You are on page 1of 16

Political Studies (1999), XLVII, 939954

Who Governs Japan? Politicians and


Bureaucrats in the Policy-making Processes
MAURICE WRIGHT
University of Manchester

The long-running debate about who governs Japan has been given a new twist by
`rat-choicers' who argue that Japan has been governed for the last thirty years or
more neither by bureaucrats nor by a `conservative coalition' of bureaucrats,
politicians and businessmen but by the Liberal Democratic Party alone. This article
examines their arguments and sets them in the context of other competing and
conicting explanations. It is argued that more relevant and researchable questions
are what is governed and how, an approach calling for a more nuanced analysis of
policy making in order to observe the impact on dierent policies and policy-
processes of the role of the state and its institutional structures and their embedded
`collective identities'.

How the Conservatives Rule Japan (Nathaniel Thayer, 1969).


`Japan's a system of bureaucratic rule' (Chalmers Johnson, MITI: The
Japanese Miracle, 1982).
`No one is ultimately in charge' (Karel Von Wolferen, The Enigma of
Japanese Power, 1989).
`Members of the Liberal Democratic Party, not bureaucrats or judges or
even LDP faction leaders, control Japan's policy-making process' (J. M.
Ramseyer and F. M. Rosenbluth, Japan's Political Marketplace, 1993).
Japan, the world's second economic power and its largest creditor nation, has
been in crisis for almost a decade. The collapse of the `economic bubble' in
1990 triggered the longest recession in Japan's post-war history. After the
briefest of respites the economy plunged into a still steeper decline in 1997. A
decade earlier, such was the apparently unstoppable progress of its economy,
the strength of its currency, and the breadth of its expansionism through direct
investment in the USA, Europe and the emerging markets of China and East
Asia, that Japan's capacity and willingness to play the hegemonic role in global
trade and nancial markets assumed by the USA since World War II was
seriously debated.
At that time a mixture of admiration and fear of the Japanese `economic
miracle' sparked o a wide-ranging debate among American economists and
political scientists about its principal characteristics and causes; and con-
sequences for the US economy. It was conducted against the back-drop of an
apparently secular decline of the US economy, attributed partly to the loss of
comparative economic advantage to Japan in several high-tech sectors such as
computers, semi-conductors and micro-processors, chronic Federal scal
decits and debts, and a savings and loans crisis undermining its nancial
# Political Studies Association 1999. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main
Street, Malden, MA 02148, USA.
940 Review Section

system. The competitive challenge was accompanied by the rapid overseas


expansion of Japanese corporations and nancial institutions, symbolized by
the purchase of the Rockefeller Center in New York and the Pebble Beach Golf
Club in California.
Since the end of the Allied Occupation of Japan in 1952, research and debate
about the role and inuence of bureaucrats and politicians in Japanese policy
making has been conducted among (mainly American) scholars within a
political context encompassing some of the major issues of the evolving special
relationship between the US and Japan issues of national security and defence
as well as those of trade relations and industrial structure and performance. The
questions posed and the answers given commonly have a normative bias, most
academic studies concluding with policy prescriptions and advice for US
Governments, legislators and businessmen. Analysis is not only of Japanese
policy and policy making but also for US policy and policy making. Who is
perceived to be inuential in the policy processes, in the roles of entrepreneur,
broker, mediator, gate-keeper, has been signicant for the design and
implementation of US policies. Conversely, policies and policy making in
Japan have been inuenced, at times decisively so, by perceptions of the
attitudes and behaviour of US governments and congressmen. Changes in
Japanese policies for trade and trading relations have not infrequently been
justied domestically and carried through on the grounds of the need to bow
before the demands of `foreign pressure' (gaiatsu) of other governments and
international organizations, but especially to pressure from the USA.1
Two main schools of thought have dominated the debate about the `secret' of
Japan's post-war economic success: those who believed that Japanese economic
behaviour was market-rational, explicable almost wholly in terms of conven-
tional Western neo-classical economics, and the so-called `revisionists' who
argued that Japan, its institutions, its political economy, and economic
behaviour were fundamentally dierent, even unique, the dening and disting-
uishing characteristics of its `economic miracle' partly or mainly culturally
determined. The most basic issue raised by the debate was the extent to which
Japan had a long-term economic strategy directed by the state. Neo-classicists
argued that while Japanese corporations had such objectives, they were not
synthesized and directed by state institutions. By contrast, revisionists argued
that Japan had a state-directed economic strategy, but were divided whether the
motivation was defensive to achieve economic self-suciency and industrial
equality with the West or oensive, to achieve economic domination by
destroying American industry. The two explanations produced radically dier-
ent policy prescriptions. Revisionists argued that Japan posed a major security
threat to the USA, and recommended that the USA should respond with a set of
trade, industrial and other policies to defend the American economy, and
contain Japanese economic imperialism. Neo-classicists urged that such action
was unnecessary because expansionism would slow as Japan closed the
economic and technology gap with the West.
In fact market forces halted that expansionism, but in ways that were wholly
unanticipated. The collapse of the `bubble economy' left in its wake grossly
over-valued and unrealizable land and property assets, bankruptcies, and banks

1
For a critical discussion of gaiatsu and its limitations, see L. J. Schoppa, Bargaining with Japan:
What American Pressure Can and Cannot Do (New York, Columbia University Press, 1997).
# Political Studies Association, 1999
Review Section 941

and nancial institutions with bad debts and an overhang of non-performing


loans totalling 87 trillion yen, some 17% of nominal GDP. It triggered the
longest recession in Japan's post-war history and precipitated the most serious
nancial and scal crisis since the 1930s.
The reversal in Japan's economic fortunes combined with the end of the
LDP's 38 years of rule in 1993, and the political turbulence which ensued, raises
questions about the continuing eectiveness of Japanese institutions. This
article reviews the arguments and considers some of the evidence adduced by
the main protagonists in the continuing academic debate about the relationships
between politicians, bureaucrats and businessmen which underpin those institu-
tions. The evidence is mainly drawn from the work of American scholars and
those Japanese whose work has been published or translated into English. The
work of other scholars writing in Japanese is normally accessible through both.
We begin with Chalmers Johnson's concept of the `developmental state' in
which bureaucrats are accorded the pre-eminent role in policy making, the
trailblazer and the inspiration or provocation for much that followed.2 We then
look briey at the qualication to theories of bureaucratic dominance to
accommodate the evidence of competition and conict in the policy processes
provided by the neo-pluralists. The work of area and sector specialists is briey
reviewed in the general context of the thrust of a `counter-revisionist' challenge,
emphasizing the interdependence of relationships of politicians, bureaucrats
and societal interest groups in sub-governments or networks. This leads on to a
more detailed and critical review of the recent work of the `new institutionalists'.
Adopting and adapting theories of rational choice, they have attacked both
revisionists and counter-revisionists, arguing that the relationship between
politicians and bureaucrats is one of principal and agent, bureaucrats acting
merely as the instruments of politicians.

Bureaucratic Dominance in the `Developmental State'


Revisionists reject both behaviouralism and the tenets of neo-classical econo-
mics, especially trade theory, as inappropriate in analysing and accounting for
the success of Japan's post-war political economy. They argue that Japan is
dierent from Anglo/American models of capitalism; its institutions, the role of
the state, and its mercantilist economic policies are unorthodox. Hence those
who pointed out those dierences were `revisionists'. Central to the argument
was the role of the state, in which the dominant role was played by elite
bureaucrats; by implication the ruling Liberal Democrat Party was accorded a
subordinate role.
In MITI and the Japanese Miracle Johnson explains the origins and causes of
the `economic miracle' which occurred in Japan, roughly from the mid 1920s to
the mid 1970s. The centrepiece of his model of the `developmental state' is the
bureaucracy; the most important part of it, the `economic bureaucracy' of MITI
and the Economic Planning Agency (EPA). The elements of the model, briey
summarized, emphasize the central and crucial role of the bureaucracy in the
Japanese political system: rst, the developmental state is `plan-rational', as
opposed to the `market-rational' liberalism of regulatory states such as the

2
C. J. Johnson, MITI and the Japanese Miracle: the Growth of Industrial Policy, 19251975
(Stanford CA, Stanford University Press, 1982).
# Political Studies Association, 1999
942 Review Section

USA. Secondly, there is an explicit industrial policy aimed at producing and


sustaining `high speed economic growth'. The pursuit and achievement of that
growth is part of an overarching set of goals for society, upon which there is
general consensus the third element. Fourthly, a `natural corollary of plan
rationality' is the `existence of a powerful, talented and prestige-laden economic
bureaucracy'. The role of this bureaucracy is explicitly instrumental. Although
inuenced by interest groups, `the elite bureaucracy of Japan makes most
decisions, drafts virtually all legislation, controls the national budget, and is the
source of all major policy innovations in the system'.3
Power within the developmental state is not exercised by a bureaucratic
monolith. There are many competing sources among the dierent ministries,
and conict between them occurs within the policy processes. However,
Johnson argues that `the centre that exerts the greatest positive inuence is the
one that creates and executes industrial policy'. That centre is dominated by
MITI. Johnson's conclusion from his analysis of the empirical evidence of the
fty-year period, is explicit: `Japan's is a system of bureaucratic rule'.4
Johnson's explanation of the role of the bureaucracy became the bench-mark
by which to judge those who oered alternative explanations. Later, as
empirical evidence of policy making in other sectors accumulated, Johnson
modied his position. He himself was aware and wrote about the decline of
MITI's power and inuence after the liberalization of the Japanese economy in
the 1960s. Its `golden era' was over with the abandonment of many of its
controls of allocation, licensing, and foreign currency, and its success in re-
establishing the basic industrial structure. He now conceded that the LDP had
begun to play a more prominent and inuential role in policy making from the
early 1970s onwards. He repeated his earlier claim that until about 1975 there
was no question who governed Japan: it was the ocial state bureaucracy.
However, during the 1970s a subtle combination of events started an
apparent decline in the power of the bureaucracy, and a concurrent rise in
the power of the LDP . . . It seems wise to speak of an apparent trend
because none of the evidence about it is conclusive. There are still powerful
ideological pressures in Japan to pretend that the LDP is more inuential
than it actually is; and the relationship between the bureaucracy and the
politicians has historically been cyclical rather than linear, with the
bureaucracy regaining power during times of crisis.5
Nevertheless he concluded that the trend away from a `bureaucratic leadership
structure' (kanryo shudo taisei) towards a `party leadership structure' (to shudo
taisei) was real. Although the bureaucracy's monopoly of policy-making powers
remained `virtually complete', political rather than bureaucratic interests
prevailed in some domestic, non-industrial sectors such as education, defence
policy and agricultural subsidies. `Most important policies still originate within
a ministry or agency, not within the political or private sectors'.6

3Johnson, MITI, pp. 2021.


4Johnson, MITI, p. 320.
5
C. J. Johnson, `Tanaka Kakuei, structural corruption and the advent of machine politics in
Japan', The Journal of Japanese Studies, 2, (1986), 2067.
6
C. J. Johnson, L. D'A Tyson and J. Zysman (eds), Politics and Productivity: How Japan's
Development Strategy Work (New York, Ballinger, 1989), p. 182.
# Political Studies Association, 1999
Review Section 943

The Neo-pluralists: Competition and Conict in the Policy-making Processes


The rst challenge to theories of a dominant bureaucracy was mounted by a
number of, mainly Japanese, scholars arguing from a variety of pluralist and
neo-pluralist perspectives. While still biased towards a `strong bureaucracy',
they identied and incorporated into their explanations the countervailing
pressures exercised by other social groups: the LDP and its organizations, both
formal and informal; sectional interest groups clustered around fragmented
policy areas; and opposition parties in the Diet. `Patterned pluralism' qualied
the assumption of a normally dominant and monolithic bureaucracy with two
institutional characteristics of Japanese policy making: jurisdictional competi-
tion and conict within the bureaucracy, and the incorporation of minority
opinion (mainly the opposition parties in the Diet) into decision making, either
directly by negotiation or indirectly through anticipated reaction. The `LDP
made [the] nal decision on political competition between groups and the
bureaucracy or between agency-group coalitions . . . Ministerial bureaucracies
propose, and the LDP decides'.7 Subsequently, Muramatsu argued that the
bureaucracy became more defensive in the 1980s in response to the impact of
administrative reform and American pressures to open the economy and expand
domestic demand, and as individual ministries struggled to maintain their
jurisdictional boundaries and their policy-making authority.8
These and other neo-pluralist theories constructed along similar lines
acknowledged the polycentric character of Japanese bureaucracy, with its
inherent and endemic jurisdictional competition and conict.9 This became
more evident, and intractable, it was argued, as policy making became less
orientated to national goals associated with the high-growth era, and ministries
and agencies competed for jurisdictional authority to control policy making in
new areas associated with social and welfare objectives and the environment,
and the development of new technologies.
The analysis of patterns of conict and methods of conict resolution became
an issue of central concern for some analysts of policy making within the
emerging critique of theories of dominant bureaucracy. Thus while Pempel
(1982) continued to emphasize `consensual and bureaucratic policy making'
within the conservative coalition of the bureaucracy, LDP and business, he
conceded that conict could arise but only from the intervention of foreign
actors and/or the opposition parties. Campbell argued the now more familiar
proposition that conict could and did occur within the conservative coalition
itself, in those circumstances where policy issues transcended the boundaries of
ministerial jurisdiction, or where jurisdiction was disputed between `sub-
governments' founded on policy areas.10 Later Samuels went further still,

7
M. Muramatsu and E. S. Krauss, `The Conservative Policy Line and the Development of
Patterned Pluralism', in K. Yamamura and Y. Yasuba (eds), The Political Economy of Japan, Vol. 1:
The Domestic Transformation (Stanford CA, Stanford University Press, 1987), p. 60.
8
M. Muramatsu, `Patterned Pluralism under Challenge' in G. D. Allison and Y. Sone (eds),
Political Dynamics in Contemporary Japan (Ithaca NY, Cornell University Press, 1993), pp. 5071.
9 T. Inoguchi, Gendai Nihon seiji Keizai no kozu: Seifu to shijo (Tokyo, Toyo Keizai shimposha,

1983); Y. Murakami, Shin chukan taishu no jidai (Tokyo, Chuo koron Sha, 1983); S. Sato and T.
Matsuzaki, Jimint Seiken (Tokyo, Chuo Koronsha, 1986).
10
J. C. Campbell, `Policy Conict and its Resolution within the Governmental System', in E.
Krauss, T. Rholen and P. Steinho (eds) Conict in Japan, (Honolulu, University of Hawaii Press,
1984), pp. 294334.
# Political Studies Association, 1999
944 Review Section

arguing that a permanent state of conict and contestation may be the natural
order of Japanese politics, paradoxically serving to stabilize the system and to
reinforce the values to which all participants must appeal. Through confronta-
tion and political struggle, they construct `ever denser networks of obligation
and reciprocity tacit compacts and protocols of reciprocal consent'.11
The increasing inuence of the LDP in the policy-making processes in the
past two decades has been attributed to three factors: excessive sectionalism
and jurisdictional inghting within the bureaucracy which presented politicians
with opportunities to support one ministry against another, or to assume a
leadership role when policy making became stalled or deadlocked; and
secondly, to the growth of policy expertise among senior LDP politicians.
The third factor was a shift in the recruitment and preferment of LDP leaders
away from former bureaucrats towards long-serving politicians from rural
constituencies. Through its formal organs, most notably the LDP's Policy
Aairs Research Council (PARC) and its Divisions, Special Committees and
Research Societies, and informally through the emergence of powerful policy
tribes (zoku-giin) but also through the attempts to inuence particular policies
by more active and interventionist Prime Ministers the LDP played a more
active role in policy making. Most accounts date this greater inuence to the
early 1970s, when the LDP responded to the challenge to its political hegemony
with welfare and compensatory policies.
That conventional wisdom has in turn been challenged from a wide-ranging
historical perspective of several policy areas by, among others, Kent Calder,
who demonstrated that the occurrence of politico-electoral crises before the
early 1970s, provoked the LDP to respond with compensatory policies.12 Other
more narrowly focused studies of individual policy sectors have found
comparable evidence of LDP inuence earlier still, in the 1950s and 1960s, to
deny or qualify theories of bureaucratic-led or dominant policy making.13
Silberman roots LDP inuence more explicitly in the historical development
of the state in the early part of the twentieth century.14 Political decisions were
constrained by the adoption of Weberian specialized and scientic rules of
administration, and political parties subordinated to the bureaucracy. The
conditions in which political parties might re-establish and maintain control of
the legislature and make the bureaucracy accountable were spelled out by Prime
Minister Hara (191821): majoritarian politics in the Diet; and a combination
of distributive politics and economic development designed to produce an
appearance of an economically rational strategy for development which neither
the bureaucracy nor private interests could resist. Those conditions proved
impossible of fullment in the inter-war period.
Nevertheless, Silberman argues, that earlier vision was implicit in the
reconstruction of party politics after World War II. The LDP became the ruling
majoritarian party and achieved and sustained a dominant position for 38 years

11
R. Samuels, Rich Nation, Strong Army (Ithaca, Cornell University Press, 1994), p. 334.
12K. E. Calder, Crisis and Compensation: Public Policy and Political Stability in Japan
(Princeton, Princeton University Press, 1988).
13 D. T. Yasumoto, The Manner of Giving (New York, Lexington, 1986); J. C. Campbell, How

Policies Change: the Japanese Government and the Ageing Society (Princeton, Princeton University
Press, 1993); K. E. Calder, Strategic Capitalism (Princeton, Princeton University Press, 1993).
14
B. S. Silberman, `The continuing dilemma: bureaucracy and political parties', Social Science
Japan, 7 (1996), 35.
# Political Studies Association, 1999
Review Section 945

by the combination of distributive politics and economic development. The


LDP's continuing monopoly of legislative power obliged the bureaucracy to
work with the party in the pursuit of economic development goals. As the LDP
successfully maintained itself in power, the bureaucracy became increasingly
politicized. In short, Silberman argues that the relationships between bureau-
crats and politicians are embedded in the long-term historical evolution of the
institutional structures of party politics, the bureaucracy and the role of the state
in economic development; the appearance of bureaucratic domination belied by
the reality of the LDP's structural power. That conclusion, but not the premises
from which it derived, is similar to the arguments of `historical institutionalists'
who believe that rationality is a product of institutional circumstances. We
look at their work below, but rst we examine the challenge of the `counter-
revisionists'.

The `Counter-revisionists': the Sectoral Challenge to the Theories of


Bureaucratic Dominance
A more sustained attack upon the characterization of Japan as a `strong state',
according a pre-eminent role to the bureaucracy, was mounted by the grow-
ing number of scholars who argued that whatever might have been the case in
the period of high growth, the political system had (before the end of the LDP's
rule in 1993) changed, and that the role and inuence of politicians in the
policy processes had increased. `Unbundling Japan Inc.', Pempel concluded
that Japanese policy making was far more complex and less coherent than two
decades earlier, with `a relative decrease in the hegemonic powers of bureau-
cratic agencies, and a rise in the inuence of the LDP and its parliamentary
members'.15 More emphatically Haley declared that `the dominance of the
Japanese bureaucracy in the political process has been grossly exaggerated. Not
only has bureaucratic inuence rarely been as signicant as generally perceived,
but also what powers the bureaucracy has exercised have declined steadily
during the post-war period'. The capacity of government ocials to determine
and implement policy was constrained by a variety of institutional factors. Any
assessment of their inuence `must be particularised to specic issues and
programs, requiring careful analysis of the circumstances of each case'.16
More narrowly, Johnson's theory of the `developmental state', with its
centrepiece of economic and industrial policy making dominated by the
bureaucracy, has been confronted with the accumulating evidence of empirical
analysis of individual industrial sectors, policies and processes. Like Johnson,
the `counter-revisionists' too emphasized the historical context or `setting', but
whereas his approach and model of the `developmental state' was top down,
they proceeded `bottom up'. Collectively their work comprises a powerful and
persuasive critique of Johnson's thesis of bureaucratic dominance. From
dierent perspectives, using dierent analytical frameworks, they arrive at

15 T. J. Pempel, `The Unbundling of ``Japan Inc.'': the Changing Dynamics of Japanese Policy

Formation', in K. B. Pyle (ed.), The Trade Crisis: How will Japan Respond? (Seattle, University of
Washington, 1987), p. 152.
16
J. Haley, `Governance by Negotiation: a Reappraisal of Bureaucratic Power in Japan' in Pyle,
The Trade Crisis, pp. 17791.
# Political Studies Association, 1999
946 Review Section

similar conclusions about the constraints on the exercise of bureaucratic


power.17
Several recent empirical studies of sectoral industrial policies provide further
support for that conclusion. Writing as an unabashed `counter-revisionist'
Callon argues that MITI's industrial policy regime collapsed as the Japanese
economy was transformed from that of a `catch-up' follower to a `caught-up'
economic superpower in the period 197585.18 The paradigm of coherence and
co-operation that formerly marked the relationships between MITI, private
companies and other bureaucratic actors was replaced by one of competition
and conict. MITI's industrial policy for micro-electronics was neither co-
operative nor successful. The dominant position achieved by its companies
occurred despite not because of MITI. Callon's claim that the whole basis for
Japan's post-war industrial policy has disintegrated since the 1970s is con-
testable. While it is true that a signicant disjuncture apparently occurred in the
making and carrying out of industrial policy between 197585, it is still unclear
whole role MITI and the other institutions of the economic bureaucracy really
played before that. The explanation of the economic miracle of the high-growth
era now appears to be more complex than Johnson and the `revisionists' once
argued. Other institutions in both the public and private sectors, and other
processes such as the formation of industrial credit might have contributed as
much or more to it than MITI bureaucrats and their industrial policies. The
assumption that the Japanese state automatically acted strategically and
purposefully has been increasingly questioned. Empirical ndings of research
at the sectoral level have revealed the structural complexity and fragmentation
of state institutions and processes, and the diculty in moving the Japanese
state to purposeful action.
Johnson earlier argued that credit allocation had been one of the state's most
important instruments for achieving the strategic transformation of the
industrial structure. Zysman had gone further, arguing that the nancial system
centred on credit relationships was `the eyes and hands of the state's industrial
brain'.19 Calder's study of government credit and industrial capital, which
questions the respective roles of the public and private sectors in bank-rolling
the economic miracle in the 1950s and 1960s is therefore a critical test of the
concept of the `development state'. His concept of `strategic capitalism' is a
`hybrid public-private system, driven pre-eminently by market-created private-
sector calculations, but with active public sector involvement to encourage
public spiritedness and long-range vision'.20 The implications of the detailed
analysis of that system for the discussion of policy making here are: rst, to cast
yet more doubt on the dominant role of the bureaucracy, even in the era of high
growth. `Japanese bureaucrats had more diculty attaining their sector-specic
objectives than commonly thought', other than in the provision of industrial
infrastructure.21 Private sector actors, here banks, trading companies, industrial

17
For a discussion, see S. Wilks and M. Wright (eds), The Promotion and Regulation of Industry
in Japan (London, Macmillan, 1991) pp. 3945.
18 S. Callon, Divided Sun: MITI and the Breakdown of Japanese High-Tech Industrial Policy,

19751993 (Stanford CA, Stanford University Press, 1995).


19
J. Zysman, Government, Markets and Growth: Financial Systems and the Politics of Industrial
Change (Ithaca NY, Cornell University Press, 1983), p. 308.
20
Calder, Strategic Capitalism, p. 16.
21 Calder, Strategic Capitalism, p. 19.

# Political Studies Association, 1999


Review Section 947

associations and rms, were more inuential. Secondly, while state structures
matter, they are not undierentiated, as implied in Johnson's MITI-centric
account. State institutions and structures vary sectorally and the amount of
regulatory authority which they possess and dispose aects the coherence of
policy outcomes. Thirdly, the complexity and fragmentation of state structures
at the sectoral level undermined the possibility of achieving broad gauge, cross-
sectoral targeting of the kind conventionally ascribed to the Japanese state.
Fragmentation and decentralization provided the means of access and of the
opportunity for inuence of private sector actors. As a result, fourthly, there
was a pluralist bias to the allocative processes. Relationships between public and
private actors with shared interests in the nancial policy area were institution-
alized within the policy processes in `circles of compensation', through which
the Japanese state allocated benets.
In the struggle for strategy, `one cannot assume the rationality and eective-
ness of industrial policy in developmentally transforming an economy from the
statements and actions of industrial bureaucrats alone'.22 In a similar vein,
Tilton argues that `to understand the true scope of industrial policy, one must
look beyond ocial state-sponsored policies to unocial policies initiated and
implemented by trade associations'.23 Their role in four basic but declining
industries reected a paradigmatic system of private-interest-governance of
Japan's political economy. Bureaucrats did not usually attempt to impose their
will on business; they worked with it to solve common industrial policies in ways
that served national policy goals. Typically MITI provided the encouragement
and sometimes the initiative for the organization of cartels which the rms
wanted but were unable to co-ordinate on their own. Its aim was neither
forward looking nor eciency oriented. It did not try to ease companies out of
inecient declining industries, but to maintain self-suciency in those vital for
national security.
By contrast with these and some other recent studies,24 Vogel's comparative
study of the transformation of the relationships between governments and
markets in Japan, the USA and Western Europe is a compelling re-statement of
the thesis of bureaucratic dominance.25 It challenges the conventional wisdom
that the processes of globalization, privatization and, above all, de-regulation
have led to less government intervention and control. In the Japanese case (as in
the UK) more competition meant more government control, as markets were
`re-regulated'. In the policy-making processes, state actors not only held
autonomous preferences, they `acted upon those preferences, and inuenced
outcomes in ways that we can not understand by focusing on private interests
alone'. But private interests also helped shape policy outcomes; `both state and
societal actors matter'.26 Vogel demonstrates the leading role played by
bureaucrats in policy making, arguing that the state is relatively autonomous
of societal interests. Competition and conict between dierent interest groups

22
Calder, Strategic Capitalism, p. 20.
23M. Tilton, Restrained Trade: Cartels in Japan's Basic Materials Industries (Ithaca NY, Cornell
University Press, 1996), p. 205.
24 E.g., J. E. Vestal, Planning For Change: Industrial Policy and Japanese Economic Development,

19451990 (Oxford, Oxford University Press, 1994); R. Uriu, Troubled Industries: the Political
Economy of Industrial Adjustment in Japan (Ithaca NY, Cornell University Press, 1996).
25
S. Vogel, Freer Markets, More Rules (Ithaca NY, Cornell University Press, 1994).
26 Vogel, Freer Markets, p. 268.

# Political Studies Association, 1999


948 Review Section

leave state actors as the interpreters and arbiters of those interests. To that task
they bring their own specic ideological biases and institutional capabilities, or
pursue a particular conception of the public interest, which partly transcends
that of private interest groups. How state actors dene and interpret the public
interest and how they pursue it is shaped by the `ideological and institutional
context' in which they operate.

`Governance by Negotiation'
While the high ground of strategic industrial policy making has been claimed
(mainly) by the counter-revisionists, the challenge to theories of a dominant
bureaucracy has continued to advance on a broader front encompassing a
variety of both domestic and international policy sectors, institutions and
processes; the list is now too long to give in detail here. Most such studies accept
the need `to engage in extensive historical analysis to determine concrete
instruments, institutions and processes of national goal setting' in order to
understand `how actual public and private organisations operate at the micro-
level'.27 They provide explanations of particular policy outputs, or changes in
policy outputs, and show why attempts to change them succeed or fail. Using
dierent methods, and developing dierent frameworks for analysis their
characterization of the policy processes is nevertheless remarkably similar.
McKean provides an overview, rejects conventional arguments about the
leadership role of the Japanese state, and concludes that Japan does not have a
strong state: `Rather, the state follows when it can, co-ordinates when it must,
and de-regulates when it cannot co-ordinate'.28
Horne's masterly study of Japanese nancial markets can stand as exemplar
for the genre. It analyses the interaction between the LDP, MOF, public and
private nancial institutions, and other corporate and non-corporate bodies
with an interest in the development and implementation of regulatory policy for
Japanese nancial markets in the 1970s and early 1980s.29 It shows rst, that
public policy making for the regulation of nancial markets is complex, rich and
varied. Disaggregation of the sector reveals ve dierent but related policy
areas, for each of which the number and the range of participants and the nature
of their interaction in the policy processes diered. In four of them the LDP had
little or no direct involvement; policy was dominated by the MOF. In postal
savings, however, where there were important electoral implications, the LDP's
view prevailed; MOF was `relatively helpless in the face of eective political
alliances built within the LDP'.30 Secondly, it shows that the role and inuence
of the participants varied with the nature of the policy-making activity. Thus the
LDP was uninterested in the implementation of regulations within settled policy.
`Few politicians within the LDP saw any need for themselves to become
involved explicitly in matters concerning nancial regulations'.31 Thirdly, the
LDP could and did inuence the general context of policy making, without

27Calder, Strategic Capitalism, p. 8.


28M. M. McKean, `State Strength and the Public Interest' in Allison and Sone Political
Dynamics in Contemporary Japan, pp. 72104.
29
J. Horne Japan's Financial Markets: Conict and Consensus in Policy-making (Sydney, Allen
and Unwin, 1985).
30
Horne, Japan's Financial Markets, p. 212.
31 Horne, Japan's Financial Markets, p. 213.

# Political Studies Association, 1999


Review Section 949

explicit intervention, like Adam Smith's `invisible hand' or the `ghost in the
machine', by establishing the parameters within which policy options could be
initiated and discussed by MOF. Fourthly, the inuence of the LDP in shaping
policy and its implementation was also evident but not visible, in its support for
the aims of particular interest groups. Fifthly, the institutional structures of
political party, bureaucracy, and private sector were complex, polycentric and
dierentiated. For example, within the MOF there are dierent interests and
objectives among competing bureaux, with dierent jurisdictions, appealing to
dierent constituencies of interest groups.

Rational Choice: `Politicians Reign and Rule'


There is little support in the work of `counter-revisionists' for a general
interpretation of policy making which accords a consistently dominant role to
politicians, although from time to time in dierent policy sectors the evidence
shows that the LDP's inuence could be decisive, and in a few policy areas such
as agriculture, public works, defence, and some education issues with ideo-
logical overtones, the policy process has been traditionally politicized. Beyond
that, most sectorally-based accounts agree only that the LDP's inuence relative
to the bureaucracy has increased since the early 1970s, although as we have seen
even that generalization is qualied by some authors.
The most provocative and full-blooded attack on the dominant bureaucracy
thesis has come from those schooled in rational choice approaches. Ramseyer
and Rosenbluth's claim that the LDP dominates policy making, and did so
before the 1970s, is a singular and, as presented, deliberately iconoclastic
interpretation of Japanese policy making, which, if tenable, not only subverts
Johnson's thesis but runs counter to the general thrust of most of the sectoral
and area studies.32 Their `choice-theoretic' approach prescribes political actors
as principals and agents competing in a political market, manipulating the
institutional framework of government to their private advantage. Sets of actors
behave rationally to maximize their self-interests. The dominance of the LDP
( principal), it is claimed, is secured through its control of the bureaucracy
(agent). The leadership delegates power to bureaucrats who make and carry out
policies consistent with the aims and strategies of the party. Bureaucrats predict
or anticipate the policy preferences of the party and act accordingly. Their
apparent dominance occurs because of the acquiescence of the leadership in
their assumption of the initiative, for example in drafting legislation. In practice
they are allowed to do so because the leadership has the means to monitor and
control their actions, and where necessary punish them if they act inconsistently
with the leadership's politico-electoral interests. Because bureaucrats know
more about the work they have been `contracted' to do than their principals,
information is asymmetrical. It is vital then for principals to monitor and
sanction agents to ensure that they comply with the conditions of the contract.
Unless this is done eectively, agents can pursue interests other than those

32 J. M. Ramseyer and F. M. Rosenbluth, Japan's Political Marketplace, (Cambridge MA,

Harvard University Press, 1993). For an extended, sober and critical review of this and, more
generally, the merits and limitations of rational choice theories, and their application to Japanese
politics, see J. P. Gownder and R. Pekkanen, `The end of political science? Rational choice analyses
in studies of Japanese politics', Journal of Japanese Studies, 22 (1996) 36384.
# Political Studies Association, 1999
950 Review Section

specied by their principals. Thus they can veto bureaucrats' policy proposals,
legislation, and actions; control their careers through promotions and postings;
and, their post-retirement careers through the practice of amakudari, the
`descent from heaven' into lucrative and prestigious private (and public) sector
jobs. Thus bureaucrats acting in their own self-interest provide the LDP with
those policies aimed to maximize their self-interest in holding on to power. In
brief, Japanese bureaucrats are nothing more than the agents of LDP
politicians; bureaucratic dominance is an illusion.
However, Ramseyer and Rosenbluth agree with the `counter-revisionists' on
the need for rigorous empirical testing of their model, and draw upon a wide
range of secondary material in support of their argument. While it provides
some support independently for each of the building blocks used in the
construction of the argument, it fails as a coherent explanation of policy making
when tested against the empirical evidence of dierent policy sectors and issues.
Like all successful political parties, the LDP's conduct in oce is shaped by
both short- and longer-term electoral considerations. Until electoral reform in
1994 the SNTV in multi-member constituencies obliged LDP Dietmen in
dierent factions to compete for votes and build and maintain local personal
networks in which votes and organizations for the mobilization of those votes
are traded for tangible community and personal benets jobs, infrastructure,
environmental amenities, business opportunities and contracts. Bureaucrats
understand the need to take such considerations into account. But it is only one
factor, whose signicance varies between policy sectors, and within them from
one issue to another, and over time. To assert that all policy making is driven by
such a politico-electoral imperative is not supported by the empirical evidence
of the sectoral studies reviewed earlier in this paper. Nor does it follow that the
behaviour of all LDP Dietmen is governed all of the time by such self-interest.
Indeed Ramseyer and Rosenbluth acknowledge the tension between the self-
interest of rank-and-le LDP Dietmen and the collective interest of the party as
a whole represented by the leadership, which they argue may dictate restraint or
policies which contradict or vitiate those immediate local interests.
The application of principal-agent theory to Japanese politics is modelled
closely on earlier applications to Congressional politics in the USA, which
revealed a `dominant legislature'. Moe's famous critique of that model applies
equally to Ramseyer and Rosenbluth's interpretation of Japanese policy
making in which the Diet is accorded a similarly dominant role: for example the
failure to identify and explain the preferences of bureaucrats themselves;
the under-estimation of `agency slack' available to bureaucrats, especially in the
implementation of policies.33 As an explanation of how policy is made in Japan
it has been criticised by many Japanologists as reductionist, over-simplied and
misleading.
Two further comments may be made briey. First, the policy making which
the LDP supposedly dominates is treated by Ramseyer and Rosenbluth as an
undierentiated activity. No distinction is drawn between initiation, formula-
tion, legitimation and implementation all identiable activities undertaken in
policy processes, involving dierent mixes of public and private organizations
and players interacting in dierent `arenas' ministry, Diet, local authorities,

33
T. Moe, `An assessment of the positive theory of ``congressional dominance'' ', Legislative
Studies Quarterly, 4 (1987) 475520.
# Political Studies Association, 1999
Review Section 951

the courts for example. Sectoral studies of social policy,34 education,35 overseas
development aid,36 and defence,37 for example reveal the complexity of the
process of initiating and formulating policy options to which many dierent
societal groups contribute. Studies of policy making of industrial capital and
credit,38 and the regulation of nancial markets,39 point to the lack of interest of
the LDP in some policy-making activities, and its minor role in others. In many
policy areas where the LDP is active, it is interested and involved mainly or
wholly at the stage of the carrying out of policy, for example overseas
development aid, public works, and takes little or no part in the earlier stages of
initiating or revising policy. Thus the second criticism is that an interpretation of
policy making which postulates a relationship of principal and agent does not
allow for a plurality of interests, other than those of the LDP and the
bureaucracy, whose interests may conict with either or both. The potential to
exercise power, dened as the possession of resources of authority, information,
expertise and money, is distributed more widely than Ramseyer and Rosenbluth
allow, and includes a variety of private sector actors, intermediary structures,
and quasi-governmental bodies. It diers between policy sectors, and changes
over time. It is deployed dierently, and with dierent outcomes within policy
sectors, according to issue, circumstance and context.
Ramseyer and Rosenbluth's explanation of policy making drew heavily upon
work then in progress of a number of historical institutionalists.40 McCubbins
and Noble use two key institutional variables electoral system and regime-
type to explore and explain dierences in policy making in the USA and
Japan, drawing a sharp distinction between the `abdication' of authority by
politicians to bureaucrats and `managed delegation'.41 Central to the abdication
thesis ( for which read dominant bureaucracy) is the possession of `hidden
knowledge' of information and expertise by bureaucrats, and their control of
the agenda. The appearance of bureaucratic power is however belied by the
reality of policy making in which, by contrast to abdication, politicians delegate
authority and manage it so that in `equilibrium' there is balance between what
legislators expect of their bureaucratic agents and what those agents deliver.
The model applied to Japanese policy making and the relations between
politicians and bureaucrats allows little qualication: either there is abdication
or there is delegation; bureaucrats have `hidden knowledge' or they do not;

34 S. J. Anderson, Welfare Policy and Politics in Japan (New York, Paragon House, 1993);

Campbell, How Policies Change.


35
L. J. Schoppa, Education Reform in Japan (London, Routledge, 1991).
36
A. Rix, Japan's Foreign Aid Challenge (London, Routledge, 1993); D. Yasutomo, The New
Multi-lateralism in Japan's Foreign Policy (London, Macmillan, 1995); D. Arase, `Public-private
sector interest co-ordination in Japan's ODA', Pacic Aairs, 67 (1994), 17199.
37 M. W. Chinworth, Inside Japan's Defence (Washington, Brassey's, 1992); J. R. Keddell, The

Politics of Defence in Japan (Armonk NY, Sharpe, 1993).


38
Calder, Strategic Capitalism.
39
Horne, Japan's Financial Markets; F. M. Rosenbluth, Financial Politics in Contemporary Japan
(Ithaca NY, Cornell University Press, 1989); F. M. Rosenbluth, `Financial Deregulation and
Interest Intermediation' in Allison and Sone, Political Dynamics in Contemporary Japan, pp. 107
29.
40 P. R. Cowhey and M. D. McCubbins (eds), Structure and Policy in Japan and the United States

(Cambridge, Cambridge University Press, 1995).


41
M. D. McCubbins and G. W. Noble, `The Appearance of Power: Legislators, Bureaucrats and
the Budget Process in the US and Japan', in Cowhey and McCubbins, Structure and Policy in Japan
and the United States, pp. 5680.
# Political Studies Association, 1999
952 Review Section

control the agenda or not; the legislature is important or not; ministers and
cabinets ineectual or not. In rejecting dominant bureaucracy tout court, an
unqualied thesis of delegation from principals to agents ascribes all policy-
making authority to politicians. In practice, as the empirical evidence of sectoral
studies demonstrates, for some kinds of policy issues in some kinds of policy-
making activity, ministers may or do abdicate authority, may be or are willing to
leave bureaucrats with discretionary authority; in others, they may themselves
prescribe specic policy options.
McCubbins and Noble's application of rational choice theories to Japanese
budgeting provide Ramseyer and Rosenbluth with much of the empirical
underpinning of their explanation of the dominant role of politicians in the
policy-making process. Using disaggregated budget data for the period
195286 they demonstrate that conventional notions about budgeting, incre-
mentalism, fair shares, and non-retrenchment all supposedly indicators of
bureaucratic rather than political inuence are incorrect, and that the LDP
through its control of the Diet had both the constitutional authority and the
electoral incentive to govern budget making, and hence to determine the
priorities of expenditure. However, they concede that MOF is a crucial player,
and that it self-evidently wins some battles. But in so doing it exercises only
delegated authority: `it does not act on its own, and it is not free to follow its
own objectives at the expense of the goals of those of the ruling party, whether
particularistic or collective'.42
The size, distribution and variation in budget outputs are a function of the
base-line used and how `expenditure' is dened. Their reliance on the data
provided by planned expenditure understates the amount of additional (revised)
spending in-year through regular and substantial Supplementary Budgets. Thus
the apparent retrenchment in the main planned budget which they detected was
in reality no more than a slow-down in the rate of growth of the budget,
measured by what was actually spent in-year (out-turn). For example, their
conclusion that `spending on public works was cut drastically in the 1980s' is
contradicted by the evidence of in-year spending.43 Measured by the outturn of
planned and revised expenditures, public works budgets actually increased
throughout the decade, and by very substantial mounts.44 Secondly, their
reliance on the data of the main planned budget and Special Accounts seriously
understates the total amount of resources available to fund public policies. The
Fiscal Investment Loan Programme (FILP), the `second budget', provided
substantial `o-budget' funds for capital investment programmes, and was used
by MOF as an alternative source of funding for some programmes in the main
budget. Throughout the period 197497, FILP grew much faster than the main
budget.45 Thirdly, McCubbins and Noble's analysis of disaggregated data
by budget `items' fails to capture the organizational dimension in determining
budget outputs, the distribution between ministries and agencies. Nor does it
reveal the extent of discrimination between policy groups and their constituent
programmes. Where programmes are shared organizationally the principles of
42 M. D. McCubbins and G. W. Noble, `Perceptions and Realities of Japanese Budgeting', in

Cowhey and McCubbins, Structure and Policy in Japan and the United States, p. 108.
43
McCubbins and Noble, `Perceptions and Realities of Japanese Budgeting', p. 105.
44
M. Wright, `The Outputs of the Budgetary System: Who Wins, Who Loses', Comparative
Budgetary Systems Working Paper Series, Japan, No. 24 (1998).
45 M. Wright, `The Outputs of the Budgetary System'.

# Political Studies Association, 1999


Review Section 953

`balance' and `fair shares' are more signicant in determining budget outputs.
For example, the budget for public works was shared among the Ministry of
Construction, Ministry of Transport and MAFF in proportions that scarcely
changed over three decades.46
But in any case proving or disproving so-called conventional norms of
budgetary allocations is not indisputable evidence for or against bureaucratic
or political dominance in the budget processes. Non-retrenchment which
McCubbins and Noble claim is a characteristic of bureaucratic inuence is
equally a characteristic of LDP behaviour in some budget programmes, notably
defence spending, agricultural subsidies and above all public works. From the
late 1970s onwards MOF's main scal aim was to cut the scal decit, reduce
government borrowing, and restore a balanced budget. The LDP apparently
acquiesced in that, but in practice exempted and excluded favourite programmes
from the operation of budget controls. A contrary conclusion to that of
McCubbins and Noble may be drawn: politicians inuenced budget-outputs to
slow-down or prevent the retrenchment policies pursued by MOF bureaucrats.
Above all, what this example illustrates is the complexity of the budgetary
processes, and the danger of ascribing dominant inuence to either bureaucrats
or politicians generally in that and other policy areas.

Towards Analytic Convergence


Revisionists, counter-revisionists and rat-choicers share some important
explanatory ground in their reaction against theories of neo-classical econo-
mics, and their rejection of culturally-determined explanations of Japanese
policy making. They also broadly agree on the importance of institutional
structures and the historical `setting' or context in which the dynamics of both
institutional and policy changes can be captured longitudinally. Each diers in
the `one big idea' but both the macro-level approach favoured by most
revisionists and rat-choicers and the `meso-level' of the policy area and sector
specialists are necessary to a fuller understanding of how and why policy is
made, and by whom; they are essentially complementary. This points to a more
nuanced analysis of policy making in order to observe the impact on dierent
policies and policy processes of the role of the state, including its transnational
dimension, and its institutional structures and their embedded `collective
identities'; and to distinguish the `appearance' of power from the `reality' of
its use.

A `Regime Shift'
The theories and hypotheses discussed here were largely constructed and rened
when both the political and economic systems were stable. The volatile
conditions of the `regime shift' which occurred following the collapse of the
`1955 political system' in 1993 will provide a searching test of their robustness
and parsimony.47 The longer-term consequences of that shift for the roles and

46
M. Wright, `The Role and Signicance of Public Works Expenditure in the Budgetary
Process', Comparative Budgetary Systems Working Paper Series, Japan, No. 20 (1998).
47
T. J. Pempel, `Regime shift: Japanese politics in a changing world economy', Journal of
Japanese Studies, 23 (1997), 33361.
# Political Studies Association, 1999
954 Review Section

relationships of politicians and bureaucrats in the new political and, less


certainly, economic and administrative orders slowly emerging at the end of the
twentieth century will not be apparent for some time to come. While the
uncertainty, complexity and delay which characterized the policy-making
processes in the period of multi-party government proved temporary, the return
of the LDP to (mainly) one-party government after October 1996 was no
restoration of the status quo ante. The processes of the de-alignment, dissolution
and re-alignment of political parties continued, stimulated by the LDP's search
for allies in the House of Councillors where it lacked an overall majority.
Relationships between politicians and bureaucrats in the policy-making pro-
cesses deteriorated, partly the result of a deliberate and sustained campaign of
criticism and denigration waged by the LDP against the bureaucracy in general
and the Ministry of Finance in particular. Its call for more political control of
policy making was strengthened by the mounting evidence of bureaucratic
incompetence in several policy areas, above all in MOF's management of the
economy and its handling of the banking crisis; and, by the public mood of
cynicism and distrust of a bureaucracy, many of whose senior ocials were
implicated in a succession of highly publicized episodes of bribery and
corruption. Seizing the policy initiative from a defensive, beleaguered and
demoralized bureaucracy, the LDP claimed publicly the inspiration of pro-
grammes of reform in the administrative, economic, nancial, scal and welfare
systems statutorily enacted between 1996 and 1999. How and by whom they are
implemented in the rst decade of the twenty-rst century will determine
whether that assertion of political power in the policy-making processes is
matched by the reality.

# Political Studies Association, 1999