Chapter Eight
Rewards

In order to attract and keep qualified men and women, bureaucracy needs to have an adequate system of rewards. The rewards may be either tangible or intangible, immediate or deferred. Tangible rewards include salary and fringe benefits, and intangible rewards encompass a sense of service, the prestige and power of office, and the like. If, in addition to all this, service in government bureaucracy can facilitate subsequent career advancement, then it generates long-term rewards.

In this chapter we shall examine both the immediate and the deferred rewards of Japanese bureaucracy. We shall first note a few salient aspects of the compensation system. Second, we shall touch briefly on the question of corruption among Japanese civil servants. Third, we shall discuss the reemployment of retired higher civil servants. Finally, we shall try to place the Japanese case in comparative perspective.

Compensation

Salary Schedules

As we saw in chapter 4, Japan's national public employees are divided into a number of different categories; for purposes of compensation, the two most important categories are (1) employees who are subject to the regular compensation law (kyuyoho shokuin ) and (2) those governed by the special compensation law (kyuyo tokureiho shokuin ). Because the

latter consist mainly of postal employees, however, we shall be concerned primarily with the former.

The regular-compensation-law employees are divided into seventeen salary schedules (hokyuhyo ). Table 31 displays selected attributes of each schedule. First, by far the largest is the Administrative Service (gyosei shoku ) I, which accounts for nearly half of all public employees covered by the regular-compensation law. As we have previously noted, the higher civil servants with whom this study deals belong for the most part to this category.

Second, in terms of average age, the Designated Service (shitei shoku ) surpasses all the others, because this category consists of administrative vice ministers and their equivalents, bureau chiefs, university presidents, directors of research institutes and hospitals, and other high-ranking officials. Given the rule of lifetime employment, it is not surprising that average age and average length of service are highly correlated.

Third, the proportion of college graduates varies widely from schedule to schedule. The top three categories are the Designated Service; the Medical Service (iryo shoku ) I, which is made up of physicians and dentists; and the Educational Service (kyoiku shoku ) I, which consists of college professors. Overall, the level of educational attainments is quite high; one of every three civil servants is a college graduate, and well over half have had at least two years of college.

Finally, the proportion of women, too, varies widely from schedule to schedule. Although there are no women in the Marine Service (kaiji shoku ) II, they all but monopolize the Medical Service III, which is made up of midwives, nurses, and nursing assistants.

Each of these salary schedules encompasses a number of grades, each of which in turn subsumes a number of steps. Until 1985 the Administrative Service I salary schedule consisted of eight grades. Effective 1 July 1986 the number of grades increased to eleven. Table 32 presents basic data on the revised structure of the Administrative Service I. Whereas, in the old system, the lowest grade was 8 and the highest 1, the new system has reversed the order: the higher the grade level, the larger the number.

The number of steps within grades ranges from fifteen to twenty-eight, which accounts for the wide range in monthly salaries within each grade. Step 16 in grade 1, for example, pays more than step I in grade 3; step 28 in grade 4 pays more than step 1 in Grade 10; and step 24 in grade 6 pays more than step 1 in grade 11, the highest grade. This type of pay structure is designed to benefit "noncareer" civil servants, whose

promotional opportunities are limited. Noncareerists tend to stay in the same grade substantially longer than "career" civil servants, hence they have a chance to climb to the highest step within each grade. Thus a veteran noncareerist who is a subsection chief usually draws a higher salary than a careerist section chief to whom he reports, and a veteran assistant section chief may be better off financially than a careerist assistant bureau chief.^[1]

A few other aspects of the compensation system emerge from a review of statistics on civil-servants' compensation in two different years, 1958 and 1976.^[2] First of all, if we control for the level of educational attainments, we find a linear relationship between seniority and pay. Although no comparable data are available for the 1980s, there is no reason to assume that the situation has changed to any appreciable degree. Second, in both 1958 and 1976, women lagged behind men in compensation. What is striking is that the discrepancy appears after the effects of education have been neutralized. Becuase the salary schedules per se make no distinctions between the sexes, the data

[1] The Designated Service Salary Schedule, to which bureau chiefs and administrative vice-ministers belong, has twelve steps and the monthly salary range of 468,000 yen and 1,065,000 yen. Jinji-in geppo 440 (Sept. 1987): 28.

[2] Jinji-in, Nenji hokokusho , 1958, pp. 48-49; Jinji-in, Kyuyokyoku, Kokka komuin kyuyoto jittai chosa hokokusho , 15 Jan. 1976, pp. 38-39.

suggest that women are concentrated on the lower rungs of the ladder and receive stepwise promotions within grades at a slower pace than men.

On the other hand, the gender gap in compensation became somewhat narrower by the latter part of the 1970s than it was in the late 1950s. More important, the gap was closed completely for women with two years or less of service. It needs to be stressed, however, that the data on which these observations are based pertain to the civil service as a whole. If we look at the Administrative Service I only, the gender gap remains. Although no data are available for 1958, those for 1976 show that women lagged behind men in all categories of seniority; the first two categories nonetheless showed the smallest gaps: 0.98 for one year or less and 0.964 for one to two years.^[3] The lack of more recent data prevents us from knowing whether there has been further progress on this front.

Finally, it should be pointed out that the monthly salary figures shown in table 32 and discussed in the preceding paragraphs pertain only to the basic salary (honpo ). When the various allowances are added, the figures increase by 10 to 20 percent. The allowances include the dependents' allowance (fuyo teate ), the housing allowance (jukyo teate ), the commuter allowance (tsukin teate ), and the adjustment allowance (chosei teate ) paid to civil servants assigned to areas where living costs are unusually high.^[4] In addition, there are overtime pay and bonuses.

Although higher civil servants express embarrassment at their relatively meagre income, financial reward does not appear to figure prominently in their incentive structure. When Koitabashi Jiro, a free-lance journalist, interviewed twenty section chiefs in the various ministries and agencies, all belonging to the postwar generation, in 1985, he found that very few could cite reasonably accurate figures for their salaries, an indication of the relative unimportance of the issue. Most reported, however, that their income was between a half and two-thirds of what their peers (former college classmates) in major private firms were making. According to a section chief in the Transport

[3] Ibid., pp. 40-41. Because comparable data are not available for 1958, all civil servants, instead of those in the Administrative Service (I), were used as the data base for the preceding discussion.

[4] "Komuin kyuyo no gaiyo" [An Outline of the Compensation of Civil Servants], Jinji-in geppo 402 (July 1984): 24 and 26; "Kokka komuin kyuyoto jittai chosa" [An Investigation into the Actual Conditions of National Civil Servants' Compensation and Related Matters], ibid., 420 (Jan. 1986): 7-11.

Ministry, bureaucrats sometimes earned extra money by writing magazine articles using the name of their superior, such as a bureau chief, not to enrich themselves but to generate funds with which to purchase meals at night for their unit, which is a far cry from the affluent world of private-sector managers who have the perquisities of expense accounts and business-related entertainments.^[5]

The Recommendation System

A novel feature of the compensation system for Japanese civil servants has to do with the role of the National Personnel Authority (Jinji-in) in the determination of salary schedules. Both the national public-service law (kokka komuinho ) and the law concerning the compensation of regular service employees (ippan shoku no shokuin no kyuyo ni kansuru horitsu ) require the National Personnel Authority to conduct surveys to determine whether salary rates for public employees are appropriate and to report simultaneously to the Diet and the cabinet on its recommendations for revision should an adjustment of 5 percent or more either upward or downward be deemed necessary.^[6]

The right, or obligation, to submit recommendations directly to both the Diet and the cabinet is significant, for no other agency belonging to the executive branch of the Japanese national government enjoys such a privilege. This was designed to bolster the independence of the National Personnel Authority in carrying out its mandate; the principal rationale for the recommendation system, which was initially imposed by the American Occupation authorities on the Japanese government, was to compensate for the restrictions on the exercise of constitutional rights by civil servants. Although article 28 of the Japanese constitution explicitly guarantees "the right of workers to organize and to bargain and act collectively,"^[7] civil servants were forbidden to bargain collectively for their salary rates and, especially, to engage in strikes. It was therefore considered imperative to have a mechanism under which the right of civil servants to have a fair and equitable system of compensation would be protected. The recommendation system, in other words,

[5] Koitabashi, "Sengo umare" erito kanryo no sugao , pp. 52-66.

[6] See art. 67 of the National Public Service Law and art. 2, para. 3 of the Compensation Law in Jinji-in, Nin'yokyoku, ed., Ninmen kankei horeishu , 1984 ed., pp. 25 and 52.

[7] See the English version of the constitution in Suekawa, ed., Iwanami kihon roppo , 1974 ed., p. 114.

was devised in part as a functional equivalent to wage determination by collective bargaining.^[8]

The compensatory function of the recommendation system was explicitly recognized by the Japanese Supreme Court in April 1973. The court noted that the guarantee of fundamental rights for workers in article 28 of the constitution was not an end in itself but a means to the advancement of the workers' economic position, hence it could be restricted in the common interest of all the citizens. When that occurs, however, "compensatory measures" (daisho sochi ) become necessary. According to the court, detailed regulations regarding civil servants' status, appointments, dismissals, compensation, and other conditions of work and the establishment of a central personnel administrative agency endowed with quasi-judicial powers to oversee the implementation of these regulations, particularly those relating to compensation, constituted such compensatory measures. Two judges held in a supplementary opinion that civil servants had a right to engage in limited "dispute behavior" (sogi koi ) to the extent that the preceding compensatory measures failed to perform their intended functions. All the responsible authorities had an obligation to make sincere efforts within the limits of law and available resources to make the system work, they added.^[9] This latter principle was reaffirmed in a more recent ruling by the Tokyo Higher Court, which held in November 1985 that the nonimple-mentation of the National Personnel Authority's recommendation on compensation might under certain circumstances justify strikes by civil servants.^[10]

Under the recommendation system the National Personnel Authority has conducted surveys of compensation in the private sector every year with the aim of ascertaining whether government pay rates are

[8] Asai Kiyoshi, "Omoide no ki: Kokka komuinho no seitei to dai ikkai kyuyo kankoku" [Reminiscences: The Enactment of the National Public Service Law and the First Recommendation on Compensation], in Jinji-in, Jinji gyosei sanjunen no ayumi , pp. 38-42; Sakaiya Taiichi and Utsumi Hitoshi, "Taidan: Komuin kyuyo o kangaeru" [Conversation: Thinking About Civil Servants' Compensation], Jinji-in geppo 402 (July 1984): 4-5.

[9] For excerpts from the Supreme Court decision, see Jinji-in geppo 402 (July 1984): 11-13. For a discussion of the case and its implications, see "Shinso o saguru: Ima magari kado ni tatsu Jinji-in" [In Search of Truth: The National Personnel Authority at a Turning Point], Kankai , Oct. 1984, pp. 134-43. Written by the editorial department of the magazine, this article construes the drastic cuts in the authority's recommended pay increases in the early 1980s as a sign of its decline in power.

[10] Asahi shinbun , 20 Nov. 1985 (evening ed.). For excerpts from the decision, see ibid., p. 2.

comparable to private-sector pay rates for the same levels of work. In 1985, for example, the authority collected data on compensation from 40,660 places of work (jigyosho ) with fifty or more full-time workers that were part of enterprises employing at least one hundred full-time workers. All told, data were collected on more than half a million persons spanning ninety-one different occupational categories.^[11] As was the case in previous surveys, the data showed that government pay rates lagged behind private-enterprise pay rates.

On the basis of detailed comparisons, controlling for type of jobs, position level, education, and age, the authority recommended an average pay increase of 5.74 percent for civil servants, effective 1 April 1986. In December 1985 the cabinet formally approved the recommendation on condition that the new rates enter into force on 1 July instead of 1 April 1986.^[12] This meant, in effect, that the authority's recommendation would be diluted by 25 percent.

The record of the authority in having its recommendations implemented is somewhat mixed. Since the adoption of the recommendation system in 1948, the authority's recommendations were totally rejected twice, in 1949 and 1982, and revised six times. During the early 1950s the revisions were relatively minor: 1 percent reduction in 1950, 11 percent reduction in 1951, and 5.1 percent reduction in 1952. In 1953, there was an unprecedented upward revision, although the increase was negligible (0.02 percent).

It was in the 1980s that the authority suffered major setbacks. As noted, its recommendation was rejected in toto in 1982. The authority's proposal in 1983 for a 6.47-percent increase was scaled down to 2.03 percent, a reduction of 69 percent. In 1984, the recommended rate was cut by 48 percent, from 6.44 to 3.37 percent. The year 1985 saw a de facto reduction of 25 percent. On the other hand, during the remaining years the authority's recommendations on the upward adjustment of pay rates were implemented in full, although some of its other recommendations were occasionally amended. In August 1986 the authority recommended that an average increase of 2.31 percent be granted to civil servants, retroactive to 1 April. Although this was the lowest rate of increase ever recommended by the authority since 1960,

[11] "Kyuyo ni kansuru hokoku to kankoku (zenbun)" [Report and Recommendations on Compensation (Full Text)], Jinji-in geppo 416 (Sept. 1985): 50.

[12] For reports on the handling of the National Personnel Authority's recommendations on compensation, see Asahi shinbun , 23 Aug., 13 Sept., 14 and 22 Oct., 5 and 6 Nov. (editorial), and 6 Dec. 1985.

the record was broken in 1987, when it recommended and the Diet approved an average increase of a mere 1.47 percent. The increase was implemented retroactively to I April 1987. The rates of increase recommended by the authority have ranged from none (1954) to 29.64 percent (1974).^[13]

Corruption

Before turning to the civil servants' retirement benefits and postretirement employment, let us examine briefly the issue of corruption. Logically, compensation and corruption are interrelated, for inadequate compensation can increase incentives for illicit financial gains. Empirically, however, the linkage between the two remains to be established. According to statistics published by the National Personnel Authority, the frequency of corruption among Japanese civil servants is remarkably low.

Table 33 presents data on two categories of behavior that pertain to corruption: embezzling public funds and accepting bribes. Other types of behavior, even when they involve public funds, were excluded from the table because they involve mistakes rather than intentional misappropriation of funds. Note also that the table pertains to all civil servants; in fact, the absence of any breakdown by rank in these data constitutes a major deficiency.

The four types of disciplinary action taken against the offending civil servants correspond with the relative severity of the breaches involved. If we look at those cases that led to the dismissal of the offenders, we find not only that their numbers are extremely small but also that they have declined over the years. On the other hand, the proportion of these most serious cases to the total cases has steadily grown: whereas they constituted less than half of the total in the 1960s, they exceeded three-quarters in the 1980s. The last column in the table shows that the decline in the frequency of corruption has not only been absolute but relative as well: whereas, in 1960, 6 of every 10,000 civil servants were disciplined for corrupt behavior, by 1975 the ratio had dropped to 2 of every 10,000, stabilizing at that level in subsequent years.

[13] For basic facts regarding the disposition of the Authority's recommendations on composation, see Jinji-in, Komuin hakusho , 1987, pp. 198-27; "Komuin seido no shikumi" [The Plan of the Civil-Service System], Jinji-in geppo 365(temporary, expanded issue; 1981): 36-37. For the 1987 recommendation, see "Kyuyo kankoku no kosshi" [The Gist of the Recommendation on Compensation], Jinji-in geppo 440 (Sept. 1987): 2, and Asahi shinbun , 12 Dec. 1987.

Because we are primarily interested in the higher civil servants, however, we need to know whether this overall picture can be extrapolated to that subset of the civil service. The available evidence bearing on this question is fragmentary. One study conducted by a research institute attached to the Ministry of Justice shows that a total of 204 civil servants were either indicted or received suspended indictments (kiso yuyo shobun ) in 1976 for bribery; of these, 31 were national civil servants. Of the latter, 8 were section chiefs or above. In

other words, the higher civil servants constituted about a quarter of the total. Of the 173 prefectural and local civil servants, however, 65 (37.6 percent) held the rank of section chief or higher.^[14]

Another indication that corrupt behavior among the higher civil servants is rare is that it literally makes headlines when it does occur. When a section chief in the Ministry of International Trade and Industry (MITI) and a subsection chief in the Medium and Small Enterprises Agency, one of the external bureaus (gaikyoku ) of MITI, were arrested in March 1986 on suspicion of having accepted bribes from an industrial organization, the news became the lead story in Asahi shinbun's evening edition on the same day. The newspaper noted that this was the first time in twelve years that a leading official of a key ministry or agency (kikan kancho no kanbu ) had been arrested on charges of corruption. The significance attached to the event was reflected in the prompt public apology by the chief government spokesman: on the same day Chief Cabinet Secretary Gotoda Masaharu apologized for "having produced arrestees," pledging to step up efforts "to establish discipline among civil servants and to restore their credibility."^[15] Although the arrested civil servants faced a prolonged legal process before their guilt or innocence could be conclusively determined, the government was clearly embarrassed by the episode.

In sum, notwithstanding their modest compensation, Japanese civil servants are remarkably free of corruption, and this is particularly true of the higher civil servants. It should be stressed that what has been discussed above pertains to personal corruption among the bureaucrats. As Chalmers Johnson points out, however, "structural corruption" has plagued the Japanese political system throughout the postwar period, and among the many causes of the phenomenon is the reemployment of retired bureaucrats by private firms, a topic to which we now turn.^[16]

Retirement And Reemployment

Monetary compensation for civil servants is only part, albeit a significant part, of the rewards they receive. They also receive various psychological rewards as well as deferred ones. The latter include separation pay and pension and also opportunities to find postretire-

[14] Sase Minoru, Nippon komuin jijo [The Condition of Japanese Civil Servants] (Tokyo: Nihon Jitsugyo Shuppansha, 1978), p. 108.

[15] Asahi shinbun , 26 Mar. 1986 (evening ed.).

[16] Johnson, "Tanaka Kakuei," pp. 1-28.

ment careers—the so-called "second careers" (daini no jinsei ). Two facts make such reemployment opportunities particularly important in the Japanese context: (1) Japanese higher civil servants retire relatively early and (2) their retirement benefits are seldom sufficient.

An Overview Of Resignations

Although a mandatory retirement system did not take effect until March 1985, most civil servants, particularly higher civil servants, retired at relatively early ages. Before focusing on the latter, however, let us first look at the overall picture. Table 34 contains some relevant statistics. Note that this table pertains to the Administrative Service I only. If the other salary schedules were added, the total numbers of resignations in the last column would increase about fourfold in the years covered.

Several things are apparent from the table. First of all, disproportionately large numbers of women resign from the civil service before age twenty-nine. Although, overall, women constitute between 15 and 20 percent of all those who resign, they account for between 36 and 55 percent of those who resign in their twenties. This is no doubt related to the difficulty of combining marriage and career in the early years.^[17] On the other hand, not only has there been a steady decline in the proportion of younger resignees as a whole, but women's share of the below twenty-nine group has dwindled markedly.

Also noteworthy is that in the 1970s the single largest age group among resignees was the thirty-fifty-four-year group. Further breakdown is available for recent years only, and it shows that, although resignations occur more frequently between the ages of forty-five and fifty-four, they also occur in significant numbers between the ages of thirty and forty-four. Another notable trend has to do with a steady increase in the proportion of resignees in the fifty-five-fifty-nine-year group. The proportion of resignees in the over-sixty group remained fairly high until 1985, when a law mandating retirement at age sixty entered into force. If we look at the age distribution of incumbent civil servants in the Administrative Service I, we find that the proportion of the over-sixty group has declined precipitately from 11.9 percent in 1973 to 0.9 percent in 1985.^[18]

In the pre-1985 period, early retirement was a function of cultural norms, pressure from superiors, and financial incentives. A key cultural norm is a strong sense of equality among cohorts: all members of the same entering class in the same ministry share such a keen sense of equality that they cannot function in any hierarchical relationship among themselves. As we have previously noted, when one member of an entering class reaches the pinnacle of the career civil service, administrative vice-ministership, all of the remaining classmates feel compelled to resign. Given the limited opportunities for advancement, however, "noncareer" civil servants seldom face similar constraints.

[17] A survey of 1,598 civil servants who resigned before reaching the age of 30, conducted by the National Personnel Authority in 1964, found that 55.4 percent of women (N = 960) cited marriage as the reason for their resignation. "Marriage" also included childbirth, rearing children, and other family-related matters. The most important reason cited by men, on the other hand, was alternative employment in the private sector (31.2 percent). See Kondo Masaru, "Shokai: Jakunen taishokusha no jittai" [Introduction: the Realities of Young Resignees], Jinji-in geppo 186 (Aug. 1966): 6-9.

[18] Jinji-in, Nenji hokokusho , 1973, in Kanpo [Official Gazette], extra ed., no. 72, 19 Aug. 1974, p. 15; idem., Komuin hakusho , 1987, p. 46.

Pressure from superiors, on the other hand, affects all civil servants, particularly those who are not on the elite track. Known officially as taishoku kansho (encouraging retirement) and unofficially as kata tataki (a tap on the shoulder), this procedure typically involves not only a suggestion that the civil servant in question should retire but also assistance in arranging reemployment for him.^[19] There is an additional inducement: the lump-sum retirement allowance increases by one-third for those who retire in this fashion.^[20]

Available evidence suggests that nine out of ten civil servants who retire after the age of fifty-five do so after receiving kata tataki .^[21] With the introduction of a mandatory retirement system in 1985, however, the proportion of "encouraged resignations" will probably decline sharply. As noted, the new system designates sixty as the mandatory retirement age for most civil servants. Those who are exempt from this rule include physicians and dentists (whose retirement age is set at sixty-five), guards (whose retirement age is set at sixty-three), and other persons who are specifically exempted by the rules of the National Personnel Authority.^[22] Although some observers have expressed fear that the mandatory retirement system may lead to a de facto increase in the retirement age of higher civil servants,^[23] its actual impact on higher civil servants may well be minimal.^[24]

Retirement Of Higher Civil Servants

The single most important factor affecting the retirement patterns of higher civil servants will no doubt continue to be the culturally rooted norm that decrees that one should avoid the embarrassment of taking

[19] Asahi Shinbun "Kanryo" Shuzaihan, Kanryo: Sono seitai , pp. 90 and 130-31.

[20] For example, those who retire after twenty years of service receive 28.875 times their monthly salary at the time of retirement, whereas the regular rate is twenty-one times the monthly salary. For those with thirty years of service, the difference between "encouraged retirement" and retirement at personal convenience is 54.45 times the monthly salary versus 41.25 times the monthly salary. See "Komuin no taishoku teate to nenkin" [Civil Servants' Retirement Allowance and Annuity], Jinji-in geppo 402 (July 1984): 22.

[21] The National Personnel Authority has published statistics on "encouraged resignations" for the years 1968 through 1972 only. They show that their proportion ranged from 89 to 96 percent. See Jinji-in, Nenji hokokusho , 1969-1973.

[22] Jinji-in, Komuin hakusho , 1982, pp. 44-45.

[23] Kato, Kanryo desu, yoroshiku , pp. 180-81. Kato is a former higher civil servant in the Home Ministry.

[24] This assessment was given to the author by Nakajima Sachiko, a counselor (sanijikan ) in the Appointments Bureau (Nin'yokyoku ) of the National Personnel Authority in an interview on 7 June 1985.

orders from one's peers. With rare exceptions, therefore, the promotion of a member of one's entering class to the position of administrative vice-minister will continue to trigger resignations by his classmates; by the same token, the failure to be promoted to the position of bureau chief or its equivalent with one's classmates is a signal that the time for retirement has arrived, even if no kata tataki may have occurred.

For these reasons the average age of higher civil servants at various stages in their career suggests when most resignations are likely to occur. Table 35 shows that the average age has steadily increased across the board over the years. The sole exception is section chiefs: according to the table, their average actually decreased by 1.4 years between the early 1970s and the mid-1980s. The average age of senior section chiefs in 1986, however, was 2.2 years higher than that of all section chiefs.

If we look at bureau chiefs, we find that their average age increased by nine years between 1949 and 1986. Equally striking is the steady aging of administrative vice-ministers: whereas they were still in their late forties in 1949 and 1954, they were in their late fifties by 1986. More relevant for our purposes is the average age of newly appointed administrative vice-ministers, for that is the age when the remaining

members of their respective entering classes retire. We have data for three time periods: 1972-73: 53.6 (N = 34), 1981-83:56.5 (N = 42), and 1984-87:56.0 (N = 50).^[25]

What all this suggests is that the average retirement age of higher civil servants has steadily increased over the years. As we saw in chapter 5, nearly all careerists advance to the rank of section chief, hence the moment of truth arrives just before promotion to the rank of bureau chief. For, given the nature of the bureaucratic pyramid, only a handful of positions are available at this exalted level. Those who are passed over must leave, and they are in the majority. It is not surprising, then, to learn that the average retirement age for all higher civil servants approximates the average age of bureau chiefs: it was 52.9 in 1972-73 and 55.2 in 1985.^[26]

Reemployment Patterns

Even if retirement benefits were adequate, those who retire in their fifties would find it necessary to seek reemployment or something useful to do. For most people, however, the retirement benefits are far from adequate. Until 31 March 1986, for example, a civil servant with thirty years of service was entitled to a separation allowance amounting to about four and half years of pay and an annuity totaling 55 percent of salary.^[27] Effective 1 April 1986, a new system of annuities went into effect; under its extremely complicated provisions the retirement benefits of civil servants were expected to increase slightly.^[28] Most higher civil servants, therefore, seek reemployment either by choice or by necessity.

[25] The figure for 1972-73 is from B. C. Koh and Jae-on Kim, "Paths to Advancement in Japanese Bureaucracy," Comparative Political Studies 15, no. 3 (Oct. 1982): 304, table 6. The figures for the 1980s were calculated from a data set compiled by the author from the following sources: Kankai , May 1981-Sept. 1987; Asahi shinbun , May 1986-Sept. 1987; Yomiuri nenkan bessatsu, bun'ya betsu jinmei roku [Separate Volume Supplement to Yomiuri Yearbook, Who's Who by Fields], 1985 (Tokyo: Yomiuri Shinbunsha, 1985); Seikai kancho jinji roku, 1987-nenban ; ibid., 1988-nenban .

[26] See Koh and Kim, "Paths to Advancement,' p. 304, for the 1972-73 average; for the 1985 figure, see Jinji-in geppo 425 (June 1986): 5.

[27] "Komuin no taishoku teate to nenkin," p. 22. The computational formula for annuity consists of multiplying the retiree's last salary by a rate determined by his length of service. For the first twenty years the rate is 0.4. Each additional year thereafter earns a credit of 0.015. The maximum rate is 0.7. The amount of separation allowance is a function of both length of service and the reason for separation. A civil servant who resigns voluntarily after twenty years of service will receive twenty-one times his last monthly salary, whereas a thirty-year veteran will get 41.25 times his last monthly salary.

There are a number of options for those seeking postretirement jobs. One is to find employment in the private sector. This move from the ostensibly exalted position of a higher civil servant to an employee of a private, profit-making firm is known as amakudari (descent from heaven).^[29] As Gerald L. Curtis notes, however, what happens may not necessarily be descent but ascent; the former bureaucrat may actually enjoy more affluence and prestige than ever before.^[30]

The second option is to land a job in a public corporation, known collectively as "special legal entities" (tokushu hojin ). Unlike the first option, this does not involve any legal restrictions. That is to say, there is no need to seek permission. The third option is to become a member of the Diet, which of course requires winning an election.^[31] Although there are other possibilities, such as teaching, research, and working for a local-government body, the three options noted above seem most important. Let us examine them in more detail.

Amakudari

As noted, amakudari in a narrow sense is subject to some legal restrictions.^[32] Article 103 of the national public-service law stipulates a

[28] For a comprehensive review of the new system, see Kokka Komuin Nenkin Seido Kenkyukai, ed., Kokka komuinto no shin kyosai nenkin seido no shikumi [The Plan of the New Cooperative Annuity System for National Civil Servants and Others] (Tokyo: Zaikei Shohosha, 1986) and Jinji-in, Taishoku Teate Nenkin Seido Kenkyukai, ed., Komuin no tameno taishoku teate nenkin gaidobukku [Retirement Allowance and Annuity Guidebook for Civil Servants] (Tokyo: Okurasho, Insatsukyoku, 1986). Under the new system, civil servants are subject to a dual system of annuities. First, they are required to enroll in kokumin nenkin (citizens' annunity plan) along with all other citizens; this requires monthly payment of premiums. Second, the civil servants continue to participate in kyosai nenkin (cooperative annuity plan), which, too, requires monthly contributions on the part of its participants. The formula for computing benefits, however, has become so complicated that at least five different tables need to be used; in addition, the calculation of the average monthly salary during one's entire civil-service career, which becomes the constant in the equation, is beyond the capability of individuals. It will be done by computers in the Social Insurance Agency. In computing the average monthly salary, adjustments are made to neutralize the effects of inflation and other factors.

[29] For an informative analysis of this and related matters, see Johnson, "The Reemployment of Retired Government Bureaucrats in Japanese Big Business," pp. 953-65.

[30] If that is the case, then a more accurate term would be amaagari (ascending to heaven). See Gerald L. Curtis, "Big Business and Political Influence," in Vogel, ed., Modern Japanese Organization and Decision-Making , p. 44, no. 32.

[31] Johnson, "The Reemployment of Retired Government Bureaucrats," pp. 953-54; Ino Kenji and Hokuto Man, Amakudari Kanryo: Nihon o ugokasu tokken shudan [Bureaucrats Who Descend from Heaven: A Privileged Group That Moves Japan] (Tokyo: Nisshin Hodo, 1972), pp. 91-92.

two-year moratorium on the reemployment of retired civil servants in profit-making enterprises if the government organs in which they have served during the preceding five years have had a "close connection" (missetsu na kankei ) with such enterprises. However, the article adds that this prohibition shall not apply should a recommendation for exemption by the government organ concerned be approved by the National Personnel Authority.^[33]

Depending on how it is interpreted and applied, then, article 103 has the potential of preventing conflicts of interest. In 1963, the Diet appended a requirement to article 103: the National Personnel Authority was enjoined to report annually to both the Diet and the cabinet regarding the number and details of all the exemptions it granted in the preceding year. Officially known as the "Annual Report on Approval of Employment in Profit-Making Enterprises" (Eiri kigyo e no shushoku no shonin ni kansuru nenji hokoku ), the report has been dubbed the amakudari white paper by the mass media. Table 36 presents summary statistics from these reports in selected years.

Note that the numbers shown in the table refer to cases rather than individuals. From time to time, an individual may seek approval for employment in more than one position; occasionally, therefore, the number of cases exceeds that of individuals. Even though table 36 displays statistics for selected years only, we shall also refer, in the following discussion, to the years not covered in the table.

First, we can see a marked increase in the total numbers over the years. The 1986 total is almost double that of 1965. There was actually a notable decline in 1986 from the previous year, when the total had climbed to 320. Second, it is remarkable that a handful of ministries have dominated the scene throughout the period; although the five ministries have continued to account for an overwhelming majority of all the cases, their degree of dominance actually decreased in the 1970s compared with the late 1960s. Beginning in 1977, the Ministry of Posts and Communications joined the five ministries listed in the table as a

[32] In a narrow sense, amakudari refers to the reemployment of retired government officials in private, profit-making enterprises only. Their reemployment in public corporations is called yokosuberi (side slip). However, the broad meaning of amakudari encompasses both. Although Johnson makes the distinction, Ino and Hokuto do not. For a reference to yokosuberi , see "Shinso or saguru: Kareinaru amakudari no urade ugomeku yokubo" [In Search of Truth: The Desire That Squirms Behind the Splendid Descent from Heaven], Kankai , May 1982, p. 140.

[33] See paragraphs 2 and 3 of art. 103 of the law in Jinji-in ed., Ninmen kankei horeishu , 1984 ed., p. 33.

major player in the game. In 1986 it surpassed both MITI and the Ministry of Agriculture, with twenty-eight exemptions.

Third, the Ministry of Finance has been the indisputable leader: in all eight years covered in the table, it was number one, accounting for an average of 24 percent of the total exemptions. Although the number-two position goes to MITI, the other ministries are not very far behind. What all these ministries have in common is their close linkage with the strategic sectors of the Japanese economy. The close patterns of interaction between them and their respective clientele groups make retiring higher civil servants valuable assets to their prospective employers in the private sector: not only do the retirees possess managerial ability and substantive expertise but they can also measurably facilitate interactions with the government bureaucracy. To ascertain whether the National Personnel Authority actually performs the function of obviating or, at least, minimizing conflicts of interest, we need to know details of specific cases, which are not available. Fragmentary evidence suggests, however, that the authority is not a rubber stamp: in the 1970s, it rejected between 3.0 and 7.4 percent of the requests for exemption.^[34]

Table 37 discloses another notable trend in amakudari : the growing number of technical officials. Whereas they accounted for only four of

[34] Murobushi Tetsuro, "Kokyu kanryo—riken no kozo" [Higher Civil Servants: The Structure of Interests], Sekai , Feb. 1980, p. 56.

every ten cases in 1965, their share nearly doubled by 1970. After declining slightly in 1975, the proportion of technical officials surpassed that of administrative officials in 1980. Their lead widened in 1985. All this suggests that a major reason that former government bureaucrats are hired by private firms may be the substantive expertise they bring to their jobs in addition to other assets.

As noted, even though all the cases included in the annual amakudari white papers pertain to former bureaucrats with the rank of section chief or its equivalent and above, only a handful of them actually occupied key positions in the headquarters of the ministries and agencies. Of the 318 persons who were granted exemptions by the National Personnel Authority in 1985, for example, only 49 (15.4 percent) were in that category (honshocho kacho shoku ijo ). Of this number, 13 held the rank of bureau chief or above.^[35] In 1986, 38 (15.3 percent) of the 248 persons receiving exemptions were in the same category, of whom only 7 had held the rank of bureau chief or above.^[36]

To examine briefly the destinations of the 20 highest-ranking retirees in these two years, the largest proportion of the group (7, or 35 percent)

[35] Asahi shinbun , 28 Mar. 1986.

[36] "Showa 61-nen eiri kigyo eno shushoku no shonin ni kansuru nenji hokoku" [1986 Annual Report on Approval of Employment in Profit-making Enterprises], Jinji-in geppo 437 (June 1987): 6.

went to banks, all but one being appointed as advisers (komon ). The lone exception became vice president (fuku todori ). In fact, komon is the title most commonly given to these people, accounting for 8 of the 20 titles. Technically, they are not of the directorial rank (yakuin ); therefore, approval of reemployment is believed to be almost pro forma. In most cases, they will be elevated to directorial positions after a lapse of two years. Should a change in status occur earlier, however, approval by the National Personnel Authority is required de novo. Exactly half of the group landed yakuin positions immediately: one presidency (torishimari yaku shacho ), one vice-presidency (fuku todori ), two managing directorships (senmu torishimari yaku ), one executive directorship (jomu torishimari yaku ), four directorships (riji ), and one consulting directorship (torishimari yaku sodan yaku ). Also noteworthy is that a recently privatized company, Nippon Telegraph and Telephone (formerly the Japan Telegraph and Telephone Corporation, or Nihon Denshin Denwa Kosha), hired 5 of the top group, 1 of them as an executive director and 2 as directors.^[37]

Inasmuch as the Ministry of Finance is the single largest source of amakudari , let us examine the destinations of its high-level retirees. In 1984, 29 persons at the rank of section chief or above in the ministry (honsho kacho-kyu ijo ) retired. Of this total, only 1, a former administrative vice minister, did not seek immediate reemployment. The destinations of the remainder were as follows: government-affiliated financial institutions, 6; public corporations, 9; private banks, 2; private firms, 8; licensed tax accountants (zeirishi ), 3. In terms of position titles, 16 became directors, and 5 became advisers. On the other hand, of those who entered the private sector, only a few landed jobs in what may be described as first-rate companies, which included Mitsui Trust Bank, Japan Air Lines, and Mitsubishi Shoji.^[38]

The number of retirees in the same category in 1985 was 35. Of this total, 1 died and 1 was preparing to run for the House of Representatives in the next election. Five became licensed tax accountants, and the remainder went to either public corporations or private banks and firms. Although 11 became directors, none was hired by a really prestigious firm. Officials in charge of finding reemployment for the Finance Ministry's retirees complained of a growing scarcity of suitable landing spots; noting they had to "lower their heads" to find reemployment

[37] For destinations of these and other high-ranking retirees, see Asahi shinbun , 28 Mar. 1986, and 28 Mar. 1987.

[38] Ibid., 19 Feb. 1985.

positions for their retirees, they pointed out that it was difficult to call their ministry a "first-rate government agency" (ichiryu kancho ) any longer.^[39]

The situation in 1986, however, was far from gloomy. All but two of the Finance Ministry's 32 high-level retirees found employment in that year. The 2 who did not apparently chose not to seek immediate reemployment; one of these was former administrative vice-minister. Although the destinations of the 30 varied widely, the largest number, 7, went to public corporations. Three went to private banks, 2 to insurance companies, and 5 to other private firms. Three went into private practice as either an attorney or a licensed tax accountant. One entered the Judicial Training Institute with the aim of entering the legal profession. Among the more notable destinations were the Tokyo Stock Exchange (executive director), the Mitsubishi Trust Bank (adviser), the Japan Foundation (executive director), and the Sumitomo Life Insurance Company (adviser).^[40]

All in all, the record of the Finance Ministry in finding postretirement employment for its elite-track bureaucrats seems solid. It is plain that the ministry remains the premier government agency insofar as amakudari is concerned. No other ministry or agency, for example, has produced so many directors and presidents of banks.^[41] As we shall see below, moreover, the Finance Ministry is the leading source of candidates for the Diet as well.

Yokosuberi

The movement from ministry or agency to a public corporation is known as a "side slip" (yokosuberi ).^[42] Most writers, however, use the term amakudari to encompass this form of reemployment by retired officials as well.^[43] Unlike amakudari as strictly defined, yokosuberi allows the retired government bureaucrat to remain in the public sector; what happens is equivalent to a transfer from the mainstream of the

[39] Ibid., 16 Feb. 1986.

[40] Ibid., 8 Mar. 1987.

[41] For examples of and statistics pertaining to amakudari of retired Finance Ministry bureaucrats to the various banks and financial institutions, see Takamoto, Okura kanryo nokeifu , pp. 159-95; Asahi shinbun , 22 May 1986; "Kasumigaseki konhidensharu," Bungei shunju , July 1986, p. 163.

[42] Johnson, "The Reemployment of Retired Government Bureaucrats," p. 953.

[43] Although Chalmers Johnson differentiates between the two terms in his Asian Survey article cited above, he blurs the distinction in his study Japan's Public Policy Companies (Washington: American Enterprise Institute for Public Policy Research, 1978). In the latter Johnson defines amakudari as the "practice of employing retired government officials as chief executives or members of boards of directors of public and private corporations" (p. 5, emphasis added). Elsewhere in the study, however, he mentions the difference between the two terms (p. 102). For another definition of amakudari that subsumes yokosuberi , see "Shinso o saguru: Kareinaru amakudari," p. 144.

government to its periphery, for he will be reemployed by one of the hundred-odd "special legal entities." In 1980, there were 111 such entities, of which 3 were kosha (public corporations), 16 were kodan (public units), 19 were jigyodan (enterprise units), 10 were koko (public finance corporations), 2 were tokushu ginko (special banks), 2 were kinko (depositories), 1 was eidan (corporation), 11 were tokushu kaisha (special companies), and the remainder were called by a wide variety of names. These entities employed nearly a million persons, of whom fewer than 800 were in executive positions (yakuin ).^[44] In 1985, however, 2 of the 3 kosha were privatized, and the third was scheduled to follow suit in 1986.^[45]

Because yokosuberi is not subject to any legal restrictions, the government does not publish any comprehensive data on it. Nonetheless, a labor federation, to which labor unions comprising employees of special legal entities belong, collects data on its own and publishes annual reports. According to its report published in 1987, 379 of 489 yakuin in eighty-three special legal entities it studied in 1986 were retired higher civil servants. This amounted to 77.5 percent of the total. MITI had produced the largest number of the subgroup (N = 54), followed by the ministries of Agriculture, Forestry, and Fisheries (N = 46), Finance (N = 34), and Construction (N = 30).^[46]

Yokosuberi differs from amakudari , narrowly construed, in another sense: the same individual can experience it more than once. The retired higher civil servant who moves from one yokosuberi post to another, collecting generous separation allowances in the process, is known as wataridori (migratory bird). According to information disclosed by opposition-party members in the House of Representatives in May 1969, of 363 retired higher civil servants serving as yakuin in 108 special legal entities at the time, 75 (20.7 percent) had experienced yokosuberi three times or more. Of the latter, 15 had experienced it four times, and I five times. Separation allowances collected by these people during each transition were equal to 65 percent of their total earnings

[44] For a complete list and definitions of these "special legal entities," see Gyoseikanri benran [Handbook of Administrative Management], 1980 (Tokyo: Gyosei Kanri Kenkyu Senta, 1980), pp. 28-39. The English translations were taken from Johnson, Japan's Public Policy Companies , pp. 5-7 (Glossary of Japanese Terms).

[45] See "Shinso o saguru: NTT, min'eika no sono ato" [In Search of Truth: Nippon Telegraph and Telephone in the Wake of Privatization], Kankai , May 1986, pp. 164-73.

[46] Asahi shinbun , 5 Apr. 1987.

during their tenure in each job. Although the percentage was lowered to 45 in 1970 and to 36 in 1978, the comparatively high salaries of yakuin in public corporations make the separation allowance quite generous.^[47] In 1986, about a quarter of the 379 former bureaucrats occupying leadership positions in the special legal entities were wataridori .^[48]

A leading example of wataridori is Funayama Masakichi, who retired as administrative vice-minister of the Finance Ministry in 1953. He first served as a director of the Bank of Japan and then became the deputy governor of the Japan Monopoly Corporation. His next job was with the Japan Export-Import Bank, where he was deputy governor. Finally, he was appointed the governor of the Smaller Business Finance Corporation. The separation allowances he had collected prior to his last appointment amounted to over 27.5 million yen, a considerable sum in the 1950s and 1960s.^[49]

Entering Politics

The option of running for a seat in the Diet is available only to the select few among retired higher civil servants. Whereas amakudari , broadly defined, is typically arranged by the prospective retiree's ministry or agency, running for election is something one must arrange on one's own. Because virtually all retired higher civil servants run as candidates of the Liberal-Democratic party, they find it necessary to affiliate themselves with one of the factions within the party in order to win official endorsement. Factional affiliation is also necessary to help finance the campaigns, which cost astronomical sums.

As we saw in the preceding chapter, in the double election of 6 July 1986, a total of 80 former higher civil servants were elected to the House of Representatives, accounting for 23.3 percent of the 300 successful LDP candidates. Of the 72 LDP candidates who were elected to the House of Councillors in the same election, 24 (33.3 percent) were former bureaucrats. Another former bureaucrat who ran as a Democratic Socialist party candidate was also elected to the upper house.^[50]

[47] Murobushi, "Kokyu kanryo," pp. 59-60. The formula for calculating the amount of separation allowance is as follows: monthly salary at the time of separation × number of months served × 0.36. Sekiguchi Takeshi, Komuin tengoku! ! [Civil Servants' Heaven] (Tokyo: Aro Shuppansha, n.d.), p. 161. Although this book does not show any publication date, it appears to have been published in 1978. The copy that I used had been purchased by the U.S. Library of Congress on 14 July, 1978.

[48] Asahi shinbun , 5 Apr. 1987.

[49] Ino and Hokuto, Amakudari kanryo , p. 22; for other examples of wataridori , see ibid., pp. 22-24, and Sekiguchi, Komuin tengoku! !, pp. 162-63.

[50] Asahi shinbun , 7 July 1986 (evening ed.) and 8 July 1986.

The broad picture given above, however, pertains to all former higher civil servants. Table 38 is therefore aimed at showing what happens to those retired civil servants who elect the political option. It is plain that throwing one's hat into the political arena is not as easy as it sounds: in the 1986 election, only 4 of every 10 former bureaucrats who had not previously been elected to the Diet achieved their goals. The probability

of success varied inversely with age: those in their fifties were more than twice as likely to succeed as those in their forties and thirties.

Although the relationship between rank and the probability of electoral success is somewhat murky because of missing data, the two nonetheless seem to covary to an appreciable degree. Whereas 7 of the 9 persons who had attained the rank of division chief or higher won their bids, only 2 of the 10 former section chiefs did so. On the other hand, the success rate of those who had not even attained the section-chief rank is quite good, although because of the small numbers involved, caution is called for in drawing any conclusions.

Finally, it is interesting to note that the ministries that contribute most to amakudari also produce the largest number of candidates for the House of Representatives. The success rates of the top two, Finance and MITI, however, are mediocre at best.^[51]

Let us now consider briefly the situation regarding the House of Councillors. Of the twenty-five former bureaucrats elected to the upper house on 6 July 1986, six were freshmen. A comparison of the backgrounds of the six freshmen with those of the eleven novice members of the House of Representatives yields some interesting results. To begin with similarities, all are Liberal-Democrats. And, by a remarkable coincidence, all are graduates of the University of Tokyo. However, the upper-house members are considerably older: their average age is 58.2; in fact, half are in their early sixties. By contrast, the average age of the lower-house group is 50.4, and none is in his sixties. Another striking difference has to do with their former bureaucratic ranks: whereas only three of the eleven representatives had attained the rank of bureau chief or above, all of the councillors had done so. Half of the latter had retired as either administrative vice-minister or director general (chokan ).^[52]

An Assessment Of Reemployment Patterns

As noted, the need for reemployment of retired higher civil servants is rooted in the twin realities of early retirement and insufficient retire-

[51] For a discussion of the motives of young Finance Ministry bureaucrats who resign to run for elective offices, see Kuribayashi, Okurasho shukeikyoku , pp. 203-24.

[52] Biographical data on the candidates, both successful and unsuccessful, were gleaned from Tsuchiya Shigeru, "Kanryo shusshin rikkohosha sotenken" [A Complete Examination of Candidates Who Are Former Bureaucrats], Kankai , Apr. 1986, pp. 160-71; May 1986, pp. 142-51; and June 1986, pp. 145-57; Asahi shinbun , 7 July 1986 (evening ed.) and 8 July 1986. Of the twenty-five former bureaucrats elected to the upper house in July 1986, nine had served as administrative vice-ministers and two as directors general (chokan ). Only one had retired as a section chief. The remainder had been bureau chiefs. In terms of educational background, twenty were graduates of the University of Tokyo, four were graduates of Kyoto University, and one was a graduate of a technical high school.

ment benefits. Of the two, the former may be a more potent cause of amakudari , broadly defined, than the latter. Unless the practice of early retirement ceases, therefore, the need for reemployment will persist. However, there are no signs that the practice will change to any notable degree.

Hallowed by tradition, early retirement is sustained by the interaction of organizational dynamics and cultural norms. The progressive diminution of positions in the upper levels of the organizational pyramid dictates a pruning of the ranks among the old-timers. Moreover, a strong sense of equality among peers defined by the year of entry expedites resignations by those who fall behind leaders in the unacknowledged but nonetheless real race toward administrative vice-ministership.

Amakudari in a narrow sense, that is, reemployment by a private firm, can occur in three distinct ways. The most common mode is for the ministry or agency to take charge and arrange a landing spot for the retiree. Another mode takes the form of what the Japanese call "scouting." In this mode it is the private firm that takes the initiative: the firm seeks out a higher civil servant who has the kind of expertise, experience, and contacts that it needs most. In a variant of this mode, a firm may ask a ministry or agency to recommend a suitable candidate. In a third and final mode, the civil servant is rewarded for past services rendered to a firm; the initiative in such cases usually emanates from the retiring civil servant himself. All three modes may require approval by the National Personnel Authority.^[53]

No matter what form it may take, amakudari entails mixed consequences. On the positive side, it contributes to the optimal utilization of talent, facilitates communication between government bureaucracy and private business, and "enhance[s] the effectiveness of administrative guidance."^[54] On the negative side, amakudari may

[53] Ino and Hokuto, Amakudari kanryo , pp. 169-71. According to Ojimi Yoshihisa, a former administrative vice-minister of MITI, "From the point of view of private companies, there is a need for these men [retired higher civil servants]. Requests frequently come to the personnel division of the ministry. Thus, placement is usually taken care of by the personnel division." See Ojimi, "A Government Ministry," p. 110.

[54] Johnson, "The Reemployment of Retired Government Bureaucrats," p. 964. See also Curtis, "Big Business and Political Influence," p. 45; Kazuo Noda, "Big Business Organization," in Vogel, ed., Modern Japanese Organization and Decision-Making , p. 133; Akimoto Hideo and Kanai Hachiro, "Taidan: Komuin rinri to 'amakudari' mondai o megutte" [Conversation: Concerning the Ethics of Civil Servants and the Problem of "Descent from Heaven"], Jinji-in geppo 377 (June 1982): 4; Hayashi, Nihon Kanryo kenkoku ron , pp. 61-71.

compromise the independence and integrity of government bureaucracy, breed corruption, and confer unfair advantages on the firms that hire retired higher civil servants. It also has demoralizing effects on those members of the private firms whose promotional opportunities are undercut by the lateral entry of outsiders. To a large extent, however, the potential for corruption is offset by the modus operandi of Japanese organizations, which, by accentuating consensual decision making and diffusing authority and responsibility, makes it difficult for any individual to do favors for clientele.^[55] As we have seen already, this is borne out by the actual record. Not only is corruption among incumbent higher civil servants rare, but the behavior of retired officials who have landed amakudari positions has not been marred by any scandals.^[56]

The entry of former bureaucrats into the political arena via the ballot box, too, can be viewed from the standpoint of resource utilization. The expertise and experience they bring to their roles as members of the Diet can theoretically enhance the latter's capability to formulate policy and monitor policy implementation. On the other hand, the necessity to plan ahead—to find a patron, to align oneself with a faction, and to build a political base, no matter how rudimentary it may be—may conceivably interfere with a dispassionate discharge of bureaucratic responsibilities. The potential for compromising the bureaucrat's objectivity and neutrality is ever present.

On balance, then, the reemployment of retired higher civil servants is neither an unmitigated evil nor a cause for satisfaction. It represents a pragmatic response to real needs of individual bureaucrats, the government bureaucracy as a whole, and private business.^[57] So long as the needs remain, the practice is likely to persist. The real challenge for the Japanese government remains how to prevent flagrant abuses, how to minimize patent conflicts of interests, and how to retain the trust and confidence of its citizens in the basic integrity of higher civil servants.^[58]

[55] Akimoto and Kanai, "Taidan: Komuin rinri," p. 5.

[56] This last point is stressed by a former bureau chief in the National Personnel Authority. See Shima Yotsuo, "Kaisobun Amakudari mondai' arekore" [Reminisciences: Aspects of the "Problem of Descent from Heaven"] in Jinji-in, Jinji gyosei sanjunen no ayumi , p. 394.

[57] Johnson, "The Reemployment of Retired Government Bureaucrats," p. 965.

[58] In a survey of a national sample of 2,445 adults conducted by the prime minister's office in 1973, 65 percent of the respondents thought retired civil servants were either well off or relatively well off, citing their annuity, separation allowance, and reemployment in that order. Thirty-seven percent disapproved of amakudari , and 31 percent endorsed it. Of those who disapproved of the practice, 63 percent said it should not be allowed even after a lapse of several years. See Sorifu, Naikaku Sori Daijin Kanbo, Komuin ni kansuru seron chosa, Showa 48-nen 9-gatsu [Opinion Survey Concerning Civil Servants, September 1973] (Tokyo: Sorifu, Naikaku Sori Daijin Kanbo Kohoshitsu, 1973), pp. 10-13 and 20-21.

A Comparative Perspective

To what extent, if any, are the patterns sketched above unique to Japan? The basic structure of compensation appears to be broadly similar among all industrialized democracies. The stratification of civil-service employees is reflected in the stratification of pay schedules. The discrepancy between the public and private sectors, the former perennially lagging behind the latter, is a universal phenomenon. Equally universal is the concern for narrowing the gap, for approximating the idea of "comparability" between government and private-sector salaries.

Both the United States and the United Kingdom have mechanisms for periodic review of government pay schedules. Under the Postal Revenue and Federal Salary Act of 1967, a nine-member commission is appointed every four years in the United States by the president, the chief justice of the Supreme Court, and the speaker of the House of Representatives; the commission submits a report to the president, who in turn makes his own recommendation to Congress. Until 1977 the president's proposal automatically took effect thirty days after transmittal to Congress "unless in the meantime either house had formally voted disapproval or Congress had enacted a statute establishing alternative pay rates." The law was revised in 1977 to require roll-call votes by Congress. That is, unless the president's proposal is explicitly enacted into law, it does not take effect.^[59]

In Great Britain the government conducts periodic surveys of compensation, but the task of making specific recommendations for higher civil servants' pay adjustments is assigned to a commission appointed by the prime minister. Unlike the situation in both the United States and Japan, however, the commission's recommendations to the prime minister are implemented without revision.^[60] Although a major rationale for empowering the National Personnel Authority to make annual recommendations on civil servants' pay was the restrictions on civil servants' right to engage in collective bargaining and in strikes,

[59] Robert W. Hartman and Arnold R. Weber, eds., Rewards of Public Service: Compensating Top Federal Officials (Washington: Brookings Institution, 1980), pp. 76-87.

[60] Hashidachi Eiji, "Sho gaikoku komuin no kyuyo seido no gairyaku" [An Outline of Civil Servants' Compensation Systems in Various Foreign Countries], I, Jinji-in geppo 297 (Nov. 1975): 15.

such restrictions are by no means confined to Japan. In fact, none of the four Western democracies with which Japan has been compared in this study permits strikes by their civil servants.^[61]

Provisions regarding the retirement of civil servants vary somewhat in the four countries. The most common age for mandatory retirement seems to be sixty-five. Eligibility for annuity typically requires the attainment of sixty years of age. Interestingly, the practice of encouraging retirement that we noted earlier is found in Britain as well: it is applied to civil servants who have passed the age of sixty.^[62]

Retirement annuities in these countries are a function of the civil servant's salary and length of service. In the United States, a federal employee becomes eligible for voluntary retirement with full annuity after either thirty years of service and attaining the age of fifty-five or twenty years of service and attaining the age of sixty. Since 1970, federal employees have contributed 7 percent of their entire salary to the retirement system. The amount of annuity is computed by multiplying "the highest average salary during any three consecutive years of employment" by credits based on the number of years of service. "Each of the first five years of service earns a 1.5 percent credit for the worker; each of the next five earns him or her 1.75 percent, and each year of service above ten adds another 2 percent." The maximum annuity is 80 percent of the average salary. Under the preceding formula, a federal employee with thirty years of service receives credits of 56.25 percent.^[63]

A British civil servant with at least ten years of service who attains the age of sixty is eligible for an annuity, which is equal to the average of his last three years of salary × one-eightieth × the number of years of service. A retiree with thirty years of service will thus receive 37.5 percent of his average salary. Although this is substantially less than what his American counterpart would get, the gap is reduced to some extent by a lump-sum separation allowance equal to three times the annuity. The British civil servant contributes about 8 percent of his salary to his pension scheme.^[64]

[61] "Kakkoku komuin seido no hikaku (10): Fukumu, shokuin dantai" [A Comparison of Civil Service Systems in Various Countries (10): Duties and Employee Organizations], Jinji-in geppo 112 (June 1960): 8-10.

[62] Yoshida Kozo, "Eikoku ni okeru komuin taishoku seisaku no genjo to tenbo" [The Present Situation and Future Prospects of Britain's Policy toward Civil Servants' Retirement], Jinji-in geppo 345 (Nov. 1979): 6-10.

[63] Robert W. Hartman, Pay and Pensions for Federal Workers (Washington: Brookings Institution, 1983), pp. 16-17.

[64] Shimoda Akiko, "Igirisu komuin no nenkin seido" [British Civil Servants' Annuity System] I, Jinji-in geppo 420 (Jan. 1986): 29-30.

In the French civil service one must have at least fifteen years of service to be eligible for pension; the minimum age for drawing pension is either fifty-five or sixty, depending on circumstances. The amount of annuity is computed by multiplying the preretirement salary by two-one-hundredths times the number of years of service; in no case, however, may the amount exceed 75 percent of the last salary. Thus a civil servant with thirty years of service will receive 60 percent of his salary; he will have contributed 6 percent of his salary to the pension fund during his civil-service career.^[65]

A distinctive feature of the West German pension system is that it is completely state-financed; the civil servant need not make any contributions during his career. At least ten years of service are required, however, before he is eligible to draw pension. A credit of 35 percent is awarded for the first ten years of service, and a credit of 2 percent is added for each of the next fifteen years. Beyond that, each additional year earns a credit of 1 percent. In no case may the total exceed 75 percent. A civil servant with thirty years of service will thus receive 70 percent of his salary upon retirement.^[66]

In comparative terms, then, the annuity of Japanese civil servants is on a par with that of their American counterparts. If the lump-sum separation allowance is added, it is probably comparable to that of the French civil servant as well. It is substantially better than that of the British civil servant. The clear winner is the West German civil servant, who enjoys the most-generous benefits at no cost.

What of the reemployment of retired civil servants? Although no statistics are available, it does not appear to be a major problem in Britain and West Germany. France, however, has its own version of amakudari . Known as pantouflage , this involves primarily the movement of elite administrators, mostly products of the Ecole Nationale d'Administration (ENA) and relatively young, from the public to the private sector. Those who choose this option are either "men with great family wealth who have entered the [government] service primarily for experience in managing large affairs" or "men who simply find private salaries irresistibly attractive."^[67] Unless they have fulfilled their obligation to serve in the government for ten years, ENA graduates who move

[65] Nomura Nario, "Furansu Kyowakoku kanri no onkyu seido" [The Pension System of the Officials of the French Republic] I, Jinji-in geppo 315 (May 1977): 4-8.

[66] Okada Jin, "Kakkoku komuin no onkyu seido" [The Pension System of Civil Servants in Various Countries] II, Jinji-in geppo 86 (Apr. 1958): 4-6.

[67] Armstrong, The European Administrative Elite , p. 221.

to the private sector must reimburse the government the entire expense of their three-year education. In most cases the firms that hire them will pick up the tab.^[68]

In the United States there is an extraordinarily high degree of mobility in all aspects of employment, both public and private. The closest thing to amakudari may therefore be the phenomenon of the revolving door—the movement of managerial-level personnel from government to private business and vice versa. Although it occurs in all government agencies, it is particularly prevalent in the Department of Defense. According to the New York Times , about a thousand people pass through the revolving door at the Pentagon each year. To illustrate, in 1985 an assistant secretary of defense for manpower and logistics resigned to assume a new position at Raytheon Company, a manufacturer of missiles, and a senior vice president at Northrop Corporation, a major defense contractor, became under secretary of defense for research engineering.^[69]

All four countries have laws regulating the reemployment of former government officials. In the United States, former Federal employees are forbidden by law to return to "their agencies as supplicants for their new employers. But the restrictions are narrowly drawn to catch clear conflicts of interests. There are few prosecutions."^[70] Former British civil servants must obtain approval from the government if, during two years following their retirement, they wish to work for firms with which the government is involved in any kind of financial transaction. This means that the British law is much more restrictive than its Japanese counterpart; the latter applies only to cases where the civil servant's own agency, not any government agency, has had a close connection with a prospective employer. Approval is required as well for reemployment in profit-making firms in both France and West Germany.^[71]

As for entering the political arena, civil servants in France and West Germany enjoy the extraordinary privilege of running for elective

[68] Shimizu Kunio, "Kanryo okuni-buri (6), Furansu: Banjaku no cho-erito shugi" [Bureaucrats in Different Countries (6), France: An Unshakable Super-elitism], Kankai , Mar. 1981, p. 87.

[69] New York Times , 25 Aug. 1985.

[70] Ibid. Efforts to enact a new law that would "for two years prohibit officials leaving the Pentagon from going to work for a contractor over which they had 'personal and substantial' decision-making responsibility' have thus far been unsuccessful. Although passed by the House, the bill faced opposition in the Senate; the opponents argued that "it would hurt recruitment by the Pentagon, reduce opportunities for officials usefully forced out of Government and damage a helpful 'cross-fertilization' among military experts." See "Close the Pentagon's Revolving Door," ibid., 6 Sept. 1986, editorial.

[71] "Kakkoku komuin seido no hikaku (10): Fukumu, shokuin dantai," pp. 6-7.

offices without first resigning from their jobs, and a sizable number of them do run for their respective parliaments. If successful, they are allowed to take what amount to leaves of absence to serve in parliament.^[72] However, the phenomenon of retired bureaucrats running for national legislatures in any conspicuous numbers is not duplicated outside Japan.

In sum, although Japanese higher civil servants are not alone in suffering from relatively meager financial rewards, particularly in relation to their counterparts in the private sector, their custom of early retirement may be unique in its scale and clocklike regularity. Likewise, the pattern of reemployment by retired Japanese bureaucrats does seem to have some unusual features.

[72] Ibid., pp. 5-6.