Chapter Four
Recruitment

Of all the variables affecting the performance of a government bureaucracy, none is as universally important as the caliber of its personnel. No matter how well endowed it may be in terms of funds, legislative mandate, and constituency support, a bureaucracy's operating effectiveness is likely to be severely curtailed unless it is staffed by competent and dedicated people. How to recruit and retain such people, then, becomes a task of the highest priority for nearly all bureaucracies.

In the case of the Japanese-government bureaucracy, however, a number of factors magnify further the importance of recruitment. First, the custom of lifetime employment underscores the imperative necessity of selecting the right persons initially. Second, the persistence of decentralized recruitment despite the creation of a central personnel agency, coupled with an extremely low degree of lateral mobility—that is, interagency transfers—helps to foster both competition and caution in the recruitment process. Third, the multiple-track system of recruitment, under which the future leaders of government ministries and agencies must be identified and selected at the outset, raises the stakes of initial recruitment. Moreover, the virtual absence of probationary appointment further enhances the importance of selectivity in recruitment.

In this chapter we shall first note a few general attributes of the Japanese civil service that may help us to understand its system of recruitment. We shall then scrutinize the modes of recruitment, with

particular attention to the higher civil-service examination. Next, we shall examine the patterns of recruitment. That is, what are the actual results obtained during the postwar period? Statistical data will be analyzed with the aim of ascertaining continuity and change in key areas. Finally, we shall compare the Japanese experience with those of four industrialized democracies in the West.

An Overview Of The Japanese Civil Service

To place the discussion of recruitment in perspective, it is necessary to note a few outstanding aspects of the Japanese civil service. The preceding chapter has already delineated its basic legal and institutional framework. The national public-service law spells out its basic standards, principles, and procedures. The law sets up the National Personnel Authority as one of the two "central personnel administrative organs" (chuo jinji gyosei kikan ).^[1] Until 1984, the Bureau of Personnel in the prime minister's office was the other organ. In July of that year, however, the Bureau of Personnel was transferred to the newly created Management and Coordination Agency (Somucho).^[2] In practice, the authority functions as. the principal personnel agency in the Japanese government.

At the policy making level, the authority consists of three commissioners (jinjikan ) who are appointed by the Cabinet for four years with the approval of both houses of the Diet. No two commissioners may belong to the same political party or be graduates of the same faculty of the same university. They may be reappointed but their maximum term of office is twelve years. One of the commissioners is designated by the cabinet as president (sosai ) of the authority. The commissioners may not be removed from office except through impeachment. The daily operations of the authority are in the hands of a staff of about seven hundred persons headed by the secretary general (jimu socho ).^[3]

[1] hapter 2, articles 3 through 25 of the national public-service law in Jinji-in, Ninmen kankei horeishu [Collection of Laws and Ordinances Concerning Appointments and Dismissals], 1984 ed. (Tokyo: Okurasho, Insatusukyoku, 1984), pp. 14-19.

[2] yosei Kanricho, Gyosei kanri no genjo: Gyosei kaikaku no doko , pp. 26-28. In the 1984 reorganization the Administrative Management Agency (Gyosei Kanricho) was abolished and most of its functions were taken over by the newly created Management and Coordination Agency.

[3] or details regarding the organization and functions of the National Personnel Authority, see Kato Hisabumi, Jinji-in: Nihon no shihai kiko [The National Personnel Authority: An Organization That Controls Japan] (Tokyo: Rodo Junposha, 1966); Kyoikusha, ed., Kaikei Kensa-in, Jinji-in, Naikaku Hoseikyoku [The Board of Audit, the National Personnel Authority, and the Cabinet Legislation Bureau] (Tokyo: Kyoikusha, 1979), pp. 69-116; and Sato Tatsuo, Kokka komuin seido [The National Public-Service System] (Tokyo: Gakuyo Shobo, 1975), pp. 13-16. See also the text of the National Public Service Law (kokka komuinho ) in Jinji-in, Ninmen kankei horeishu , 1984 ed., pp. 12-45.

In 1987, the president of the authority was Utsumi Hitoshi. Born in 1917, Utsumi graduated from the law faculty of the University of Tokyo in 1941 and entered the Home Ministry in the same year. He retired from the civil service after serving as the administrative vice-minister of the Defense Agency. He was appointed to his current position in 1984 by his former classmate in both Todai and the Home Ministry, Prime Minister Nakasone Yasuhiro. The other two commissioners were Sano Hiroyoshi (b. 1916) and Ishizaka Seiichi (b. 1922). A 1937 graduate of Aoyama Gakuin University who majored in English literature, Sano had worked as a journalist for Asahi shinbun and NHK and as a political commentator before being appointed to his present position in 1985. Ishizaka graduated from Todai's engineering faculty in 1944, where he also did graduate work, and served in the Ministry of International Trade and Industry (MITI) for twenty-five years. He was appointed a commissioner in 1986. Traditionally, one of the three commissioners has always been a former journalist and another has had a scientific or engineering background.^[4]

The basic principles guiding the Japanese civil service may be summed up in two words: democracy and efficiency. Democracy implies that unlike the Emperor's officials in prewar Japan, the public servants in postwar Japan must be true to their name: as article 15 of the 1946 constitution puts it, "all public officials are servants of the whole community and not of any group thereof."^[5]

To help attain this goal, the public service must be opened up to the entire people: the principle of open and equal access to civil-service appointments needs to be implemented. This does not mean that all citizens are guaranteed government jobs, which no government can afford to do, but that they are assured of an equal opportunity to compete for such jobs. This is where the other principle comes into play:

[4] aniai Kenzo, "Kankai jinmyaku chiri: Chuo shochohen: Jinji-in no maki" [Who's Who in Central-Government Ministries and Agencies: The National Personnel Authority], Kankai , Aug. 1987, pp. 39-41.

[5] or the English text of the Japanese constitution, see Suekawa Hiroshi, ed., Iwanami kihon roppo, Showa 49-nenban [Iwanami Six Fundamental Laws, 1974 Edition] (Tokyo: Iwanami Shoten, 1973), pp. 110-16.

that of efficiency. In other words, the guiding criterion of civil-service appointments shall be merit—the objective qualifications of candidates as demonstrated in their educational attainments, experience, and performance in examinations. The mechanism through which and the degree to which these twin principles are translated into reality will be the focus of our inquiry in this chapter.

Let us now attempt to clarify the bewildering array of categories in the Japanese civil service. As of 1 July 1987, there were a total of 4,506,725 public employees (komuin ) in Japan, of whom 1,172,797 belonged to the national government and the remainder to the various local governmental units. As figure 1 shows, the national public employees are divided into two broad categories: the general service (ippan shoku ) and the special service (tokubetsu shoku ). The latter, which comprised 328,507 employees in July 1987, encompasses such diverse positions as the prime minister, ministers of state, commissioners of the National Personnel Authority, deputy director general of the cabinet secretariat, director general of the cabinet legislation bureau, parliamentary vice-ministers, judges and other employees of the courts, members of the Diet and their secretaries, and employees of the Defense Agency.^[6]

The ippan shoku public employees are in turn divided into (1) those governed by the regular compensation law (kyuyoho shokuin ), (2) those governed by the special compensation law (kyuyo tokureiho shokuin ), and (3) public prosecutors. In July 1987 there were 505,791, 336,332, and 2,167 persons respectively in the three categories. Nearly 90 percent of public employees who are subject to the special compensation law are postal employees; the remainder are engaged in such specialized occupations as forestry, printing, and engraving.^[7]

Finally, the ippan shoku public employees who are subject to the regular compensation law are subdivided into seventeen categories, each of which carries its own salary schedule. They are (1) the Administrative Service I, (2) the Administrative Service II, (3) the Specialized Administrative Service (senmon gyosei shoku ), (4) the Taxation Service, (5) the Public Security Service I, (6) the Public Security Service II, (7) the Marine Service I, (8) the Marine Service II, (9) the Educational

[6] ee art. 2 of the national public-service law in Jinji-in, Ninmen kankei horeishu , 1984 ed., p. 13. All the statistics cited here are taken from Jinji-in, Komuin hakusho , 1988, esp. p. 251.

[7] bid., pp. 39 and 168.

Fig. 1.
Number of National Public Employees

Source: Jinji-in, Komuin hakusho, 1988 (Tokyo: Okurasho, Insatsukyoku, 1988), p. 251.

Service I, (10) the Educational Service II, (11) the Educational Service III, (12) the Educational Service IV, (13) the Research Service, (14) the Medical Service I, (15) the Medical Service II, (16) the Medical Service III, and (17) the Designated Service (shitei shoku ).^[8]

By far the single most important of these categories in terms of both size and relevance for this study is the Administrative Service I. In 1987 it consisted of 227,725 or 46.8 percent of all persons covered by the seventeen salary schedules. It includes most people whom one ordinarily associates with civil service. The Administrative Service II salary schedule, on the other hand, covers drivers of automobiles, operators of all mechanized equipment, guards, and other people who are engaged in similar activities. The Specialized Administrative Service, which was newly established in 1985, encompasses inspectors and judges of the Patent Agency, inspectors of ships, those who work in the air-traffic-control field, and others whose duties require specialized knowledge and skills. The Educational Service I schedule covers the faculty and staff of colleges and universities. With 52,619 persons under its rubric, it was the second-largest category in 1987. The schedule for the Designated Service, on the other hand, is one of the smallest, subsuming under its rubric such positions as administrative vice-minister, head of

[8] bid., p. 168.

an external bureau (gaikyoku )—an entity that is affiliated with but not an integral part of a ministry or an agency—bureau chief, dean of a college, head of a laboratory or a research institute, and director of a hospital or a sanatorium. In 1987 these positions numbered 1,357.^[9]

Modes Of Recruitment

Among the continuities between the prewar and postwar civil-service systems is the reliance on both competitive examinations and evaluation (senko ) in recruiting government officials. The national public-service law makes the use of competitive examinations the rule and permits the use of evaluation only in those cases where the National Personnel Authority has certified that the use of competitive examination would be inappropriate (article 36). Competitive examinations are administered by the National Personnel Authority. There are four general categories of exceptions: (1) top-level executive positions such as administrative vice-minister that call for managerial and supervisory abilities and competence to deal with external groups; (2) positions requiring licenses or certificates issued by the state, such as physicians, dentists, and nurses; (3) positions for which the candidates have already demonstrated their suitability by performing satisfactorily either in equivalent positions or in those requiring a greater degree of expertise and responsibility; and (4) special positions for which there is a scarcity of qualified candidates.^[10] It should be noted that evaluation does not preclude but frequently entails some kind of written examination.

Although the first civil-service examination for the recruitment of national public employees under the national public-service law was not held until January 1949, there were other examinations both prior to and following that date. In 1946 and 1947 a modified version of the higher examination (koto shiken ) was given three times.^[11] As noted in the preceding chapter, competitive examinations were also held in 1948 to recruit staff members of the temporary National Personnel Authority. Then in 1950 there were "S-1 examinations," which was a unique experiment designed to further the Occupation's policy of democratiza-

[9] bid.; "Sanko shiryo (kyuyo kankei)" [Reference Material (Related to Salary)], Jinji-in geppo 440 (Sept. 1987): 30.

[10] Jinji yogo haya wakari: senko [A Quick Guide to Personnel Terms: Evaluation], Jinji-in geppo 87(May 1958): 17. For a more technically oriented explanation, see "Senko ni yoru shokuin no saiyo" [Selection of Public Employees by Evaluation], ibid. 94 (Dec. 1958), 20-23.

[11] ada, "Bunkan nin'yo seido no rekishi," III, p. 13.

tion. Meanwhile, the regular civil-service examinations under the new law have been held every rear since 1949. Although they have undergone many changes over the years, their principal characteristics are nonetheless worth noting.

First, there are three distinct levels. The highest level (jokyu until 1984, isshu since 1985) is aimed at graduates of four-year colleges; the intermediate level (chukyu until 1984, nishu since 1985) at graduates of junior colleges; and the lower level (shokyu until 1984; sanshu since 1985) at high-school graduates. From 1960 to 1984 the jokyu examinations were divided into two types: A (koshu ) and B (otsushu ).^[12] During this period, then, there were in effect four levels of examinations, each of which led to appointment at a different grade. Of the four levels, the type-A higher examination was equivalent to the prewar higher civil-service examination, and the lower examination approximated the ordinary civil-service examination. The remaining two—the type-B higher exam and the intermediate exam—must be viewed as postwar innovations.

It is noteworthy that none of these examinations has ever had a formal educational requirement. The type-A higher civil-service examination, for example, was open to anyone between the ages of twenty-one and thirty-three regardless of educational attainments. In practice, however, the educational level of candidates taking the civil-service examinations has been quite high, reflecting the rising level of educational attainments in the general population. Between 15 and 16 percent of applicants for the type-A (or I) higher civil-service examination in recent years have had postgraduate education. The proportion increases to about 40 percent among the successful candidates. Moreover, the success rates of candidates with graduate education seem much higher than those of university graduates. For graduates of junior colleges and high schools, the probability of passing the higher examination seems negligible. Most striking of all, perhaps, are the statistics pertaining to the intermediate examination. Although geared to junior-college graduates, it has attracted growing numbers of applicants with four years of university education and more. By the 1980s, the latter accounted for

[12] he type-B higher civil-service examination was originally designed to serve as a supplemental means of recruiting scientists and engineers into the higher civil service. So far as administrative personnel were concerned, it was intended to be a gateway for future midlevel managers. However, until 1966 nearly all of those who passed the type-B examination in the administrative fields were those who had failed the type-A higher civil-service examination, which applicants were allowed to take in tandem. In 1967 this privilege was rescinded. Jinji-in, Jinji gyosei nijunen no ayumi , p. 123.

more than 90 percent of the successes in the examination. Even the lower examination has attracted a sizable number of university and junior-college graduates. These two groups together have accounted for about 17 percent of the successes in the lower examination in recent years.^[13]

It was partly in response to the changing character of the intermediate examination that the examination system was reorganized in 1985. The intermediate examination having effectively been turned into a vehicle for recruiting university graduates, there was a need to come to grips with that reality. Specifically, its prestige needed to be upgraded in order that the national government might compete successfully with local governments and private firms in the recruitment arena. Moreover, the demand for graduates of the type-B higher civil-service examination had steadily decreased over the years. On the other hand, the implementation of a mandatory retirement system beginning in 1985 would pave the way for generational change in the middle echelons of the national-government bureaucracy, making it necessary to increase the supply of capable middle-level administrators. Given these considerations, the National Personnel Authority decided to abolish the type-B higher examination, upgrade the intermediate examination, and rename all three examinations. The type-A higher examination became simply type I, the intermediate became type II, and the lower examination became type III.^[14]

A second noteworthy feature of the civil-service examinations pertains to specialization, for all candidates are required to choose a specific field of specialization. In the higher (type I) examination, candidates must choose one of the following twenty-eight fields: public administration, law, economics, psychology, education, sociology, mathematics, physics, geology, information engineering, electrical engineering, electronics and communications, mechanical engineering, civil engineering, architecture, chemistry, metallurgy, energy engineering, pharmacy, agriculture, agricultural economics, agricultural chemistry, agricultural engineering, animal husbandry, forestry, fishery science, erosion control (sabo ), and landscape design (zoen ).^[15]

[13] or relevant statistics, see Jinji-in, Komuin hakusho , 1985, p. 28; ibid., 1986, pp. 29-31; ibid., 1987, pp. 35-36.

[14] Saiyo shiken taikei no saihen seibi ni tsuite" [On the Reorganization and Adjustment of the Recruitment Examination System], Jinji-in geppo 411(Apr. 1985): 10-17.

[15] inji-in, Komuin hakusho , 1987, p. 29.

From 1949 to 1954, candidates for the higher examination opting for the field of public administration were allowed to choose an additional field of specialization. Now candidates for the type-II examination are given a choice of twelve fields: public administration, library science, physics, electrical engineering and electronics, mechanical engineering, civil engineering, architecture, chemistry, energy engineering, agriculture, agricultural civil engineering (nogyo doboku ), and forestry. All but three of these fields are also available for type-III-examination candidates. The exceptions are library science, physics, and energy engineering. Two fields that are unique to the type-III examination are postal affairs (yusei jimu ) and taxation.^[16]

Table 5 presents some statistics regarding the fields of specialization represented in the hiring of those who have passed the higher civil-service examination (HCSE) in the 1960-86 period. Note that this table pertains not to the successful applicants but to those who were actually hired by the various ministries and agencies. As we shall see below, only about half of the successful applicants receive appointments. The table shows that those choosing law and the social sciences as their fields of specialization in the HCSE, most of whom become administrative officials (jimukan ), have always been a minority. In fact, until 1972 they were eclipsed by those choosing engineering and the natural sciences. In subsequent years, although "law and the social sciences" has become the largest of the three categories, it has nonetheless remained well below the 50-percent mark, hovering around 40 percent.

A third feature of the examination system is that all examinations have multiple stages. The first stage in the higher examination consists of two sets of multiple-choice tests. The first set, known as kyoyo shiken (general-culture test), is made up of sixty questions designed to test general knowledge, aptitude, and intelligence; it lasts three hours. The second set, called senmon shiken (specialized test), contains the same number of questions and is designed to measure the level of expertise in the candidate's chosen field of specialization and lasts three and a half hours. Between 90 and 95 percent of candidates are eliminated in this initial phase.^[17]

[16] or the type-Ill examination, public administration is subdivided into administrative affairs (gyosei jimu ) A and B. See ibid., pp. 30-31 and "Saiyo shiken taikei no saihen," pp. 14-15.

[17] uken Shinpo Henshubu, Komuin shiken mondai to taisaku: Jokyu shiken, '85-nenban [Questions and Strategies for Civil-Service Examinations: The Higher Examination, 1985 Edition] (Tokyo: Hogaku Shoin, 1984), pp. 2-206.

In the second phase, candidates are again subjected to two sets of tests, one specialized and the other general. But unlike the previous tests, these consist of essay-type or problem-solving-type questions. Those opting for the field of public administration, for example, must choose three subjects out of five (constitutional law, political science, social policy, public administration, and principles of economics) and answer one essay question in each subject. They are given three hours in which to compose three essays. To give some examples of questions: "discuss equality and discrimination in democracy" (political science), "discuss the similarities and differences between public administration and private management" (public administration), "discuss measures for dealing with the aging of the work force" (social policy), and "discuss

the microeconomic foundations of the Keynesian macroeconomics" (economics).^[18]

The general test (sogo shiken ) in this phase typically requires the applicant to answer two questions in two and a half hours. Its aim is to test the candidate's ability to synthesize ideas and his judgment, analytic ability, and ability to think. A frequently used device is to present long excerpts from an article, often in English, and then ask the candidate to summarize and evaluate the main arguments.^[19] There is then a brief oral examination (jinbutsu shiken ) during which the candidate is questioned about his motives for choosing government service, his interests and hobbies, and other matters. About half of the candidates surviving the first round are eliminated in the second round of competition.^[20]

A fourth and final feature of the examination system that needs to be explained is the decentralization of hiring. Officially, recruitment activities should not commence until early October after the final results of the higher examination have been announced. The National Personnel Authority compiles rosters of successful candidates (saiyo kohosha meibo ) for each field of specialization; names are listed in the order of the scores earned. Upon receipt of requests from the ministries and agencies, the authority forwards lists of eligible candidates, whose numbers typically exceed the number of vacancies by a ratio of five to one. Actual hiring is supposed to begin in early November.^[21]

Unofficially, however, the process begins in early August immediately after the completion of the second phase of the higher examination. Even though half of them will be eliminated in the competition, the candidates nonetheless begin to pound the pavement in Kasumigaseki—a section of Tokyo's Chiyoda district where government offices are located—in a ritual known as kancho homon (visits to government

[18] omuin Shiken Joho Kenkyukai, ed. 88 nendo-han komuin saiyo shiken shirizu: Isshu kokka komuin shiken [1988 Edition, Civil-Service-Examination Series: Type-I National Public-Employee Examination] (Tokyo: Hitotsubashi Shoten, 1986), p. 46. All of these questions are from the tests intended for applicants taking the public-administration option.

[19] or examples of questions in the general test, which are too long to be translated here, see ibid., pp. 382-99.

[20] ashiro Ku, "Nihon gyoseikan kenkyu, III: gyoseikan no kyaria keisei katei (1)" [A Study of Japanese Public Administrators, Part III: The Making of Administrator's Career (1)], Kankai , Dec. 1981, pp. 69-71; Juken Shinpo Henshubu, ed., Watashi no totta kokka jokyu shiken toppaho, 86-nenban [How I Passed the National Higher (Civil-Service) Examination, 1986 Edition] (Tokyo: Hogaku Shoin, 1984). The last-mentioned source is a collection of essays by nineteen persons who passed the higher examination in 1984.

[21] Komuin shiken gokaku kara saiyo made" [From Passing the Civil-Service Examination to Being Hired], Jinji-in geppo 103 (Sept. 1959): 14-15.

agencies). It is widely believed that candidates from top-rated universities enjoy an edge, particularly if they have strong academic records and are likely to receive high scores in the examination. A participant-observer notes that grades received at Todai's Law Faculty carry more weight than scores in the higher examination, for, in his view, the former is a more reliable gauge of the candidate's intellectual ability than the latter. Inasmuch as the level of difficulty is about the same in both the higher examination and examinations taken in Todai's Law Faculty, he states, how one has performed over a two-and-a-half-year period ought to mean more than one's score in a single examination.^[22] Another observer notes, however, that one's rank in the higher examination is far more important than one's grades in college. He speculates that to have a good chance of being selected by the Ministry of Finance, which hires only twenty-five or so elite-track civil servants a year, one must rank among the top fifty successful candidates in HCSE.^[23]

Government ministries and agencies are known to seek out outstanding candidates from time to time, utilizing their kone (connections). Senpai (seniors) from the same university, especially from athletic or other clubs, and professors can play the role of match-maker. A candidate who had earned the third highest scores in the higher examination in the mid-1960s, for example, received a telephone call from a senpai , a former member of Todai's English Speaking Society. Over lunch, the senpai made a pitch for the Finance Ministry, which the candidate eventually chose over the Ministry of International Trade and Industry.^[24]

During the unofficial visits to government ministries and agencies, some candidates receive signals about possible appointment. Typically, a candidate visits a number of ministries and agencies and then narrows his choice to a few. After repeated visits to a ministry or agency, he may even receive a signal about a possible offer of appointment. Some candidates report having received an informal decision (nainaitei ) as early as late August—a full month before the final results of the HCSE are published and formal ministerial examinations (oral) begin.^[25]

As we shall see later in the chapter, there are signs that the gap

[22] ato, Kanryo -desu, yoroshiku , pp. 39-40.

[23] Kuribayashi, Okurasho shukeikyoku , p. 30. Kato served in the Home Ministry, whereas Kuribayashi is referring to the Finance Ministry.

[24] bid., pp. 19-22. In this case the recruitment activity took place after the final results of the higher civil-service examination were announced.

[25] uken Shinpo Henshubu, ed., Watashi no totta kokka jokyu shiken , pp. 120-21.

between Todai and the other universities is slowly being narrowed and that the disadvantages of private-university graduates are becoming increasingly less severe.

Table 6 presents some statistics about the higher civil-service examinations. It is plain that they are keenly competitive. As column D shows, only a small proportion of the applicants pass the examinations. What needs to be stressed is that the number of applicants is invariably greater than the number of those who actually take the examinations. During the years for which relevant statistics are available, about 18 percent of those who submitted their applications failed to show up. This means that the numbers in column D overstate the extent of competition during the years 1967 and 1973-1988. Nonetheless, during most years, over 95 percent of those who took the examinations failed. The number of applicants peaked in 1979 and leveled off in subsequent years, whereas the number passing increased steadily during the same period. This led to a notable increase in the proportion of successful candidates (column D), although it still remains well below 10 percent even after allowance is made for the discrepancy between the number of applicants and the number of those taking the examinations.

The table also demonstrates that passing the highly competitive examinations by no means ensures an appointment to a civil-service position. As column E shows, only about half of the successful candidates are actually hired. This does not mean, however, that the remainder are simply bypassed. In the 1976-84 period, three out of ten candidates who had passed the type-A higher examination and four out of ten candidates who had passed the type-B higher examination either withdrew their names from the rosters of eligibles or failed to respond to further communication from the government. In the 1981-86 period, the withdrawal rate of type-A candidates increased sharply, to about 37 percent. The proportion of type-B candidates withdrawing from the roster also increased, from 33 percent in 1981 to 52 percent in 1984.^[26]

The higher withdrawal rate of type-B candidates reflected their somewhat limited career prospects as compared with those of type-A candidates. Not only did the former have relatively limited choice in terms of potential employers, but they were not expected to rise to the apex of the bureaucratic pyramid. Type-A candidates, on the other hand, faced a wider range of options as well as fairly rapid advancement in the civil service. Although all graduates of the HCSE theoretically

[26] he rates of withdrawal or nonresponse were calculated from the data in Jinji-in, Komuin hakusho , 1978-1986.

qualified as kyaria (career civil servants), in practice the prestige and upward mobility implied by the term kyaria were not equally shared by all of them. In Japan, kyaria refers to elite-track bureaucrats, and nonkyaria (noncareer) refers to all the remainder.^[27]

Since 1971, the largest proportion of type-B graduates was hired by the National Tax Administration Agency (Kokuzeicho). In that year the proportion was 51.6 percent, and by 1975 it had risen to 59.0 percent. In 1985, the last year for which type-B graduates were eligible for civil-service appointment, the proportion stood at 77.8 percent.^[28] In the Finance Ministry, of which the NTAA is an external bureau (gaikyoku ), informal distinctions are maintained among honsho kyaria ("career" civil servants hired by the ministry proper), kokuzei kyaria ("career" civil servants hired by the NTAA), zaimu kyaria ("career" civil servants hired by the Financial Bureau [Zaimukyoku], and zeikan kyaria ("career" civil servants hired by the customs). Journalists who cover the Finance Ministry sum up the differential speeds of promotion among the various groups of civil servants as follows: "The honsho kyaria rides the jet airplane, the kokuzei kyaria the bullet train (shinkansen ), the zaimu kyaria the special express (tokkyu ), the zeikan kyaria the express (kyuko ), and the nonkyaria the slow train (donko )."^[29]

The successful applicants who either voluntarily withdraw their names or fail to fill out forms required for further consideration do so because they have other options. Some have offers from private firms. Others may choose to pursue graduate studies. Still others may have passed the judicial examination and opted for entry into the Judicial Training Institute. If we add up those who are hired and those who withdraw, then we are left with only 20 to 25 percent of the successful applicants who fail to receive offers of appointment from government ministries and agencies. No data are available regarding what happens to those people.^[30]

[27] ashiro Ku, "Nihon gyoseikan kenkyu, VII: kyaria to nonkyaria" [A Study of Japanese Public Administrators, Part VII: Career and Noncareer], Kankai , Apr. 1982, pp. 68-79.

[28] he National Tax Administration Agency (NTAA) began to hire type-B graduates in 1968; however, the number remained small (17 in 1968, 16 in 1969, and 26 in 1970) until 1971, when 197 were hired. Jinji-in, Nenji hokokusho , 1961-1986 (Tokyo: 1962-87).

[29] Kuribayashi, Okurasho shukeikyoku , p. 43.

[30] he proportion of the successful candidates who neither withdrew nor were hired in selected years is as follows: 1960, (type A) 15.4 percent, (type B) 26.0 percent; 1965, (A) 26.1 percent, (B) 27.3 percent; 1970, (A) 20.9 percent, (B) 18.2 percent; 1975, (A) 25.0 percent, (B) 7.1 percent; 1980, (A) 23.2 percent, (B) 10.0 percent; 1985, (type I) 19.6 percent. Jinji-in, Nenji hokokusho , 1961, 1966, 1971, 1976, 1981, and 1986 (Tokyo, 1962-1987).

Patterns Of Recruitment

Examination Vs. Evaluation

The available statistics suggest a number of noteworthy patterns in the recruitment of civil servants during the postwar period. In the first place, although competitive examination was intended to be the principal mode of recruitment, evaluation (senko ) has played and continues to play a very imporant role. As table 7 shows, if we look at the civil service as a whole, it is evaluation, not competitive examination, that has served as the primary mode of recruitment. Until the early 1960s, fewer than one in four new appointees had taken civil-service examinations. Although the proportion has increased steadily since then, it has yet to reach the 50-percent mark.

The overall picture, however, is misleading, for it is affected, to a large extent, by the recruitment patterns of specialized personnel, such as educators and health-care specialists. When such personnel are excluded, that is to say, when we focus on civil servants in the Administrative Service I only, the picture changes appreciably.^[31] In 1957, almost seven of every ten Administrative Service I bureaucrats were recruited through competitive examinations, and by 1965 the proportion rose to 78 percent. In the 1980s, eight of every ten civil servants who were appointed to Administrative Service I positions each year were examination-certified.

Because these numbers pertain to new hirings only, however, we need to take note of the situation concerning the incumbent civil servants. So far as the total population of the Administrative Service I was concerned, those who had been recruited through the various competitive examinations constituted 60.7 percent of the total in March 1986, a substantially lower proportion than that for the new appointees in recent years.^[32] It should also be stressed that only a very small proportion of the examination-qualified appointees had passed the higher civil-service examinations. When both types A and B are considered, the proportion has averaged about 9 percent. The average drops to about 6 percent when type A alone is taken into account.

Another glimpse of the patterus of recruitment is provided by table 8. It must be borne in mind that whereas table 7 pertained to new appointees only, this table presents data on incumbent civil servants covered by the Administrative Service I salary schedule. To make our task manageable, however, only the occupants of grades I through 3 will be analyzed.^[33] Focusing on grade 1 first, we see a clear trend toward institutionalization, in which the higher civil-service examination has emerged as the dominant mode of entry into the elite positions. Whereas in 1970 only one-quarter of the grade 1 occupants had passed the higher civil-service examination, by 1972 over one-third of them had done so. By 1974, higher-examination-qualified incumbents were in the major-

[31] As mentioned previously, the Administrative Service (I) accounts for nearly half of general-service civil servants and includes most people who perform clerical, administrative, and managerial functions. The other categories of civil servants encompass revenue and custom agents, crews of state-owned vessels; teachers, professors, researchers, and educational administrators; and health-care personnel.

[32] Jinji-in, Komuin hakusho , 1987, p. 46.

[33] These grades refer to the system that was in use until 20 Dec. 1985. On 21. Dec. 1985, a new system went into effect. See "Kyuyoho no kaisei ni tsuite" [On the Revision of the Compensation Law], Jinji-in geppo 421 (Feb. 1986): 6-47.

ity, and by the early 1980s the higher civil-service examination had become the source of 80 percent of grade-1 officials.

Note also that since 1974 those who had passed the intermediate civil-service examination began to appear in grade 1. Although their actual number is quite small—3 each in 1974, 1976, and 1978, 7 each in 1980 and 1981, 16 in 1982, and 27 each in 1983 and 1984—it is nonetheless significant that products of the intermediate-level examination can and do rise to such high positions. What is more, in 1981 an alumnus of the lower civil-service examination reached grade 1 for the first time in the postwar period. The number has steadily increased in the ensuing years: 2 in 1982, 5 in 1983, and 18 in 1984.

Turning to grade 2, we see the same general trend: a steady increase in the proportion of examination-qualified incumbents over the fifteen-year span. Unlike the situation in grade 1, however, it was not until the end of 1983 that examination had produced a bare majority of incumbents. While the proportion of those who passed the higher civil-service examination appears to have stabilized at the 40-percent mark since 1974, that of the intermediate-examination graduates has shown a marked growth throughout the period. We also find a steady increase in the number of lower civil-service-examination graduates: 2 in 1976, 3 in 1978, 40 in 1980, 53 in 1981, 80 in 1982, 120 in 1983, and 143 in 1984.

As for grade 3, the situation is notably different. Not only has evaluation eclipsed examination by a ratio of seven to three, but the rate of change has been rather low. Alumni of both the intermediate and lower examinations have nonetheless made steady progress. In sum, table 8 demonstrates several distinct trends: (1) access to the positions of assistant bureau chief, division chief, senior-level section chief, and their equivalents is gradually becoming limited to those who have passed the higher civil-service examinations; (2) a tiny fraction of those who entered the civil service by way of the intermediate examination can attain these high-level positions; and (3) that goal is within the reach of even alumni of the lower examination. Realistically, however, the highest position level that the latter can aspire to appears to be that of junior-level section chief or its equivalent.

A statistical breakdown among the higher-examination-certified civil servants is available for the years 1982 through 1984 only. It shows that graduates of the type-B higher examination are concentrated in lower grades: whereas 51 (1982) to 53 percent (1984) of the type-A higher-examination graduates occupied the top three grades, only 12 (1982) to

14 percent (1984) of the type-B higher-examination graduates did so. The actual number of type-B graduates in grade I was negligible: 3 each in 1982 and 1983 and 7 in 1984. This means that there were more lower-examination graduates in the top grade than there were type-B graduates (18 compared with 7 in 1984). In grade-2 positions, lower-examination graduates outnumbered type-B graduates by 80 to 30 in 1982, 120 to 46 in 1983, and 143 to 56 in 1984.^[34]

University Background Of Successful Candidates

In chapter 2 we saw that the prewar Japanese bureaucracy was dominated by graduates of Tokyo Imperial University in its higher echelons and that the phenomenon was a function of the preponderance of Todai men among the successful candidates in the higher civil-service examinations. Table 9 presents statistics that suggest the continuation of that age-old pattern. At first glance, the phenomenon of Todai dominance in the higher civil-service examinations appears to be less pronounced in the postwar period than it was during the earlier era. Even though Todai has never ceased to be the single largest supplier of successful candidates in the higher examinations, its share of the pot has declined somewhat. The precipitate decline in 1972, however, (41-percent decline in absolute numbers and 45-percent decline in proportional terms) is an aberration triggered by an unusual circumstance: because of disruptions caused by campus unrest, no students were admitted to the University of Tokyo in 1968, which in turn led to a sharp decline in the number of its graduates in 1972. In the succeeding years, Todai has supplied between three and four out of every ten successful candidates.

Kyoto University (Kyodai) has retained its position as the number-two supplier of senior bureaucrats in the postwar period. Although its share has generally remained well below one-half of Todai's share, there has consistently been an unbridgeable gap between Kyodai and all other universities. In the 1966-84 period, the number-three position alternated among three schools Hokkaido University, Tohoku University, and Tokyo Institute of Technology (Tokyo Kogyo Daigaku). Hokkaido

[34] hese statistics are not displayed in table 8; they were taken from Jinji-in, Komuin hakusho , 1984, p. 36; ibid., 1985, p. 44, and 1986, p. 44.

University retained the third spot continuously from 1977 to 1984.^[35] Even in its best years, Hokkaido University, which is clearly the strongest of the three contenders for the number-three position, did not come close to matching Kyodai's performance, even though it did break through the 100 barrier in 1979 for the first time, repeating the feat in 1984.

Significantly, all of these universities are national universities (kokuritsu daigaku ). In 1985, however, Waseda University, a private institution of higher learning, captured the number-three position for the first time: it produced eighty successful candidates in the higher examination, accounting for 4.8 percent of all successes. Even though Waseda's total was identical to its 1984 total, it turned out to be one more than Hokkaido's, which had suffered a nearly 20-percent decline in the number of successful candidates.^[36] In 1986 Waseda, which had increased its total to eighty-two, was overtaken by Tokyo Institute of Technology, which produced eighty-nine successes.^[37] In 1987 Tohoku University had ninety-two successful applicants, and in 1988 it had ninety-five. Waseda had eighty-five successful applicants in 1987 and ninety-one in 1988.^[38]

As table 10 shows unmistakably, national universities have consistently outperformed the other two types of universities, public (koritsu daigaku ) and private (shiritsu daigaku ). In all five years for which data are available, we see substantially higher success rates (ratio of successful to all applicants) for national-university graduates than for those of the other two types of universities. In those years, national-university graduates were between 2 and 4.5 times more likely to pass the type-A higher examination than those of public universities and between 7 and 20 times more likely to pass the examination than those of private universities. Graduates of public universities were better off than those of private universities by a substantial margin: the former's chances of passing the examination were between 1.7 and 7.7 times greater than the latter's. If we look beyond the aggregate figures, we find

[35] Jinji-in geppo 197 (July 1967): 15, and 209 (July 1968): 15; Sato Tomoyuki et al., Todaibatsu , p. 15; Asahi shinbun , 22 Aug. 1972, 9 Sept. 1974 (evening ed.), 1 Nov. 1975 (evening ed.), 26 Oct. 1976 (evening ed.), 25 Oct. 1978, 15 Oct. 1983, and 1 Oct. 1984 (evening ed.); Mainichi shinbun , 11 Sept. 1973, 15 Oct. 1979, 15 Oct. 1980, 15 Oct. 1981, and 15 Oct. 1982.

[36] Asahi shinbun , 1 Oct. 1985 (evening ed.).

[37] Ibid., 1 Oct. 1986 (evening ed.). In 1987, undergraduate enrollments at selected universities were as follows: Tokyo, 14,379; Kyoto, 11,553, Hokkaido, 10,167; Tohoku, 10,075; Kyushu, 9,978; Tokyo Institute of Technology, 3,578; Waseda, 40,532; Keio, 22,839; and Chuo, 30,745. Association of International Education, Japan, ed., Japanese Colleges and Universities 1987 (Tokyo: Maruzen, 1987).

[38] Nihon keizai shinbun , 15 Sept. 1987; Asahi shinbun , 7 Sept 1988 (evening ed.).

wide variations among individual universities within each type. Unfortunately, a breakdown by university is available for 1967 only. As the notes in table 10 show, the success rate of Todai graduates was the highest (33.3 percent) and that of Kyodai trailed not too far behind (25.7 percent). The top performers among public and private universities were virtually indistinguishable in terms of their success rates: Osaka City University and Tokyo Metropolitan University in the former group, with 7.3 and 6.8 percent, respectively, were only a fraction of a percentage point ahead of Keio (7.0 percent) and Waseda (6.7 percent) in the latter group.

As we have seen previously, however, those who have passed the higher civil-service examinations do not ipso facto become civil servants. To get a more meaningful picture, therefore, we need to ascertain whether the pattern of dominance by graduates of national universities and, particularly, by a few elite universities continues in the hiring stage. Table 11 gives us a breakdown by type of universities for recent years. It reveals that the advantages of national-university graduates have begun to diminish at the hiring stage. Only in 1975 did they enjoy a substantial edge over graduates of public and private universities. In 1980, the latter registered higher rates of appointment than the former, and the same pattern was repeated in 1983, 1984, 1985, and 1986.

What is especially noteworthy is the performance of private-university graduates: their hiring rates in the 1980s have eclipsed those of national-university graduates by a substantial margin. When this is viewed in conjunction with a steady increase of private-university graduates in the proportion of successful candidates in the higher examination (for example, 8.5 percent in 1982, 10.1 percent in 1983, 12.8 percent in 1984, 11.1 percent in 1985, 12.6 percent in 1986, 13.1 percent in 1987, and 13.4 percent in 1988),^[39] we can anticipate the dawning of the day when attending a private university is no longer perceived as a handicap in civil-service employment. The slight decline in the share of private-university graduates in 1985 was offset by an increase in absolute numbers from 167 in 1984 to 184 in 1985. In 1986, their share nearly returned to the 1984 level in proportional terms and reached a new high (N = 217) in absolute numbers. The year 1987 saw the best record in twenty-one years for private universities in absolute (N = 221) and proportional (13 percent) terms alike. In 1988 the record improved further: 243 private university graduates passed the

[39] Yomiuri shinbun , 16 Oct. 1983, and 15 Sept. 1987; Mainichi shinbun , 1 Oct. 1984 (evening ed.); Asahi shinbun , 1 Oct. 1985 (evening ed.), 1 Oct. 1986 (evening ed.), and 7 Sept. 1988 (evening ed.).

higher examination, accounting for 13.4 percent of all sucesses. A look at absolute numbers, on the other hand, helps us to appreciate a key fact: notwithstanding the increase in the rates of hiring, the number of private- and public-university graduates entering the higher civil service remains low. Between eight and nine out of every ten new appointees in the six years covered by table 11 came from national universities.

The continued dominance of national universities becomes plain when we turn from aggregate statistics to more specific data, of which only fragments are available. First, we have the testimony of Tashiro Ku, a former bureau chief in the National Personnel Authority, who writes that between 110 and 115 of the 160 candidates hired from the 1980 roster of type-A eligibles in the field of law were graduates of Todai's law faculty. That accounted for 69 to 72 percent of the total. Tashiro reports that this was not an atypical year.^[40]

[40] Tashiro, "Nihon gyoseikan kenkyu, VII: kyaria to nonkyaria," pp. 89-90.

Second, we have data regarding the university background of the new hiring by the Ministries of Finance and Home Affairs. Looking at the Finance Ministry in selected years first, we find the predominance of Todai graduates; they constituted 77.3 percent (17 out of 22) of the 1970 entering class, 76.9 percent (20 out of 26) of the 1975 class, 82.6 percent (19 out of 23) of the 1980 class, and 84.6 percent (22 out of 26) of the 1985 class. About 75 percent of Todai men hired by the Finance Ministry during these years were graduates of its law faculty. Non-Todai graduates came principally from Kyodai and Hitotsubashi University. In 1970, 1975, and 1980, these three universities were the exclusive sources of Finance Ministry recruits. In the twenty-three-year period from 1960 to 1982, only 8 graduates of private universities were hired by the Finance Ministry on the career track: 4 from Keio, 3 from Waseda, and 1 from Chuo.^[41]

Turning to the Home Ministry, we find a similar situation. The proportion of Todai graduates among all new appointees in selected years was as follows: 1950: 100.0 percent (3 out of 3); 1955: 62.5 percent (5 out of 8); 1960: 75.0 percent (9 out of 12); 1965: 76.9 percent (10 out of 13); 1970: 80.0 percent (12 out of 15); 1975: 81.3 percent (13 out of 16); 1980: 88.2 percent (15 out of 17); and 1985: 93.7 percent (14 out of 15). All but 3 of the 81 Todai graduates hired by the Home Ministry in these years came from its law faculty. Of the 18 non-Todai men, 10 were Kyodai graduates, all of them in law, and 3 were graduates of Tokyo Institute of Technology. Only 2 came from private universities.^[42]

How may one account for the slow increase in the number of private-university graduates in the higher civil service? If one may hazard a speculation, a number of factors may be cited. One is the changing perception of the higher civil service: the view that it is the exclusive preserve of graduates of a few elite national universities is being increasingly rejected by students of private universities. And the success of the latter in scaling the wall of the higher civil service will undoubtedly help reinforce this trend. Another factor may be the

[41] The data for 1960 through 1982 are from Jin Ikko, Okura kanryo: Cho-erito shudan no jinmyaku to yabo [Finance Ministry Bureaucrats: the Personal Connections and Ambitions of a Super-Elite Group] (Tokyo: Kodansha, 1982), appendix: "Okurasho zen kyaria ichiran" [Complete List of Career Civil Servants in the Finance Ministry], unpaged. The 1985 data are from Shukan Yomiuri [Weekly Yomiuri], 16 Dec. 1984, p. 171.

[42] Jin Ikko, Jichi kanryo [Home Ministry Bureaucrats] (Tokyo: Kodansha, 1986), pp. 232-57. These pages list all new appointees from 1947 to 1985.

narrowing of the gap between national and private universities in terms of the caliber of students. That is to say, the gap between those who enter the top-rated national universities and those who enter the top-rated private universities may not be as great as is generally believed.^[43] A third, and probably the most important, factor is the apparent change in the attitudes of the hiring agencies. Many of them appear to be making a conscious effort to diversify the background of new elite-track appointees.

Another dimension of the educational background of new recruits into the higher civil service is their level of education. As we have already noted, the educational level of applicants for all types of civil-service examinations has increased appreciably over the past two decades, and this is bound to affect the situation at the hiring stage as well. Nearly four out of ten persons who were hired as higher civil servants in 1975, 1980, and 1983, for example, had done graduate work. Scientists and engineers show the highest level of educational attainments, whereas "law and the humanities" graduates display the lowest. Indeed, it is striking that between 65 and 76 percent of scientists and engineers hired during the three years mentioned above had done graduate work.^[44]

All this is clearly linked to the rapid increase in the graduate population. Whereas the number of university students increased 2.9 times between 1960 and 1980, that of graduate students increased 3.4 times during the same period. What is more, the number of graduate students in the natural sciences and engineering grew 6.7 times in the twenty-year period, as compared with a mere two-fold increase in the number of graduate students in the humanities and social sciences.^[45] It

[43] of the applicants who take entrance examinations to both Todai and a top-rated private university, a sizable proportion is known to pass the former but fail the latter. In 1984, for example, 30 (or 24 percent) of the 127 graduates of Yoyogi Seminar, a leading preparatory school in Tokyo, who passed the entrance examination to Todai's Law Faculty failed the entrance examination to Waseda University's Faculty of Politics and Economics. Overall, 32 percent of the 275 Yoyogi graduates who were admitted to Todai's humanities faculties (bunka kei ) were rejected by Waseda's Faculty of Politics and Economics. Of the Yoyogi graduates who passed the entrance examination to Todai's Faculty of Medicine, reputed to be the most competitive in Japan, about 20 percent failed the entrance examination to Keio University's Faculty of Medicine. "Juken sensen de kigyo kyoso de ima 'Todai shinwa' no hokai ga hajimatta!" ["The Myth of the University of Tokyo" Has Begun to Crumble on the (University) Entrance Examination Front and in the Competition Among Business Enterprises], Shukan gendai , 2 Mar. 1985, pp. 22-27.

[44] These statistics were obtained from the National Personnel Authority during the author's visit to that agency in June 1985.

[45] Monbusho, ed., Waga kuni no kyoiku suijun [The Educational Level of Our Country], 1980 (Tokyo: Okurasho, Insatsukyoku, 1981), appendixes 66 and 67.

is noteworthy that the door to the higher civil service is not totally closed to those lacking university diplomas: a total of 15 such persons were hired in 1975, 1980, 1983, 1984, 1985, and 1986 (see table 11).

On balance, however, the picture that emerges from the preceding survey is the continuing dominance of a few elite universities, particularly Todai, in both the examination and hiring stages of higher civil servants. On the other hand, the steady increase in the proportion of private-university graduates, both among the successful candidates in the higher examination and among those who are actually hired by government ministries and agencies, suggests that the process of incremental change may have already begun.^[46]

Hoka Banno

Let us now turn to another age-old pattern in Japanese bureaucracy: the dominance of law graduates (hoka banno ). Table 12 displays some data suggestive of how law has fared as a field of specialization in the higher civil-service examination in the postwar period. In a word, it has become the single most important field, accounting for between 14 and 27 percent of the successful candidates during the 1955-87 period. The total number of fields during the period has ranged from twenty-four to twenty-seven. The number and proportion of successful candidates in the next four fields (type A or type I only) in selected years are as follows:

1955: economics, 142 (10.8 percent); public administration, 111 (8.4 percent); civil engineering, 100 (7.6 percent); mechanical engineering, 93 (7.1 percent).

1965: mechanical engineering, 165 (10.2 percent); civil engineering, 160 (9.9 percent); physics, 149 (9.2 percent); chemistry, 100 (6.2 percent).

1975: civil engineering, 150 (12.4 percent); chemistry, 84 (7.9 percent); economics, 83 (6.9 percent); physics, 80 (6.6 percent).

[46] Another sign of change worth mentioning is that in 1986 the Ministry of Trade and Industry (MITI) hired a graduate of a foreign university as a career civil servant for the first time in its history: a 24-year-old Japanese who had graduated from Brown University with a major in econometrics. MITI's chief of the Minister's Secretariat said that the appointment reflected MITI's goal of securing "persons with rich international experience" in view of the growing internationalization. See "Kasumigaseki konhidensharu" [Confidential Report from Kasumigaseki], Bungei shunju , Jan. 1986, p. 171.

1985: civil engineering, 212 (12.8 percent); economics, 110 (6.6 percent); physics, 110 (6.6 percent); electronics, 110 (6.6 percent).^[47]

The preceding data indicate that the gap between law and the other top fields is fairly wide: only in the 1980s has the next field begun to

[47] Jinji-in, Nenji hokokusho , 1955, 1965, 1975, and 1985 (Tokyo, 1956, 1966, 1976, and 1986).

challenge law in number and proportion. In 1983, civil engineering surpassed law for the first time, with 228 successful candidates, 7 more than law. In 1985, however, law recaptured its paramount position, and the situation remained unchanged in 1986 and 1987.

More important, about a quarter of the successful candidates who are actually hired have law as their field of specialization. Comparison of columns G and H in table 12 shows that the proportion of successful candidates in law who are actually hired has consistently surpassed that of all candidates by 15 to 20 percent, reaching an all-time high of 29 percent in 1986. Moreover, law offers the widest range of options to the successful candidate in terms of government employment, for virtually every agency recruits law specialists every year. The other fields that offer relatively abundant options to the candidates are economics and public administration, but neither can match law.

It should also be noted that the hiring rates for successful candidates in engineering and the natural sciences are relatively low. Because of attractive alternatives in the private sector, a large proportion of these candidates withdraw, and this has in turn prompted the government to pass a larger number of candidates in the technical fields than would be the case otherwise. The hiring rates for candidates who passed the 1986 higher examination in selected scientific and engineering fields were as follows: physics, 23.6 percent; chemistry, 26.8 percent; information engineering, 18.7 percent; electrical engineering, 25.8 percent; electronics, 25.9 percent; mechanical engineering, 25.0 percent; civil engineering, 25.9 percent; and metallurgy, 15.0 percent. Among the technical fields, those related to agriculture have the highest hiring rates, which reflects the relative scarcity of private-sector alternatives in them.^[48]

The preponderance of law as a field of specialization in the higher civil-service examination tends to accentuate the influence of elite universities. As table 13 shows, Todai supplies about seven out of ten successes in the field of law. Comparison of this table with table 9 indicates that Todai's share of successful candidates in law is about twice its share of all successful candidates. When law, economics, and public administration are combined—the three fields have traditionally supplied a disproportionate share of top administrators in Japanese bureaucracy—Todai's share declines slightly.

[48] These rates were calculated from the data in Jinji-in, Komuin hakusho , 1986, p. 25, and 1987, pp. 40-41. The hiring rates for agriculture-related fields in 1986 were as follows: agriculture, 56.0 percent; agricultural economics, 38.7 percent; agricultural chemistry, 42.2 percent; agricultural engineering, 50.0 percent; animal husbandry, 48.0 percent; and forestry, 62.9 percent.

As we shall see in the following chapter, the importance of law increases further when promotion patterns are examined. The nature of the educational experience that candidates in law undergo, therefore, will be a major component of our inquiry into the socialization of Japanese higher civil servants. At this point, however, it is necessary to explain briefly what specialization in law signifies in the Japanese context. The most important thing to bear in mind is that being a law graduate (that is, a graduate of a faculty of law in a university) is not the same thing as being a lawyer. All faculties of law in Japan are undergraduate institutions. Although their graduates receive a substantial amount of legal training as well as law degrees (hogakushi [bachelor of law]), only a small proportion of them take the judicial examination, and still a smaller proportion pass it.

In 1986, for example, only 1 in 49 applicants passed the examination. Of the 486 successful candidates, 97 (20.0 percent) came from Todai, 85 (17.5 percent) each from Waseda and Chuo universities, 32 (6.6 percent) from Kyodai, and 25 (5.1 percent) from Keio University. Interestingly, three of the top five universities are private—Waseda, Chuo, and Keio. The judicial examination is said to be so difficult that only a handful of those who pass it do so on their first try. The average age of the successful candidates in recent years has hovered around twenty-eight, about five years higher than the average age of graduation from a law faculty. In 1986, a record number of women (59 or 12.1 percent of all successes) passed the judicial examination.^[49]

Those who pass the judicial examination must undergo two years of postgraduate training at the state-run Judicial Training Institute (Shiho Kenshujo). At the end of the two years they must take another examination, consisting of both written and oral tests. Those who pass the second examination are officially certified to enter the legal profession, as judges, public prosecutors, and attorneys. Of the 450 persons who successfully completed their training in April 1986, 70 became judge trainees (hanjiho ), 34 became public prosecutors, and the remainder opted for careers as attorneys.^[50] In a sense, the Judicial Institute can be equated with a law school in the United States. Given that only five hundred or fewer qualified lawyers are produced in this fashion each year, it is not surprising that the number of lawyers in Japan is exceedingly small. In 1983 there were an estimated 12,500 attorneys in private practice, approximately 1 per 10,000 population. This contrasts sharply with the situation in the United States, where the ratio was about 1 lawyer per 400 population (about 600,000 attorneys).^[51]

For all but a handful of the people involved, then, graduation from a faculty of law does not foreshadow a career in the legal profession; it simply signifies completion of what amounts to general education. Law graduates, in other words, are generalists, not specialists. As such they

[49] Asahi shinbun , 1 Nov. 1986, and 21 May 1987. For a description of the Judicial Examination, see Tokyo Daigaku shinbun (weekly), 5 July, 1976, p. 3.

[50] Although nearly all trainees of the Judicial Training Institute pass the second examination, it is by no means automatic. In 1986, eight trainees were accused of cheating in the examination and failed to graduate with their classmates. Asahi shinbun , 27 Mar., 3 Apr. (evening ed.), 4 Apr. and 5 Apr. 1986.

[51] "Land Without Lawyers," Time , 1 Aug. 1983, pp. 64-65. "Although the Japanese have a highly developed sense of individual rights, social harmony, not personal justice, is the basis of their law. Litigation, never common, has actually decreased over the past 15 years." Ibid., p. 64.

enjoy a wide range of options. A study of some 24,000 graduates of Todai's law faculty about whom information was available in 1958 showed that the largest proportion, 44 percent, were either in banking or in private enterprises. The second-largest group, 23 percent, was in the civil service. Next, about 13 percent were in the legal profession (7 percent as judges or prosecutors and 6 percent as attorneys). The remainder were scattered in such fields as education, journalism, and politics.^[52] Data on the destinations of Todai law graduates in the 1970s and 1980s confirm that the preceding trend continues without any notable modifications.^[53]

Male Dominance

Prewar Japan was a male-dominated society. Politically, this was symbolized by the lack of franchise for women. It was natural that the government bureaucracy should mirror the inferiority of women in the larger society. Legally, however, women were placed on an equal footing with men in the competition for civil-service appointments as early as 1909, when they became eligible for the higher civil-service examination. However, they did not actually end male monopoly until 1928, when three women passed the judicial section of the higher civil-service examination. In 1931, a women passed its administrative section for the first time but was never offered an appointment to the higher civil service. Although more women passed the judicial section, no other women passed the administrative section before it was abolished after the Japanese defeat in World War II.^[54] As two American observers wrote in 1946: "The status system shows little scope for female service. The only women who have invaded the higher civil service are a few physicians and one labor executive in the Welfare Ministry. Even among ordinary officials, there are very few women; generally the exclusion is quite complete."^[55] To what extent has the situation changed since 1946?

The earliest available information regarding the sexual composition of the civil service pertains to 1958. A government survey of general-service employees covered by the regular compensation law revealed

[52] Shimizu Hideo, Tokyo Daigaku Hogakubu: Nihon erito no manmosu kichi [The University of Tokyo Faculty of Law: A Mammoth Base of Japan's Elite] (Tokyo: Kodansha, 1965), p. 48.

[53] See Tokyo Daigaku shinbun , 28 June 1976, and Hata, Kanryo no kenkyu , p. 189.

[54] Hata, Kanryo no Kenkyu , pp. 103-4.

[55] MacDonald and Esman, "The Japanese Civil Service," p. 215.

that 62, 128 (17.2 percent) of the 360,328 civil servants were women. However, only 802 of the women civil servants had college degrees. In other words, only 1.6 percent of all civil servants with college degrees were women. The proportion increased sharply—to 14.7 percent—when graduates of junior colleges were examined.^[56]

In terms of average monthly salaries, female college graduates in the civil service earned 41.9 percent less than their male counterparts, and female graduates of junior colleges earned 30.5 percent less than their male colleagues with the same amount of education. When the length of service is held constant, the gender gap in compensation narrows somewhat. Women college graduates with less than one year of service earned 5.6 percent less than men with the identical background, but the disparity nearly quadrupled for women with fifteen to twenty years of service.^[57]

By 1976, some progress had been registered in the overall picture. The proportion of women in the civil service increased slightly, to 19.5 percent, and the proportion of women among college graduates in the civil service increased fourfold, to 6.8 percent. Meanwhile, the gender gap in compensation became markedly narrower. This was especially true when comparison was made between men and women with the same level of education and the identical length of service; in fact, women with less than two years of service earned slightly more than their male counterparts. However, the gap for those with longer experience remained, although it had been substantially reduced.^[58]

Table 14 shows that women have by no means been shut out of the higher civil-service examination. On the contrary, between 4 and 8

[56] Jinji-in, Nenji hokokusho , 1958, p. 50. To a large extent, the scarcity of college-educated women in the Japanese civil service reflected the scarcity of college-educated women in Japanese society as a whole. Of 1,858 persons who graduated from four-year colleges in 1950, only 31 (1.7 percent) were women. The situation had improved markedly by 1955:13,544 (14.3%) of the 94,735 university graduates that year were women. The proportion of women among university graduates increased steadily thereafter: 16.2 percent in 1965, 21.6 percent in 1975, and 24.7 percent in 1980. Among graduates of junior colleges, women began to surpass men in 1953, and by the late 1960s, eight of every ten junior-college graduates were women. In 1980, 155,200 (91.3 percent) of the 169,930 persons who graduated from junior colleges were women. Monbusho, ed., Waga kuni no kyoiku suijun , 1980, appendixes 137 and 138.

[57] Jinji-in, Nenji hokokusho , 1958, pp. 48-49.

[58] The gap for women with 15 to 20 years of service had been reduced from 19 percent in 1958 to 8.3 percent in 1976. See Jinji-in, kyuyokyoku, Kokka komuin kyuyoto jittai chosa hokokusho [Report on an Investigation into the Actual Conditions of National Civil Servants' Compensation and Related Matters], 15 Jan. 1976 (Tokyo: Okurasho, Insatsukyoku, 1976), pp. 4-5 and 38-39. Chap. 8 discusses further the question of the gender gap in compensation.

percent of the successful candidates in recent years have been women. The peak for women was reached in the early 1960s, when they constituted 8 to 9 percent of the successful candidates. From 1966 to 1982, the number of successful women candidates remained well below 100—in fact, close to half that number—and their proportion consistently fell short of the 5 percent mark. In 1983, however, women registered a substantial gain in both absolute and proportional terms. In

1984 not only did women sustain the same rate of gain as the previous year's, but, for the first time in the history of the higher civil-service examination, they received the highest grade in four fields: sociology, pharmacy, agriculture, and agricultural economics.^[59] In 1985 women broke the 100 barrier for the first time in twenty years, and their proportion among all the successful candidates, too, was the highest since 1964. Women were top-scorers in two fields, chemistry and pharmacy.^[60] In 1986 women continued their upward trend, gaining twenty-three in absolute numbers and 1.2 percentage points in their share of the successes over the previous year.^[61] In 1987 although women registered a slight decline both in absolute numbers and in proportional terms, they produced top scorers in four fields for the second time in three years: sociology, psychology, education, and pharmacy. In 1988 women had their best performance in twenty-five years: 150 successes (8.3 percent of all successes). They also produced top scorers in three fields: public administration, psychology, and animal husbandry.^[62]

Another notable pattern has to do with the ratios of types A and B. As we saw in table 6, since the 1970s successful type-A candidates have outnumbered successful type-B candidates by the ratios of between ten and fifteen to one. Among women, however, the ratios until 1984 hovered around four to one. In fact, when women's share of successful type-A candidates alone is examined, there is a slight deterioration in the picture. Except for 1963, 1964, and 1984, women's share of the total type-A successes is about a half percentage point below that of type-B successes. To sum up, in the category that really counts in terms of upward mobility, women still have a long way to go.

How do women compare with men in terms of their rates of success in the higher civil-service examination? Table 15 presents the relevant statistics for 1983 and 1984. It shows that although women lag behind men in their success rates in all major categories of specialization, the gap is the narrowest in "law and the humanities," where the difference is only a half percentage point. The widest gap is found in "science and engineering," where men's success rates were 2.6 to 2.7 percentage points higher than women's. Although women's success rates improved

[59] Mainichi shinbun , 1 Oct. 1984 (evening ed.).

[60] Asahi shinbun , 1 Oct. 1985 (evening ed.).

[61] Ibid., 1 Oct. 1986 (evening ed.).

[62] Yomiuri shinbun , 15 Sept. 1987; Asahi shinbun , 7 Sept. 1988 (evening ed.).

somewhat in 1984 compared with 1983, so did men's, thus leaving the overall gender gap intact.

One bright spot in table 15, so far as women are concerned, is that nearly half of women's successes are found in "law and the humanities," compared with only about a quarter of men's successes. This is good news for women, primarily because candidates in "law and the humanities" enjoy an edge in upward mobility. In terms of specific fields, as opposed to groups of fields, the top five for women in 1983 were pharmacy (13), law (11), education (6), sociology (6), and agricultural chemistry (6). These five fields together accounted for 57.5 percent of all successful women candidates. In 1984, three fields tied for the fifth position. The top fields were law (12), agricultural chemistry (11), pharmacy (9), chemistry (7), psychology (5), education (5), and sociology (5). These seven fields together accounted for 62.8 percent of all successful women candidates.^[63]

[63] These statistics were obtained from the National Personnel Authority during the author's visit to that agency in June 1985.

As noted, success in the higher civil-service examination does not automatically translate into an appointment to the higher civil service. Hence we need to know how many women actually are appointed to the higher civil service. Table 16 presents two kinds of data pertinent to the question. First, column B displays the number of new women appointees to Administrative Service I positions who have passed either the type-A or type-B higher civil-service examination. Second, column A displays the number of all higher-examination-qualified women civil servants who occupy Administrative Service I positions in the years indicated. Looking at column B first, we see that the number of new women appointees is quite small. Column D indicates, moreover, that their proportion is equally small. Only in 1985 and 1986 did women in the type-A category exceed 4 percent of the total new appointees. So far as the less prestigious type B is concerned, women's proportion surpassed the 10-percent mark three times, in 1976, 1982, and 1985. In fact, in all of the twelve years covered in the table, type-B women surpassed type-A women in proportional terms, though the record is mixed in terms of absolute numbers.

Turning to column A, we find that with the notable exception of 1977, the total number of higher-examination-certified women in Administrative Service I positions increased steadily until 1984. Although 1977 registered a 23-percent decline, the following year saw an 18-percent increase, and by 1979 the 1976 level was all but restored. In subsequent years we see a steady increase until 1985, when the number declined by 12 percent over the previous year. In proportional terms the picture has remained relatively stable: at no time did the combined totals of types A and B exceed 4 percent. When we focus on type A only, we find that women's share is below the 3-percent mark; even though a breakdown by types A and B is not available for the years 1975 through 1981, we can surmise that the same general pattern holds. It should be stressed that table 16 displays data for women in Administrative Service I positions only. If we were to examine all women who are in the higher civil service, the picture would improve somewhat. The number of women who are graduates of the type-A (or type-I) higher examination, for example, doubles, and their proportion increases by 1 to 2 percent.

If we focus on the administrative elite in the narrow sense, that is, those civil servants who have attained at least the position of section chief or its equivalent, we find that women are even more underrepresented than the preceding statistics suggest. In 1987, one could count

only twelve such women serving in the headquarters (honsho ) of the twelve main ministries. In fact, only six ministries had women in elite positions: Labor (4), Health and Welfare (3), MITI (2), Agriculture (1), Justice (1), and Education (1). By including those serving in the "field" (gaikyoku ), we could increase the total to seventeen and the number of ministries represented to seven (with the Finance Ministry being added). Inclusion of the other government agencies would add eight more: four from the National Personnel Authority and two each from the prime minister's office and the Economic Planning Agency. Actually, one of the two women in the prime minister's office had been temporarily detached from the Labor Ministry.^[64]

Of the twenty-five women, two were bureau chiefs, one was chief of a prefectural bureau, ten were section chiefs, three were chiefs of offices (shitsu cho ), two were counselors (sanjikan ), two others were directors of research institutes, and the rest held miscellaneous titles that were equivalent to section chiefs. Their age ranged from 41 to 59, the median age being 47. In terms of educational background, twelve were graduates of Todai, three were from International Christian University, and the remainder had gone to Tohoku, Keio, Waseda, Ochanomizu, and other universities. Two, both nurses, had not attended college. In terms of field of study, the largest number (five) had studied general education (kyoyo ). Law and the humanities tied for second place with four each. Economics and pharmacy followed with two each. The remaining women had majored in various fields ranging from mathematics to nursing.

The two highest-ranking women were employed in the Ministries of Labor and Health and Welfare, respectively. Sato Ginko, a 1958 graduate of Todai's general-education faculty, was chief of the Labor Ministry's Women's Bureau. Since 1947, when the bureau was first created, it has always been headed by a woman.^[65] The other senior woman, Nagao Ritsuko, was promoted to her position as chief of the Children's and Families' Bureau of the Health and Welfare Ministry in September 1987. She thus became not only the first woman bureau chief in her ministry but also the first woman to attain a bureau-chief rank

[64] These data are based on a review of all the higher civil servants listed in Seikai kancho jinji roku, 1988-nenban [Who's Who in Politics and Government, 1988 Edition] (Tokyo: Toyo Keizai Shinposha, 1987).

[65] Sano Mitsuko, Josei kanryo: Sono ishiki to kodo [Women Bureaucrats: Their Consciousness and Behavior] (Tokyo: Jihyosha 1983), p. 397.

outside the Labor Ministry. She is a 1958 graduate of Todai's literature faculty and a twenty-nine-year veteran in her ministry.^[66]

The third-highest-ranking woman in 1987 was Matsumoto Yasuko, chief of the Labor Ministry's Bureau of Labor Standards in Saga prefecture. A 1959 graduate of Waseda University (politics and economics), she had previously served as a section chief in her ministry and as chief of a women's-affairs office in the prime minister's office.^[67]

Another woman who merits brief mention is Sakamoto Harumi. A 1962 graduate of Todai's law faculty, she became, in the same year, the first woman ever hired by the Ministry of International Trade and Industry (MITI) on its elite track. After compiling a distinguished record in MITI, she was appointed chief of the Sapporo Bureau of International Trade and Industry in June 1986 and subsequently drew much attention and praise for her handling of her new job. In June 1987, however, she abruptly resigned her position. Shortly thereafter she became an adviser to the Daiichi Kangyo Bank, Japan's largest bank.^[68]

Most of the women who opt for careers in the higher civil service do so in the belief that in so doing they are likely to encounter markedly less discrimination than they would in the private sector. For many women, there is no other viable choice. According to a 1976 Todai graduate who entered the Finance Ministry in the same year, even graduates of Todai's economics faculty are not in much demand in the private sector if they happen to be women; she chose public employment "by the process of elimination."^[69] In the case of Sakamoto Harumi, however, an explicit commitment of equal treatment was said to have been obtained prior to her entry into MITI.^[70]

Nearly all women, nonetheless, do run into difficulties. Some discover that they are expected by their male superiors to perform such traditionally feminine tasks as serving tea, cleaning up the office, and answering telephones for other people. One woman, Matsumoto

[66] Asahi shinbun , 26 Sept. 1987. For a profile of Nagao, see Sano, Josei kanryo , pp. 193-218.

[67] For a profile of Matsumoto, see Yoshihara Atsuko, Sukato o baita kokyu kanryo [Higher Civil Servants Who Wear Skirts] (Tokyo: K. K. Kanki Shuppan, 1986), pp. 107-33.

[68] Asahi shinbun , 19 June 1987; "Kasumigaseki konhidensharu," Bungei shunju , Feb. 1986, p. 192, and Oct. 1987, p. 180. For a profile of Sakamoto, see Sano, Josei kanryo , pp. 277-302 and Yoshihara, Sukato o haita kokyu kanryo , pp. 35-106.

[69] Sano, Josei kanryo , p. 122.

[70] Yoshihara, Sukato o haita kokyu kanryo , pp. 23 and 67. Sakamoto had another offer from a major private firm, which gave her some bargaining power.

Yasuko, upon being asked to do such chores at a prefectural office of the Labor Ministry in her second year, flatly refused to take turns as a telephone operator, even though she did serve tea. When her superior tried to persuade her in front of all other women civil servants, none of whom was on the career track, she countered by proposing that in return for women's serving tea, men should clean up the office. Her proposal was tacitly accepted.^[71] Her intransigence apparently did not hurt her career, for, as we have seen, she later served as a section chief and chief of a prefectural bureau.

A number of women also report incidents in which, when they answer telephones or greet visitors, the callers or visitors ask to talk with the "person in charge," refusing to believe that they are talking to such persons. The most difficult of all, however, is the challenge posed by children. Since most do not get home until 9 or 10 P.M. on most days and, during the budget-preparation period, the working hours are extended further, they need to find housekeepers and baby-sitters. Many end up enlisting the aid of their mothers.

Women in the elite track try very hard not to deviate from the unwritten social norms of Japanese bureaucracy. These entail having drinks with colleagues and, especially, subordinates after work and playing mah-jongg, a favorite pastime of Japanese bureaucrats, with them.

In sum, although Japanese bureaucracy is no longer the bastion of male supremacy that it once was, it still has a long way to go before women achieve a semblance of equality or, at least, reasonably fair representation, in its high echelons. The handful of women who have either attained or are on their way toward attaining senior ranks appear to have educational and other credentials that either equal or surpass those of their male counterparts.

Recruitment Of Diplomats

Recruitment of foreign-affairs personnel merits brief discussion, because it is handled separately in Japan, as is true in most other countries. It was not until Japan regained its sovereignty in 1952 that the Ministry of Foreign Affairs formally took over the responsibility of conducting examinations, of which there were two until 1962: the examination to select diplomats and consuls and the examination to

[71] Ibid., pp. 107-33.

select foreign-affairs clerks. The former was equivalent to the higher civil-service examination. From 1963 to 1976, there were three types of examinations: the higher-level examination to select foreign-affairs civil servants, the intermediate examination to select foreign-affairs civil servants, and the examination to select language trainees for the Ministry of Foreign Affairs. The last-named examination was also an intermediate one aimed at recruiting graduates of junior colleges. In 1977 the two intermediate examinations were consolidated into the examination to select specialist employees of the Ministry of Foreign Affairs. And in 1985 the higher-level examination was renamed the type-I examination to select foreign-affairs civil servants, in conjunction with the revision of the civil-service examinations in other fields.^[72]

Table 17 displays some data about the higher-level (or type-I) foreign-service examination. Although the data are incomplete, it is nonetheless striking that an extraordinary number of applicants fail either to show up for or complete the first stage of the examination. Their proportion is twice as high as that for the higher civil-service examination. This appears to be a function of the difficulty of the examination; many applicants are so intimidated by it that they either have second thoughts about taking the examination or give up after sampling the first few subjects.

What does the examination actually consist of? Open to Japanese citizens between the ages of twenty and twenty-eight regardless of their educational background, the examination has two stages. In the first stage, applicants are tested in (1) general education (ippan kyoyo ), (2) a foreign language, (3) constitutional law, (4) international law, (5) principles of economics, (6) diplomatic history, and (7) two elective subjects, to be chosen from administrative law, civil law, public finance, and economic policy. The general-education test consists of sixty multiple-choice questions, which must be answered in three hours. This is a key test, for all applicants who do not pass it are automatically disqualified no matter how well they may do in the other subjects. Aimed at testing the applicant's general knowledge and analytic ability, the questions in this test encompass the humanities and the social and natural sciences; they also include problems in mathematical reasoning and quantitative data analysis. In 1986, for example, the general-education test contained questions on Western philosophy, existential-

[72] Jinji-in, Nenji hokokusho , 1948-49 through 1952, 1963, and 1964; Jinji-in, Komuin hakusho , 1978 and 1986.

ism, famous battles in world history, modern Korean history, Japanese poetry, world geography, logic, analysis of trade statistics, and interpretation of economic graphs.^[73]

For the foreign-language test, the applicant must choose one of the following: English, French, German, Russian, Spanish, Chinese, and Arabic. He is then asked to translate texts in the language of his choice into Japanese and vice versa. The tests for the remaining subjects typically consist of two essay questions (or topics) that must be answered in two hours. To give some examples from the 1985 examination:

[International Law]

1. If the fundamental human rights of a private individual who resides in a foreign country are violated, what are the internationally recognized remedies at his disposal?

2. Discuss the immunities of consuls and other consular personnel from the judicial and police jurisdictions of their receiving state.

[Principles of Economics]

1. Discuss the role of public enterprises in a free-market economy from the standpoint of economic theory.

2. Discuss the theoretical advantages and disadvantages of the gold-standard system in comparison with the floating-exchange-rate system.

[Diplomatic History]

1. World War II and the relations among Japan, the United States, and the Soviet Union (1939-45).

2. The movement toward the integration of Europe in the wake of World War II (1945-58).^[74]

Those who pass the first stage of the examination must take the second round of tests, which comprise (1) a physical examination; (2) oral tests in constitutional law, international law, and principles of economics; (3) a foreign-language test (composition and conversation); (4) a comprehensive written test; and (5) a character test (jinbutsu shiken ). The comprehensive test (sogo shiken ) consists of writing an essay on a given topic, such as "Japan and the security and prosperity of Asia," "on security" (anzenhosho o ronzu ), and "Japan and Southeast Asia." The candidate is expected to demonstrate his mastery of essential information, analytic acumen, and ability to think independently. The

[73] Juken Shinpo Henshubu, Gaikokan shiken mondaishu [Collection of Questions in the Foreign-Service Examination] (Tokyo: Hogaku Shoin, 1987), pp. 9-28 and 61-62.

[74] Ibid., p. 327.

character test has two components: an interview and a group discussion. In the latter, eight to ten applicants are assembled in a room, given a problem (usually a diplomatic one), and asked first to state their individual views and then to engage in discussion among themselves. Successful candidates stress the importance of striking a balance between being able to articulate one's own views clearly and showing respect for the views of other participants.^[75]

To return to table 17, we see in column C that only about two dozen or so of the applicants pass the examination each year; hence the competition rate is quite high (see column E). What is more, unlike the situation in the higher civil-service examination, nearly all successful candidates are hired. In fact, all of the handful of the successful candidates who were not hired in the period covered by the table had voluntarily withdrawn from further consideration.

What of the university background of successful candidates? In the period from 1945 to 1955, 78.1 percent of the successful candidates (100 out of 128) came from Todai. The remainder came from Tokyo University of Foreign Languages (N = 6, 4.7 percent), Kyodai (N = 5, 3.9 percent), Tokyo College of Commerce [the predecessor of Hitotsubashi University] (N = 4, 3.1 percent), Harbin University (N = 4, 3.1 percent), and a handful of other institutions.^[76] If we look at a later period, 1965-74, we find a marked decline in the share of Todai graduates—49.4 percent (121 out of 245)—and an increase in the share of the other universities—Hitotsubashi: 15.5 percent (N = 38); Kyodai: 11.8 percent (N = 29); Keio: 8.2 percent (N = 20); and Tokyo University of Foreign Languages: 4.5 percent (N = 1).^[77]

In 1978, Todai maintained its top position with 44.4 percent (12), followed by Hitotsubashi, 14 percent (4). Kyodai tied for third place with Tokyo University of Foreign Languages with 11.1 percent each (3 each).^[78] In 1982 Todai remained at the top with 53.8 percent (14), and Hitotsubashi and Kyodai retained their traditional second and third spots, with 15.4 percent (4) and 11.5 percent (3), respectively.^[79] In 1986 Todai produced 71.4 percent (20) of the successes, and Waseda and

[75] Ibid., pp. 63, 75-90, 194-97.

[76] Nagano Nobutoshi, Nihon gaiko no subete [All About Japanese Diplomacy] (Tokyo: Gyosei Mondai Kenkyujo Shuppankyoku, 1986), p. 328.

[77] Nagano Nobutoshi, Gaimusho kenkyu [A Study of the Ministry of Foreign Affairs] (Tokyo: Saimaru Shuppankai, 1975), p. 42.

[78] Mainichi shinbun , 28 Oct. 1978.

[79] Ibid., 23 Oct. 1982.

Kyodai tied for second place with two successes each. Four other universities supplied one successful candidate each: Hitotsubashi, Keio, Aoyama Gakuin, and Harvard.^[80] In 1987 only four universities supplied all 26 successful candidates—Todai: 19 (73.1 percent); Kyodai: 3; Hitotsubashi: 3; and Waseda: 1. In 1988 the number of universities supplying successful candidates remained the same—Todai: 17 (68.0 percent); Hitotsubashi: 4; Kyodai: 2; and Keio: 2.^[81] In the eleven-year period 1978-88, Todai produced 60.3 percent (179) of all successful candidates, and Hitotsubashi produced 11.4 percent (34). Kyodai placed third with 9.4 percent (28) of the total successes.^[82]

In sum, not only is the phenomenon of Todai dominance replicated in the recruitment of diplomats, but its degree is somewhat accentuated. Hitotsubashi, previously known as Tokyo College of Commerce (1920-44) and Tokyo College of Industry (1944-49), is a prestigious national university with a tradition of excellence in business and economics.^[83] All in all, the pattern of dominance by national universities continues in the foreign-service field as well.

How have women fared in this exclusive domain? Until the mid-1970s, only two women had passed the higher-level foreign-service examination, but beginning in the latter half of the 1970s, there were one or two women in each year's career-track entering class in the Foreign Ministry. In the nine-year period 1978-87, a total of sixteen women passed the higher-level foreign-service examination. The grand total of all women who have passed the examination in the postwar period (1949-88), however, is a mere twenty. In 1980 and 1986, there were three women among the successful candidates in the examination. In 1983, however, there were none. In 1987 one of the twenty-six successful candidates was a woman. The twenty-five successful candidates in 1988 also included a woman. Of the sixteen women who passed the examination in the 1978-87 period, five were graduates of Todai and two others had done graduate work there. Five were graduates of Hitotsubashi. The remaining four had studied abroad—two at Colum-

[80] Nihon keizai shinbun , 6 Oct. 1986 (evening ed.).

[81] Ibid., 15 Sept. 1987; Asahi shinbun , 7 Sept. 1988 (evening ed.).

[82] Nagano Nobutoshi, "Za Kasumigaseki (9): Gaimusho: Nihon gaiko o ugokasu pawa" [Kasumigaseki Series (9): The Ministry of Foreign Affairs: The Power That Moves Japanese Diplomacy], Kankai (Mar. 1986): 39; Yomiuri shinbun , 7 Oct. 1985 (evening ed.); Nihon keizai shinbun , 6 Oct. 1986 (evening ed.), 15 Sept. 1987, and Asahi shinbun , 7 Sept. 1988 (evening ed.).

[83] Sogo Gakusei Mondai Kenkyujo, ed., Nihon daigaku taikan [An Overview of Japanese Universities] (Tokyo: Nihon Gakujutsu Tsushinsha, 1973), p. 117.

bia University, one at Harvard University, and one at Oxford University in England.^[84]

In the entire postwar period, no woman in the career foreign service has ever achieved ambassadorial rank; the highest rank attained by a woman was that of minister. Although two women have served as ambassadors, neither was a career diplomat.^[85] Given the need for extended service abroad at periodic intervals, careers in the foreign service would be even more difficult to combine with the burdens of a family life than those in the civil service in general. Hence the long-term prognosis for women in the foreign service does not appear to be good.

International Comparisons

How does Japan's record in the recruitment of its higher civil servants compare with those of other industrialized democracies, specifically, the United States, Great Britain, France, and West Germany? Each country has a different procedure for recruiting its higher civil servants.

Although the picture in the United States is somewhat complicated by the existence of multiple channels of recruitment, including political appointment, the inauguration in 1977 of the Presidential Management Intern Program (PMIP) has spawned a method that comes closest to those employed by the other industrialized democracies in selecting their respective elite-track civil servants.

Presidential Management Interns (PMI) are selected through a four-stage process. In the first stage, deans and directors of graduate programs in "general management with a public focus" nominate a limited number of their top students. Next, the U.S. Office of Personnel Management (OPM), the successor to the U.S. Civil Service Commission, "reviews each nominated application to determine if the nominee meets the basic eligibility requirements." Then follows the regional screening process, during which eligible nominees participate in "group and individual problem-solving exercises, through which they will be evaluated on their leadership, oral communication and interpersonal

[84] Nagano, "Za Kasumigaseki (9): Gaimusho," p. 99; Nihon keizai shinbun , 6 Oct. 1986 (evening ed.); Yomiuri shinbun , 15 Sept. 1987. The Harvard graduate was pursuing her second bachelor's degree at Todai's Law Faculty when she passed the examination in October 1986.

[85] Shukan Yomiuri , 19 June 1985, p. 165; Asahi shinbun , 13 Nov. 1985 and 2 Feb. 1986; Nagano Nobutoshi, "Kankai jinmyaku chiri—chuo shochohen: Gaimusho no maki" [Who's Who in Government: Central Ministries and Agencies: The Ministry of Foreign Affairs], Kankai , Apr. 1987, pp. 40-54.

skills, and their organizing, planning, problem-solving and decision-making abilities." In addition, nominees must complete a writing sample. They are divided into panels of not more than eight students, each of which is rated by a panel of two to four public managers. These panels are convened at approximately thirty-five sites across the country. In the final stage of the competition, all nominees who have received qualifying ratings in the third stage have their writing sample and credentials evaluated by the OPM, and about 250 finalists are selected for the PMIP each year.^[86]

The responsibility of obtaining an actual appointment falls on the shoulders of the finalists. Although all of them find jobs, competition for certain agencies and locations is intense. PMIs start at the grade-9 level and can expect to be promoted to grade 11 after completing one year. After completion of the two-year program, the interns are eligible to advance to grade 12 with career or career-conditional status.^[87]

In Great Britain, until the late 1960s, induction into the administrative class was required of aspiring higher civil servants, and this was done by two different methods of competitive examination. Following the publication of the report of the Fulton committee in 1968, however, the administrative class was abolished. Under the new system, higher civil servants are recruited as administrative trainees (AT). Applicants must undergo a three-stage screening process. First, they must take the qualifying test, which "consists of a series of written papers. Typically, a candidate must: reduce an article of 1,500 words to a summary of 300 to 350; give advice on an imaginary government problem . . .; analyse some tables of social statistics; and answer tests on verbal intelligence, comprehension and data sufficiency."^[88]

About three-quarters of the candidates are eliminated in this initial stage. Those who survive then face two and a half days of grueling examination by the Civil Service Selection Board, popularly known as the Cizbee. Candidates are divided into groups of five or six candidates, each of which is evaluated by a panel of three judges—"a chairman, who is usually a retired permanent secretary; a psychologist; and an observer, who is a youngish civil servant who passed Cizbee a few years ago." The tasks the candidates are required to perform include writing

[86] U.S. Office of Personnel Management, The Presidential Management Intern Program, 1980-1981 (Washington: Government Printing Office, 1981), pp. 10-11.

[87] Ibid., pp. 12 and 2.

[88] Peter Kellner and Lord Crowther-Hunt, The Civil Servants; An Inquiry into Britain's Ruling Class (London: MacDonald Futura Publishers, 1980), p. 120.

essays and position papers, participating in group discussion, engaging in "committee exercises," undergoing personal interviews, and taking intelligence and general-knowledge tests. Approximately half of the candidates pass the Cizbee, and then must go to "a short formal interview with the Final Selection Board (FSB). Almost 80 percent of the candidates who reach the FSB are offered jobs as ATs or equivalent posts in the Diplomatic Service or in the Inland Revenue Service.^[89] Between 250 and 300 ATs are appointed each year, of whom a substantial number (about 100) are "internal candidates," that is, incumbent civil servants in executive-level positions who want to speed up their promotion process.^[90]

The gateway to the French higher civil service is admission to the Ecole Nationale d'Administration (ENA). There are two types of entrance examination to ENA, one for external candidates and another for internal candidates. External candidates must be twenty-five years of age or less and graduates of a university or a grande école . Internal candidates (those who are already in the civil service) must be aged 30 or less, graduates of a university or a grande école , and have at least five years of experience in a public-sector job. Between 20 and 30 percent of the successful candidates come from the internal competition. In addition, one or two students are admitted without examination from the Ecole Polytechnique.^[91] From 1982 to 1985, ten places were set aside for "union leaders and local elected officials, such as small-town mayors, who took different—and, some said, easier—examinations from those required of other applicants." This method of admission, called the "third way," was abolished in 1986.^[92]

There are two main stages in the entrance examination to ENA. In the first stage, candidates must take a series of written examinations, each lasting from three to five hours. Depending on whether one

[89] Ibid., pp. 124-36.

[90] Oki Yu, "Igirisu seifu shokuin no nin'yo seido" [The Appointment System of Government Employees in England], II, Jinji-in geppo 292(June 1975): 29; Tashiro, "Nihon gyoseikan kenkyu, VII," p. 77. In the British civil service the term "executive" does not carry the same exalted connotation as it does in other countries. Executive positions refer to middle-level positions occupied primarily by nonuniversity graduates.

[91] Nomura Nario, "Furansu kanri no nin'yo seido" [The Appointment System of French Officials], II, Jinji-in geppo 220(June 1969): 14 and 24; Henry Parris, "Twenty Years of l'Ecole Nationale d'Administration," Public Administration (London) 43 (Winter 1965): 395-411. The proportion of external candidates has steadily declined over the years. It was 52 percent at its peak (1953). The age limit of external candidates is not strictly enforced.

[92] New York Times , 24 Aug. 1986.

chooses law or economics as the field of specialization, the subjects of the examination may vary. Applicants in law, for example, are tested in public law; analysis of legal material; social problems involving the application of law, social policy, or international problems; and economics. They must also take an essay test in a subject chosen from business law, economic geography, modern history, linguistics, sociology, psychology, mathematics, statistics, and other fields. There is then a common five-hour-long essay test for both law and economics candidates covering contemporary political, economic, social, or international problems. Only those who pass the written examinations are allowed to proceed to the second stage, which consists primarily of oral examinations. In addition to the subjects covered in the first stage, they include foreign-language examinations. In addition, there is a physical-fitness test consisting of a short-distance dash, a long jump, a discus throw, chin-ups, and swimming. Although 160 candidates succeeded in clearing all these hurdles and were admitted to the prestigious ENA in 1985, the number was scheduled to be reduced by 50 percent by 1987.^[93]

In West Germany, aspirants for a career in the higher civil service must undergo even more stringent processes of selection and training. First, they must pass a state examination that is equivalent to a university graduation examination. Those who pass may then begin two and a half years of preparatory service in a government agency. During this period the trainees do not receive a salary but modest stipends for living expenses. Upon completion of such service, they are allowed to take the second state examination. Successful candidates are certified as full-fledged jurists (Volljurist ) and earn the title of assessor. An "Assessor can apply to a public authority of his choice for an administrative post in the higher grade." If accepted, he becomes a trainee. Only after a probationary period, which may last three years, will he be appointed as a permanent official.^[94] In other words, a university graduate must pass two separate state examinations and two

[93] Nomura, "Furansu kanri no nin'yo seido," II, p. 19; Suzuki Kazuo, "ENA (Furansu Kokuritsu Gyosei Gakuin) no kyoiku seido" [The Educational System of ENA (the French National School of Administration)], II, Jinji-in geppo 321(Nov. 1977): 23-24; Asahi shinbun , Aug. 1986; New York Times , 24 Aug. 1986.

[94] Nevil Johnson, State and Government in the Federal Republic of Germany: The Executive at Work (Oxford: Pergamon Press, 1983), p. 183; Sumitomo Tadashi, "Nishi Doitsu ni okeru kanri nin'yo seido" [The Appointment System of Officials in West Germany], Jinji-in geppo 131(Jan. 1962): 7.

probationary periods spanning five and a half years to gain entry into the higher civil service on a permanent basis.^[95]

What these methods of recruiting higher civil servants have in common are their selectivity and reliance on extended training periods. In both absolute and relative terms, the number of people selected is substantially fewer in these countries than is the case in Japan. As we shall see later, moreover, the length and content of training received by graduates of the type-I (or type-A) higher civil-service examination cannot even begin to compare with those of their counterparts in the Western democracies. In one sense, the Japanese system appears to be more open than those of America and Western Europe. For there is no formal educational requirement in Japan. The practical significance of this formal flexibility evaporates quickly, however, when the actual track record is examined.

Moreover, the rigor with which the track system is adhered to in Japan has no parallel elsewhere. Internal candidates in both Britain and France, are allowed to compete with external candidates for admission to the elite corps. In France, in particular, between a quarter and a third of successful candidates in the entrance examination to ENA are internal candidates. As for the United States, the apparent rigidity of the PMIP's admission requirements is diluted by a number of considerations. First, there is an abundance of opportunities, in terms of the sheer number of programs and their flexibility, for obtaining the necessary credentials even for incumbent government officials. Second, the manner in which the selection process operates is calculated to ensure equality of opportunity to a very high degree. Third, there is a great deal of mobility, both vertical and lateral, in the American federal bureaucracy. On the other hand, West Germany seems to come close to matching Japan's rigid stratification, for it provides relatively little opportunity for those who have not undergone the arduous process of state examinations and probationary periods to attain the upper rungs of its bureaucratic ladder. However, the rigidity of the West German

[95] For descriptions of the West German system, see, in addition to the work cited above, David Southern, "Germany" in F. F. Ridley, ed., Government and Administration in Western Europe (New York: St. Martin's Press, 1979), pp. 137-43; Klaus von Beyme, The Political System of the Federal Republic of Germany (New York: St. Martin's Press, 1983), pp. 149-54; Hans Joachim yon Oertzen, "Republic Personnel Management in the Federal Republic of Germany," International Review of Administrative Sciences 49, no. 2 (1983): 210-17; Sumitomo Tadashi, "Nishi Doitsu renpo kanri no nin'yo seido" [The Appointment System of West German Federal Officials], 2 parts, Jinji-in geppo 215(Jan. 1969): 6-9, and 217(Mar. 1969): 12-15.

system is tempered by an extensive use of political appointment in the top echelons of its bureaucracy.^[96]

What of the phenomenon of domination by elite universities? Both Britain and France manifest a similar phenomenon. From 1948 to 1963, Oxford and Cambridge Universities provided 81 percent of successful candidates in methods I and II competition for direct entrants to the administrative class. But unlike the situation in Japan, the two top universities split the successes fairly evenly, Cambridge trailing behind Oxford by the ratio of four to six.^[97] After the reforms of the late 1960s, the Oxbridge dominance seemed to slacken slightly. In the mid-1970s, about 65 percent of external candidates for administrative trainees who passed the Final Selection Board came from the two universities. Oxbridge's share decreased in the late 1970s to slightly over 50 percent.^[98]

With regard to France, between 70 and 90 percent of successful external candidates in the entrance examination to ENA are products of Instituts d'Etudes Politiques (IEP), of which the institute situated in Paris supplies over 90 percent. When successful internal candidates are added, the share of the IEPs drops to about 60 percent. The IEPs, which are state institutions, in effect serve as preparatory schools for ENA entrance examinations; as such they attract students and graduates of both universities and grandes écoles .^[99]

Turning to the United States, we find nothing that can match the elitism of Japan, Britain, and France. A 1959 survey found that of the 7,640 career civil servants holding the rank of GS 14 or higher, no more than 3 percent came from any single university. In fact, it was not an Ivy League university but George Washington University and the City College of New York that shared the top positions, with 3 percent each. The top ten universities together supplied only 21 percent of the career civilian executives, and the top thirty universities brought the proportion to a mere 42 percent. Only in the Foreign Service did Ivy League universities appear at the top. But Harvard, Yale, and Princeton together were the source of only 14 percent of senior Foreign Service

[96] Southern, "Germany," p. 142.

[97] Geoffrey K. Fry, Statesmen in Disguise: The Changing Role of the Administrative Class of the British Home Civil Service, 1853-1966 (London: Macmillan, 1969), p. 435. The statistics were calculated from table 5.

[98] Kellner and Crowther-Hunt, The Civil Servants , pp. 120-23 and 137.

[99] Suzuki, "ENA no kyoiku seido," III, Jinji-in geppo 322(Dec. 1977): 19; Ezra N. Suleiman, Politics, Power, and Bureaucracy in France: The Administrative Elite (Princeton: Princeton University Press, 1974), p. 54; New York Times , 24 Aug. 1986.

officers.^[100] The PMIP became even more egalitarian: 250 finalists chosen in 1978 were spread among 127 different graduate schools, and the same pattern was repeated in subsequent years.^[101] In West Germany the problem does not arise at all, because students are allowed to take courses in any of the state universities and many of them attend several universities during their academic careers. There is no one university that is dominant in terms of reputation.^[102]

Is the predominance of law graduates replicated outside of Japan? Here West Germany resembles Japan closely, which is not surprising in view of the German influence on the prewar Japanese bureaucracy. Although hard data are not available, the legalistic bias in German administration is well known. In the words of Herbert Jacob: "German administration . . . went much further in its obeisance to legal formulas than most administrative organizations, for its officials were specifically trained in the law . . . . The university training received by German administrators concentrated heavily in law."^[103] Speaking of the postwar period, Lewis J. Edinger writes that "a civil servant, who wants to rise above a medium-grade position, no matter what his task will be, must usually have a degree in law."^[104] Finally, Nevil Johnson estimated in 1983 that about two-thirds of West German higher civil servants might be lawyers.^[105]

It is worth noting that, as is true in Japan, a law degree in and of itself has no professional significance in West Germany. In order to enter the legal profession, the candidate must undergo the same procedure as the candidate for the higher civil service: two state examinations and two prolonged periods of postgraduate training. What sets West German law graduates apart from their Japanese counterparts, however, is that the former must, in effect, qualify for a professional career in law before they can become full-fledged members

[100] Lloyd W. Warner et al., The American Federal Executive (New Haven: Yale University Press, 1963), p. 372, table 5lB.

[101] U.S. Office of Personnel Management, The Presidential Management Intern Program: Anniversary Report, 1978-1980 (Washington: Government Printing Office), pp. 3 and 21-23.

[102] Tashiro Ku, "Nihon gyoseikan kenkyu, IV: Gyoseikan no kyaria keisei katei, 2," Kankai , Jan. 1982, p. 92.

[103] Herbert Jacob, German Administration Since Bismarck: Central Authority Versus Autonomy (New Haven: Yale University Press, 1963), p. 203.

[104] Lewis J. Edinger, Politics in Germany: Attitudes and Processes (Boston: Little, Brown, 1968), p. 46.

[105] Johnson, State and Government in the Federal Republic of Germany , p. 183. Another scholar writes that depending on how one defines a "lawyer," the proportion can range from 46 to 85 percent. See Southern, "Germany," p. 139.

of the higher civil service. Interestingly, there is even a German phrase that is comparable to hoka banno , namely, Juristenmonopol (monopoly by lawyers). The Japanese attempt to alleviate the problem has also been duplicated in West Germany; there are laws and regulations that are designed to place those who have majored in economics, finance, and sociology on a par with law graduates in their competition for the higher civil service.^[106]

As already noted, candidates for admission to the French ENA must choose between two fields: law and economics. About 70 percent of them opt for law.^[107] In the United States, where a law degree is all but synonymous with a license to practice law, it is not a significant factor in the higher civil service. The study by Warner and his associates cited earlier found that less than 8 percent of the senior career civil servants had law degrees. Overall, only 12 percent of all federal executives, both civilian and military, in the sample (1,582 out of 12,929) had law degrees. So far as the sources of the law degrees were concerned, only three law schools had produced more than 5 percent: George Washington (18.6 percent), Georgetown (6.9 percent), and Harvard (6.6 percent). The fact that two of these law schools are situated in Washington, D.C., suggests that many Federal executives may have earned their law degrees after entering the government service.^[108]

The PMIP, the most recent vehicle of elite recruitment in the U.S. federal government, is aimed explicitly at attracting candidates with graduate training in management. Sixty-eight percent of the 250 interns selected in 1978, for example, had graduate degrees in public administration. The remainder were spread among business administration (8 percent), urban studies (6 percent), international studies (3.2 percent), public policy (2.8 percent), planning/resource management (2.8 percent), and other related fields (such as general management, criminal justice, social work, and political science).^[109]

The British administrative class was traditionally dominated by those who had majored in history and classics. Fifty-eight percent of direct entrants to the class in the 1948-56 period (methods I and II combined) and 51 percent of them in the 1957-63 period had taken their degrees in these two subjects. The proportion of those who had taken their degrees

[106] Sumitomo, "Nishi Doitsu renpo kanri no nin'yo seido," I, p. 9.

[107] Nomura, "Furansu kanri no nin'yo seido," II, p. 25.

[108] Warner et al. The American Federal Executive , p. 369, table 48B.

[109] U.S. Office of Personnel Management, The Presidential Management Intern Program: Anniversary Report, p. 3.

in law during the two periods were 3 and 5 percent, respectively.^[110] The post-Fulton reforms have not produced any substantial change; although the proportion of graduates in history and classics has declined, the majority of successful candidates for administrative-trainee positions have nonetheless continued to be "arts" (that is, humanities) graduates.^[111]

The underrepresentation of women in the higher echelons of government bureaucracy is a universal phenomenon. In none of the Western industrialized democracies have women achieved parity with men in the higher civil service. However, the degree to which women are under-represented varies widely. In general, women have been making notable progress in the United States. Warner and his colleagues found in their 1959 survey that only 145 of the 10,851 civilian federal executives who responded to their questionnaires were women, "about one in seventy-five of the civilian federal executives."^[112] Two decades later the first class of the PMIs contained 116 women, accounting for 46 percent of the total. Twenty-one percent of them were nonwhites.^[113]

In France, the proportion of women among those who are admitted to the ENA has steadily increased. It was about 13 percent in the late 1970s and 25 percent in the mid-1980s.^[114] In Britain, women first gained entry into the administrative class during World War I on a temporary basis and then became eligible to compete for AC positions in the mid-1920s. From 1925 to 1945, however, the proportion of women among successful candidates for the administrative class remained at or below 7 percent. During the postwar period, the proportion increased to about 10 percent, a situation that has remained unchanged in the post-Fulton era.^[115]

With respect to the other aspects of recruitment, there are certain parallels and contrasts between Japan and the other countries. The 20-percent no-show rate in Japan's higher civil-service examinations,

[110] Fry, Statesmen in Disguise , p. 436. Percentages were calculated from table 6.

[111] Kellner and Crowther-Hunt, The Civil Servants , pp. 120-23; Oki Yu, "Igirisu seifu shokuin no nin'yo seido," I, p. 18.

[112] Warner et al., The American Federal Executive , p. 177.

[113] U.S. Office of Personnel Management, The Presidential Management Intern Program: Anniversary Report , p. 3.

[114] Suzuki, "ENA no kyoku seido," I, p. 25; New York Times , 24 Aug. 1986.

[115] Kelsall, Higher Civil Servants in Britain , pp. 169-77; Kellner and Crowther-Hunt, The Civil Servants , p. 12. Regarding West Germany, the proportion of women in the federal civil service as a whole is said to be about 21 percent. However, their percentage in the higher civil service is unknown. Oertzen, "Public Personnel Management in the Federal Republic of Germany," p. 212.

for example, is replicated in England and France. The effective rate of competition also appears to be equally high in all three countries, although the Japanese rate, particularly in recent years, remains unsurpassed. When it comes to the rate of appointment, however, the picture changes markedly. In general, the rate is much higher in the other countries than it is in Japan. Nearly all successful candidates receive civil-service appointments in the Western democracies, whereas only half of them do so in Japan.

In sum, international comparisons suggest that most of the phenomena that characterize Japanese patterns of recruiting higher civil servants are not confined to that country. Nonetheless, the variations among the five countries in terms of the salience, pervasiveness, and intensity of some of the phenomena should not be overlooked. In general, the Japanese system emerges as somewhat more closed and less egalitarian than its Western counterparts. Before we can make an overall comparative assessment, however, we need to examine the other dimensions of the Japanese higher civil service.