agency decisions. The principal-agent perspective (see, for example, McCubbins and Schwartz Reference McCubbins and Schwartz1984; Ramseyer and Rosenbluth Reference Ramseyer and Rosenbluth1993) suggests that ACRB members should be eager to uphold agency decisions because members are appointed by governors, who have considerable leeway to dismiss ACRB members after they have been appointed.Footnote 5 By contrast, the social background perspective on judicial behavior suggests that judges’ professional backgrounds strongly shape their rulings (see, for example, George Reference George2001; Barton Reference Barton2007; Miller and Curry Reference Miller and Curry2023), a point that can easily be extended to ACRB members, who perform quasi-judicial functions of determining facts and applying the law. From this standpoint, ACRBs with more lawyers should rule against agencies more often than those with fewer lawyers because lawyers in Japan are professionally socialized to be attentive to claims being made by the average citizen and thus may be sympathetic to their views.
We test these competing claims on an original dataset of prefectural appeals decisions in Japan between 2016 and 2019. Japan’s prefectures share similar institutional structures, allowing us to test the effects of different percentages of lawyers as ACRB members while controlling for potentially confounding factors. Although not necessarily causal, we find that prefectural ACRBs with more lawyers are associated with more frequent decisions against agencies than those with fewer. Consistent with the socialization hypothesis, we also find preliminary evidence that ACRBs with lawyers who have had longer careers are associated with more decisions against agencies. We further show that more recently elected governors are likely to appoint more lawyers to ACRBs. Although governors are likely to appoint lawyers who are sympathetic to their governments, our findings suggest that, once appointed, lawyers are not constrained from siding with the appellant against the prefectural government, despite the weak institutional independence of the ACRBs. We find limited evidence that partisanship matters for administrative appeals decisions or the appointment of lawyers to ACRBs.
Our finding that a professional background as a private attorney shapes decisions in quasi-judicial settings yields broader implications for the study of judicial behavior beyond the Japanese context. Whereas the effects of professional background are often overshadowed by the strong effects of partisanship in the US courts, our findings suggest that the effect of professional background may be stronger in political contexts that are less polarized than the United States. Whereas previous studies have pointed to judges’ prosecutorial backgrounds as driving their rulings, our study suggests that private attorney backgrounds might also shape judges’ decisions. This is an especially important point, given that in many developed democracies, private legal practice is a common pathway to judicial or quasi-judicial positions. Our finding that governors who have served fewer terms tend to appoint more lawyers suggests that, all else equal, governors who are less secure in office may appoint more lawyers because they are likely to rule on the basis of legal-technical reasons than partisan motivations, even if the governor is ousted from power. As governors become more secure in power, they may deem it less necessary to worry about how the ACRB will rule if and when they are voted out of office.
The Japanese administrative appeals system in comparative perspective
Administrative adjudication systems vary considerably across different developed democracies. For instance, the independence of review panels varies, as do the professional backgrounds of the reviewers. Although this variation creates challenges for cross-country comparisons of these systems, Michael Asimow (Reference Asimow2015) identifies five distinct models of administrative adjudication around the world, as summarized in Table 1. Note that these are broad generalizations, with many exceptions, even within the same country.
Varieties of administrative appeals systems
Source: Asimow 2015.
The first model, which is typical of the United States, consists of a combined-function agency that makes the initial decision and also reviews its own decision, which may then be further reviewed by a generalized court. New evidence, arguments, or reasons may not be introduced during the judicial review phase, and the court relies on the agency’s judgment for matters of law, fact, and discretion. The second model, which is typical of the European Union’s (EU) administrative adjudicatory process, is similar to the first model, except that review by the combined-function agency is inquisitorial rather than adversarial as it often is in the United States.Footnote 6 In the third model, employed by the United Kingdom (UK) and Australia, among others, initial review is conducted via an adversarial process in a tribunal that is external to the agency that made the initial decision that is being challenged. Challenges against tribunal rulings may be brought to the courts, but, as in the first and second models, new evidence or arguments may not be introduced. The fourth model, adopted by Argentina and China, is similar to the second model in employing a combined-function agency that reviews its own decision, but challenges to the review may be brought to the courts, at which stage the claimants may introduce new arguments and/or reasons. Finally, the fifth model, which is typical of France and Germany, is similar to the second model in that initial review is conducted via an inquisitorial process in combined-function agencies, but challenges to the agency’s review decision are brought to specialized administrative courts.
Until 2016, when the 2014 reform went into effect, Japan adopted the fourth model. The same agency—indeed, often the same official—that had made the initial decision also served as the hearing officer to review their own decision. Not surprisingly, law scholars and private attorneys had long criticized the system for its lack of independence (see, for example, Miyazaki Reference Miyazaki1990; Bito Reference Bito2007; Matsuzawa Reference Matsuzawa2007; Abe Reference Abe2009, ch. 10). Despite the system’s flaws, administrative appeals were much more popular among the Japanese public than administrative litigation. For instance, in 2006, more than one hundred thousand administrative appeals were filed against national and local government decisions (Somusho 2025) as opposed to 3,734 new administrative cases filed at the district court level (Supreme Court of Japan 2007). Note that, under the existing system, for some categories of cases, such as those involving tax disputes, citizens were required to file an administrative appeal to the relevant agency before filing an administrative lawsuit (Ohno Reference Ohno2014). Yet the vast majority of citizens who filed administrative appeals did not bring their case to the courts when their claims were denied.Footnote 7
The 2014 reform of the administrative appeals system was originally initiated by the Ministry of Internal Affairs and Communications (Somusho) in 2005 after a subcommittee of the ruling Liberal Democratic Party (LDP) urged for a reform of Japan’s quasi-judicial procedures (Honda Reference Honda2007). The passage of the reform bill was greatly delayed by the LDP’s fall from power in 2009 and the subsequent fall of the Democratic Party of Japan (DPJ) government in 2012. The final reform bill combined ideas from the pre-2009 LDP bill and the pre-2012 DPJ bill. These draft bills were built largely on the recommendations that were made by different expert panels that consisted primarily of legal scholars as well as including private attorneys and local and national bureaucrats (Somusho, n.d.b; Nikkei Shimbun Reference Shimbun2011).
As noted earlier, the 2014 reform retained the system of internal review but required agencies to assign a different official to review the decision than the one who had made the initial decision. It also required the internal review decision to be further reviewed by a third-party tribunal—the ACRB—unless requested otherwise by the appellant.Footnote 8 Thus, Japan’s current system now combines elements of the third and fourth models, with combined-function agencies reviewing their own decisions subject to further review by a third-party tribunal via an inquisitorial process.Footnote 9 Although the new system retained the process of internal review that the Japan Federation of Bar Associations (JFBA) had opposed (Nihon Bengoshi Rengokai 2007), it offers greater independence from the administrative apparatus than the prior system. ACRB rulings for some laws may be re-appealed.Footnote 10 Citizens may also appeal ACRB rulings to the courts, at which point the parties may introduce new evidence and/or arguments.
The 2014 reform also eliminated the requirement of having to file an administrative appeal before filing an administrative lawsuit for most categories of cases, greatly expanding the realm of cases for which citizens could file either an administrative appeal or an administrative lawsuit (Kobayakawa Reference Kobayakawa2014). Thus, in effect, the reform of the administrative review system made it easier for citizens to file not only administrative appeals but also administrative litigation. Nevertheless, most citizens have opted for the administrative appeals route.
Figure 1 shows the number of administrative appeals filed before and after the 2014 reform. At present, post-reform data is only available for the years 2016, 2018, and 2019. Although the limited number of data points makes it difficult to assess the long-term impact of the reform, the available evidence suggests that the number of administrative appeals has increased since the reform. With the exception of 2014, 2019 saw the largest number of claims that were filed during this period.Footnote 11 In contrast, although the reforms made it easier to file both administrative litigation and administrative appeals, administrative litigation fell markedly. Figure A1 in the Appendix shows that the number of new administrative cases filed at the district court level declined from 2,486 cases in 2015, to 1,892 in 2018, and to 1,692 in 2020 (Supreme Court of Japan 2020), although the COVID-19 outbreak beginning in 2020 complicates the task of assessing the precise impact of the 2014 reform.
Number of administrative appeals filed at the national and local levels, 2005–19
Note: Figures show the number of administrative appeals cases filed on the basis of the Administrative Complaint Review Act. Data is only available for the years shown.
Source: Ministry of Internal Affairs and Communications, n.d.a.
This study focuses on administrative appeals decisions at the Japanese prefectural level. An important study by Ayako Hirata (Reference Hirata and Hoshakaigakkai2023) analyzes the determinants of Japanese administrative appeals decisions by focusing on the characteristics of the internal reviewers and of individual cases. This study builds on her work by focusing on the ACRB members’ professional backgrounds.
Although Japan has a unitary, as opposed to federal, system, decentralization reforms over the last few decades have led to a considerable expansion in the powers of prefectures (Kitamura Reference Kitamura2002). Some of the important powers of prefectures include the authority to determine eligibility for welfare payments and some categories of childcare assistance; prefectures also levy various categories of tax. Not surprisingly, the expansion of prefectural powers has led to an increase in the number of appeals against prefectural decisions. The 2014 reforms accelerated this trend. As shown in Figure A2 in the Appendix, between 2014, which was just before the administrative appeals reform, and 2019, the number of new appeals filed increased roughly by 23 percent—from 19,427 cases to 23,903 cases. In 2019, the largest percentage of cases—or 43.4 percent of new appeals against prefectural government decisions—concerned decisions relating to welfare, while another 17.6 percent involved freedom of information decisions, and 9.3 percent concerned decisions regarding elderly health insurance (Somusho 2025).
Large numbers of appeals are filed when agencies implement cuts in welfare benefits. Benefit cuts between 2013 and 2015 and, again, between 2018 and 2020 led to a flood of appeals being filed (see, for example, Itagaki Reference Itagaki2022). Significant numbers of cases are also filed that challenge agency decisions that deny welfare benefits. For instance, an appellant in Osaka Prefecture filed an appeal against an agency decision that partially suspended welfare disbursements on the grounds that the appellant had failed to comply with an agency order to sell their car as a condition for receiving welfare benefits (Osakafu Gyosei Fufuku Shinsakai 2018). Although detailed socioeconomic information on the appellants is not disclosed, the large percentage of welfare-related cases suggests that the Japanese administrative appeals system provides an especially important avenue for low-income citizens to seek administrative redress.
Because of the small number of available data points, it is difficult to say whether the reforms have led to a systematic increase in decisions in favor of appellants (see Figure A3 in the Appendix). But, since the reform, Japan’s forty-seven prefectures have varied considerably in the extent to which the decisions have favored the claimant. The mean “win rate” for appellants over the 2016–19 period was 4.2 percent, but the figures range from 0 percent in Kochi Prefecture in 2018 to 33.3 percent in Kagawa Prefecture in 2016 (Somusho 2025).Footnote 12
In all prefectures, ACRB members are appointed by the governor, who is popularly elected, and appointments do not require the consent of the assembly. Different prefectures vary in the number of ACRB members, so the statutory number of reviewers ranges from three (for example, in Miyazaki Prefecture) to twelve (in Tokyo Prefecture). The more populous prefectures such as Tokyo typically have more members than do the less populous ones. The mean number of reviewers per prefecture across the 2016–19 period is 6.14 (standard deviation = 2.26). In the majority of prefectures, reviewers serve three-year terms, but in sixteen of the forty-seven prefectures, including Tokyo, they serve two-year terms.Footnote 13 All prefectures for which information was available stipulate that ACRB members must have “outstanding credentials in law or administration.”Footnote 14 Because ACRBs are a third-party tribunal, no active member of government may serve as a member, although governors often appoint retired prefectural bureaucrats. Members serve on a part-time basis.
Data compiled by the author reveal that governors most often select lawyers, academics, and former local bureaucrats to serve on ACRBs. Table 2 shows that, from 2016 to 2021, lawyers made up 35.9 percent of all ACRB members, followed by academics, who comprised 32.4 percent, and former local bureaucrats, with 10.6 percent. Most of the former local bureaucrats are retired local bureaucrats who had spent their entire careers as local government officials. Over two-thirds of the academics are legal scholars, who comprised 21.9 percent of all ACRB members.Footnote 15 Lawyers, academics, and former local bureaucrats together accounted for 78.9 percent of all ACRB members between 2016 and 2021. For the remaining 21.1 percent, governors appointed individuals from a variety of different backgrounds. For instance, several prefectures, including Aichi and Kagawa, appointed an accountant; Wakayama Prefecture appointed the head of the prefectural association of social workers and a leader of a women’s rights organization. Over the 2016–21 period, 41.3 percent of all members were women.
Professional backgrounds of ACRB members, 2016–21
Source: Data compiled by the author from the ACRB websites of forty-seven prefectures. National Diet Library, n.d.; Somusho, n.d.a.
Theoretical framework
Decisions by ACRB lawyers
Are lawyers who serve as ACRB members more or less likely to favor upholding agency decisions than others? Not only does the inclusion of legal professionals represent a major break from the prior Japanese administrative appeals system, but lawyers are also the most numerous professional group to hold positions on the ACRBs. Moreover, assessing the legality of agency decisions requires legal training, which means that ACRB members with legal expertise—specifically, lawyers and legal scholars—are likely to have a strong voice in the ACRB deliberations. In this section, we discuss the motivations of lawyers vis-à-vis non-lawyers for serving on ACRBs.
To begin with, the principal-agent framework suggests that governors should appoint lawyers who are likely to uphold agency decisions (see, for example, McCubbins and Schwartz Reference McCubbins and Schwartz1984; Ramseyer and Rosenbluth Reference Ramseyer and Rosenbluth1993). Also, given the limited independence of ACRBs, incentives for reappointment and retention may play a significant role in shaping members’ behavior (Boyea Reference Boyea2010). For instance, in Tokyo Prefecture, governors may dismiss ACRB members for physical or mental health reasons, for violating the responsibilities of ACRB members, or if the governor determines that the member is otherwise unfit to serve.Footnote 16 Other prefectures, such as Kanagawa Prefecture, have similar provisions. To date, no governor has removed an ACRB member from their position mid-term, but the knowledge that removal is possible could constrain members’ behavior. Even if governors do not dismiss ACRB members outright, they may not renew their terms, so members seeking reappointment may try to remain on good terms with the governor by upholding agency decisions as much as possible. Even if lawyers who serve on an ACRB are unconcerned about their own reappointment, many are likely to have a professional interest in making sure that lawyers as a professional category continue to be appointed to ACRBs,Footnote 17 so lawyers on ACRBs should generally show deference to agency decisions.
Yet ACRB lawyers’ behavior may also be driven by other factors. To begin with, ACRB members serve on a part-time basis and compensation is fairly modest, so removal from the board does not threaten their livelihoods. Hyogo Prefecture, one of the few prefectures that discloses renumeration for ACRB members, sets the pay at 12,500 yen, which is roughly eighty dollars, per day.Footnote 18 Ranked ninth out of forty-seven prefectures in terms of fiscal health in 2021 (Somusho 2021), Hyogo is among the wealthier prefectures, and, thus, this rate is likely on the more generous end of the scale. For comparison, a recent survey of Japanese lawyers shows that a private attorney typically charges roughly thirty thousand yen, or two hundred dollars, per hour for their services (Nihon Bengoshi Rengokai 2021), so even if prefectures compensated ACRB members at higher rates than they are compensated in Hyogo, a lawyer’s day would still be spent more lucratively doing other work. Indeed, the relatively high incomes of Japan’s private attorneys allows them to take on various pro bono activities for little or no pay, including serving on ACRBs.
Patricia Steinhoff (Reference Steinhoff and Patricia2014, 4) has noted Japan’s high percentage of “cause lawyers,” and a large number of studies show that lawyers have often spearheaded social change in Japan (see, for example, Upham Reference Upham1987; Kidder and Miyazawa Reference Kidder and Miyazawa1993; Arrington Reference Arrington2014, Reference Arrington2019, Reference Arrington2021; Foote Reference Foote and Patricia2014; Arrington and Moon Reference Arrington and Moon2020; Sala Reference Sala2024). Accordingly, surveys show that Japanese citizens do not view private attorneys as simply elites who are inattentive to the average citizen. In a 2015 survey where ordinary Japanese were asked whether they agreed with the statement: “Private attorneys care about the average citizen,” the mean response was 3.1 on a scale of one to five,Footnote 19 suggesting that, although lawyers are not viewed as being “friends” of the general public, they are not viewed as uniformly serving the wealthy either (Ota Reference Ota, Sato and Hamano2015, 222–23). Japanese lawyers’ active involvement in cause lawyering is facilitated by the relatively high incomes that Japanese lawyers enjoy, which, in turn, is driven by the very small size of the bar. Japan has 6,008 citizens per lawyer, compared to 434 citizens per lawyer in the United States, 619 in Germany, and 1,281 in France (Barzilai Reference Barzilai2007, 250–51). The JFBA has long resisted an expansion of the bar on the grounds that it would undermine lawyers’ capacities to undertake public interest work (Feeley and Miyazawa Reference Feeley and Miyazawa2007).
Recent reforms have expanded the number of lawyers, causing a fall in their incomes (Nihon Bengoshi Rengokai 2021, 12–13). Lawyers typically do not represent appellants in administrative appeals cases; for instance, as of 2016, only 7.9 percent of all appeals filed against Tokyo Prefecture were represented by lawyers (calculated from Shimizu and Akiyama Reference Shimizu and Akiyama2018, 28). But they do represent individuals in administrative litigation, and, thus, in the increasingly competitive legal market, serving on ACRBs could, in theory, help lawyers attract more clients, especially those who wish to challenge agency decisions in court. Lawyers may certainly view serving on ACRBs as part of their broader career strategy, but there are better ways to expand their business than serving on ACRBs. Surveys show that lawyers view administrative cases as being among the most socially meaningful, but economically unprofitable, type of case (Sato Reference Sato, Sato and Hamano2015, 16). Our own data also finds that the average ACRB lawyer has roughly twenty years of professional experience, which suggests that these lawyers are generally fairly well-established members of the professional community.
Thus, although lawyers may worry about retaining their ACRB positions, they do not depend financially on ACRB work and are thus unlikely to go out of their way to defend agency decisions. This opens the way for factors other than reappointment incentives to shape ACRB members’ decisions. Because ACRB members perform a quasi-judicial function—determining facts and applying the law—we draw here on the large literature that points to judges’ socioeconomic backgrounds as shaping judicial behavior. Existing studies point to judges’ gender (Boyd, Epstein, and Martin Reference Boyd, Epstein and Martin2010; Haire, Moyer, and Treier Reference Haire, Moyer and Treier2013; Kleps Reference Kleps2022), age (Manning, Carroll, and Carp Reference Manning, Carroll and Carp2004; Kaheny, Szmer, and Christensen Reference Kaheny, Szmer and Christensen2020), career stage (Kaheny, Brodie Haire, and Benesh Reference Kaheny, Haire and Benesh2008; Boyea Reference Boyea2010), ethnicity (Choi, Harris, and Shen-Bayh Reference Choi, Andrew Harris and Shen-Bayh2022), and race (Kastellec Reference Kastellec2013; Morin Reference Morin2014; Boyd and Rutkowski Reference Boyd and Rutkowski2020; Kleps Reference Kleps2022) as shaping their rulings. Particularly relevant to this study is the literature’s finding that judges’ professional backgrounds tend to shape various facets of their behavior. Beginning with Stuart Nagel (Reference Nagel1962), a large number of studies have found that prosecutors-turned-judges are less likely to favor defendants than judges of other professional backgrounds (Ulmer Reference Ulmer1973; Tate and Handberg Reference Neal and Handberg1991), less likely to rule in favor of claimants in employment cases (Shepherd Reference Shepherd2021), and more likely to rule against a motion to suppress evidence in search-and-seizure cases (Miller and Curry Reference Miller and Curry2023). In addition, legal scholars-turned-appellate judges in the United States are more likely to issue single-authored judicial opinions than those without academic careers in law (George Reference George2001). Underlying these studies is a socialization view of judicial behavior that holds that individuals become socialized into particular professional norms and practices, which in turn shape their behavior once they become judges.
Studies also show that individuals with lawyer backgrounds behave differently than those with other backgrounds, even after they leave lawyering for a new career. For instance, attorneys who are elected to the US Congress or state legislatures are more likely to support bills that extend tort law than legislators of other professional backgrounds (Matter and Stutzer Reference Matter and Stutzer2015); lawyer-legislators in the United States are also more likely to support legislation that removes impediments to filing lawsuits and less likely to support caps on awards for damages (Bonica Reference Bonica2020).
We build on these literatures to argue that lawyers who serve on ACRBs are more likely than those of other professional backgrounds to rule against agencies. As noted earlier, lawyers in Japan have historically been known for their public interest work. Those who serve on ACRBs and make decisions on behalf of prefectural governments are unlikely to fit the stereotype of a “cause lawyer” whose mission is to challenge and confront the state, but, still, lawyers in Japan are generally likely to view themselves as champions of the ordinary, private citizen. We draw here on data from the “Fact-finding Survey on Economic Foundation of Attorney Practices,” a survey of lawyers that was conducted by the JFBA in 2010 and which included a question that asked whether they had ever served on a local government commission or advisory committee. The survey was conducted before the inception of the ACRBs and also does not specifically ask which local government commission or advisory committee the lawyer served on, so the figures should be interpreted with caution, but it offers a glimpse into the Japanese legal community in general and also sheds light on the type of lawyer who has served on local government commissions like the ACRB compared to those who have not.
Table 3 shows that lawyers who have served on a local government commission or advisory committee are more likely to have handled at least one civil legal aid case and to have offered free legal counseling during the previous year than those who have never served. But note that, even among lawyers who have never served, a majority have offered free legal counseling, and close to half have handled at least one civil legal aid case. Moreover, as shown in Table A1 in the Appendix, outside the two largest cities of Tokyo and Osaka, the percentage of lawyers who have handled at least one civil legal aid case and/or offered free legal counseling is even higher, and the difference between lawyers who have served on local government commissions or advisory committees and those who have not largely disappears. Outside of Tokyo and Osaka, lawyers who have never served on a local government commission or advisory committee were actually slightly more likely to have handled at least one civil legal aid case in the previous year than those who have served, although the difference is small. This suggests that, outside of Japan’s largest cities, lawyers who have served on local government commissions like the ACRB are not particularly different from those who have not and that both are committed to assisting the disadvantaged and to taking the perspective of the average citizen.Footnote 20
Japanese lawyers’ professional activities
Source: Nihon Bengoshi Rengokai 2010. Figures are for previous year.
More generally, in their day-to-day work, lawyers are likely to come into contact with the average citizen, like those who file appeals against agency decisions. Table 3 shows that lawyers who have served on local government committees spent 58 percent of their time on individual, as opposed to corporate or government, clients, compared to 49.3 percent for lawyers who had never served. Of course, many of these individual clients may be wealthy citizens, but, given the high percentage of lawyers who handle civil aid cases and/or offer free legal counseling, a considerable amount of time is likely to have been spent with the less wealthy as well. Indeed, according to the “Fact-finding Survey on Economic Foundation of Attorney Practices,” the average lawyer in Japan had handled 9.7 civil legal aid cases in 2009, and, outside Tokyo and Osaka, the figure was 19.8 cases, which is well over one case a month (calculated from Nihon Bengoshi Rengokai 2010).Footnote 21 Thus, although lawyers who serve on ACRBs may not fit the bill of a typical “cause lawyer,” they are likely to be socialized to be attentive to the views of ordinary citizens and to take their perspective, even when they are engaged in contexts other than client work. As will be shown later in this article, this socialization may distinguish lawyers from the legal scholars who serve on ACRB panels.
A comparison with former local bureaucrats, the third largest professional group on the ACRBs, is also useful. In contrast to lawyers, former local bureaucrats are likely to have been socialized over the course of their careers to take the perspective of agencies (Robinson Reference Robinson2012). Thus, compared to lawyers, former local bureaucrats should be more likely to show deference to agency decisions.
This is not to say that there is no self-selection into private attorney careers or to say that lawyers on ACRBs do not make judgments on the basis of their professional expertise. Recall that ACRBs only rule in favor of prefectural governments in a small percentage of cases. Lawyers are, first and foremost, legal professionals, and they are likely to scrutinize cases from a professional perspective. Indeed, the literature on how judges’ socioeconomic backgrounds impact their rulings stresses that judges’ backgrounds matter on the margins precisely because the law sometimes does not provide unequivocal answers to cases at hand (Harris and Sen Reference Harris and Sen2019, 242). Similarly, in cases where lawyers’ professional expertise leads them to favor either the claimant or prefectural governments, their socialization as champions of ordinary people may come into play. Thus, we hypothesize:
Hypothesis 1: All else equal, ACRBs with larger representations of lawyers are more likely to rule in favor of appellants than those with fewer lawyers.
Selection of ACRB members
If ACRBs with more lawyers are indeed more likely to rule against agencies, why would some governors appoint more lawyers than others? This is an important question because it speaks to the issue of whether governors appoint more lawyers because they expect them to rule more often against agencies or whether they appoint them for other reasons, and the higher incidence of decisions against agencies is an unintended consequence. We draw here on the literature on technocratic appointments and on judicialization to generate our hypotheses.
Studies of technocratic appointments are useful for illuminating Japanese governors’ motivations for appointing lawyers to ACRBs because, as with “technocrats,” ACRB lawyers are non-elected experts who perform important administrative functions. For example, this literature has examined why the appointment of non-elected economists to Cabinet positions increased across Europe in the wake of the 2008 financial crash (Wratil and Pastorella Reference Wratil and Pastorella2018; Alexiadou and Gunaydin Reference Alexiadou and Gunaydin2019), and why presidents in Latin American countries often appoint non-elected economists as finance ministers (Centeno Reference Centeno1997; Dargent Reference Dargent2015; Kaplan Reference Kaplan2017). Of course, “technocrats” need not be economists, and the literature has also examined health experts (see, for example, Dargent Reference Dargent2015; Koch and Durodié Reference Koch and Durodié2022), climate change scientists (see, for example, Grundmann Reference Grundmann2007), and, most importantly for our purpose, lawyers (see, for example, Centeno Reference Centeno1997; Kaltenegger and Ennser-Jedenastik Reference Kaltenegger and Ennser-Jedenastik2022).
The literature offers two competing perspectives as to why political leaders may appoint experts to important administrative positions (Alexiadou and Gunaydin Reference Alexiadou and Gunaydin2019). On the one hand, leaders may appoint experts to legitimate unpopular decisions. Thus, for instance, studies find that economists are more likely to be appointed to ministerial positions during economic crises because governments often need to adopt unpopular policies, and appointing individuals with expertise will help to gain the public’s acceptance of those policies (see, for example, Wratil and Pastorella Reference Wratil and Pastorella2018; Emmanuelle et al. Reference Emanuele, Improta, Marino and Verzichelli2023). If the policies are still unpopular, experts who serve as ministers can help politicians avoid blame (Wratil and Pastorella Reference Wratil and Pastorella2018; Emanuele et al. Reference Emanuele, Improta, Marino and Verzichelli2023). When unpopular policies are needed, politicians may be especially motivated to appoint experts in competitive political environments where electoral volatility, and, in turn, the risk of losing power, is higher (Emanuele et al. Reference Emanuele, Improta, Marino and Verzichelli2023). The issue of legitimacy is particularly important for quasi-judicial institutions like the ACRB, which, like courts, lack a direct means of enforcement and often make decisions against plaintiffs’ claims (Gibson Reference Gibson2012).
Studies of judicialization also point to the importance of competitive party systems as a precondition for the appointment of individuals with expertise. Both the introduction of the ACRBs and the appointment of lawyers to the ACRBs may be viewed as an instance of judicialization of governance, or the expansion of judicial or quasi-judicial powers and procedures in government decision making (see, for example, Domingo Reference Domingo2004; Ginsburg Reference Ginsburg, Ginsburg and Albert2008; Hirschl Reference Hirschl, Epstein, Grendstad and Weinshall2023). This is because the ACRB is a quasi-judicial body, and lawyers bring the legal expertise that is needed for the quasi-judicial procedures to be implemented in practice. The judicialization literature’s political insurance perspective suggests that more competitive party systems are more likely to lead to judicialization than less competitive party systems (see, for example, Ramseyer Reference Ramseyer1994; Ginsburg Reference Ginsburg2003; Stephenson Reference Stephenson2003; Finkel Reference Finkel2005; Aydin Reference Aydin2013; Dixon and Ginsburg Reference Dixon, Ginsburg, Erin and Dixon2018). This, it is argued, is because when no party is assured a victory in the next election, all parties will prefer to limit the power of the majority and, as a result, to expand the powers of minoritarian institutions such as courts. In effect, this means that when ruling parties are in danger of losing power, they are likely to prefer institutions that rule on the basis of expertise than political loyalties. By the same logic, when party systems are competitive, governors are likely to appoint more legal professionals to the ACRB so that its rulings will be based more on legal-technical considerations than partisan motivations.
In sum, both the technocratic appointments literature and the judicialization literature expect that governors will appoint more lawyers in prefectures where gubernatorial elections are competitive, both because they believe that they and/or their party may be ousted from power and because they view lawyers as neutral arbiters of the law. The competitiveness of gubernatorial elections may be operationalized in at least two ways. First, governors who have been re-elected fewer times are likely to be less secure in power than those who have been elected more times.Footnote 22 Second, governors who were elected with narrower vote margins relative to the runner-up candidate are likely to be less secure in power than those who won office with larger vote margins. Thus, we hypothesize:
Hypothesis 2: All else equal, governors who have served fewer terms should appoint more lawyers to ACRBs than those who have served more terms.
Hypothesis 3: All else equal, governors who won office with smaller vote margins vis-à-vis the runner-up candidate should appoint more lawyers to ACRBs than those who were elected with larger vote margins.
Another variant of the political insurance perspective focuses on executive-legislative relations. Party systems may be more competitive during periods of “divided government” at the prefectural level, where the parties that support the governor do not hold a majority in the legislature and, also in such cases, governors may appoint more lawyers. Thus, we hypothesize:
Hypothesis 4: All else equal, governors who lack a majority in the legislature should appoint more lawyers to the ACRB than those with majorities.
Second, governors may also appoint lawyers because of their policy orientations or political convictions. The literature on technocratic appointments shows that left-leaning politicians often appoint mainstream economists who are committed to macroeconomic stability in order to send credible signals to international financial markets (see, for example, Kaplan Reference Kaplan2017; Alexiadou and Gunadyin Reference Alexiadou and Gunaydin2019). Note here the conceptual distinction between expertise and policy convictions. In the case of the ACRB as well, governors may appoint lawyers because of their political convictions. Given, as noted earlier, that lawyers have often spearheaded social change in Japan, governors may view lawyers as being generally left leaning.Footnote 23 From this perspective, governors who are endorsed by the conservative LDP may appoint fewer lawyers, whereas those who are endorsed by leftist parties may appoint more for two reasons. First, the more conservative leanings of LDP governors should lead them to prefer more agency decisions to be upheld, whereas leftist governors should favor more agency decisions to be overturned. Second, because most governorships and prefectural assemblies have historically been dominated by the LDP, local ordinances in most prefectures have typically been passed under the heavy influence of the LDP. LDP governors should thus have an interest in seeing agency decisions based on those ordinances to be upheld, whereas leftist governors should be more eager to see more agency decisions overturned. Thus, we hypothesize:
Hypothesis 5: All else equal, governors who belong to or are endorsed by the LDP should appoint fewer lawyers to ACRBs, whereas those who belong to, or are endorsed by, left-leaning parties should appoint more. Those who belong to, or are endorsed by, both should fall somewhere in between.
Data and methods
We compiled an original dataset of prefectural ACRB members between 2016 and 2021 by searching the websites of all forty-seven Japanese prefectures. Most prefectures’ websites disclose the names and occupations of current ACRB members. For past members, we consulted the Japanese National Diet Library’s Web Archiving Project, which stores past Japanese government websites, including those of prefectural governments (National Diet Library, n.d.). For the small number of prefectures that did not reveal the ACRB members’ names on their websites, we searched the Database of Administrative Appeals Decisions (Gyosei Fufuku Shinsa Saiketsu Toshin Kensaku Detabesu), where prefectures post ACRB decisions (Somusho, n.d.a). Most prefectural ACRB decisions are signed by ACRB members, so the database serves as an additional source of information as to who the ACRB members were and for which periods. Three prefectures (Ibaraki, Hiroshima, and Yamaguchi) were excluded from the study because no information on ACRB members was available either on the prefectural government website or in the above database. Although less than ideal, it was necessary to code the ACRB members annually since only annual data is available on the number of decisions and the number of cases for which the ACRB sided with agencies or the claimant. When new members were appointed in June or earlier in a particular year, they were coded as having served for that year; when they were appointed in July or later, they were coded as having begun their terms in the following year.
We also compiled data on the win rates by appellants by prefecture-year. This data is available from the website of the Japanese Ministry of Internal Affairs and Communications (Somusho 2025). At present, data is only available for the years 2016, 2018, and 2019. We employed win rates not only for cases that were decided by the ACRBs but also for those that were reviewed by internal reviewers and did not reach the ACRB stage because the internal reviewers are likely to anticipate the decisions of ACRBs and to make decisions based on the expected ACRB rulings. Because of the limited sample size, the results of our analyses for the correlates of win rates should be interpreted with caution.
For Hypothesis 1, the dependent variable is operationalized as the percentage of cases won by appellants among the total number of cases decided, by prefecture-year. Unfortunately, there is no publicly available data on the judgments of individual ACRB members on each case that they review, so we relied on aggregate prefectural-level data. Our main variable of interest is the percentage of lawyers on an ACRB panel, which is obtained by dividing the number of lawyers on the ACRB by the total number of ACRB members in each prefecture, by year. We also included a number of covariates. Appellants’ win rates may be affected by the percentage of former local bureaucrats and professors who serve on ACRBs so we included those percentages in the model. Because appellants’ win rates are likely to be influenced by the win rates from the previous year, we also employed a one-year lagged dependent variable. Finally, we also included the total number of reviewers. Social-psychological studies of jury deliberation as well as on small-group deliberation more generally show the influence of higher-status individuals (such as lawyers) to be greater in larger groups than in smaller ones (see, for example, Devine et al. Reference Devine, Clayton, Dunford, Seying and Pryce2001). We thus expected ACRBs with larger numbers of members to be more likely to rule against agencies than those with fewer. We ran two separate models: first, a two-way fixed effects model with prefecture and year fixed effects and, because our model includes a lagged dependent variable, we also ran a model with prefecture fixed effects only.
For Hypothesis 2 through Hypothesis 5, the dependent variable is the percentage of lawyers in the ACRB in each prefecture-year. To test Hypothesis 2, we examined whether governors who have served fewer terms are more likely to appoint more lawyers to ACRBs than those who have served more terms. We tested Hypothesis 3 by assessing whether governors who were elected with smaller vote margins relative to the runner-up candidate are likely to appoint more lawyers than those who were elected with larger vote margins. For Hypothesis 4, we tested whether governors who lack a majority in the assembly are more likely to appoint more lawyers than those who enjoy legislative majorities. We employed a variable coded 1 if the governor enjoys a majority within the assembly and 0 otherwise. Most governors during the period under study were endorsed by a combination of several different parties, so we coded whether the parties endorsing the governor combined had a majority in the legislature.
For Hypothesis 5, the variable of interest is the party endorsement of governors. Following Yoshihiko Takenaka and Masahisa Endo (Reference Takenaka and Endo2020), we created a dummy variable coded 1 if the governor ran from and/or was endorsed by the LDP and 0 otherwise and another dummy variable for leftist parties if a governor ran from, and/or was endorsed by, the DPJ, the Constitutional Democratic Party of Japan (CDP), the Social Democratic Party (SDP), or the Japanese Communist Party (JCP). Note that endorsement by the LDP and the leftist parties are not mutually exclusive; Japanese gubernatorial candidates are often endorsed both by one or more conservative party and one or more leftist party. Table A2 in the Appendix presents the summary statistics.
Because the percentage of lawyers may be affected by the number of reviewers on ACRBs, we include the number of reviewers as a covariate for tests of Hypothesis 2 through Hypothesis 5. Because the percentage of lawyers on an ACRB is likely to be driven by the percentage of lawyers in the previous year, we also included a one-year lagged dependent variable. As with Hypothesis 1, we ran both a two-way fixed effects OLS model with prefecture and year fixed effects and an OLS model with prefecture fixed effects only.
It is possible that appellants behave strategically and adjust their decisions to file depending on the composition of ACRBs. To probe this possibility, we tested the correlation between the number of cases filed in 2014, just before the ACRB reform, against 2019 figures. If appellants adjusted their behavior strategically in light of the new ACRB system, we should see a marked change in the number of cases filed between 2014 and 2019. We find, however, that the correlation between the number of cases filed in 2014 and 2019 was very high: 0.964. Later appellants may have adjusted their based on ACRB tendencies from 2016–19, but, for the period under study, it is reasonable to assume that there is limited endogeneity.Footnote 24
Results
Figure 2 shows the results of the tests for Hypothesis 1, with and without year fixed effects. For all figures in this section, coefficients for the year fixed effects and lagged dependent variable are not shown. The regression outputs for all analyses are presented in the Appendix in Tables A3–A20. As shown below, in both models, all else equal, ACRBs with higher percentages of lawyers are associated with significantly higher rates of decisions against agencies compared to ACRBs with lower percentages of lawyers. Although not strictly causal, we thus find support for Hypothesis 1. All else equal, a 10 percent increase in the percentage of lawyers on an ACRB is associated with a roughly 5 percent increase in decisions against agencies.
Determinants of administrative appeals rulings in favor of claimant, 2016, 2018, and 2019
Note: N = 86. Horizontal bars show 95 percent confidence intervals.
ACRBs with higher percentages of professors and local bureaucrats are also associated with higher rates of decisions against agencies, although the effects for former local bureaucrats are smaller than for lawyers, and the standard errors for professors are quite large. Consistent with our expectations, ACRBs with larger numbers of reviewers also see more decisions against agencies, with or without year fixed effects.
Figure A4 and Tables A4 and A5 in the Appendix show the results of tests for Hypothesis 2 through Hypothesis 5 tested separately. Although not necessarily causal, Figure A4 (a) shows some support for Hypothesis 2, namely, that governors who have served fewer terms and, thus, are less secure in power are likely to appoint more lawyers than those who have served more. The effects miss statistical significance at the p = 0.05 level but are significant at the p = 0.1 level. Figure A4 (b) reveals limited support for Hypothesis 3, that governors who are elected with smaller vote margins are likely to appoint more lawyers than those who were elected with larger vote margins. Figure A4 (c) presents weak support for Hypothesis 4, that governors with legislative majorities are likely to appoint fewer lawyers to ACRBs. Figure A4 (d) shows some support for Hypothesis 5, namely, that leftist governors do generally appoint more lawyers. The coefficients miss statistical significance at the p = 0.05 level, but they are significant at the p = 0.1 level, with or without year fixed effects. Yet governors who belong to, or are endorsed by, the LDP do not systematically appoint fewer lawyers. Thus, support for Hypothesis 5 is mixed. In all four specifications, the number of reviewers exerts little effect over the percentage of lawyers on ACRBs, with or without year fixed effects.
Finally, Figure 3 shows the results of Hypothesis 2 through Hypothesis 5 tested together. We continue to find some support for Hypothesis 2. The coefficient for the number of governors’ terms is significant at the 5 percent level with year fixed effects and just misses statistical significance at the 5 percent level without year fixed effects. Contrary to Hypothesis 3, we find no evidence that governors with smaller margins of victory appoint more lawyers. Contrary to Hypothesis 4, we find that governors with legislative majorities are more, rather than less, likely to appoint more lawyers. Contrary to Hypothesis 5, we find no evidence that governors’ partisanship affects the percentage of lawyers on ACRBs.Footnote 25 We continue to find limited evidence that the number of reviewers impacts the percentage of lawyers on ACRBs.
Determinants of percentage of lawyers on ACRBs, 2017–21
Note: N = 173. Horizontal bars show 95 percent confidence intervals. For governor partisanship, the reference category include governors who are not endorsed by any party.
Testing the mechanisms
In order to probe the mechanisms via which more lawyers on ACRBs are associated with more decisions against agencies, we tested whether the percentage of law professors, compared with other professors, might affect the incidence of decisions against agencies. Like lawyers, law professors have a professional grasp of the legal issues that are involved in making an informed ruling. Testing whether ACRBs with more law professors are also associated with more decisions against agencies is therefore useful for assessing whether the results that we have found thus far are driven by legal expertise or by socialization effects.Footnote 26 The results are shown in Figure 4 and in Table A11 in the Appendix. The upper graph in Figure 4 and Models 1 and 2 in Table A11 replace the variable for professors with a variable for the percentage of law professors on the ACRB by prefecture-year. The lower graph in Figure 4 and Models 3 and 4 in Table A11 test the professor variable and the law professor variable concurrently. We show both results because the high degree of correlation between the two variables (r = 0.672) may confound our results.
Determinants of administrative appeals decisions in favor of the claimant, 2016, 2018, and 2019, with the percentage of law professors in the model
Note: N = 86 for both figures. Horizontal bars show 95 percent confidence intervals.
Consistent with Hypothesis 1, we continue to find that, in all four models, ACRBs with higher percentages of lawyers are associated with higher rates of decisions against agencies. But three of the four models in Figure 4 and Table A11 show that having more law professors on ACRBs reduces the incidence of decisions against agencies, and the remaining one model shows no effect. As in Figure 2, all else equal, ACRBs with more professors, local bureaucrats, and more reviewers on ACRBs still see more decisions against agencies.
Although not strictly causal, these results lend tentative support to the socialization hypothesis, that lawyers may be more likely to rule in favor of appellants not only because of legal expertise but also because they are socialized into taking the perspective of individuals who are similar to those who are appealing agency decisions. At times, lawyers may represent individuals similar to appellants in court. Law professors also have legal expertise, yet they are much less likely to encounter the kinds of individuals who bring appeals against agency decisions, let alone to represent those individuals.
We further probe the socialization hypothesis by examining the extent to which lawyers with different career lengths rule against agency decisions. If ACRBs comprised of lawyers with longer careers are associated with more decisions against agencies, this would lend preliminary support to the socialization hypothesis that lawyers gradually become socialized into professional norms of siding with the average citizen who wishes to challenge agency decisions. To test this claim, we collected information on the year in which lawyers serving on ACRBs were admitted to the bar. This information was compiled from the JFBA’s database of lawyers (Nihon Bengoshi Rengokai, n.d.) and, when information was not available there, from the websites of the lawyers’ law firms, which often provided information on the year of bar admission. We took the mean of the years in which lawyers on the same ACRB were admitted to the bar. Because we did not test the effect of individual lawyers’ career lengths, the results of the analyses are merely suggestive and should be interpreted with caution.
We tested the effect of career length in two ways. First, we added to our models the mean year in which lawyers were admitted to the bar by prefecture-year. Second, because the mean year in which lawyers were admitted to the bar spanned just over thirty years, between 1979 and 2013, we split the mean years of bar admission into three variables of roughly ten years each: between 1979 and 1990 (inclusive), after 1990 and in or before 2000, and after 2000. As shown in Table A17 in the Appendix, we find no linear relationship between mean length of lawyer career and the likelihood of decisions against agencies. But, as shown in Figure A10 and Table A18 in the Appendix, we find that ACRBs with lawyers with longer careers—specifically, when the mean year of bar admission was in or before 1990, were associated with more decisions against agencies. Because we tested the effect of lawyers’ career lengths at the ACRB level rather than at the individual level, our results are tentative. Again, this is not to say that self-selection with respect to private attorneys is not present. But our results are consistent with the socialization hypothesis that lawyers become more likely to side with appellants as they gain experience.
Robustness checks
Further tests of Hypothesis 1
We conducted a series of robustness checks. First, studies of American judicial politics have consistently found that judges’ ideologies and/or partisanship exert the strongest effects over their decisions (see, for example, Segal and Spaeth Reference Segal and Spaeth1993, Reference Segal2002; Sunstein et al. Reference Sunstein, Schkade, Ellman and Sawicki2006; Epstein, Landes, and Posner Reference Epstein, Landes and Posner2012; Cohen and Yang Reference Cohen and Yang2019). The literature on the determinants of ALJ decisions in the United States also finds strong effects of partisanship (see, for example, Taratoot and Howard Reference Taratoot and Howard2011; Taratoot Reference Taratoot2014a). Because we lacked direct measures of ACRB members’ partisanship, following Lee Epstein and colleagues (Reference Epstein, Martin, Segal and Westerland2007), we use the partisan orientation of the governor that appointed the ACRB members as a proxy. If decisions against agencies occur more often just after the partisanship of the governor has changed, this would provide support for the view that members vote on the basis of ideology. To test this hypothesis, we coded changes in governorships as 1 if a LDP-endorsed governor is replaced by a governor who is endorsed by leftist parties or vice versa and 0 otherwise. As shown in Figure A8 and Table A15 in the Appendix, we find limited evidence that changes in governors’ partisan orientations is associated with changes in the incidence of decisions against agencies. This finding may reflect the less polarized nature of Japanese compared to American politics. Consistent with Hypothesis 1, we continue to find that ACRBs with more lawyers are associated with higher rates of decisions against agencies.
Second, relatedly, we tested whether changes in governors are correlated with changes in the frequency of decisions against agencies. In the American context, scholars have found evidence of “loyalty effects” in which judges decide in favor of agencies less frequently when the president who appointed them leaves office (Epstein and Posner Reference Epstein and Posner2016). As shown in Figure A9 and Table A16 in the Appendix, we find limited evidence of “loyalty effects” in the Japanese ACRB context; changes in governors are not associated with systematic changes in the incidence of decisions against agencies. We continue to find support for Hypothesis 1 that higher percentages of lawyers on ACRBs are correlated with higher rates of decisions against agencies.
Third, we tested whether caseload affects the incidence of decisions against agencies. Jeffrey Budziak (Reference Budziak2016) shows that US Courts of Appeals judge panels with heavier caseloads are more reluctant to rule against agencies because reversals typically require greater justification than affirmations. We tested this argument using the number of decisions handed down divided by the number of ACRB members, by prefecture-year. As shown in Figure A7 and Table A14 in the Appendix, we find some evidence that panels with larger caseloads are associated with more decisions that uphold agency decisions. The coefficients miss statistical significance at the 5 percent level but are significant at the 10 percent level. Here again, consistent with Hypothesis 1, the percentage of lawyers is positively associated with the rate of decisions against agencies.Footnote 27
Further tests of Hypothesis 2 through Hypothesis 5
We also tested the robustness of our findings for Hypothesis 2 through Hypothesis 5. First, we examined whether changes in the governor’s partisanship are associated with increases or decreases in the appointment of lawyers by adding two new dummy variables: one that is coded 1 if a conservative governor is replaced by a leftist governor and 0 otherwise, and another that is coded 1 if a leftist governor is replaced by a conservative governor and 0 otherwise. For the period under study, there were only sixteen cases of leftist governors replacing conservative governors and four cases of conservative governors replacing leftist governors, so our findings should be interpreted with caution. As shown in Tables A19 and A20 in the Appendix, changes in governors’ partisanship are not associated with systematic shifts in the percentage of lawyers on ACRBs. We still find support for Hypothesis 2, namely, that governors who have served more terms appoint fewer lawyers, but we continue to find limited support for Hypothesis 3 through Hypothesis 5; governors who won with smaller vote margins, governors without legislative majorities, and leftist governors did not systematically appoint more lawyers.
Second, if lawyers view serving on ACRBs as being beneficial, they may lobby for more positions on the panels. Bar associations may be politically more influential in prefectures where there are more lawyers per population. We thus tested whether governors appoint more lawyers in prefectures with more lawyers per population. As shown in Tables A7–A9 in the Appendix, we find that governors in prefectures with more lawyers per population did not systematically appoint more lawyers to ACRBs. We continue to find some support for Hypothesis 2, namely, that governors who have served fewer terms are likely to appoint more lawyers to ACRBs. The effects miss statistical significance at the 5 percent level but are significant at the 10 percent level. We also continue to find limited support for Hypothesis 3, namely, that governors who won office with smaller vote margins are more likely to appoint more lawyers. Contrary to Hypothesis 4, governors who lack legislative majorities were not more likely to appoint more lawyers than those with majorities.
We also find mixed support for Hypothesis 5. In Table A7, governors who are endorsed by leftist parties are likely to appoint more lawyers to ACRBs. But we fail to find that LDP-endorsed governors are likely to appoint fewer lawyers. The number of lawyers per population has little effect over whether more lawyers will be appointed to an ACRB.
Figure A4 and Table A8 show results of Hypothesis 2 through Hypothesis 5 tested together. Consistent with Hypothesis 2, we find governors who have served more terms to be less likely to appoint more lawyers to ACRBs. We continue to find limited support for Hypothesis 3 and Hypothesis 5, but, for Hypothesis 4, we find that, contrary to expectation, governors with legislative majorities are more likely to appoint more lawyers than those without. The reason for this is unclear, but it deserves scrutiny in future studies.
Conclusion
Across the developed world, citizens bring many more administrative appeals than administrative litigation, yet we know little about the determinants of administrative appeals decisions, especially outside of the United States. To address this gap in the literature, we built on previous studies that found that judges’ professional backgrounds have had an important bearing on judicial behavior. Drawing on the case of Japanese ACRBs, we have shown that professional careers in private legal practice affect quasi-judicial decisions. We find that, ceteris paribus, ACRBs with more lawyers are associated with more decisions against agencies than those with fewer lawyers. Although not necessarily causal, a comparison with law professors shows that these effects are not driven by legal expertise per se, and we have also found preliminary evidence that ACRBs with lawyers with longer careers are associated with more decisions against agencies. These findings are consistent with a socialization perspective. Although serving on ACRBs may elevate lawyers’ professional status, the gains in compensation are limited, so the risk of falling out of favor with the governor does not deter them from ruling against agencies, despite the limited institutional independence of Japan’s ACRBs. Although this is not to deny the possibility of selection effects, lawyers routinely see, and at times represent, individuals who wish to contest government claims and thus may be socialized to take their perspective. Whereas existing studies of US courts and tribunals have pointed to the strong effects of partisanship, we find that, in the less polarized Japanese context, professional background more strongly shapes decisions than partisanship.
Consistent with political insurance theory, we find some evidence that governors who have served fewer terms appoint more lawyers than those who have served more. Although the political insurance perspective has typically been applied to judicial settings, our findings provide new evidence for the theory in quasi-judicial contexts. By contrast, we fail to find that ideological concerns drive governors’ appointment of lawyers. Our findings thus suggest that the association between more lawyers on ACRBs and the higher incidence of decisions against agencies is not driven by left-wing governors who stack the ACRBs with lawyers in hopes that the latter will rule more often against agencies but, rather, is an unintended consequence of politically insecure governors attempting to legitimize ACRB decisions and to shield themselves from blame for unpopular decisions. Our finding that governors who have served fewer terms appoint more lawyers suggests that, all else equal, governors who worry about being ousted from office may appoint more lawyers because of the likelihood that they would rule more on the basis of legal-technical reasons than partisan motivations, even after the governor has left office. By contrast, governors who are more secure in power may be less concerned about “insuring” themselves against what may happen after they leave power.
Due to limited data availability, our analyses of the determinants of administrative appeals decisions rely on three years of data. Because of the small sample size, as well as the aggregate nature of our data, our results should be interpreted with caution. Moreover, as ACRB members gain in experience, it is possible that their decisions could change over time. We hope to follow up on the extent to which the findings from this study travel to subsequent years as more data becomes available.
Recent work by Adrienne Sala (Reference Sala2024) shows that the courts in Japan may have more influence over policy making and/or administrative enforcement than has been commonly assumed, and this article shows that the same pattern may be true of quasi-judicial procedures as well. The precise causal pathways by which this operates present a fruitful avenue for future study. In particular, future work should also further probe the mechanisms by which more lawyers on ACRBs lead to more decisions against prefectural governments. We have found preliminary evidence to support the socialization hypothesis that ACRBs with lawyers with longer careers are associated with more decisions against prefectural agencies. But we have also found that the effect was not linear over time. Future studies should assess the broader applicability of this finding as well as the mechanisms as to why this may be. In-depth, qualitative case studies would be well suited for this research.
The generalizability of our findings beyond the Japanese context also presents an important question for future inquiry. Administrative adjudication systems operate in markedly different institutional contexts in different countries, and the composition of administrative appeals review panels also varies considerably across different countries. In the United Kingdom, for instance, typically over 60 percent of tribunal judges are non-barristers (Ministry of Justice 2023b). Perhaps because Japanese prefectural politics are not very ideologically polarized, in contrast to studies of US administrative judge decisions (Taratoot and Howard Reference Taratoot and Howard2011; Taratoot Reference Taratoot2014a), we did not find ideological concerns to drive the ACRB decisions. Future studies should probe whether, as in Japan, the impact of ideological orientations on ACRB appointments, and, in turn, administrative appeals decisions, is lower in relatively less polarized contexts. Moreover, because lawyers typically serve on a variety of different judicial and quasi-judicial settings in many developed democracies, the applicability of findings beyond the ACRB presents a fruitful path for future inquiries.
Beyond the Japanese context, our finding that a private attorney background shapes administrative appeals decisions also yields broader implications for the study of judicial behavior. In Anglo-American systems, judges are often appointed from the ranks of private attorneys. Between 1789 and 2010, 20.7 percent of all US Supreme Court justices have been appointed straight from private legal practice, and, indeed, along with federal appellate judge, this is one of the most common professional backgrounds for serving on the Supreme Court (Hurwitz and Lanier Reference Hurwitz and Noble Lanier2012, 82). British judges are also commonly selected from among barristers and solicitors, and the higher up in the judicial hierarchy, the smaller the percentage of non-barristers; at the High Court level, only 7 percent of judges do not have barrister backgrounds (Ministry of Justice 2023b). Even in civil law countries, which typically rely heavily on career professional judges, it is not unusual for private attorneys to become judges at the highest court level. For instance, two of the twelve current justices of the German Constitutional Court, including the president, have private law experience (Federal Constitutional Court 2023), and four of the fifteen current Japanese Supreme Court justices are former private attorneys (Supreme Court of Japan, n.d.). Although the effects of partisanship are typically stronger in the US context, our findings suggest that private attorney background, and professional background more generally, may be stronger in contexts that are less polarized than in the United States. In this context, it is important that lawyers also perform different tasks and functions in different countries, and Japan’s lawyers perform relatively narrow tasks relative to those of other developed democracies (Rueschemeyer Reference Rueschemeyer1986). How the effects of a professional legal background may vary in countries where lawyers perform a broader range of functions poses a promising avenue for further research.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/lsi.2025.10103
Acknowledgments
Earlier versions of this article were presented at the 2023 Annual Meeting of the Socio-Legal Studies Association, the 2023 World Congress of the International Political Science Association, and the 2024 APSA Virtual Research Meeting. I thank participants at these panels, as well as the three anonymous reviewers, for helpful comments. Part the analysis in this article draws on the “Fact-finding Survey on Economic Foundation of Attorney Practices, 2010” (Depositor: Japan Federation of Bar Associations), which is available upon request by the Social Science Japan Data Archive, Center for Social Research and Data Archives, Institute of Social Science, University of Tokyo (Survey no. 1038). Replication files for the main empirical results are available on Harvard Dataverse at https://doi.org/10.7910/DVN/1CB8YMl. This research was generously supported by JSPS Kakenhi Grant nos. 19K01441 and 24K00212.
Open access
