Law's dominion and the market for legal elites in Japan

Milhaupt, Curtis J

ABSTRACT

In this Article, we present data on legal elites in Japan-legally trained university graduates poised to pursue successful careers either as fast-track bureaucrats or lawyers handling sophisticated business transactions. The data show a marked shift in employment patterns over the past decade: increasingly, Japan's most elite university graduates are forsaking the bureaucracy for law.

We find that changes in Japan's underlying economic, political, and legal institutions are a primary cause of this shift. We argue that this trend is not a temporary phenomenon, but reflects a more fundamental transfer of authority in Japan from the bureaucracy to the legal system. The evidence sheds new light on two longstanding debates: the impact of law and lawyers on economic success, and the bureaucracy's role in the governance of the Japanese economy.

The data we examine are hard to square with the widespread view of Japan as "Exhibit A" for the proposition that societies encourage economic growth by steering their most talented youth away from "redistributive legal careers. " Rather, the data indicate that in Japan (as elsewhere), talented college graduates pursue positions of power, prestige, and profit. While those positions were once located in the elite economic bureaucracy, they are now migrating to the legal system. Contrary to the evidence of stagnation in the economic and policy environments flowing out of Japan in recent years, close examination of the career choices of Japan's most highly regarded youth reveals a society in transition.

I. INTRODUCTION

The role of lawyers in the economy has been a continuing source of controversy. A decade ago, economists produced a spate of statistical research indicating that lawyers (and implicitly law) are a drag on the economy, and bad for society to boot.1 Even some of America's most prominent academic lawyers publicly supported this view.2 This research and commentary suggested that societies the world over have a choice: they can either nurture engineers and other innovators who produce wealth or they can churn out lawyers and other rent seekers who will redistribute wealth and contribute to the complexity and adversarial nature of human interaction.3 Legal scholars retorted in a variety of ways; most interesting was the argument that lawyers, far from functioning as an albatross around the economy's neck, are "transaction cost engineers" who add value to deals by reducing the uncertainties inherent in imperfect market exchanges.4

Common to many entries in this debate was reference to Japan. Coincidentally or otherwise, at the height of the academic controversy and amidst a booming economy, Japan had very few lawyers-fewer than 13,000 in 1985. For law skeptics, "Japan was Exhibit A, displaying the inverse relation of lawyers to economic vigor."5 Law agnostics criticized the "armchair comparativists" for drawing simplistic conclusions from the Japanese data.6

In addition to serving as "Exhibit A" in the foregoing debate, Japan was also widely discussed in the realm of political science; Japan specialists sometimes used comparisons between lawyers and the economy to advance their views on the distinctiveness of Japanese economic governance. The most famous example is Chalmers Johnson's depiction of Japan as a "plan rational" state led by capable bureaucrats, as opposed to a regulatory, or "market rational," state such as the United States.7 Consider Johnson's description of the distinction:

In Japan the developmental, strategic quality of economic policy is reflected within the government in the high position of the so-called economic bureaucrats, that is, the officials of the [elite economic] ministries . . . . These official agencies attract the most talented graduates of the best universities in the country, and the positions of higher-level officials in these ministries have been and still are the most prestigious in the society. . . . [T]he elite bureaucracy of Japan makes most major decisions, drafts virtually all legislation, controls the national budget, and is the source of all major policy innovations in the system. Equally important, upon their retirement . . . these bureaucrats move from government to powerful positions in private enterprise . . . .

In market rational systems such as the United States, public service does not normally attract the most capable talent, and national decision-making is dominated by elected members of the professional class, who are visually lawyers, rather than by the bureaucracy.8

While Johnson's argument has been challenged by scholars as misleading and outdated,9 it remains representative of many views on Japan.10

Today, the contours of the debate have shifted. Over the past several years, a formidable literature has developed around the central idea that the quality of law and its enforcement (which presumably implicates lawyer access and quality) has major effects on such economically central areas as corporate structure, capital formation, and the ability of a system to rebound from exogenous shocks.11 To date, however, insights from this paradigm shift in economics have not been reflected in the literature on lawyers and economic growth.

As the debate on law has shifted, so too has "Exhibit A" changed in ways both obvious and subtle. The Japanese economy has suffered an entire decade of low or negative growth, asset deflation, financial crises, and rising unemployment. Other changes have been less visible to casual observers but equally significant. For instance, not only have many markets been deregulated, but the bureaucracy has suffered considerable setbacks in policy and public opinion while the demand for legal services has grown.

By taking a close look at "Exhibit A," we thus join two distinct debates: the impact of law and lawyers on Japan's economic success and the bureaucracy's role in the governance of the Japanese economy. To do so, we explore the evolving roles of law and bureaucratic regulation in Japan by examining changes in the market for legal elites during a time of dramatic economic and institutional transformation. We define Japanese "legal elites" as the top two to three hundred law graduates in a given year who could successfully pursue highly compensated careers in either the upper echelons of the most prestigious ministries (such as the Ministry of Finance) or in Japan's largest and most prestigious law firms.12 They are, in short, fast-track career bureaucrats and transactional lawyers.13 These legal elites, many of whom would have in the past opted for a bureaucratic career, have begun to join the bar in ever-increasing numbers.

Our central claim is that the shift in employment patterns of legal elites is primarily caused by changes to the legal institutions and political economy of Japan, reflecting a conscious move toward more transparent mechanisms of governance based in law. To be sure, the underlying institutional setup in society is not the only factor that affects individual career decisions; idiosyncratic considerations such as the guidance of a mentor, the choices of one's peers, or the desire to follow in (or avoid) a parent's professional footsteps may also be critical in individual cases. Nonetheless, we find considerable evidence that the relative attractiveness of careers in government and law has been affected by new legal rules designed to strengthen market discipline over Japanese firms and to reduce bureaucratic discretion-whether the original source of that discretion was a dominant bureaucracy (as Chalmers Johnson claimed) or in political slack (as other observers have argued more recently).

In making this claim, we explore further two themes from our earlier work on institutions and law's domain in the Japanese economy. First, in an earlier article, we showed how various rights-enforcement gaps resulting from Japanese institutional design are filled by the entrepreneurial impulses of organized crime syndicates, in a phenomenon we called "the dark side of private ordering."14 Here, we show that in an analogous fashion, the governance void created by the decline of the Japanese bureaucracy has opened avenues of opportunity for elite lawyers. Although hardened law skeptics may disagree, we view this as a decidedly "bright side" development for private ordering.

Second, in other work, we have also noted the shift toward a more law-oriented society in Japan.15 Increasingly, Japanese commentators are analyzing the shift as well. A recent book by one of Japan's most prominent business lawyers argues that government downsizing and deregulation is leading to a more "legalized" Japanese society in which rights are "retuned from the bureaucrats to the people."16 Kyoto University Law Professor Michio Muramatsu argues that "relations among groups, and between groups and the bureaucracy, are becoming increasingly determined by law, not bureaucracy."17

To date, however, no commentator has attempted systematically to test the validity of these assertions. We do so here by examining previously unexplored career data that shed new light on an age-old question-Japan: Who Governs?18

This Article proceeds as follows. Part II briefly discusses the controversial theory and data on law, lawyers, and economic success, and links that debate to the question of who governs the Japanese economy. Part III presents empirical data on the career choices of Japanese legal elites, which reveals that legal elites are forsaking the bureaucracy for law. In Part IV, we explain how institutional change underlies this reallocation of talent. Part V analyzes implications from our findings for the debates over Japanese governance and the impact of talent allocation on economic growth.

II. LAW, LAWYERS, AND ECONOMIC SUCCESS: JAPAN AS EXHIBIT A

Until recently, economists' view of lawyers as rent seekers who serve as a drag on economic growth was firmly entrenched in the literature and, as noted in the Introduction, was persuasive even to some high profile members of the legal profession. While this view took support from various theoretical constructs (many of which can be traced to Mancur Olson's The Logic of Collective Action and The Rise and Decline of Nations19), one important strand of the argument was the talent diversion theory. That is, rent seeking shifts talent from productive pursuits into the legal profession. Former Harvard University president Derek Bok states the case bluntly: the attractiveness of law school in the U.S., he claims, results in "a massive diversion of exceptional talent into pursuits that often add little to the growth of the economy. . . ."20

Arguably the most prominent and controversial studies in this genre are by two teams of economists: Stephen Magee, William Brock, and Leslie Young (Magee et al.),21 and Kevin Murphy, Andrei Shleifer, and Robert Vishny (Murphy et al.).22 Magee et al. frame their basic inquiry as whether "redistributive lawyering reduces GNP growth rates."23 They find a negative correlation between the ratio of lawyers per doctor (used to normalize countries at different stages of development) and the growth of GNP per capita for a sample of 34 countries for the period 1960-1980 and conclude that the answer is a resounding yes.24

Murphy et al. sought to test the talent-diversion theory more directly by examining the ratios of law students and engineering students to total college population across a large sample of countries. They find that economic growth is negatively correlated to the law student ratio and positively correlated to the engineering student ratio.25 They conclude that "[l]awyers are indeed bad, and engineers good, for growth" because of the diversionary effect of the legal profession, since they find no direct effect on levels of investment, which theoretically could be influenced by rent-seeking activities.26 Their logic is that "if law is an attractive major, the quality of rent seekers is higher, and hence, indirectly, the quality of entrepreneurs is lower, and technological progress and income growth are smaller."27 Accordingly (writing in 1991), they speculate that "the allocation of talent to the rent-seeking sectors might be the reason for stagnation in much of Africa and Latin America, for slow growth in the United States, and for success of newly industrializing countries where these sectors are smaller."28

These studies are not without problems. The number of lawyers per doctor is an imprecise measure at best, and undergraduate law majors in Japan and elsewhere often pursue non-legal careers, some of which may be rent-seeking, some of which are not. More fundamentally, as researcher Charles Epp has pointed out, until now the talent-diversion theory has lacked a sound basis for empirical investigation, because "[w]e do not know what occupations most lawyers would have chosen had they not entered law, nor could we easily determine whether the alternative occupations are more productive than law."29

Fortunately, if somewhat ironically, Japan-"Exhibit A" in the debate-allows for precisely such a test because, for reasons we explain below, we do know with a high degree of certainty what occupations most Japanese lawyers would otherwise have chosen, and we have a fair idea whether the alternative occupation is more productive than law. To be sure, there are significant limitations to our test. First, we have data on only one country, albeit one that is uniquely positioned in this debate. Second, we have only imperfect measures of the "productivity" of the alternative occupation. Despite these limitations, our test offers a fresh insight into this longstanding debate.

We are doubly fortunate because the data also allow us to contribute uniquely to the second age-old debate: the question of what actor has the upper hand in the regulation of the Japanese economy. The lines of argument in this debate are widely known, and we outline them here only in skeletal form.

The most prevalent view, championed by Chalmers Johnson and still common in popular accounts of Japan, is that bureaucrats dominate, or at least orchestrate, policy with input from big business.30 This view has always been open to challenge by those who saw a more pluralistic governance structure in operation,31 and scholarly support for this view has waned considerably in recent years. The revisionist argument, drawn from political science theory on the behavior of U.S. federal agencies, holds that the evidence typically presented in support of the bureaucratic dominance argument32 is equally supportive of a more plausible interpretation: bureaucrats are the faithful agents of the politicians, who are actually, if indirectly, in control of policymaking.33 The scope of bureaucratic discretion is thus determined by the policy preferences of politicians, which are in turn determined by electoral interests.

Two important and related points from this brief survey deserve highlighting. First, most accounts of Japanese governance are relatively static and do not focus on the possibility or avenues of structural change. Thus, accounts developed in the 1980s and early 1990s to explain Japanese economic regulation may not be accurate today. Second, in sharp contrast to accounts of economic governance in the United States, in most accounts of Japanese economic governance, lawyers (and by extension, courts and law) are left out completely.34

III. CAREER CHOICES OF LEGAL ELITES

Elite law students in Japan have two basic, and until very recently, mutually exclusive career options: the bureaucracy or the legal profession.35 Each involves very specific and patterned career paths, marked by admission to law departments of elite public and private universities, cram school preparation, and passage of difficult state-sponsored standardized tests. After briefly describing each stage, we show changing career patterns over time. We present several measures of these changes, and while no single measure alone necessarily captures the phenomenon adequately, the entirety of the data strongly suggest a shift in social stature, authority, and income from the bureaucracy to the bar.

A. The Path of Legal Elites

University students who excel at academics become elite bureaucrats and lawyers by passing through a similar gatekeeper-the standardized test. In this section, we describe the two tests: for lawyers, the bar exam (shiho shiken), and for bureaucrats, the civil service exam (kokka komuin isshu shiken).

1. Lawyers: Shiho Shiken

With a handful of statutory exceptions, all legal professionals in Japan (i.e., lawyers, judges, and prosecutors) must take and pass the bar examination, which consists of three parts.36 Part I is a multiple-choice test composed of 60 questions, equally divided among constitutional law, civil law, and criminal law. In 2002, 6,457 of 45,622 candidates (14%) passed this 3 1/2-hour test, and thus earned the right to sit for Part II in that year.37 Part II is the most difficult part of the bar exam, consisting of essays in the fields of constitutional law, civil law, criminal law, commercial law, and civil and criminal procedure. In 2002, 1,244 candidates (3%) passed this test, earning the right to sit for Part III that year or the following year.38 Part III is a four-day oral test of knowledge of constitutional law, civil law, criminal law, and civil and criminal procedure. Candidates spend fifteen to twenty minutes each day fielding questions from two examiners. In 2002, 1,183 candidates (3%, about one-fourth of whom are women) passed this portion of the test, making it one of the easier hurdles as a percentage matter (95% of Part II passers).39 Candidates who pass this exam are admitted to the Legal Training and Research Institute (LTRI), where they receive eighteen months of formal legal training before being admitted to the legal profession. Because they are technically classified as civil servants during that period, they are paid a monthly base salary of 208,300 yen, or about $25,000 annually including bonuses, by the state.40

We have discussed the scarcity of lawyers in Japan and its negative consequences elsewhere,41 and we will not repeat those arguments here. Suffice it to say that Japan has few lawyers. In 2002, Japan had slightly more than 18,000 lawyers, up from 14,000 in 1990, 11,600 in 1980, and 9,000 in 1970. Even with recent expansion in the ranks of the bar, as of April 2000, of the 253 court districts in Japan, 71 were "zero-one districts," with no lawyers or only one, as 75% of the bar is concentrated in Tokyo and Osaka.42 The number of transactional lawyers in large firms remains quite small, perhaps 600 to 800.43

2. Bureaucrats: Kokuichi

By definition, all fast-tract elite bureaucrats have taken and passed the First-Tier Civil Service Examination.44 Although the exam is issued in several fields, including such law-related fields as administration and economics, "law dominates. For first-class bureaucrats who expect to rise to the top of the most prestigious agencies, with virtually no exception, the field is law (horitsu-shoku)."45 The national civil service exam has three tiers, with each tier reflecting a different level of the bureaucracy.46 For potential elites, however, the first tier is all that matters.

The First-Tier Civil Service Examination, which is divided into two parts, is heavily geared toward testing legal knowledge.47 Part I is a three-hour, fifty-five-question multiple-choice test of general knowledge. Twenty-five of the questions test reading comprehension and logic and thirty questions, from which the candidate must choose twenty, test science, literature, and social sciences. Part II is a three and one half-hour essay test of legal knowledge, followed by a two-hour essay test of general knowledge in which the candidate answers one of two questions.48 After passing the civil service exam, elite ministries make employment decisions based on candidates' national ranking. Unlike bar examinations in the United States, where all that matters is obtaining a score above a predetermined passing threshold, Japanese civil service exam takers compete for the highest scores, for only the best will actually be hired. Hiring is determined by the personnel departments of individual agencies and ministries. In 2000, of 38,841 total candidates for all First-Tier positions, 7,937 candidates took the exam in law, 255 passed, and only 149 eventually were offered and accepted positions (leaving 110 passers without positions). Of these 149, 8 went to the Ministry of Finance, 9 to the Ministry of Justice, 26 to the Ministry of Public Management, Home Affairs, Posts and Telecommunications, 14 to the Ministry of Economy, Trade and Industry (formerly known as MITI) and 13 to the Ministry of Education.49 We examine hiring trends in the next section.

Passing the exam qualifies a passer for employment at twenty-eight ministries and agencies, but legal elites are primarily only concerned with the most prestigious ministries. The top choices among elite University of Tokyo students are the Ministry of Finance, METI (the former MITI), the former Ministry of Home Affairs, and the National Police Agency, the so-called "four dynasties."50 Hiring by these ministries, however, does not necessarily lead to a law-related career. Although legal recruits are trained in and tested on the law, upon arrival at the ministry, they are in fact rotated among departments and not assigned distinctly legal tasks.51 As Japanese legal scholar Setsuo Miyazawa puts it, "Law graduates are hired essentially because they are considered to be smarter than others."52

B. Test Data

In this section, we analyze long-term and recent trends in examination-taking, examination-passing, and employment decisions of young legal elites. We begin by looking at national statistics. To get a more precise understanding of developing phenomena, we then look at data from elite universities. We examine the phenomenon of double-passers before turning to data on quasi-lawyers.

1. National data

The nationwide data on the civil service exam and the bar exam show a marked shift. We begin with the data on the bar exam. Figure 1 shows the number of individuals who took the bar exam from 1960 to 2002.53 As the figure shows, with the exception of a slight decrease in takers in the 1980s (when Japan's private sector boomed in the bubble economy), both the number of those who took the bar exam and failed and those who took the bar exam and passed have risen significantly over time. The rise in the past decade is particularly significant. In 1991, 22,596 candidates sat for the bar exam. By 1998, that number topped 30,000, and in 2002, the number was 45,622.

The number of passers is determined not by the overall performance ability of that year's exam takers, but by a predetermined formula. Prior to 1991, the LTRI admitted approximately 500 students each year. Although the number was firmly grounded in the bar's desire to limit competition, it also reflected the physical constraints of the LTRI, which was said to be capable of housing only 500 students.54

After 1991, the LTRI gradually raised the number of passers to the approximately 1,000 students that are now admitted. Still, the test remains statistically difficult to pass; because of the increase in takers, the pass rate remains at about 2.5%. The average age of passers has fallen from nearly 29 to just over 27 from 1990 to 2002, but even in 2002, 38.5% of passers needed more than five years of study to pass.55 Of 2001 passers, only 53, or 5.4%, passed on their first try.56 Future changes to the system, including a tripling of the annual 1,000 figure and an introduction of U.S.-style graduate law schools, are discussed in Part IV.

As the number of bar takers and bar passers has increased, the number of students who sit for and pass the elite civil service exam has fallen.57 Figure 2 shows the number of takers and passers for the twenty-five-year period from 1978 to 2002. As the figure shows, the number of exam takers has declined twice in the period. The first decline comes in the late 1980s, as the Japanese bubble economy boomed. This decline roughly tracks the decline in the number of bar takers in the same period (see Figure 1), a correlation that strongly suggests that potential test-takers of both types chose other profitable careers, at least for the years in question. After the bubble burst, the number of civil service exam takers increased, but not to the level of the early 1980s. The number of takers then declined in 1997, falling to a twenty-five year low in 1998. Although the number has risen incrementally in the last three years, the average number of takers in 1998-2002 remains much lower than the same figure for the previous five years. Since the decline comes in the absence of either unusual economic prosperity or a similar downward trend in the bar exam, more complex, and perhaps less transient, forces may be at work. Still, the absence of a definitive long-term trend over the period suggests that additional data examination is necessary, and we do so in the next subsection with university-specific data.

Assuming arguendo that the number of civil service exam takers is decreasing in the long term, fewer persons might opt to take the civil service exam if the pass rate were lower. But the data in Figure 3 on civil service exam passers suggests that the drop in exam takers is not a response to grading, as the number of passers has actually increased steadily, if slightly, since 1997. The post-1997 pass rate is significantly higher than the pass rate for the early 1980s, when the sit rate was high. In fact, as Figure 3 shows, in percentage terms, it has become easier to pass the civil service exam than the bar exam. Yet applicants for the civil service exam do not show the same long-term increase as bar exam applicants.

The recent and apparent long-term decrease (or at least lack of a long-term increase) in civil service exam takers is also interesting in light of relatively constant employment data. Our interviews with those in hiring suggest that the number of persons that elite ministries wish to hire has remained relatively constant over time. While data on this point are not publicly available, we do have data on actual hires. Table 1 shows the number of law-educated persons hired by the elite bureaucracy within the twenty-month period following the exam for each exam year (for example, data for the 1999 exam year are for hires through March 31, 2000).58

As the table shows, both the raw number of hires and the number of hires as a percentage of exam passers has fallen in the last decade.59 But the decrease has been minimal; the raw number hired by 2002 from the 2000 test was only sixteen persons fewer than the figure from the 1990 test. The number of persons eventually hired as a percentage of exam takers remains at about two percent. This is only slightly lower than the rate at which candidates pass the bar exam, which virtually guarantees employment.

It is plausible that the small decline in hiring might affect the number of candidates for the exam. Candidates have access to information, as the number of anticipated hires in each category is announced before each year's test. But current university students preparing for the exam told us that they were either unaware of the historical decline or that they considered it unimportant, and in fact the recent slight increase in takers is consistent with that explanation.

Recent developments suggest that the bureaucracy itself views these trends as indicating a decline in popularity of the profession and not a transient phenomenon. Specifically, the bureaucracy has adopted two institutional strategies, and is considering a third, in response to competition from the bar for legal elites. First, until 2000, the Lawyers' Law prohibited attorneys from serving as government employees; the only exceptions were certain enumerated high-level positions such as prime minister.60 But in response both to increased competition from the bar for high-quality candidates and to the shift to a more law-oriented society, the Diet passed a special law in late 2000 that, among other things, amended the Lawyers' Law to allow lawyers to serve in certain short-term positions of less than five years in the bureaucracy.61 A handful of lawyers (in 2002, one such lawyer told us that he estimated the number to be five) now work in the elite bureaucracy, primarily in finance-related positions, and the number is expected to increase.

Second, in 2001, in direct response to competition for candidates from the expanding bar, the government introduced a plan to double the number of civil service exam passers.62 Finally, in spring 2002, the National Personnel Agency and other related entities began to debate a new policy by which a graduate of one of Japan's graduate law schools (to open in April 2004, see infra) would be given identical standing for hiring purposes as a law civil service exam passer.63 While it remains unclear how these current and proposed changes might affect career decisions of candidates or hiring decisions of individual ministries and agencies, the direct threat posed by the bar for bureaucratic jobs is apparent.

2. Data from Elite Institutions

To further investigate the phenomena suggested by the national data, we obtained data from specific elite universities.64 Over 100 schools offer undergraduate degrees in law. But as suggested above, bar exam takers and passers are dominated by a handful of schools. In 2001, of the 38,930 students sitting for the bar, 19,726, or about half, were from seven schools: Waseda, Chuo, Tokyo, Keio, Meiji, Kyoto, and Doshisha universities.65 Many law departments have never graduated a single bar-passer, and some prestigious schools only have a handful of passers.6 Table 2 lists the ten schools with the most bar exam passers in each of 2001, 2000, and 1999.67

As the table shows, the top ten schools in 2001 accounted for 792 of the 990 passers, or 80%, and while the rankings may change, the schools included in the list are unchanged from year to year.

As Table 3 shows, the civil service exam data show a remarkably similar hierarchy.68 In 2001, the University of Tokyo accounted for 44% of passers, the top four schools accounted for 73%, only 18 schools had more than one passer, and only 30 had any passers.69 These patterns allow us to examine data on two specific schools with a high degree of confidence that they accurately represent a very large percentage of the population of legal elites.

First, we examine the data from the University of Tokyo (commonly referred to as Todai). Todai is widely viewed as the breeding ground for the elite of the elite, and with good reason. Historically, elites have come overwhelmingly from Todai, and admission to Todai, especially its law department, is a virtual guarantee of career success.70 Because Todai comprises such a large portion of the sample, we asked administrators in the Todai law department to provide us with a list of the number of Todai law students in every year for the past ten years who (a) passed the bar exam whether as a student or a graduate, (b) passed the bar exam while still a student (possible after two years of university), (c) matriculated to the LTRI upon graduation,71 (d) passed the law civil service exam while still a student, or (e) entered the bureaucracy upon graduation. The results, from an average graduating class of about 600 to 700,72 are in Table 4.

The data in Table 4 are consistent with the nationwide data. Although the trend is perhaps not as readily observable as in the national data, the number of bar exam passers who are Todai graduates has increased over time. The number of Todai students who pass the bar while still enrolled at the university has remained relatively steady over time, a reflection, perhaps, of the difficulty of that feat, or perhaps of recent changes in the grading system.73 As the LTRI number shows, with some allowance for yearly overlap, almost all passers attend the LTRI upon graduation. One recent survey taken by the Todai campus newspaper shows that the LTRI is now the number one destination of Todai graduates;74 another shows that more than two-thirds of Todai law students take the exam or plan to do so.75

If the decline in the popularity of a bureaucratic career were not clear from the national data,70 the Todai data offer more convincing evidence. Both the number of Todai students passing the civil service exam and the number of students choosing to join the bureaucracy upon graduation have significantly declined over time, and by a magnitude that greatly surpasses the overall national decline shown in Table 1. Combining data from Tables 1 and 4 shows that while 62% of passers in 1991 were Todai undergrade, in 1998 the figure was 34%, 37% in 1999, and 39% in 2000. Complementary data from various issues of Gakunai Koho, Todai's internal weekly newsletter, show a marked decline in the number of law students who upon graduation became "civil servants," including those who take local bureaucratic posts: in 1990, 176 of 619 students, in 1995, 182 of 705, and in 2001, 82 of 601.77 If, as is often said, Todai law students are the elite of the elite, the bureaucracy is no longer getting the most elite students in anywhere near the numbers it once did.

If hiring practices change over time to favor or disfavor a particular school such as Todai, the above data might be driven not by student choice but by hiring practices. Because there in fact is evidence of political pressure to decrease the number of Todai hires, we gathered additional hiring data. We were able to obtain passing and hiring data from the National Personnel Agency for the 1999, 2000, and 2001 tests on Todai and the second highest-passing school, Waseda University (a private institution).78 Those data are in Table 5.

As the table shows, pass rates remained relatively constant for both schools. The percentage of Todai passers (including current students and graduates) hired decreased slightly in the three-year period; but so did the percentage of Waseda passers hired. While again our sample is small and far from definitive,79 the limited available evidence suggests that occasional political pressures to decrease the number of Todai hires have had little effect, and that the decrease in interest in the bureaucracy that we see at Todai is likely supply-driven, and not demand-driven.

A similar story is seen at Kyoto University (Kyodai). Kyodai is another large national university, second only to Todai in prestige, and the first choice of many students from western and southern Japan. An organization called Hoyokai, analogous to a PTA and affiliated with the Kyoto University Law Department, publishes an annual bulletin in which it notes the employment plans of the previous spring's graduates. Data are compiled by the law department's administration. Table 6 lists the career choices for Kyodai legal elites for the period 1992-2002.80

Again, the table shows results similar to the national and Todai numbers. Those headed toward legal careers comprised only 13% of the class in 1992, but 42% of the class in 2002. The bureaucrat data are especially striking. In the past, as many as 27 students headed to elite bureaucrat positions. But in 2002, only eleven did; and in 2001, only nine. While caution is of course warranted when dealing with such a small number of observations, our discussions with Kyodai faculty lead us to suspect that it is not merely a fluke that recent measures are the lowest in the decade.

3. Double-Passers

Each year, a handful of over-achievers, perhaps about twenty, pass both the civil service test and the bar exam. Taking both tests involves risks. Although one can pass the bar examination with a "low" score with relatively little consequence, passing the civil service exam without a high ranking limits one's employment chances. Taking both exams in the same year is an especially risky strategy, as it lessens the time one can devote to each. A common strategy, comparatively less risky but still quite difficult, is to pass one test in one's third year of university, and the other in the fourth year.

Some double-passers pursue only one career option. For these persons, having "been taught to define their lives according to such crucial tests," the test score is a sign of worth that "looks good on resumes."81 But because of the time, energy, and money involved, many double passers pursue the two careers consecutively. We are told in interviews that after taking two tests, the idea of throwing one option away simply seems wasteful.

Double-passers thus have two primary options for pursuing both careers. First, they can receive legal training and enter the bureaucracy later. Although technically possible, this is actually rather difficult, as ministries historically have been reluctant to accept older candidates for new positions. Alternatively, an easier option is to join the bureaucracy and quit after two or three years to become a lawyer.

We failed to find substantial data on these so-called "double passers," as such data are simply not collected systematically by any organization. Bureaucrats keep no formal count of employees who have passed the bar, and legal organizations keep no formal count of legal professionals who have passed the civil service exam. However, we did obtain some information on double passers from law firms.

With nearly 150 lawyers, Nagashima, Ohno & Tsunematsu is one of Japan's largest and most prestigious law firms. According to its Martindale-Hubbell directory entry, four of its lawyers, one partner and three associates, are former elite bureaucrats as we have defined the term. The partner, a 1986 University of Tokyo graduate, joined the Ministry of Foreign Affairs in 1986, left in 1990 to become a lawyer, and was admitted to the bar in 1993. The three associates, one from Waseda and two from the University of Tokyo, joined the firm after stints at the Ministry of Finance and MITI.

Importantly, each of these lawyers, like their counterparts at other large Japanese firms, is relatively young. The partner, born in 1963, was admitted in 1993. The associates were born in 1967, 1970, and 1971, and were admitted to the bar in 1995, 1995, and 2000. The youth of these lawyers reflects the newness of the trend among double-passers. According to our interview sources, double-passers ten years ago would have remained career bureaucrats, or would have left the bureaucracy to become a lawyer only in unusual circumstances. Now, though small in number thus far, young double-passers are comparatively more likely to choose law as their lifelong profession. They seem to join only the largest and most elite firms (which is reflected in our lack of more widespread data), but they do become lawyers. Both senior attorneys and young double-passers to whom we spoke expect the trend to continue.

4. Quasi-Lawyers

Finally, we obtained data on quasi-legal professionals. In Japan, because of the small number of lawyers, a variety of other professionals with training in law fill the gaps. Although these quasi-lawyers are not legal elites by our definition, their employment trends may shed light on changes in the market for legal elites.

We have data on two groups of quasi-lawyers. First, "judicial scriveners" (shiho shoshi) are licensed specialists in real estate and corporate registration and documents submitted to courts. Scriveners are required to take a difficult government-issued test to be licensed; pass rates hover around three to four percent. Although scriveners come from a very different pool than elite lawyers, some potential candidates may be affected by similar labor market factors. In fact, we find that the number of scrivener test-takers has increased dramatically, from about 16,000 takers in the 1980s to 21,475 in 1998, 21,839 in 1999, 21,475 in 2000, and 23,190 in 2001, while the pass rate remains below three percent.82

A second group of quasi-lawyers on which we have data are employees of legal departments (homubu) of large corporations. While we lack definitive data on the supply of legal elites into these jobs, we have some data on industry demand. Candidates for these positions are primarily hired directly out of college through interviews and not standardized tests (although some students take the multiple choice portion of the bar exam as an additional credential with which to impress recruiters). The most elite students generally do not take corporate jobs; our interviews strongly suggest that they choose between the bar and the bureaucracy. But at least in the case of prestigious multinational corporations, some candidates come from the same pool of students as lawyers and bureaucrats.

Recent survey data show that the average number of employees in the legal department per corporation is increasing, from 5.2 in 1990, to 6.1 in 1995, to 6.4 in 2000.83 The same survey shows that the total number of legal employees hired by firms mid-career is increasing, from 181 persons at 115 of 888 surveyed firms in 1995 to 491 persons at 290 of 1,008 firms in 2000.84 Another survey showed that in the next five years, 47% of firms planned to increase the size of their legal staff, while only 4.2% planned to shrink it.85

While not definitive, the evidence on these two groups of quasi-lawyers shows remarkable consistency with the evidence on lawyers, and contrasts with the evidence on elite bureaucrats. This combination of trends suggests that the career decisions of legal elites are not idiosyncratic, but instead are made at least partially in response to more fundamental changes in the Japanese economy and society.

C. Salary and Employment Data

In this section, we analyze salary and employment data for the two professions. We find that, in general, compensation for elite lawyers has increased, while lifetime compensation packages for elite bureaucrats are declining. Competition for elite lawyer jobs is increasing, while competition for elite bureaucrat jobs is decreasing.

1. Elite lawyers

Very little information is publicly available on elite lawyer salaries. At least in part because of the relatively small size of the organized bar, there is no Japanese equivalent of the American Lawyer magazine to publish salaries.85 The closest substitute is the decennial survey of attorneys, recently named the Bengoshi [Attorney] Census, conducted by the Japan Federation of Bar Associations. The most recent, conducted in 2000, was sent to all 17,416 then-members of the bar. Only 5,560 responded, yielding a response rate of about 33%. The survey found an average annual income of 15,031,233 yen, or about $140,000. Seventy-seven attorneys, or 1.4%, reported income in excess of 70 million yen ($580,000). These, of course, are the partners at the largest law firms; in fact, only 124 responding attorneys were in firms of more than 20 lawyers, and only 64 in firms of more than 50 lawyers. Overwhelmingly (88%), respondents were from offices of five or fewer attorneys.87

To better explore compensation arrangements, we attempted to collect additional data directly from elite attorneys. Because lawyers are reluctant to discuss salary data publicly, we obtained most of our salary data through sources who asked that the identity of his or her firm remain confidential.

According to the hiring partner at one large, elite Tokyo firm (firm L), the salary range for first-year associates at major law firms is 10 to 16 million yen, or about $85,000 to $130,000. Some firms adopt fixed salaries for junior associates, while others adopt an hourly salary system that reflects that associate's billable hours. Other firms, including firm L, adopt a fixed salary for junior associates and an hourly bonus system for midlevel and senior associates. Some firms also include a bonus component based on performance.

At a smaller elite Tokyo firm (firm S), we were told that first-year associates receive a flat 10 million yen ($85,000) salary. Second-year associates receive a base salary of less than 10 million, but they receive extra payments based on billable hours and performance. Because first-year associates are required to handle many tasks that do not result in billable hours, such as recruiting and organization of parties, the flat rate is more attractive to them.

We received additional information from a Tokyo legal recruiting service that recruits Japanese lawyers for joint ventures with U.S. firms.88 Recommended salary packages for partner-level lawyers with ten years of experience are 60 to 80 million yen ($500,000 to $650,000), 20 to 35 million yen ($165,000 - $290,000) for associates with 6 years of experience and an overseas post, 10 million to 16 million ($85,000 to $130,000) for associates with 2 to 4 years of experience and an overseas post, and 7 to 10 million ($60,000 to $85,000) for new associates with less than 2 years of experience. For all packages, bonuses are negotiable.

Tokyo salaries have increased slightly in the last decade despite the dramatic increase in supply of young attorneys. A former associate of firm L tells us that starting salaries in 1990 were 8 million yen ($66,000). One relatively large firm that specializes in securities work offered 10.5 million ($90,000) to new associates in 1998, a sum that was seen as above-market at the time but now is the norm. We are informed that until 2000, firm S, which now pays 10 million to first-year associates, paid first-years 8 million, with the possibility of bonuses up to a total of 9 million.

We have limited additional data on elite lawyers outside of Tokyo. There are very few, and virtually all of them are located in Osaka. These Osaka elite lawyers work on a combination of international and domestic deals. Starting salaries for elite Osaka associates in 2002, we are told, began at 7.75 million yen for first-year associates, and were expected to increase to 10.5 million by the fourth year.89 While these figures represent slight increases over salaries of the 1990s, they apparently are not representative of Osaka lawyers who do purely domestic work. For this group, the increased number of new lawyers recently has caused some firms to lower starting salaries after several increases in the 1990s. But elite Osaka salaries remain high, we are told, in order to compete with Tokyo firms.

At the large firms in both Osaka and Tokyo, competition for elite jobs is increasing. Hiring partners whom we interviewed told us that they are seeing "larger pools of better and better candidates each year." Although some opponents of bar expansion had voiced concerns that expansion might lower quality, according to Tokyo hiring partners, the number of qualified attorneys is actually increasing, at least in part because would-be bureaucrats are becoming lawyers.

2. Elite bureaucrat salaries

Unlike lawyer salaries, bureaucrat salary data are publicly available. But a large portion of an elite bureaucrat's compensation comes not directly as a government salary, but in a post-retirement private post. We thus present data on both elements of bureaucratic compensation.

In 2000, the base starting salary for an entry-level bureaucrat in Tokyo was 206,304 yen per month for an administrative position, and 221,648 for a research position. Each works out to an annual salary of approximately $20,000, or approximately one-third to one-eighth of a junior associate's compensation. Extra payments are possible; those employees with spouses, for instance, are given an additional 16,000 yen per month. Employees are given up to 27,000 yen for rent per month, up to 50,000 yen per month for public transportation, and a bonus equal to 4.75 times the monthly salary. If a young bureaucrat received all of these bonuses, her annual salary would still be less than $40,000 annually, or no more than half of what she could earn as an elite lawyer.90

Salary increases with seniority. A vice-minister (of which there are now 36) of a Ministry, an "11th rank" officer, earns a monthly salary of 1,346,000 yen, or an annual salary of about $150,000.91 Below vice-minister are ten ranks of officers, the largest class of which, with 581 bureaucrats, is the "6th rank." These officials, which include division and department heads, earn a monthly salary of 937,000 yen, or about $100,000 annually. The retirement bonus for career bureaucrats is approximately thirty-five times monthly salary. This works out to approximately $800,000 in the case of a senior vice-minister; for a section chief, approximately $300,000.92

These figures still pale in comparison to the compensation of elite lawyers. For a career bureaucrat, however, the lifetime compensation scheme is nevertheless attractive, even in comparison to lawyers' income, because of amakudari ("descending from heaven"). As Mark Ramseyer and Frances Rosenbluth explain:

Bureaucrats collect a large part of their potential lifetime earnings only if they successfully "retire" with their reputations intact. Between the ages of 40 and 55, elite bureaucrats leave the ministries to take positions located for them by their ministry, either in private firms that deal with the ministry or in one of a variety of government-affiliated firms. Effectively, they exit their ministries through a one-way, one-time revolving door and take high-paying senior jobs at public or private firms . . . .93

The additional income from amakudari can be large: sources suggest $200,000 to $300,000 annually, not including the additional rewards of office space, company cars, boondoggle trips, and multiple retirement bonuses.94

This is nice work if you can get it, but anecdotal evidence suggests that the practice of amakudari has declined significantly in recent years:

Even officials at the powerful Ministry of Finance have virtually no chance of securing an important executive position at such prestigious banks as Mitsubishi, Sumitomo, Dai-Ichi Kangyo, or the Industrial Bank of Japan. . . . Amakudari prospects for MITI bureaucrats are becoming increasingly limited to electric power companies, steel manufacturers, and consumer electronics companies, all of which are under MITI's strong control.95

We examined several quantitative data sources to attempt to determine the validity of the anecdotal data. The official source is a report of the National Personnel Authority, known colloquially as the White Paper on Amakudari.96 Article 103 of the Civil Servant Law97 stipulates that a public official may not accept a high-level private post for two years after retirement with any firm with which he has been connected in the past five years. Any official who wishes to accept such a post can only do so with the permission of the National Personnel Authority in the form of a waiver.98 Figure 4 represents the number of waivers given by the Authority from 1988 to 2001.99

The data show a distinct and substantial decrease over time, but their significance is open to a variety of interpretations. Officials might be increasingly retiring to positions that do not require a waiver, or waiting more than two years to avoid requesting the waiver. The slight increase from 40 in 2000 to 69 in 2001 is said to be the result of (a) government reorganization and (b) decisions by Ministry of Finance and other officials to postpone their waiver requests while in the midst of a spate of scandals in 2000.100 Such manipulability makes the data difficult to interpret.

Fortunately, three other sources of data are available. First, Teikoku Databank conducts an annual survey of amakudari by public officials to banks, particularly by Bank of Japan and Ministry of Finance officials. According to their September 2001 report, 127 bank officers and directors are the result of amakudari, a figure down 7.3% from the previous year. They also report a steep decline in the number of amakudari officials who have the power to represent the corporation, declining by 38.6% to 49 officials.101

Second, in a 1995 article on amakudari,102 Ulrike Schaede attempted to measure the proliferation of amakudari by counting the number of directors at Japan's 100 largest companies who had previously held high public posts.103 She found that as late as 1991, 177 of 3,605 directors, or 4.9% of all directors at the 100 largest firms, were former public officials. Two-thirds, or 67 of the 100 firms, employed such directors.104

To compare 2001 data with Schaede's results, we conducted a similar test. We chose the 100 largest companies from the Financial Times 500 and again gathered data from the latest (2001) edition of Kaisha Nenkan.105 We found, first, that the number of board members has drastically decreased in the last decade, from 3,605 in Schaede's study to 1,111 in ours. Next, we found that only six of the one hundred firms had directors who were former bureaucrats, and those six firms had only one such director each. Tellingly, one of those directors was a former public prosecutor from the Ministry of Justice, presumably appointed for compliance and control reasons.

While these data strongly suggest a decline in amakudari, they are by no means conclusive, as bureaucrats might be retiring to firms with which they did not work directly as bureaucrats (thereby eliminating the waiver requirement), non-banks, or smaller firms. To explore these issues further, we hand-collected data from Kigyo Keiretsu Soran (Databook on Corporate Cross Shareholding and Affiliations), published by Toyo Keizai. This exhaustive publication lists the number and background of all directors and officers for all listed companies, 1,985 in 1990 (data as of August 1, 1989), and 2,430 in 2000 (data as of October 1, 1999). We found that from 1990 to 2000, the number of such directors and officers fell nearly twenty percent, from 761 to 615. The decline is apparent in both the number of companies that have such directors and officers (397 in 1999 to 366 in 2000) and the average number of such positions at companies that hire ex-bureaucrats (1.92 in 1990 to 1.68 in 2000). Interestingly, the decrease came despite a 22% increase in the total number of companies and a 13% increase in gross domestic product, either of which might have been expected to increase the potential for amakudari. It thus appears that private amakudari has declined significantly in the last decade, with negative consequences for elite bureaucrat incomes.

Still, this evidence pertains to amakudari practices in private firms. A large number of amakudari officials retire to public corporations known as tokushu hojin, or "special corporations." As of December 2001, 77 such corporations existed.106 Officials have historically been free to accept such positions because the Civil Servant Law regulates only amakudari employment by private corporations.

The available evidence suggests that amakudari to these special corporations occurs regularly. A 1992 study found that 60% of the 784 managing directors of the 92 public corporations then in existence achieved their positions by amakudari107 In 2000, in response to public criticism, the government began releasing the names and positions of senior retirees. In that year, 78 of 538 retirees joined special corporations, a number comparable to the number of private amakudari waivers.108

While the practice of amakudari to special corporations continues, the Japanese government has adopted a policy of reducing the practice in recent years. In 1996, guidelines were passed to limit the percentage of amakudari board members at such corporations to fewer than one-third.109 In 2001, under the leadership of a newly formed bureaucratic reform agency,110 the Diet passed a law specifically designed to reform and limit special corporations.111 In addition, formal plans to eliminate 17 special and approved corporations and to privatize 45 more were announced.112 A full revision of the Civil Servant Law is scheduled for 2003.113 The negative attitudes toward special corporations accompanied by these concrete policy changes make reliance on amakudari to complete one's lifetime compensation package increasingly risky.

Such factors, we are told by personnel department officials at elite ministries, have caused a decrease in the quality of the pool of applicants for elite positions. We are told by these officers that the very best and brightest, perhaps the top ten candidates in a given year, continue to join the bureaucracy in relatively high numbers, and in so doing seek "prestige and affirmation of their intelligence." But the same officials lament that a decade ago, the top one hundred would likely have been "sure hires." Beyond the top ten, quality declines. This phenomenon, officials tell us with no small measure of concern, is the result of a loss of candidates to the bar.

D. Summary

The data presented above suggest a pattern. The number of persons who take the bar, pass it, and become lawyers is increasing, while the average number of persons who sit for the civil service exam has fallen in the last five-year period in spite of increasing pass rates and relatively constant hire rates. At elite schools, the same trends apply; an increasing number of legal elites are choosing the bar over the bureaucracy. Quasi-lawyer numbers are also increasing. We have less evidence on double-passers, but at least some appear to be choosing the bar in cases in which they would not in the past. Bar compensation has increased; lifetime bureaucrat compensation is declining with the demise of amakudari practices. Though difficult to measure precisely, the mean quality of entry-level candidates to the bureaucracy appears to be declining, while the bar is admitting expanding numbers of highly talented youth.

From this evidence, we conclude that the bar is receiving an increasingly large share of legal elites, and increasingly higher compensation, to the detriment of the bureaucracy. We have considerable confidence in this assessment, given the quantitative data, our interviews, reforms enacted both to increase the number of elite bureaucrats and to allow the bureaucracy to hire lawyers, and anxious public statements by those in charge of bureaucratic reform that the bar is taking the "best candidates." But at least two other explanations might be possible.

First, perhaps the pool of candidates for the bar differs from the pool of candidates for the bureaucracy in some meaningful way. If so, the increase in the bar and the decrease in the bureaucracy might not be related. But this is simply not the case. It is common knowledge that a small group of the top undergraduate law students spend considerable time and energy determining which of these two career paths to take, and some, content on keeping all options open, take the tests for both professions. Our interviews with those in charge of hiring reinforce the conclusion that elite bureaucrats and lawyers are drawn from the same talent pool.

Second, perhaps test numbers are not an accurate reflection of career decisions. True, some legal elites take tests for resume value, or simply to prove their intellectual acumen. But our data reflect post-testing employment decisions as well. Interviews with those in charge of hiring at large firms and elite bureaucracies confirm these patterns. Law firms (and U.S. law schools with graduate degree programs that accept Japanese students) are deluged with young attorney resumes that improve each year, while ministry officials tell us that the number and quality of entry-level candidates continues to decline.

IV. INSTITUTIONAL CHANGE AND THE RE-ALLOCATION OF TALENT

What accounts for these developments? Professional career decisions are complex,114 but the changes in employment decisions of Japanese legal elites documented above correlate temporally with significant structural changes in Japan, many of which were explicitly designed to move the locus of governance from the bureaucracy to the legal system.115

A. Institutional Change

The change in employment patterns among Japanese legal elites reflects a basic change in the rules governing the Japanese economy.116 Of course, law is often responsive to, as well as the vanguard of, social and economic change. These new rules were prompted by a tangible shift in public attitudes regarding the place of law and bureaucratic oversight in the Japanese economy. This shift, which may be largely attributable to the economic debacle of the 1990s (especially the Ministry of Finance's inability to solve Japan's bad-loan problems), was expressed in a variety of ways, including election outcomes and the media. Yet a simple shift in attitudes without accompanying changes in incentive structures would not likely lead to the phenomena described above.

While we have separated several strands of institutional change for purposes of analysis and clarity of exposition, we do not claim to have a complete understanding of how these changes rank in terms of importance, or the exact chain of causation through which the labor market is evolving. Nor is our claim that every young legal elite in Japan explicitly considers these developments when deciding whether to become a lawyer or bureaucrat.117 Some might, but most simply react to their environment, acting, like Friedman's billiards player, "as if" they had formally calculated their responses.118 Our aim here is simply to describe several salient changes in that environment.

Over the past decade and at an accelerating pace, the rules of the game for economic lawmaking and enforcement have undergone a palpable change, corresponding to social and political sentiment that Japan needs to move from a system of bureaucrat-orchestrated economic management to a more market-oriented system based on principles of transparency and individual responsibility. Concrete manifestations of the shift are apparent in numerous areas.

One development is widespread deregulation in fields such as finance and telecommunications, designed to reduce government intervention in the economy and society.119 The most prominent examples are in the areas most relevant to business lawyers, including mergers and acquisitions and financial products.120 In addition, code or statutory reforms have increased organizational flexibility for corporations and banks, while expanding the menu of transactional options. Simultaneously, these new business opportunities and structures are generating novel legal risks, such as shareholder litigation and internal compliance issues. Thus, deregulation gives elite lawyers a role in transactional and advisory work they lacked under a regime replete with mandatory rules and advance governmental approvals. For example, as a prominent business lawyer notes, while banks once turned to informal regulatory guidance for "insurance," they now seek legal opinions.121

Fortifying the deregulation movement is a series of new statutes designed to circumscribe bureaucratic discretion and increase monitoring of agency action. For example, an Administrative Procedure Act formalizes rulemaking procedures and restricts the use of informal "administrative guidance,"122 a Freedom of Information Act compels disclosure of certain information held by national agencies upon request,123 and new ethics rules regulate the use of nonpublic information, the receipt of gifts, and ex parte contacts by public officials.124 In addition to these measures, some ministries are seeking to enhance self-monitoring by adopting a system in which subordinates evaluate their superiors.125 Even if, as is highly likely, each of these legal measures imperfectly accomplishes its objectives, the plain purpose and plausible cumulative effect of these new rules is to reduce bureaucratic autonomy, enhance political and private sector monitoring of agency action, and limit the informal perks of government service. It would not be surprising if this new legal environment diminished the attractiveness of a bureaucratic career. Indeed, high-level officials indicated to us that they are frustrated by the erosion of their policy autonomy over time.

At the same time the bureaucracy has been publicly criticized and subjected to enhanced monitoring, the legal system and its practitioners have received favorable attention as a vital outlet for the protection of citizens' and business interests and a critical counterweight to the declining role of agency oversight. In 1997, an administrative reform council whose report led to the restructuring of the Japanese central government ministries, stated that the " 'rule of law' constitutes an essential base for promoting deregulation, aiming at abolishing unclear advance administrative control and converting to an after-the-fact review/remedy type society."126

Even some of the most conservative sectors of Japanese society have become law optimists-indeed, they have been in the vanguard of the law expansion movement. Not coincidentally, both the Keidanren, a powerful big business lobbying group, and the ruling Liberal Democratic Party issued reports in the late 1990s strongly advocating the strengthening and expansion of the legal system.127 Of course, the change in heart did not occur in a political vacuum; shifts in public opinion prompted elected politicians to respond by reducing the scope of bureaucratic discretion. As one knowledgeable commentator notes, "it is clear that business groups and, to a lesser extent, LDP politicians, are indicating that people need more access to the law. It appears that they are trying to create a new orthodoxy in Japanese society."128

These proposals led to the appointment by the Prime Minister of a Judicial System Reform Council, which spent two years formulating its recommendations. The Council's final report in 2001, "For a Justice System to Support Japan in the 21st Century," states that the objective of the Council was to "define clearly 'what we must do to transform both the spirit of the law and the rule of law into the flesh and blood of this country . . . .' "129 The Council's recommendations are sweeping. A partial list includes taking measures to make trials more efficient and to make better use of expert testimony, expanding access to the courts by lowering filing fees and reinforcing the legal aid system, increasing the size of the legal profession and reforming legal education through the introduction of U.S.-style graduate law schools, and diversifying the judiciary. Concrete steps have already been taken to implement some of the Council's recommendations, particularly as they relate to reform of the legal profession and legal education.130

Developments in the courts further reflect and reinforce the movement from bureaucratic discretion to law. The best example is a recent shareholder derivative suit known as the Daiwa case.131 In 2000, the Osaka District Court rendered a $775 million judgment against the directors of Daiwa Bank for breach of duty under the Commercial Code. The directors were found liable for failing to institute a compliance system to detect unauthorized trades in the bank's New York branch and for failing to make timely disclosure of the trading losses to U.S. banking authorities, even though the directors were operating at the suggestion or acquiescence of the Ministry of Finance in delaying disclosure. It is hard to overstate the significance of this case, which, in substantive legal terms and shock value to the business community, has the combined impact of two well known Delaware cases, Caremark and Van Gorkom. Though ultimately settled during the appeal process for a fraction of the initial award, the District Court's theory of liability was preserved and entered as a final judgment against the directors.132

Most striking about the case is the potential shift in the balance of corporate lawmaking and enforcement power it represents. The court's finding of enormous personal liability against the directors and the state-of-the art theory on which that finding is based-a decision consistent with, if broader in some respects than current U.S. corporate fiduciary duty law-both heightens the need for sophisticated preventive law advice in the Japanese corporate setting and belies the longstanding view of the Japanese judiciary as too cautious and detached from commercial affairs to play a useful role in the resolution of complex business disputes. Finally, the court's summary rejection of a key strand of the defendants' defense-that the board was acting at the behest of a powerful ministry-is John Marshall-esque in its implicit declaration that the courls, not the bureaucracy, are the final arbiters of the law in Japan. In this last respect, the case has few parallels in existing Japanese case law.133 The specter of losing the shield of bureaucratic protection and being exposed to massive personal liability for corporate governance shortcomings had a novel impact on the business and legal communities.

V. INTERPRETATION OF FINDINGS

Together, these developments provide a compelling explanation for the new employment trends among legal elites. Quite simply, elite legal talent is migrating to its highest-valued-added uses. In what may be the beginning of broad, long-term structural transformation, the highest-value-added uses are shifting from the generalist bureaucracy to an expert-oriented legal system.

This phenomenon has two important implications for academic literature.134 First, theories about how Japan is governed need to be revised and updated. The view of Japan as a "plan rational" state orchestrated by bureaucrats may always have been exaggerated in reality, if not perception. Whatever its past validity, our data cast considerable doubt on its existence today. Even theorists who have long posited the existence of a more pluralist and accountable Japanese polity will now have to account for law, courts, and lawyers as more central and vibrant forces in the Japanese political economy.

Second, analyzing the role of lawyers in promoting or hindering economic growth is far more complex than the most prominent of the previous studies suggest. Numerous commentators, with a supportive nod to Japan, have suggested that societies encourage their most talented youth either to pursue innovation in "productive careers" (and reap economic gains) or rent seeking in law (and invite stagnation). But simple measures of the population of legal professionals, without an analysis of the institutional context in which they function and the alternative occupations available to talented youth, may be highly misleading. Nothing in Japan's experience suggests that the allocation of talent into the bureaucracy or law independently determines economic performance; rather, the experience indicates that talent follows power and profit, both of which are determined by a complex amalgam of institutional factors and social sentiment.

Since we anticipate several specific objections to these conclusions, we respond to them in some detail here. In the process, we flesh out and confirm several strands of our argument.

1. Scandal

In recent years, the Japanese bureaucracy, and particularly the super-elite Ministry of Finance, has been plagued by scandal. It might be argued that the observed employment shift results from a temporary public reaction, as legal elites avoid employment in a tainted sector, and not a long-term shift in power.

Scandal, however, is rarely exogenous to prevailing power structures. Inappropriate or unlawful conduct is often revealed only after the collapse of the regime in which it occurred without detection or punishment, as a new regime pursues a new agenda. Many of the corrupt practices in the Japanese bureaucracy were unearthed over the past five years by public prosecutors, at the same time the institutional changes described above took root. From this perspective, heightened public perception of scandal is not simply noise in the data, but another sign of increased reliance on the rule of law in Japan.

2. Other incentives to choose law

Perhaps some other factor besides institutional change is motivating the move of high-level legal elites to law. We see at least three possibilities: salary increases, job security, and internationalization of the economy.

It may well be that legal elites are moving from the bureaucracy to law because of the increased lifetime compensation differential; elite lawyers now get paid more relative to bureaucrats than they did a decade ago, and they are less likely to be plagued by scandal or downsizing that could threaten job security. But as with scandal, compensation is not exogenous to institutions. While it may be that legal elites are responding most directly to changes in relative compensation levels, those changes are a reflection of deeper structural shifts, several of which are discussed above.

In a related manner, perhaps legal elites arc joining the large, international Japanese law firms (as opposed to smaller, primarily domestic ones) not because of institutions, but because of internationalization. Due to foreign pressure, economic distress, and other factors, foreign investment in Japan has increased, raising the market for international legal services. Again, however, these developments did not occur in a vacuum. While "globalization" has undoubtedly played some role in the shift, underlying this abstract phenomenon are specific legal changes making international transactions such as cross-border mergers and acquisitions more feasible, which in turn increase the compensation of lawyers who handle such transactions. Our interviews, as well as those of prominent practitioners, show that the fastest-growing source of revenue at the elite firms is domestic work, which now accounts for as much as two-thirds of all revenues.135 Nor can "globalization" explain the apparent decline in attractiveness of careers at internationally-oriented ministries such as METI and the Ministry of Foreign Affairs.

3. Decline in court budgets

Contrary to our claim that the role of law is expanding, the percentage of the annual budget allotted to the court system has declined over time. The court system has never received more than one percent of the national budget. In 1955, its share was .91%. It has fallen ever since, to an all-time low of .36% in 1999.136 A skeptic might take this as a sign of legal system decline.

Investing these data with great significance, however, would be a mistake. First, Japanese judges are compensated as well or better than their U.S. federal counterparts, suggesting that the profession is held in high regard.137 second, the decrease is explained in part by the substantial increase in the economy and annual budget over the past half-century.138 Finally, the judiciary historically has been very weak in promoting its own case for additional expenditures. While the modest budgetary allocation to the courts may suggest a lack of political sway in fiscal affairs, it is hardly the most reliable indicator of the courts' role in society.

4. Dual rent-seeking

If both lawyers and bureaucrats are principally engaged in rent seeking or rent extraction, the change in employment patterns we have documented is insignificant, since the aggregate amount of talent diverted from productive uses in Japan remains relatively constant.

This is a plausible interpretation of our findings, but it is hard to reconcile with other facts. For example, most Japanese bureaucrats, while trained in law, actually do not perform legal work; lawyers obviously do. Therefore, the movement toward the bar reflects a major shift in the substantive work performed by one of the most elite and highly talented segments of the Japanese work force. It seems unlikely that this change is of no import. It also suggests that if both professions are engaged in rent seeking or rent extracting, they go about this task in very different ways.

Moreover, treating both professions as rent seeking is inconsistent with the premise underlying the talent diversion theory itself, since rapid growth coincided with extensive bureaucratic involvement in the economy in Japan as well as in other high-growth economies in Asia that figured so heavily in the empirical confirmation of this theory. 139

5. Inverse relation

Some might claim that our findings support the work of the law skeptics by showing a continuing inverse relation between lawyer population and economic growth in Japan. When lawyer population was low in the 1980s, the economy boomed. As the number of legal professionals increased throughout the 1990s, the economy stagnated - because, a skeptic might argue, lawyers' redistributive activities inhibited growth.

This interpretation of the data, however, is highly implausible. Lawyer population in Japan has been low on an international comparative scale since the formal establishment of the profession in the nineteenth century.140 Yet Japan's most dramatic economic growth did not occur until after World War II. Moreover, the lawyer population increased significantly in percentage terms from 1940 to 1985, coinciding with a remarkable period of economic growth in Japan. Similarly, it is implausible that the incremental increases of about 200 members per year to the Japanese bar that occurred in the 1990s "caused" the negative economic growth in that period.

In fact, while caution is of course warranted in dealing with a small sample, the Japanese data suggest no relation between lawyer population and economic growth. Figure 5 shows the number of lawyers, number of bar takers, and real GDP growth from 1960 to 2000.141 As the figure shows, the number of bar takers increases at a much higher rate than the number of lawyers, a result of the artificial numerical controls on the size of the bar. The figure also shows high growth in the 1960s and contemporaneous low numbers of lawyers and bar takers. But over the sustained period neither measure has any obvious relation with economic performance.

We found similarly ambiguous results when we used units of analysis (similar to those of Murphy et al.) that attempt to measure talent allocation among a broader population at an earlier stage in career development. We found that the percentage of Japanese university students majoring in law has declined over time, from 9.99% in 1960 to 8.18% in 1999, and that the ratio of law majors to engineering majors also fell, from about .65 in 1960 to about .42 in 1999. Again, while the short time period may make the data suspect, neither trend has any clear relation to macroeconomic indices.

Our detailed examination of "Exhibit A" suggests that no simple relation exists between the allocation of talent and economic growth. Recent calls from the Japanese business sector for more law and legal professionals (not more engineers) belie the notion that lawyers are destructive to economic growth. Moreover, the Japanese experience suggests that the alternative occupations available to erstwhile attorneys may not be the entrepreneurs and engineers that economists assume in their studies. Ultimately, the Japanese experience suggests that the allocation of talent in society follows from, rather than determines, the rules of the game for economic and social activity. We suspect that detailed analyses of other countries would reveal similarly complex and unquantifiable issues lurking behind the quantitative cross-country studies.

VI. CONCLUSION

"It is really upsetting," one law professor at a prestigious Japanese university told us, "how our graduates are not going to the bureaucracy anymore. They simply are not interested in the positions of power." We appreciate the sentiment, but in this Article, we have argued from the data that legal elites are still pursuing positions of power - it's simply that the nature of those positions has changed. A transformation in economic and political institutions is bestowing upon the bar some of the authority that once resided in the bureaucracy. Japan's experience over the past decade, while only representing a single country, provides strong support for the view that a lawyer's value added is heavily dependent upon the structure of the economy and politics as a whole. As Japan moved from an informal, ex ante model of bureaucratic management toward a more flexible, ex post, and law-based model of governance, power, prestige, and profit have shifted from the top-tier civil service to elite legal practitioners.

Japan's experience, of course, may not necessarily be representative of other national systems. But given the unique position of Japan in the debate over lawyers and economic growth, commentators would do well to understand the implications for growth and governance of that country's allocation of talent. Close analysis of Japan's experience over the past several decades suggests that the engagement of legal and bureaucratic actors in the economy was in considerable tension, though neither regime is necessarily inconsistent with economic growth. Rather, the allocation of talent in Japanese society appears to follow from, rather than determine, institutional design and performance. It seems unremarkable that in seeking to improve the quality of its legal institutions to enhance economic performance, Japan would simultaneously increase the population and stature of its legal professionals. Acknowledging that fact, however, may require revising the conventional wisdom both about who governs Japan and the role of lawyers in the economy.

* Fuyo Professor of Law and Director, Center for Japanese Legal Studies, Columbia Law School.

** Nippon Life Professor of Law and Director, Japanese Legal Studies Program, University of Michigan Law School.

Milhaupt's research was supported in part by a grant from the Abe Fellowship Program of the Social Science Research Council and the American Council of Learned Societies, with funds provided by the Japan Foundation Center for Global Partnership. West thanks the Nippon Life Insurance Company's endowment at the University of Michigan Law School. John Campbell, Masakazu Doi, John Haley, Atsushi Kinami, Shoxo Ota, Mark Ramseyer, and workshop participants at Columbia Law School, the University of Tokyo Faculty of Law, and the Research Institute of Economy, Trade & Industry provided comments and stimulating questions on earlier drafts. Thanks also to Noboru Kashiwagi, Hideaki Kubori, Setsuo Miyazawa, Michio Muramatsu, and Jeff Wexler for helpful conversations and aid in gathering data. We are especially grateful to those persons who shared information with us on the condition of anonymity.

CURTIS J. MILHAUPT* & MARK D. WEST**

Copyright Georgetown University Law Center Winter 2003
Provided by ProQuest Information and Learning Company. All rights Reserved