Conclusion
Change and Inertia in the Politics of Japanese Public Works

Even as Ozawa and Yeutter were shaking hands and smiling for photographers at the conclusion of the negotiations that produced the Major Projects Agreement (MPA), the Japanese mass media were already predicting the onset of a second round of construction friction. A dark cloud of doubt and cynicism hung over the Japanese construction establishment. As Sako Hajime, then chairman of the Japan Federation of Construction Contractors, asserted: "The Japan-U.S. construction issue has been settled. It's over. The Japanese market is now open. But they [American firms] have no real intention of coming here. We can no longer accept the criticism that our market is still closed just because they are not willing to enter it" (Japan Times , 12 May 1988).

At least in the short run, Sako's prophesy proved correct: more than a month after the accord, no American firm had applied for permission to enter the Japanese market. Meanwhile, the fears of the Japanese construction establishment regarding an onslaught of other foreign competitors, particularly South Korean firms, quickly materialized. By late August, eleven Korean firms—among them, Samsung Construction and Hyundai Engineering and Construction—had applied to MOC, causing a Japanese executive to predict, "South Korean contractors will be a grave threat to us if they bring in cheaper wage construction workers from their own country" (Japan Times , 27 Aug. 1988).

By the first anniversary of the MPA, American frustration over the lack of progress in gaining access to the Japanese market had grown. The risk of conducting business in an uncharted market, coupled with the enormous expenses of setting up shop in an expanding bubble economy, dissuaded all but nine U.S. firms from establishing a presence in Japan. While Overseas Bechtel, Schal Associates, and Tishman Realty and Construction had managed to gain a toehold in the Japanese market, American resentment began to mount because the U.S. firms were relegated to the status of minor junior partners in joint ventures with Japanese contractors.^[1] A U.S. construction company executive charged: "Japan has given a small portion of the pie to American firms to avoid basic reform of industry practices, bid-rigging and collusive relations among political, government, and industry circles" (Shimizu 1989). Japan's tendering system continued to attract criticism for impeding foreign access. For example, despite an accomplished record in construction markets around the world, Overseas Bechtel was judged unqualified to bid on a project connected with the ¥1.14-trillion Trans-Tokyo Bay Highway.

A second round of construction friction erupted in the fall of 1989. One source of concern in Washington was the gaping bilateral trade imbalance, to which the "construction gap" was a contributor: During 1989, U.S. construction firms reportedly performed $65 million in work in Japan, compared to $2.6 billion garnered by Japanese firms in the U.S. construction market. So the U.S. government turned its attack to the "structural impediments" blocking fair access and competition in the Japanese market. These barriers included a distribution system capable of effecting a 300 percent mark-up between the import agent and the customer; keiretsu groupings that excluded overseas suppliers and blocked unwanted mergers and acquisitions; and restrictions against large retail stores that might stock imported products. Also singled out for censure were the dango system of bid-rigging, Japan's weak enforcement of its antimonopoly law, and the light penalties for violations of that law.

The criticism of the dango system focused on an illegal cartel, composed of 140 Japanese contractors, that had rigged bids, on contracts for work at the U.S. Naval Base at Yokosuka between 1984 and 1987.^[2] A hidden microphone planted on a disgruntled member of the

Stars and Stripes Friendship Association (Seiyukai) enabled the Naval Investigation Service to present the Japan Fair Trade Commission (JFTC) with irrefutable evidence of the bid-rigging operation. On 14 November, the U.S. Justice Department demanded ¥5 billion in reparations for overcharges from firms in the cartel. Meanwhile, the JFTC warned the firms and fined 70 of the companies ¥290 million for violating the antimonopoly law. For its part, MOC reprimanded 7 industry associations for not doing enough to curb bid-rigging at U.S. military bases and spanked 104 association firms by suspending their right to bid on public works projects for one to two months. In the end, 99 firms agreed to pay ¥4.7 billion in damages.

The light penalties fortified Washington's determination to press for stiffer enforcement of Japan's antimonopoly law. Scarcely two months after the Yokosuka incident, Mabuchi Construction, a leading player in the Stars and Stripes Friendship Association, won designation to bid on the construction of high-rise family housing at Camp Zama, a U.S. army base. In June 1990 MOC threatened firms caught rigging bids with suspension from bidding on public works for up to nine months, while the JFTC called for raising the maximum penalty for antitrust violations from 1.5 percent to 6 percent of illegally garnered profits. In late May of the following year, the investigation into the shady dealings of the Saitama Saturday Society first tested the JFTC's supposedly bolstered resolve. In the midst of that probe, the JFTC submitted a proposal to the LDP to increase the maximum penalty for antitrust violations from ¥5 million to "several hundred million yen." That proposal met fierce opposition from major contractors as well as influential construction tribalists. Nakamura Kishiro, then acting chairman of the LDP's Investigative Committee on Antitrust Law, was one of them. On 11 March 1994 Nakamura was arrested and charged with accepting bribes from Kiyoyama Shinji, vice-president of Kajima Corporation, in exchange for inside information about the JFTC's investigation of the Saturday Society.

As the Structural Impediments Initiative (SII) talks continued, U.S. negotiators focused their attack on Japan's underdeveloped social infrastructure. Their complaints evoked enthusiastic, if tactfully suppressed, support from Japan's public works bureaucrats and their parliamentary allies, who quickly sought to exploit the situation by for-

mulating a new ten-year public investment plan. In the initial budget proposals compiled in May 1990, the spending ministries asked for ¥500 trillion in public investment, nearly double the amount spent on public works during the preceding decade. MOC requested ¥310 trillion for roads and other projects, while the Transport Ministry sought ¥50 trillion for airports railways, and ports. Ultimately, the Japanese government agreed to spend a total of ¥430 trillion on public works in the 1990s.

A third round of construction friction exploded in 1993. In mid-April, a senior U.S. Commerce Department official hinted that the Clinton administration was considering establishing numerical targets for contracts received by U.S. firms in Japan's construction market (Japan Times , 17 Apr. 1993). On 30 April, USTR Mickey Kantor stated, "Despite years of negotiations and two trade agreements, the Japanese market remains fundamentally closed to American firms" (in Bradford et al. 1993, 6). Kantor threatened to impose sanctions within sixty days if negotiations did not produce meaningful market access. Coming in the middle of the Sagawa affair, Kantor's threat shocked an already jittery Japanese construction establishment. Even those who viewed his announcement as another instance of American bullying conceded a kernel of truth in the USTR's case.^[3] As a Yomiuri Shinbun editorial put it, "To our regret, there are various areas in which Japan must improve. The current system is a hotbed of bid-rigging and cozy relations among politicians, bureaucrats, and members of the construction industry. The system is a strong barrier to foreign companies" (in Sanger 1993b).

Although talks in June 1993 bought the Japanese side some time, they failed to extinguish the flames. Citing "a significant and persistent pattern or practice of discrimination against U.S. products or services that results in identifiable harm to U.S. businesses," USTR Kantor reserved special censure for discriminatory barriers to Japan's construction market (in Hershey 1993). However, because of turmoil in Japanese politics—turmoil surrounding the sensational zenekon scandal, preparations for the Tokyo meeting of the G-7, and especially, the upcoming general elections—Kantor delayed the imposition of sanctions until 1 November. Meanwhile, in August, MOC agreed to employ "conditional open bidding" on an experimental basis for specified

public works projects. When talks resumed in September, the Americans criticized Japans negotiating tactics, rejected MOC's proposal, and demanded an across-the-board open bidding system, expansion of the MPA, strengthening of the antimonopoly law, and the adoption of criteria to measure the progress of foreign firms in gaining access to Japans construction market.

With the USTR's deadline looming, the Japanese negotiators retreated to Tokyo to devise a plan. Any temptation to stall and press for concessions was foreclosed by the exposure of the zenekon scandal and the media-led campaign for institutional reform. Chiba Prefecture took action first. On 16 August the prefectural government announced that it would experiment with an open competitive bidding system. Other local governments quickly followed, including scandal-ridden Sendai City and Ibaraki Prefecture. On 8 September MOC agreed to implement conditional open bidding systems on an experimental basis for large, centrally contracted public works projects. But the expanding zenekon scandal and U.S. pressure soon pushed MOC even further. On 27 October Chief Cabinet Secretary Takemura Masayoshi announced the government's decision to introduce open bidding for centrally funded projects valued at or over 4.5 million IMF special drawing rights (about ¥720 million) and for projects undertaken by eighty-four public corporations valued at or over 15 million SDRs (¥2.4 million). In the case of prefectures and large municipalities, open bidding would be required for projects surpassing 15 million SDRs and for design and consultancy services on national projects exceeding 450,000 SDRs (¥72 million). Prime Minister Hosokawa predicted that the new system would bring about "historic reform with economic and social implications," and some observers estimated that the new procedures would funnel about 20 percent of national projects to foreign construction firms.

This action sufficed to persuade USTR Kantor to defer retaliatory action until 20 January in order to assess the changes. When that deadline arrived, Kantor lauded the magnitude of the concessions and expressed optimism over the potential elimination of discriminatory treatment of U.S. firms. Nevertheless, on 31 March, the USTR cited construction as one of the forty-three areas affected by unfair trading practices. With this invocation of Super 301, Japan's construction es-

tablishment had six months to institute more reforms or face possible trade retaliation. Thus, while few could have imagined it, the bilateral construction conflict continued to blaze unabated a decade after its inception.

Foreign Pressure and Domestic Change

Up until the mid-1980s foreign firms and their governments showed no interest in prying open the Japanese construction market, leaving the intricate system of clientelist exchange relations to function in relative equilibrium in a sheltered domestic arena. But in an increasingly interdependent world political economy, the existence of non-tariff barriers resonated widely and powerfully among international trade partners. Political and business leaders in the United States were frustrated by the chronic U.S. trade deficit with Japan and anxious about Japan's challenge to America's economic preeminence.^[4] Efforts to "level the playing field" in Japan's construction market was but one tactic in a larger U.S. offensive to open overseas markets. The devaluation of the dollar following the Plaza Accord in 1985 only added impetus to this market access offensive.

While seemingly irrational from a short-term economic perspective, rational calculations dictated the onset of construction friction. The fact that U.S. firms possessed superior technology for the design and construction of airport terminals made the Kansai Airport a logical focal point for the aggressive unilateralism of U.S. trade monitors. Japan's lavish budget for public works intended to create an economy driven by domestic demand, was an extremely attractive target; a U.S. Commerce Department official estimated that acquiring a fair share of that market could be worth $14 billion a year to U.S. firms (Bradford et al. 1993, 7). Among the projects discussed at the onset of the bilateral friction were an $8.7-billion bridge and tunnel spanning Tokyo Bay, a $26-billion urban renewal project on Tokyo's bay front and a similar $15.2-billion project in Yokohama, $402 billion in highway construction for 1990–1995, and a master land-use plan calling for $7.6 trillion in infrastructure investments. Yet the U.S. government did not make construction a trade issue solely because of pressure from the U.S. construction industry.^[5] Rather, it was "policy entrepre-

neurs" in the U.S. Consultate in Osaka and in Congress who placed the issue on the government's agenda (Krauss 1989).

Thus the tale of U.S.-Japan construction friction illustrates how international forces influence domestic politics and vice versa. American gaiatsu ("external pressure") compounded the problems of collusive action in Japan's public works domain. By dictating a deadline for action and moving the locus of decision-making to the top levels of the political order, American gaiatsu altered the nature of decision-making in a policy arena that had been dominated by a stable subgovernment elite. Once routine structural policymaking had been transformed into a foreign policy crisis, the Foreign Ministry, the Prime Minister's Office, and senior LDP leaders outside the construction tribe were compelled to assume new places at the bargaining table.^[6]

By insisting on the eradication of exclusionary business practices, American gaiatsu increased the risks and the costs for Japanese firms of evading the government's antitrust watchdogs. The U.S. Justice Department's decision to seek treble damages from firms that rigged bids at the Yokosuka Naval Base gave Japanese contractors a taste of American-style antitrust enforcement. A subsequent episode involving an illegal cartel that allegedly colluded on bids for telecommunications projects at the U.S. Air Force Base at Yokota reinforced that lesson (Blustein 1993b). As part of its promise to take antitrust enforcement more seriously, the JFTC increased its investigative staff by nearly 50 percent and raised the maximum penalty for violating the antitrust law. During the 1991 fiscal year, 175 Japanese companies were penalized, more than double the previous record high, and investigators levied fines fifteen times greater than in 1990.

The new JFTC was hardly a snarling antitrust watchdog, but its expanded powers represented an additional obstacle for would-be bid-riggers. Investigations of informal industry associations obliged prospective conspirators to find new means of forging collusive accords. For example, the murky activities of the Management Harmony Society came under scrutiny in the aftermath of the Yokosuka incident. Rumored to be a pipeline for funneling political contributions and sharing public works information, the society abruptly disbanded in the summer of 1990. At the time, speculation ensued in the press that a similar fate awaited various local associations. Those specu-

lations proved well founded in May 1991, when JFTC investigators descended upon the offices of forty-nine member companies of the Saitama Saturday Society. According to insiders, the Saturday Society had rigged bids on public works projects for many years. In May 1992, however, the JFTC suddenly dropped its investigation, allegedly because of "insufficient evidence." Forty-three of the member firms were fined a total of ¥1 billion and were banned from bidding on public works contracts for one month, and the firms' 100,000 employees were required to sign an oath forswearing future bid-rigging. Subsequent revelations suggest that Nakamura Kishiro and other LDP construction tribalists may have exerted enough influence to halt the investigation and reduce the proposed increase in the penalty for antitrust violations (Asahi shinbun , 16 Mar. 1994).

By prodding the Japanese side to modify the government procurement system, American gaiatsu helped to magnify the difficulties of delimiting conspiratorial rings and blocking spoilers. The cosmetic adjustments effected in 1988 under the Major Projects Agreement created an affirmative action program for foreign firms in the Japanese construction market. Between 1988 and 1991, American firms garnered around $400 million for construction services in Japan—but during that same period Japanese firms secured an estimated $7.5 billion in construction business in the United States. Even though the MPA was expanded to embrace forty projects, Japan refused to modify the designated bidder system until gaiatsu threatened severe domestic economic hardship and the unraveling political scandal began to topple high-level officials. The bursting of the bubble economy in 1991 and the onset of recession further undermined the dango system's customarily stable collusive interactions, and Japanese contractors raced to outbid one another in paying off politicans in a fierce competition for government contracts.^[7]

The construction affair offers three lessons about the impact of American gaiatsu on domestic political behavior in Japan. First, gaiatsu more often succeeds when the desired change serves the interests of key domestic actors.^[8] Indeed, the issue of Japan's increase in public works spending, enthusiastically supported by all policy protagonists in the domestic subgovernment, was the only-topic at the SII talks that elicited swift agreement, though the decision to spend ¥430

trillion on public works over the course of the 1990s fell short of the requested amount. Second, foreign pressure can alter the cost-benefit calculations of domestic actors. By prompting changes in antitrust enforcement and spotlighting murky activities, gaiatsu increased the already steep costs of collusive action in Japan's public works domain. Third, gaiatsu alone cannot dictate dramatic change in established patterns of domestic political behavior. The U.S. threat of retaliation under Article 301 transformed routine decision-making into crisis decision-making, but over the course of more than five years American pressure failed to modify the Japanese government procurement system. Modest institutional reform came only when foreign pressure capitalized on needed change in a policy domain already suffering from economic hard times and political scandal.

The Imperiled Domestic Interests

While international factors transformed Japan's construction market into a trade battlefield and influenced the timing and nature of the government's response, domestic politicial concerns pitched and protracted the conflict. As a newspaper editorial observed upon the announcement of the MPA, "The key to the settlement of the [construction] issue was how to adjust conflicting interests on the domestic front, mainly those within the ruling Liberal Democratic Party, between government departments and between the government and the industries involved" (Japan Times , 11 Apr. 1988). And, as Chief Cabinet Secretary Takemura asserted, domestic political concerns, rather than American gaiatsu , motivated the government's decision to implement an open bidding system for certain categories of public works projects (Asahi shinbun , 7 Nov. 1993).

Market liberalization most directly threatened the interests of the domestic construction industry, with its six million workers, over half a million firms, and politically potent industry associations. Having long relied on the dango system for a stable income and a "fair share" of the market—a mutual insurance system against insolvency—Japan's contractors had reason to oppose any suggestion to open the public construction market to foreign competition. Liberalization was especially terrifying for the welter of subcontractors and small-scale

firms. Although somewhat less vulnerable, the internationally competitive industry giants also resented the imposition of gaiatsu to gain special concessions for U.S. firms. Having labored to establish a profitable presence in overseas markets, many large firms were loath to concede to an affirmative action program for foreign competitors.

The construction imbroglio also imperiled the interests of Japan's public works bureaucrats. MOC officials perceived the relentless efforts of the United States to pry open the Japanese market as an encroachment upon a hallowed domestic bailiwick. The technocrats who wielded virtual monopoly control over the design phase of public works projects felt especially threatened. The public works bureaucrats knew that changes in the government procurement system would subvert their promotional and postretirement security, for the designated competitive tendering system was the linchpin in a clientelist system in which contractors rewarded helpful retiring government bureaucrats with posts suited to their descent from heaven.

In addition, the construction friction threatened the secure livelihood of Japanese legislators, particularly members of the long-dominant LDP. Over time, career politicians had come to dominate the LDP's legislative contingent, and they prized their links with construction industry constituents. Reforming government procurement policies would undermine the delicate system of "honest graft" through which a percentage of revenues from public works contracts were funneled back into the coffers of influential legislators. The prospect of construction contracts being granted to foreign firms also jeopardized the opportunities for constituency service and credit claiming so important to legislators' electoral success. Most at risk were the construction tribalists and the intraparty factions whose influence came from their reputed sway over public works decisions.

Bound by ties of mutual self-interest, legislators, government officials, and contractors were willing to endure great hardship to protect the corrupt bidding system. This "union of political, bureaucratic, and business worlds" (sei-kan-zaikai no yuchaku ) operated in an environment conducive to institutionalized political clientelism, a milieu that afforded a safe haven for bid-rigging and the economically inefficient allocation of public resources. The relentless quest for a secure livelihood bound these political actors into a concrete triangle. Yet, as an

editorial in Asahi Shinbun pointed out, "Pernicious politicians and contractors can easily weather criticism from within, but it is not so easy to sidestep the issue when the bidding system so intimately connected with dango becomes the target of foreign pressure" (7 Apr. 1993).

Growth with Clientelism

For more than forty years Japan maintained economic efficiency and high growth rates in its internationally competitive sectors alongside glaring inefficiency in many domestic sectors. Although clear signs of disquietude appeared in the mid-1980s, a delicate symbiotic balance characterized interactions between the two distinct policy regimes: Expansion of the developmental state required the stable political environment and social quiescence of the clientelist state. Meanwhile, the legitimacy of the clientelist state, with its concomitant particularism, corruption, and inefficiency, depended upon the economic growth and rising standard of living generated by the developmental state.

Partly by design and partly by default, the developmental state was largely insulated from the gluttonous clientelist networks. Policymaking for the strategic sunrise industries—unlike policymaking for agriculture, small business, and public works—rarely entailed palpable distributive benefits. State-led development, dictated by international imperatives during Japan's early industrialization, left a legacy of segregating strategic policy from structural policy in high-growth sectors. But for the clientelist state to sustain itself, it had to channel distributive benefits to pivotal claimants—to the elites—rather than to the general public. Among the groups consistently excluded from the inner sanctums of policymaking were organized labor, consumers' and taxpayers' groups, and residents of large urban areas.^[9] As the wealth generated by high-speed growth enhanced the material prosperity of the citizenry, the appeal of mass-based, ideological rallies for mobilizing partisan support gradually lost their luster, further stabilizing Japan's policymaking environment.^[10]

Because of the rural bias of the LDP's support base, peripheral and relatively low-income segments of the population benefited dispro-

portionately from the Japanese economic miracle. The residents of large cities, who received rather small payoffs in policy benefits in return for their tax burden, suffered the most. While Japan's city dwellers coped with an inferior urban infrastructure, residents of the rural hinterland were treated to a brimming barrel of political pork. Robin Hood-like, the tribalists performed "a vital function for the system, redistributing income from the rich sectors to the poor ones and ensuring that high-speed growth did not benefit one group to the exclusion of others" (Johnson 1986, 20). With the benefits accruing to business interests, political clientelism under protracted single-party dominance acted as a proxy for social welfare expenditures. (At least until the mid-1970s, Japan ranked at or near the bottom among advanced industrialized countries in such expenditures.)

A second factor helps to explain the anomaly of economic growth within a clientelist political order. In addition to promoting the broad consultation and interaction that characterized government-business relations in Japan, a deeply entrenched system of informal structures and practices bound public- and private-sector elites together. Out of this grew a pervasive corruption "so highly organized and . . . so much a part of the extra-legal ways of the Japanese system that most citizens . . . do not recognize it for what it is, but accept it as part of the system" (van Wolferen 1989, 136). A general contractor, gazing in horror at the unfolding zenekon scandal remarked: "Was what we were doing really illegal?" (Asahi shinbun , 21 Oct. 1993). As long as standards of living continued to rise, systematized corruption legitimized the status quo. The cumulative effect of the corruption scandals beginning in the late 1980s revealed the materialist covenant that bound special interests, incumbent and retired government officials, and candidates for elective office in a passionate embrace.

In the structural policy market, questions of who go what, when, and how depended on who wielded the greatest political influence. Particularly during the era of high-speed economic growth, when Japan had a relatively underdeveloped social welfare system, officially sanctioned inefficiency abounded in those policy arenas in which the redistributive effect served as a proxy for social welfare policy. Although the social welfare system improved dramatically, rice growers, construction contractors, and shopkeepers continued to reap formal

and informal government protection in spite of high rates of inefficiency and declining rates of improvement in labor productivity.^[11] With regard to public works, a cordoned-off domestic market enabled the growth of an oversized and relatively inefficient domestic construction industry. Here, American gaiatsu helped create a focal point for mass-media criticism, which rallied public disenchantment over the domestic costs of Japan's "structural impediments." And, further reinforcing a widespread sense of deprivation, the SII talks spotlighted the backwardness of Japanese infrastructure vis-à-vis other advanced countries.

Pork-barrel politics, however, sustained the LDP's protracted legislative hegemony. The money that poured into the coffers of the LDP and its factions essentially covered the cost of an insurance premium for maintaining stable rule under a one-party dominant regime:

The LDP, maintaining its monopoly of power, succeeded in making the economy grow through the expansion of exports. This provided sufficient financial resources to enable the party to follow a policy of distributive politics. Constituency service was provided by appropriations that poured into the election districts. This in turn enabled the party to secure a majority of seats in the National Diet and thus perpetuate its rule. . . . In this way, a stable system that linked politics and the economy came into being.

(Kyogoku 1987, 23)

This arrangement helped to insulate Japan's strategic policy market from partisan demands that might have sidetracked developmental goals. By satisfying greedy self-interested politicians, bureaucrats, and special interests, political clientelism served an important systemic function.

In this way, the political requisites of the developmental state dovetailed with the dysfunctional clientelist system. The maladies of "part versus whole" and "constituency versus nation" contributed to the equitable distribution of national economic wealth that legitimized Japan's steep and painful developmental trajectory to the front ranks of industrialized countries. A seamless web of mutual interactions connected a system of institutionalized corruption with a government-business partnership founded upon extensive bargaining and cross-fertilization. In sum, insofar as the inefficient allocation of public resources sustained the LDP's protracted hegemony, the dysfunctions

of a political clientelist order served essential functions in maintaining developmental capitalism.

Another Japanese Export?

Japan's economic success leads one to wonder if its model for growth with clientelism is exportable. In answering this question, we cannot ignore two principal peculiarities of modern Japanese history. First is the legacy of institutionalized political rule in a developmental state, a regime in which an elite civil bureaucracy asserted itself in policymaking and in calculating the national budget. The centralized state that emerged in the latter half of the nineteenth century posed steep obstacles to partisan meddling in the budget or in industrial policy.^[12] Second, the carefully controlled extension of suffrage ensured bureaucratic dominance over civil society and expedited state patronage of favored groups and individuals. The timing and sequence of these developments was significant for Japanese economic growth. At least until the mid-1970s, the state bureaucracy managed to maintain a strategic role in economic policymaking, while political clientelism was basically confined to structural policy arenas. In this regard, the case of Japan contrasts markedly with Italy, where, under the blocco storico , a constituency created for partisan patronage emerged before one formed for bureaucratic autonomy could take shape (Shefter 1977, 441–46). Simply stated, the sequence of these two developments left Japan with an elite, meritocratic, and activist state bureaucracy, while Italy emerged with an overstaffed, patronage-ridden, and unesteemed corps of civil servants recruited in disproportionate numbers from the backward south.

In light of these historical differences in politico-economic structure, the Japanese model cannot easily be duplicated by other advanced industrialized democracies. The broader range of salient interests, the greater empowerment of consumers and other non-producer interest groups, and the weaker position of government officials (as compared with elected politicians) in the policymaking process—all make replication impossible. In addition, the calls for political reform in many advanced liberal democracies of North America and Western Europe challenge the prospects for emulation

of the Japanese model. The United States, with its laissez-faire economic ideology, stringent antitrust codes, and vocal consumers' and taxpayers' lobbies, appears particularly unsuited. Even in Italy, where systemic dualism characterizes the politico-economic order, it would be difficult to replicate the Japanese experience.Fashioning an elite and truly meritocratic state bureaucracy without any vestige of patronage from key industrial sectors would also pose a challenge.

While Japan 's domestic political and economic structures more closely resemble those of developing countries, successful emulation in those countries is unlikely as well. Comparable systems operate in South Korea and Taiwan, and some of the structures and conditions for emulation exist in Singapore, Malaysia, and Thailand, but the majority of developing countries in Asia, Africa, and Latin America have high levels of income inequality as well as unrestrained patronage and political corruption. Moreover, the world political economy is no longer hospitable to protectionism. During the first three decades of the postwar period, the heyday of Pax Americana, countries like Japan were allowed to protect domestic industries and sectors to create a stable capitalist world economy to contain the "communist menace." But the collapse of communism in Eastern Europe has eliminated the rationale for conservative solidarity and for American support of conservative regimes in countries such as Japan and Italy. Indeed, steep international opposition will challenge any attempt to replicate the Japanese model in today's increasingly intolerant world political economy, riddled with conflicting trends toward higher levels of interdependence and rising demands for aggressive unilateralism.

In sum, the Japanese model does not appear exportable: the economic, political, and social costs of transplanting and sustaining it in toto are too high.

How Institutions Matter

Although all democratic polities exhibit sporadic instances of collusion in public works, among advanced countries only Italy can rival Japan in the systematic and pervasive nature of such activity. Under the guise of a political party, the LDP, a federation of warring factions whose candidates battled ruthlessly against one another, dominated

the legislative arena for nearly four decades. While pork-barrel politics and the "personal vote" abound in other polities, the degree and prevalence of these phenomena under LDP rule contradict a central assumption of comparative political wisdom, that pork-barrel activity is lower in systems featuring multimember legislative districts. Surprising as well is how little "bacon" influential LDP legislators brought back to their home prefectures.

If we are to understand these patterns, the question is not whether institutions matter—they obviously do—but rather, how the institutional rules of the game limit courses of action and create incentives and disincentives for certain modes of behavior. In the politics of public works, the pivotal institutions in Japan are the government procurement system, the civil service employment system, and the electoral system; in addition, the political party system affects the structure of the policymaking process and the interactions among political elites.

As we have seen, the government procurement system, by limiting the pool of bidders to ten "qualified" firms, simplifies the complex task of apportioning the costs and benefits of collusive action, it reduces the colluders' transaction costs and bars entry by outsiders. In this way, the system creates a rationale for trade associations—and, on occasion, the organized underworld—to enforce the selective incentives that sustain illicit arrangements. Since firms must be designated in order to bid, there is ample incentive for contractors to construct mutually beneficial bridges to government officials and politicians. In return for their expertise and access to confidential government information, retired bureaucrats earn lucrative second careers with construction firms. And in return for invoking the "voice of heaven" in the allocation of public works contracts, influential politicians receive campaign support and political money from grateful contractors. The competition among firms seeking to outbid one another for the services of ex-bureaucrats and influential legislators inflates the cost of public construction and increases the likelihood of political corruption. In an already collusion-prone industry like construction—with its lumpy demand and high levels of government procurement—lax enforcement and weak penalties for antitrust violations are powerless to discourage price-fixing cartels.

The civil service employment system conditions the behavior of government beaucrats and their interactions with special interests and legislators. Because the majority of government bureaucrats in Japan retire between the ages of 50 and 55, they must "descend from heaven" into second careers in the private sector, quasi-governmental entities, and elective politics. Amakudari thus blurs the boundary between public and private realms even more so than the revolving-door phenomenon in U.S. Pentagon procurement. In Japanese public construction, the need to secure amakudari landing spots prompts particularistic consideration in designating bidders and in affording strategic leaks of confidential information to firms employing former bureaucrats. The lifetime employment and amakudari systems also offer fierce disincentives for would-be whistle-blowers. Since the postretirmenent prospects of government officials relate directly to the level and prestige of their final posting in the bureaucracy, officials do whatever it takes to secure promotion. To that end, bureaucrats build bridges of mutual benefit to legislators whose influence might be instrumental in securing passage of budget proposals, in greasing promotional wheels, and in securing attractive amakudari landing spots. In return for these services, bureaucrats assist in the preferential allocation of public works contracts, advance information to bolster credit-claiming, and offer various forms of informal campaign support. The civil service employment system thus promotes dubious behavior in supposedly impartial public administrators.

Electoral institutions affect the way legislators interact with voters and campaign contributors, government bureaucrats, and other legislators. Owing to the single, nontransferable vote, for example, LDP candidates were always competing against one another for the conservative vote in multimember disticts. This arrangement fostered a strong particularistic orientation in which candidates competed to provide services to personal constituents as they downplayed ideology and policy issues (Grofman, forthcoming). The combination of intra- and interparty competition drove LDP candidates to barter with special interest groups, offering access and favorable policy treatment in exchange for reliable blocs of votes and substantial political contributions. In such contests, hereditary, politicians had the advantage over newcomers, who had to create their personal constituencies from scratch. In addition, competition among LDP candidates at the dis-

trict level created a rationale for intraparty factions, which assisted in securing party endorsement, campaign funds, political posts, and constituency services. Furthermore. the SNTV system enhanced the importance of money in politics, and Japan's decentralized campaign finance law rewarded those candidates who built the most complex financial support networks.

Finally, the political party system put a distinctive spin on the behavioral patterns of self-seeking actors. Hegemonic party regimes tend to produce higher degrees of elite cohesion and stability, fewer effective veto points, and more selective interest group access than do polities governed by multiparty coalitions or party governments (Weaver and Rockman 1993, 18). In both Japan and Italy, protracted single-party rule granted disproportionate weight to the interests of governing party candidates and fostered institutionalized links between the dominant party, its support groups, and the state bureaucracy. "As far as the party affiliation of the deputy is concerned," observed an Italian bureaucrat, "the only party that counts at the moment in public administration is Christian Democracy. We find little need to pay attention to the demands or the threats of the other political parties, or their representatives, or their deputies" (La-Palombara 1987, 319). Japanese government bureaucrats have voiced similar views, calling the LDP "the party" (to ), as if a single-party dictatorship ruled Japan. Nearly four decades of single-party dominance enabled the institutionalization of information pipelines to the government bureaucracy, and through these lines policy benefits were funneled to the LDP's key support groups and denied to its opponents. In both Japan and Italy, protracted single-party hegemony enabled a de facto shift of influence over policymaking away from parliamentary committees and into the deliberative organs and factions of the ruling party

In sum, the institutionally determined rules of the public works game in Japan produced a routinized system of bid-rigging, rampant bureaucratic turf wars, electoral rivalries among partisan colleagues, intrapary factionalism, and enigmatic pork-barrel politics. The synergistic interaction of several critical institutions gave shape and substance to a concrete triangle of legislators, bureaucrats, and contractors. Therefore, meaningful reform in this policy domain would have to address all the institutions that channel the self-interested behavior

and interactions of the central political actors. Indeed, the steadfast Japanese resistance to U.S. pressure to pry open the domestic construction market, as well as the reform crisis that hastened the LDP's collapse, derived from a recognition of the vital interests served by these institutions.

Wrecking Ball or Face-Lift?

Japan's existing institutions could not harness the particularism, corruption, and inefficiency that arise when political clientelism takes precedence over the general welfare. In 1993 international and domestic factors combined to topple the LDP's legislative hegemony and pave the way for institutional reforms. The end of the cold war deprived the LDP of the scant ideological glue of anticommunism, and deteriorating economic conditions promoted dissatisfaction among voters and important business interests. Meanwhile, a succession of sensational political scandals made it impossible for LDP leaders to continue to resist policial reform.

The fall of the Berlin Wall in November 1989, signaling the end of cold war ideology, jolted mainstream conservative parties like the Christian Democrats in Italy and the LDP in Japan.^[13] With the dissolution of the Soviet Union and the collapse of communist regimes throughout Eastern Europe, the Christian Democrat and LDP candidates lost their tenuous ideological ties to their respective parties. For both parties, a distaste for communism and a general avowal of capitalist democracy had been the only common ideological denominators. The end of the cold war also allowed the United States to re-evaluate its support of conservative regimes in Western-bloc countries. The Central Intelligence Agency had once funneled clandestine funds to prop up the LDP, but now the Clinton administration felt the American interests would be better served by an open show of support for opposition parties advocating political reform.

Clear signs of domestic unrest had appeared in Japan during the mid-1980s. Soaring land prices drove the cost of home ownership far beyond the means of the average urban salaried worker. A growing gap between haves and have-nots divided a society that had long perceived itself as a monolithic middle class. Taxpayers and consumers recognized the high cost of policies created to protect domestic inter-

est groups—such as rice farmers, small shopkeepers, and construction contractors—and they complained that their standard of living and the country's level of infrastructure development were not as high as they should be. The expression "rich Japan, poor Japanese" gained popularity. Somewhat surprisingly, given the widespread repugnance for American gaiatsu , public opinion polls showed that almost half of the Japanese people supported most or all of the U.S. government's demands in the SII talks (Nihon keizai shinbun , 27 Mar. 1990). In addition, the maturation of vital industries reduced the demand for government intervention, while a pronounced gap emerged between the needs and expectations of internationally active firms and those with domestic spheres of interests. The turbulence caused by the bursting of the bubble economy and the onset of economic recession also undermined the political status quo (Ishi 1993; Wood 1992).

Against this backdrop, an unprecedented succession of corruption scandals broke, in Japan and abroad.^[14] Scandals were nothing new in Japan: between 1975 and 1993 there was roughly one major political corruption exposé a year. As it happened, however, two of the most sensational—the Recruit scandal and the Sagawa affair—broke in the wake of the international and domestic turbulence described above. The combined effect of an unpopular sales tax, opposition to agricultural liberalization, and the Recruit scandal contributed to the LDP's stunning setback in the 1989 Upper House elections, in which the party lost its absolute majority of seats. The LDP's precarious position in the Upper House obliged both the Kaifu cabinet and the subsequent Miyazawa cabinet to focus intensively on reform. As the intervals between scandals shortened and the extent of the exposés broadened, the LDP could no longer pay lip service to reform. The party's leadership would actually have to attempt to do something.

Nevertheless, no one background factor was sufficient to motivate fundamental institutional reform in Japan—or in Italy, for that matter. The Berlin Wall had fallen and Eastern Europe's communist regimes had collapsed, but cold war tensions persisted in East Asia.^[15] Although the economic turbulence that began in the late 1980s might have heightened fears of voter reprisal in Japan and Italy, both the LDP and the Christian Democrats had survived similar challenges in the past. And while the Sagawa and zenekon scandals and the Mani Pulite investigations were highly sensational and involved major po-

litical leaders, both regimes had weathered countless corruption incidents. What was distinctive was the unprecedented magnitude of the Italian scandal and the quick succession of the major Japanese scandals. Moreover, while the Italian electorate cast out the Christian Democratic rascals, the LDP emerged from the 1993 elections with one more seat than it had at the time the elections were called. That is to say, the fall of Japan's Liberal Democrats, which paved the way for the creation of a non-LDP coalition of reformers, was not dictated by disgruntled voters.

While diverse factors contributed to the timing of the collapse of single-party, hegemony and paved the way for political reform in Italy and Japan, the ultimate force for change was structural. ^[16] The need for massive infusions of money to maintain the LDP's largest faction ultimately dictated the fall of Kanemaru and the subsequent fracturing of the Takeshita faction, which triggered the end of the LDP's legislative hegemony. Kanemaru's mythical political clout had enabled him to charge off-scale rates for his brokerage services in the allocation of distributive policy benefits, but once it became clear that Kanemaru was doomed, Ozawa Ichiro and his followers had to distance themselves from their mentor to preserve their own political future. So they broke off from the Takeshita faction and established the Reform Forum 21, with Hata Tsutomu as its titular head. When the LDP failed to agree on a meaningful electoral reform bill, Ozawa left the party altogether and created the Shinsei Party.

The collapse of the LDP's legislative hegemony was dictated by the success of Ozawa's party and the other "new parties" in the July 1993 elections, and the establishment of a flimsy non-LDP coalition held together by a vague consensus concerning the need for electoral reform. Had Ozawa and his followers chosen to remain in the fold, the LDP would have prolonged its mastery of the parliamentary realm. And had the LDP's leadership correctly perceived Ozawa's desperate desire to protect and enhance his political power, steps could have been taken to keep him and his followers from defecting. In the end, the LDP could not agree on a political reform bill because the career politicians, who had labored arduously to construct personal support networks, opposed institutional reforms that threatened their incumbency.

The enabling conditions for political reform in Italy and Japan were

thus broadly similar. The deaths of their respective hegemonic regimes were caused by political trichinosis, contracted from decades of doling out ever-increasing quantities of pork. This rapacious appetite for pork and the pervasive clientelist structures in both polities were the accumulated product of rational strategies adopted by politicians, bureaucrats, and corporate officials to manipulate the institutional rules of the game for maximum benefit. The end of the cold war, economic hard times, and corruption scandals involving major political figures—these played a part, but structural imperatives eventually made conditions ripe for reform. And in both polities, exposés involving systematized clientelism—especially in government procurement and public works—generated widespread condemnation and demands for reform. No matter how self-serving their respective motives, the vocal endorsement of institutional reform on the part of established powerbrokers like Bettino Craxi and Ozawa Ichiro reinforced the mounting pressures for change. The fall of the Christian Democrats and the LDP resulted from the inability of the respective political systems to harness the spiraling ante and ever more costly by-products of systematized political clientelism.

Implications for Reform

Beginning in the summer of 1993, each of the key institutions in Japan's public works domain experienced some reform within a matter of months. Though these reforms were, at best, only partial, and though it is too soon to gauge their full effects, some signs of change have already appeared.

By the spring of 1994, all but two prefectures and all but one of the twelve largest municipalities either had introduced conditional open-bidding procedures for large-scale public works projects or had taken steps to implement such procedures on an experimental basis (Asahi shinbun , 2 May 1994). For example, nearly one-third of the sixty-eight firms competing to construct a park in Kobe City submitted bids under the minimum acceptable price. The losers in the bidding for a riparian project in Miyagi Prefecture accused the winner of "dumping" (ibid., 9 and 14 Feb. 1994). About ¥456 million separated the high and low bids submitted by twenty-two joint ventures vying for a tunnel project in Ibaraki Prefecture. In the case of a highway tunnel

project in Kagawa Prefecture, the lowest bid among the twenty-eight competing firms came from a Tokyo-based firm that had not been designated to bid in the original round (ibid., 1 Dec. 1993 and 14 Feb. 1994). An editorial in the Asahi Shinbun commented, "Contractors with no previous record are jumping into the fray and winning contracts, competition to submit the low bid is intensifying, and a succession of contractors are submitting bids below the minimum acceptable price. This state of affairs would have been inconceivable in the era of designated bidding" (14 Feb. 1994). Indeed, the changed environment prompted the chair of the Japan Federation of Construction Contractors to call for "self-restraint" in curbing excessive competition among firms vying to submit the low bid. MOC has even contemplated a policy to restrain price-dumping in an industry long characterized by, inflated bids (ibid., 22 Feb. and 4 Mar. 1994).

Other changes have included the appearance of new actors at the credit-claiming podium. For example, Socialist Party MP Koshiishi Azuma actually preceded an LDP rival in claiming credit for the construction of a segment of the Trans-Chubu Highway through Yamanashi Prefecture (ibid., 6 Feb. 1994). Naturally, such events rarely occurred during the era of LDP hegemony. Moreover, MOC's new policy of self-restraint in the securing of postretirement posts for its upper officials has produced some results. At a press conference in March 1994, Construction Minister Igarashi Kozo announced that in 1993 fewer than 20 MOC officials descended into positions with construction firms, compared with 145 individuals the previous year. And, breaking with tradition, no upper officials assumed positions with large-scale general contractors. According to Deputy Vice-Minister Ban Noboru, the self-restraint policy obliged MOC's 1993 contingent of descending angels to commence second careers outside the construction industry or, in some cases, to opt for outright retirement (ibid., 22 Mar. 1994).

Nonetheless, although the reforms of 1993 and early 1994 marked the most concentrated and significant instance of institutional transformation since the early stages of the U.S. Occupation, they have not eradicated all vestiges of the old order. The 1993 general elections toppled the LDP from the commanding heights of the legislative realm, but former LDP legislators claimed the most powerful posts in

the ephemeral cabinets of Hosokawa Morihiro and Hata Tsutomu, and Ozawa replaced Kanemaru as the shadow ruler in these "non-LDP" cabinets. Reminiscent of the "opening to the left" policy of Italy's Christian Democrats in the 1960s, the LDP became the dominant coalition partner in the administration of Socialist Prime Minister Murayama Tomiichi in June 1994.^[17]

While the rules of the electoral game were significantly altered, the Hosokawa cabinet had to offer concessions in order to secure a compromise with the LDP. The LDP had its way with regard to donations from organizations and corporations, and, more importantly, the LDP dictated the number of single-seat districts. As a local contractor cynically observed, "Even if the LDP loses power, nothing will change" (ibid. 19 Aug. 1993). Italians, too, began to question the possibility of real change, given that Silvio Berlusconi reportedly once controlled many of the politicians arrested in the Mani Pulite operation. In the wake of the investigations, bribe-filled envelopes (bustarelle ) have continued to change hands.^[18] For meaningful change to take place, elected politicians will need good reasons to refuse bribe-filled envelopes or, for that matter melon boxes filled with cash. In the case of Japan, under an electoral order in which three-fifths of the legislative seats are to be chosen by a single-vote formula that encourages clientelism, the best way to curtail such corruption is to reduce the cost of pursuing a political career by enforcing strict limits on campaign contributions and by increasing public financing for political parties.

Despite the introduction of a conditional open bidding system in the allocation of large-scale public works contracts, the procurement system for small and medium-scale public works projects has not been modified. Instances of bid-rigging have occurred even in projects using supposedly open bidding procedures.^[19] In the words of a major construction company official, "The reform [of the bidding system] is nothing but a cheap trick to deceive the general public, and it's ridiculously silly. This will not cause dango to disappear" (ibid., 21 Oct. 1993). As a local contractor mused, "Dango will continue just as it did under the designated bidder system" (ibid., 25 Sept. 1993). Many recall that the anti-dango media campaign in the early 1980s forced the disbanding of the corrupt Construction Fellowship Society, which subsequently reemerged as the equally corrupt Management Har-

mony Society (ibid., 22 Aug. 1990). And bid-rigging continued to take place even in the midst of the JFTC investigation into the activities of the Saitama Saturday Society (ibid., 10 May 1992). To minimize the by-products of bid-rigging, Japan have to initiate open, competitive bidding procedures for all public works and be vigilant in enforcing painful penalties for antitrust violations.

Predictably, MOC's promise to exercise self-restraint has not put an end to the practice of Government bureaucrats descending into sinecures with construction firms and agencies that allocate public works contracts. The promise, a partial and temporary measure, was made just as arrests of MOC officials were anticipated in the zenekon scandal. Although some MOC officials were questioned in the zenekon investigations, none was arrested or prosecuted. Once the sensational exposés and intense media scrutiny faded, public works bureaucrats again enjoyed postretirement sinecures as their deferred compensation for less-than-lucrative careers in the government bureaucracy. Piecemeal reforms cannot curb the social costs associated with a practice as deeply rooted as amakudari: what are needed are stricter regulations concerning the reemployment of ex-officials in firms that receive public contracts and an increase in the retirement age for government officials.

Entrenched patterns of elite behavior die hard: "Although wholesale change in formal rules may take place, at the same time there will be many informal constraints that have great survival tenacity because they still resolve basic exchange problems among the participants, be they social, political, or economic" (North 1990, 91). Meaningful reform requires more than cosmetic changes in the rules of the game; it demands overhauling the key institutions that provide incentives for particular patterns of behavior among the political elite. It also demands that elites develop the will and the stamina to enforce the letter as well as the spirit of the new rules. As long as the old patterns of behavior continue to perform essential functions for the political elite, those patterns will endure.