One
Construction Contractors and the Calculus of Collusion

In January 1986, the local press in Osaka carried an inconspicuous article detailing the plans of the Kansai International Airport Corporation (KIAC) to invite bids on contracts for the construction of a levee. The article came to the attention of Keith Bovetti, a commercial counselor in the U.S. Consulate in Osaka. For over a year, Bovetti had been working to generate interest among American firms in Japan's increasingly lucrative service sector, which included public works and third-sector construction projects. Along the way, KIAC informed Bovetti that foreign companies could not bid on the project outright; they could only join Japanese firms as subordinate partners. This news did not rest well with him, for he feared that failure to break through at Kansai inhibit future efforts of U.S. firms. To alert American firms of the bidding, Bovetti sent a translation of the newspaper article to the Commerce and International Trade Agency in Washington. Later, he dispatched a translation of the bidding requirements and procedures. Eventually, some forty-three American companies registered with the KIAC.^[1]

Rather than providing a cheerful end to a short drama, this incident generated a protracted trade dispute. The conflict centered around the U.S. demands for "fairness" and "reciprocity," and Japan's refusal to reform its government procurement system. Vitriolic rhetoric and promises of retaliation issued forth from both shores. U.S. Assistant Commerce Secretary H. B. Goldfield, in testimony before a Senate

subcommittee said, "It's very clear that the Japanese want to protect their own industry. At the same time, they want to participate in our open market. Mr. Chairman, that just isn't fair" (quoted in Auerbach 1986). Meanwhile, the USTR threatened to invoke Section 301 of the 1974 Trade Act.^[*] Adding vigor to these threats, Alaska's Senator Frank Murkowski and Texas Congressman Jack Brooks sponsored a bill banning Japanese firms from participating in federally funded public works in the United States. The bill passed by near unanimous margins in both houses.

Undaunted, a procession of Japanese spokesmen staved off the foreign adversary. Officials from the Ministry of Construction and the Ministry of Transport staunchly defended the government procurement system. Politicians uttered similar refrains. Transport Minister Hashimoto Ryutaro (1988) argued that foreign firms interested in bidding on public works must first accumulate a "number of years" of experience in Japan's private construction market. Pressure group officials and business leaders also took the fore. Ishikawa Rokuro, then president of Kajima Corporation and head of the Japan Chamber of Commerce and Industry, maintained that

Since public works are financed by taxpayers, it is important to ensure that they are carried out efficiently and that the work is up to standard. Thus [in Japan] bids are only accepted from designated contractors—contractors who have proven their ability to do the job right. Within this group, the lowest bid is chosen and a contract is signed. This is the designated competitive tender system—a system adopted in many countries throughout the world. If a foreign company wants a contract for a Japanese public works project, it should follow the Japanese rules.

("Conductor of Commerce" 1988)

Why did the Japanese government allow this matter to exacerbate relations with its most important trading partner? Despite the pronouncements of the apologists for Japanese policy, the rationale for the formal procurement system was to protect not Japanese taxpayers, but an institutionalized system of bid-rigging known as dango ("agreement by consultation").

Although explicitly proscribed by the Accounting, Anti-Monopoly, Criminal and Construction Industry laws, most public works contracts in Japan come about through shadowy dango arrangements. Evidence of bid-rigging has been reported in virtually every locality and most public corporations. A headline in a major Japanese daily proclaimed: "Majority of Large-Scale National Projects Decided Through 'Dango '—¥1.8 Trillion Divvied-Up by 'Friendship Club' of 170 Major Contractors" (Yomiuri shinbun , 3 Dec. 1981). Likewise, a construction company official casually noted that most contracts for public works during the course of the past several decades had been decided beforehand through dango , a practice that he views as "routine work" in the construction industry ("Ote kensetsu" 1982). As a newspaper reporter observes, "The Japanese archipelago lies in the midst of a sea of dango (Asahi shinbun , 29 Aug. 1993).

Though the term dango benignly denotes "consultation" or "conference," its actual meaning is reflected in its many aliases, which include "adjustment," "resolution through discussion," "tendering pact," "shady cartel," and "discussion for adjusting the exchange of information and communications." "Above and beyond the act of collusion among those submitting bids for a project, the term 'dango ' denotes prior mutual consent in determining the successful bid" (Nawa 1987, 34). Less formally, dango is an "independent arrangement" among contractors to decide the successful bidder, or the contractors' practice of "neatly dividing up and parceling out their construction market by a 'gentlemen's agreement.'"^[2]

Ironically, an instance of bid-rigging emerged in a project connected with the construction of the aforementioned levee at the Kansai Airport. Following the exposure of the collusive accord in a Lower House Budget Committee meeting in September 1989, the Japan Fair Trade Commission issued a cease-and-desist order against six construction companies that composed the Marine Land Reclamation

Construction Association, an illegal cartel that conspired to set the price of soil needed for reclamation. This collusive action allegedly produced a skimmed profit in the neighborhood of ¥5 billion, a portion of which supposedly went to influential politicians.^[3]

Achieving and sustaining a collusive system, it is important to note, is enormously difficult. Indeed, the literature on cartel theory overflows with lists of "conditions," "factors," and "obstacles" that block the path to successful collusive action.^[4] For instance, John McMillan (1991, 204) uses the dango system to illustrate the three "difficulties" that must be overcome for successful price-fixing to occur: creating a mechanism for dividing the spoils, self-enforcing the agreement, and eliminating competition from new entrants.

Building upon this literature, I propose that successful and sustained collusive action demands the fulfillment of five necessary and sufficient conditions. First, it is essential to delineate the membership of the conspiratorial arrangement. Second, barriers must be erected to prevent opportunistic outsiders from undercutting the agreed-upon price. Third, since conspiratorial agreements are not legally enforceable, compliance must be secured through informal means. Fourth, there must be a mechanism for apportioning the benefits and costs of collusive action. And, fifth, a successful collusive system must have the means of evading or coopting the government's antitrust watchdogs in order to survive for an extended term. Moreover, in an increasingly interdependent international economy, the patterns of collusive behavior must evade trade monitors in other countries.

The fulfillment of these conditions poses such obstables to successful and sustained collusion that the vast majority of documented attempts have been short-lived. As we will see, these five conditions were satisfied under the dango system. And, as one would expect, the equilibrium of collusive action was destabilized by the increased public scrutiny generated by the zenekon scandal and the imposition of U.S. pressure for institutional reform. By exerting pressure, the United States underscored the drastically divergent interests and capabilities of Japan's technologically advanced large-scale general contractors as compared to its petty contractors.

But before examining how the collusive dango system operates, we need to understand something of the nature and history of Japan's construction industry.

Profile of a Two-Tier Industry

The Japanese construction market is the largest in the world. In 1992 Japan's construction investment as a percentage of gross national product (18.2 percent) was the highest among advanced industrialized countries, exceeding England (12.4 percent), Germany (11.7 percent), France (10.8 percent), and the United States (8.5 percent) by a wide margin. On a per capita basis, construction spending in Japan ($3,480) was double that in the United States ($1,630) and in the European Community ($1,690). From 1984 to 1989, the Japanese construction market expanded by 27 percent—ranking second only to late-developing Spain among Western nations—and between 1987 and 1993, the Japanese market expanded by nearly 50 percent. Construction investment in 1993 was estimated to be just under ¥90 trillion, compared to ¥2.5 trillion in 1960. Throughout this period, government spending on construction played a significant role, accounting for over one-third of total construction investment, on average.^[5]

Other measures of growth include the number of construction firms and construction workers. In 1955, for example, there were fewer than 62,000 registered construction contractors; by March 1992, there were 522,450 licensed contractors registered with the Ministry of Construction. And whereas the construction industry employed 4.7 percent of the country's workforce in 1955, by 1980 construction workers outnumbered individuals employed in farming and forestry. In 1992 Japan's 6.2 million construction workers represented 9.6 percent of the workforce—rich source of campaign support for ambitious legislative candidates, as we shall see.

During the 1980s stagnant domestic demand impelled many Japanese construction firms to look overseas for new markets. (The "ice age" in domestic construction, a product of the worldwide oil shock of the late 1970s, lasted until 1986 and slowed the average annual increase in public and private investment in construction from 21 percent a year, between 1970 and 1978, to 0.01 percent a year, between 1978 and 1986.) As a presence in overseas markets, Japanese firms rose from a seventh-place ranking in 1981—behind American, South Korean, French, West German, Italian, and British contractors—to second place in 1989. Though U.S. firms retained first place in the

international rankings, their margin of preeminence declined precipitously, from $478 billion in 1981 to $181 billion in 1989. And Japanese construction in companies made great in roads in the North American market, winning 40 percent of the overseas orders in North America in 1987. Much of this work serviced the construction needs of longtime Japanese clients who invested in North America during time frenzy of foreign direct investment that followed the appreciation of the yen after the 1985 Plaza Accord. Japanese contractors also made impressive gains in the construction markets of Southeast Asian countries by providing design and consultancy services for large-scale projects, many of which were funded by generous Japanese foreign aid (Chittiwatanapong 1992).

Apart from its size and rate of expansion, time most striking facet of Japan's construction industry is its two-tier structure. At the top are a small number of large general construction contractors, time major zenekon ; at the bottom are more than 500,000 small firms.^[6] The largest zenekon possess the technological capability to design and construct innovative skyscrapers, factories equipped with "clean rooms" for the assembly of sensitive high-technology products, fast-breeder nuclear reactors, and even projects for outer space, and they are among the world's leaders in the building of tunnels and underground facilities, the use of robots in construction, and the use of prefabricated modular devices transported directly to the construction site. In 1991 four of the five largest construction companies in the world were Japanese (Levy 1993, 24). In contrast, the multitude of small-scale firms, many of which are inadequately capitalized, possess technical capacities only slightly superior to those of medieval builders (Watanabe 1982, 272). Be it in urban Yokohama or rural Shikoku, many workers at construction sites still don the traditional garb—billowing knee-length trousers and boots split into two parts between the big toe and the other toes—and shinny up flimsy wooden poles, rather than modern ladders.

The elite of Japan's general contractors are the Big Six: Shimizu Corporation, Taisei Corporation, Kajima Corporation, Takenaka Corporation, Obayashi Corporation, and Kumagai Gumi (see Table 1). Together, the aggregate net sales volume of these six firms exceeded ¥10.3 trillion in 1993. Collectively, they employed more than 73,000

full-time workers. In 1993 Shimizu, which reported sales of ¥2.17 trillion, held the top spot for sales, while Kajima employed the largest number of workers (nearly 15,000). Each of the Big Six admninister extensive overseas branch offices and subsidiaries, and several own or hold substantial interests in domestic affiliates specializing in road paving, residential land development, and other construction services.^[7]

Beneath the elite six are the "quasi-large-scale" general construction contractors. In 1993 Fujita Corporation, Toda Corporation, Hazama Corporation, Tokyo Construction, and Mitsui Construction had the largest sales volume in this second tier. Nishimatsu Construction, Sato Kogyo, and Tobishimia Corporation were among the leaders in large-scale civil engineering works. A group of a dozen or so firms, including Penta-Ocean Corporation, Konoike Construction, and Aoki Corporation round out this second tier of large zenekon . Although smaller in terms of revenues and workforce than the Big Six, these second-tier zenekon possess the technologival sophistication to construct super-high-rise buildings, clean rooms, offshore energy developments, computerized buildings, and urban redevelopment projects. All of these firms are active internationally, and some of them—notably Nishimatsu, Penta-Ocean, and, especially, Aoki—realize a major share of their sales in overseas markets. In the case of Aoki, overseas markets accounted for nearly one-third of the firm's sales during 1993.

Beyond this small circle of major zenekon are more than half a million small firms, and the gulf between the two groups is wide indeed. In 1991, for example, construction firms capitalized in excess of ¥100 million (0.09 percent of registered firms) took in over half of all revenues from completed projects, while firms capitalized between ¥50 million and ¥100 million (1.16 percent of registered firms) reaped about 8 percent of the revenues from completed projects. In contrast, roughly 80 percent of the registered construction firms were capitalized at under ¥10 million, and nearly 40 percent of all construction firms were individually owned and operated. These petty contractors operate in a Hobbesian world, lacking modern machinery, stable management systems, or cozy ties with major banks. In 1982, for example, nearly three-fourths of bankruptcies reported among construction firms involved contractors capitalized at ¥10 million or less

(Nakamura 1982, 104). In order to survive, many of these contractors depend on the prepayment system for privately as well as publicly funded projects, whereby contractors routinely receive 20 to 40 percent in advance.

Subcontracting is more prevalent and more multi-layered in Japan's construction industry than it is, for example, in the United States. On any project, "a carpenter subcontractor may subcontract rough framing to a second-tier subcontractor, who, in turn, might subcontract roof joists to a third-tier subcontractor who, in turn, may hire a fourth-tier subcontractor to distribute the joist" (Levy 1993, 14). These subcontractors tend to have lower skills and offer relatively lower wages and fewer employee benefits. Frequently, firms at the very bottom tier of these subcontracting networks employ substantial number of unskilled illegal aliens, many of them from other Asian countries ("Demand Grows" 1988). Although some formerly small-scale firms, such as Chisaki Kogyo of Hokkaido and Fukuda Corporation of Niigata, managed to expand and to establish substantial regional or nationwide operation, most tend to remain small, local concerns.

Historical Development of the Industry

To understand Japan's construction industry, one must first view it as the product of a historical evolution that reaches back to the seventh century, when Japan's earliest builders constructed the Horyuji, the world's oldest remaining wooden building. Japan's first construction firm, the forerunner of Matsui Kensetsu, was established in 1586 by Matsui Kakuzaemon, who was commissioned to build a castle for the lord of the Kaga Domain (today's Ishikawa Prefecture). Not until the Edo period (1603–1867), however, when the capital was moved from Kyoto to Edo (the present-day Tokyo) did the building trade expand greatly. By the seventeenth century, Edo had become the worlds largest city, perhaps the first to boast a population surpassing one million, and there was a great demand for carpenters to build estates for feudal lords and shops for the growing merchant class.

Subsequently, in the Meiji period (1867–1912), under the slogans "rich country and strong military" and "increase production and promote industry," the government launched an infrastructure program

that produced the country's first true construction contractors, and the Japanese construction industry began to shed its guildlike existence (T. Maeda 1988, 81). In the early Meiji period, "many of today's leading contractors came into existence, thanks to the multitude of government orders" for administrative offices, railroads, and riparian works (F. Hasegawa 1988, 5). Also during this period, foreign engineers from Britain, Holland, Germany, and the United States introduced Western construction techniques to Japan. These efforts brought forth the Western-style Tsukiji Hotel, built by Shimizu Kisuke II in 1867, and the British First Mansion and the American First Mansions, constructed by Kajima Iwakichi to house merchants in Yokohama's foreign settlement. (Nakamura 1982, 72).

The Finance Law (Kaikei Ho ) of 1890 decreed that companies would win contracts for nationally funded public works projects through open competitive tendering. Under this law, during the late Meiji period and into the Taisho period (1912–1926), "even firms without experience or the power of trust were able to secure contracts simply by submitting the lowest bid" (T. Maeda 1988, 84). During this period of open competition, Obayashi Gumi, one of today's Big Six, was established. Ironically, the open, "democratic" bidding procedures mandated in the Finance Law were not substantially revised until 1922—the high tide of the period of "Taisho Democracy"—when the construction industry succeeded in pressuring the government to established a system of designated competitive tendering. According to one observer, this institution was motivated by the fact that

blood letting orders [shukketsu juchu ], collusion, and conflict among contractors had become an everyday occurence. This [situation] spawned the emergence of professional "fixers" [dangoya ] who served as intermediaries and go-betweens. Moreover, the custom of paying a certain percentage of the tendering cost to other firms came into existence. For these reasons, the government began assessing the accomplishments of firms and designating trustworthy contractors to bid on projects.

(Nakamura 1982, 107–8)

Japan's construction industry grew with the country's imperialist expansion. Preparations for the Sino-Japanese War of 1894–95 sparked a boom in military-related works as well as railway and electric power projects. In 1900 Kajima Gumi supervised construction of

a railway line linking Seoul and Inchon in Korea. Shortly thereafter, three Japanese firms—Okura Gumi (the forerunner of Taisei Construction), Obayashi Gumi, and Hazama Corporation—combined to build a 440-kilometer-long stretch of the Trans-Formosa Railroad. However, the overseas advance that began in earnest during the late Meiji period was not always a blessing for the Japanese construction industry. In the midst of preparations for war with Russia in 1904, railway lines were constructed connecting Pusan with Seoul and Seoul with Sinuiju. Owing to allegedly inadequate and inaccurate information provided by the client, the firms that participated in this project suffered financial losses and, in some cases, bankruptcy (Hippo and Tamura 1988, 60). Still, the overseas advance continued, and Japanese firms constructed various railways lines, including the Pusan-to-Dalien link of the South Manchurian Railway, as well as hydroelectric projects and mining facilities in the puppet state of Manchukuo. By 1941 Japanese contractors were constructing roads and military works throughout all of occupied China and in various parts of Southeast Asia. At home, the war mobilization led to the disbanding of construction industry associations in favor of the Army's Military Authority Cooperative Association (Gunken Kyoryoku Kai), the Naval Facilities Cooperative Association (Kaigun Shisetsu Kyoryoku Kai), and, in the closing days of the war, an umbrella organization known as the Wartime Construction Team (Senji Kensetsudan).

The aftermath of war provided a setting for corporate reform and rapid growth. As part of the Occupation's trust-busting program, for instance, the Okura zaibatsu dissolved, and the Okura Gumi became the Taisei Construction Company. Other firms, such as Kajima and Shimizu, took the opportunity to modernize their corporate names by dropping gumi ("group" or gang), a feudalistic-sounding term often associated with organized crime syndicates. During the first decade of the postwar era, the construction industry benefited from the domestic reconstruction boom. The peak in domestic orders came in 1947; three years later, an additional boost was afforded by the admission of Japanese firms to the Okinawa market. There, joint ventures linked major Japanese contractors with Morrison Knudsen, the Bechtel Group, Pomelroy, Kiewit, and other U.S. construction companies. In addition, war reparations, beginning with the building of a hydro-

electric station in Burma in 1954, provided a springboard for recapturing overseas markets.

The construction industry built the infrastructure for the Japanese economic miracle. In this regard, the Ikeda cabinet's New Industrial Cities Policy offered incentives that benefited contractors throughout the country. The Tanaka cabinet's program to "remodel" the Japanese archipelago bequeathed a similar, albeit ephemeral, bounty. In the meantime, the construction industry reaped handsome rewards from preparations for the 1964 Tokyo Olympics and the 1970 Osaka Worlds Fair, as well as from the flood of public investment during the "era of highways" in the 1960s and '70s. Indeed, the industry rode a continuous wave of expansion until the onset of the "ice age" in the wake of the oil shocks in the late '70s. In the mid '80s the government inaugurated a program to expand domestic demand and construction began to boom in the late '80s. But this boom was short-lived, as demand in the private construction market collapsed in the wake of the bursting of the "bubble" economy and the shock waves generated by the zenekon corruption scandal at the outset of the '90s.

Dango as a Collusive System

As mentioned earlier, the first condition necessary to sustain effective collusive action is the delineation of the membership of the conspiratorial ring. Deciding on the optimal size of a price-fixing cartel is a complex problem. The larger the number of actors who stand to benefit from collusive action, the smaller the share of the gains from the action that will accrue to each. Moreover, the larger the group, the higher the transactions costs involved in orchestrating effective bargaining. The illegal nature of the undertaking requires members of the group to bargain until they agree how to share the costs and benefits, and again the difficulties in orchestrating collusive action increase as group size increases. Creating a conspiratorial ring broad enough to satisfy the interests of all potential spoilers is also a fundamental problem. For example, if all but one firm in a competitive arena agrees to a higher price, then the hold out firm stands to reap large profits by selling its product at a lower price. This generally has the effect of undercutting the cartel by driving the fixed price down to a competitive level. (Olson 1982, discusses these problems.)

In the case of Japanese public construction, two institutional devices facilitate the task of delineating a conspiratorial ring. The first is a highly structured system of industry associations at both the national and local levels. Close contact among executives of various firms, commonly facilitated by membership in industry associations, provides opportunities for greater communication, allowing participants to resolve standards of acceptable market conduct.^[8]

The second, more direct facilitator of collusive action is the Japanese government procurement system. Government procurement system come in three generic forms. In an "open competitive tendering" system, all qualified parties submit bids, and the contract goes to the firm making the lowest responsible bid. This procurement system operates in nearly all of the projects administered by the U.S. Army Corps of Engineers, and, since early 1994, in contracting for extremely large-scale public works projects and consultancy services in Japan. Under a second system of government procurement, for projects that demand highly specialized expertise, or during times of emergency, a specific firm might be awarded a "discretionary contract." This system functions in a portion of contracts granted at all levels of government in both the United States and nationally authorized projects in several Western European countries. In Italy, as exposed by the Mani Pulite investigators, discretionary contracting was the preferred procurement system for public works projects, especially in the South, where three-quarters of public works contracts were not put out to tender. According to a Milanese business executive, "They would just write their names on piece of paper and draw them from a hat. First one out would win the first contract, second one the second, and on" (New York Times , 3 March 1993).

In spite of reforms in 1994, the bulk of government procurement in Japan corresponds to a third type, "designated competitive tendering system," or "select competition" as it is known in Great Britain. In the case of projects administered by the central government, firms wishing to bid must reveal sources of funding and operational capacity. Until 1988, when intense U.S. pressure effected a minor modification, a devious catch-22 required that all would-be bidders undergo assessment based on the results of construction work during the previous two years within the Japanese market (Ministry of Construction 1988, 37–47). Of course, foreign firms had no record of per-

formance in the Japanese market and were thus barred from bidding. Yet even after complaints from foreign governments suceeded in extracting modest reforms, foreign participants remained excluded from cozy dango arrangements, and they were forced into joint ventures as junior partners or blackballed by vital subcontractors.

Under the Japanese government's procurement system, ten "qualified" firms are generally invited to submit bids for a public works project. The contract is then awarded to the firm submitting the lowest responsible bid judged according to the government's confidential ceiling price. In this way, the procurement system limits potential conspirators on a project to the small pool of designated bidders. Not surprisingly, construction contractors and their political patrons, who staunchly seek to limit the number of firms designated to submit bids, value this institutional apparatus. When the government doubled the number of firms invited to bid, in response to a media campaign following a 1981 bid-rigging scandal in Shizuoka Prefecture, pressure groups in the construction industry immediately initiated efforts to reverse the policy. Their principal effort was to promise substantial campaign contributions to influential members of the LDP's "construction tribe," one of the groups of "policy specialist" in the party. By the end of the following year, these tribalists—led by legendary political broker Kanemaru Shin—succeeded in restoring the status quo ante (Itasaka 1987, 74–75).

Thus, the Japanese system affirms the general rule that mutual recognition of interdependence among evenly matched firms in a well-defined market breaks down when the number of members exceeds ten or twelve (Posner 1970). As one student of collusive action observes, " a smaller number of firms reduces conspiratorial transaction costs by limiting the number of encounters needed for communications" (Haar 1983, 30–31). For example, a ring that has 10 members will comprise 45 one-on-one relationships, but doubling the ring to 20 more than quadruples the constituent pairs to 190, an extremely challenging and expensive coordination problem. By defining and limiting the pool of potential conspirators, the Japanese government procurement system only simplifies the coming together of collusive contractors.

The second condition for sustaining a successful price-fixing ring rests in the group's ability to exclude outsiders, both to maintain each

member's share of the profits and to shut out spoilers who might undermine the accord by charging a lower price (see, for example, Kuhlman 1969, 69–82).

As we have seen, the government procurement system itself restricts the circle of "competition" to a small subset of firms who are invited to submit bids for public works projects. In many cases, as the zenekon scandal revealed, the meddling of elected politicians influences which firms will receive a designation to bid. Construction firms that fail to cater to the appropriate bureaucratic and political parties risk being shut out of future bidding.

Thus, gift-giving is pervasive. Most competent business managers know the birthdays, hobbies, and other particulars of key government officials, politicians, and their administrative assistants (Jin 1989; Minami 1981, 17). Outpourings of generosity occur most frequently at midsummer and year's end, the customary gift-giving seasons in Japan. (A Ministry of Construction bureaucrat told me of his amazement upon witnessing the bountiful harvest of gifts reaped by senior officials in the Road Bureau as compared with the paltry takings of those in bureaus with fewer distributive benefits to dole out.) During the 1970's, a public works scandal in Fukushima Prefecture uncovered a massive pyramid of corruption in which contractors used gifts and campaign contributions to court prefectural government officials, party and interest group leaders, local politicians, and even the prefectural governor (Yoshida 1984). At the national level, large-scale general contractors routinely funnel contributions to influential politicians. The Sagawa scandal revealed that as many as twenty of Japan's largest construction contractors donated up to ¥20 million each year over a number of years to LDP Vice-President Kanemaru Shin (Asahi shinbun , 10 Apr. 1992). Finally, as in the United States, the mass media frequently features accusations of outright bribery involving contractors and public officials—generally local and municipal administrators. Of the one hundred or so bribery scandals each year in Japan, over half involve construction firms (Iinuma 1987).

Political influence, known as the "voice of heaven" (ten no koe ), often influences the selection of the anticipated low bidder. As a local contractor observed, "when the 'voice of heaven' comes down and the process of dango is completed, for all intents and purposes the winning bidder has been decided and designated competitive bidding be-

comes a mere ceremony" (Asahi shinbun , 31 Aug. 1993). In Ibaraki Prefecture, Governor Takeuchi Fujio allegedly intervened an average of six to seven times a month on behalf of some thirty construction firms (ibid., 18 Dec. 1993).

Large transfers of money tend to accompany the articulation of heaven's will. Kanemaru's heavenly voice allegedly fetched a kickback of 5 percent or more of the contracted price of the public works project. More typically, influential politicians—particularly members of the LDP's construction tribe—receive a "back margin" of 1 to 3 percent of the cost of the project (Jin 1989, 86). For example, Mayor Oyama Masahiro of Sanwa Town in Ibaraki Prefecture allegedly received ¥14 million in "gratitude money" from Hazama Corporation, the company constructing a ¥140-million sports center in his jurisdiction (Asahi shinbun , 20–21 July, 10 Aug. 1993).

From a contractor's perspective, funneling kickbacks to politicians is a necessary "insurance fee" to secure designation to bid for especially desired public works projects. As the zenekon scandal revealed, large contractors awarded letter grades to legislators based upon their perceived influence. Not surprisingly, construction kingpins like Kanemaru, Takeshita Noboru, and Ozawa Ichiro received the highest marks and, hence, the largest contributions. At the local level, contractors in Ibaraki coined the term "Takeuchi pilgrimage" to refer to the courtesy calls they paid, hefty bundles of cash in hand, to persuade the governor of the virtue of their cause (Sanger 1993a).

The cost of active complicity rendered by government bureaucrats is difficult to tally. Instances of outright bribery, especially among officials of the central state bureaucracy, are seldom proven. Officials at the Ministry of Construction, for example, are commanded to "neither accept cash nor women" (Honda 1974, 198). One bureaucrat, Kyosaka Motoji, former director of the ministry's Hokuriku Regional Bureau, failed to heed this dictum and was arrested in 1977 on charges of accepting substantial gratuities from contractors in exchange for favoritism in deciding lists of designated bidders (Kanryo kiko kenkyukai 1978, 74–77). Indirectly, however, the cost of bureaucratic mediation is embodied in sinecures for retired officials—known as "descent from heaven landing spots." Indeed, firms employing ex-bureaucrats not only benefit from their technical competence, but

also appear to be privy to strategic information leaks concerning the government's "confidential" ceiling price.

Since sub rosa arrangements cannot be formalized in a legally binding contract, enforcing the terms of the collusive accord is a third hurdle for a collusive ring. The members of the cartel have a mutually profitable course of action—loyal adherence to their agreement—yet each member faces a strong incentive to depart from that course, to cheat. If all cheat, however, the profitable agreement breaks down, and all suffer. Thus effective collusion demands recognition of the destabilizing effect of secret price-cutting and the creation of procedures to force compliance to the agreed-upon accord.^[9]

In Japanese public construction, the need to mete out selective incentives creates a role for industry associations and, on occasion, the organized underworld. In all, there are more than one hundred formal industry associations registered with Japan's Construction Ministry. The most prominent are the Japan Federation of Construction Contractors and the Japan Civil Engineering Contractors' Association. But the actual task of meting out selective incentives falls to the myriad of informal "fellowship clubs"—known as "dango organizations"—such as the Tokyo-based Management Harmony Society (Keiei Konwakai), the Tohoku Construction Industry Conference (Tohoku Kensetsugyo Kyogikai), and the Saitama Saturday Society. For the most part, these groups sponsor breakfast meetings with influential politicians and arrange golf outings intended to foster a chummy atmosphere in which to exchange information about public works projects. On occasion, however, such groups also punish wayward members.

A firm that violates the bid-rigging agreement—a transgression known as dango yaburi (meaning "to break the dango ")—is ostracized from the industry association and barred from participating in future collusive accords (Minami 1981, 45, 76). As a local construction industry official in Yamanashi observed, "If a different company tried to get a contract by entering a low bid, it would be kicked off the list of designated bidders for future projects" (Blustein 1993a). Naturally, enforcing this sort of negative selective incentive requires the cooperation of the government authorities who determine the pool of designated bidders. Firms engaging in iterated games of bid-rigging

"lend" co-conspirators the opportunity to win a contract in exchange for the right to "borrow" the contract for a future project. From the Fukushima scandal emerged a ledger detailing the lending and borrowing activities of th prefecture's electrical contractors' association over an extended period (Yoshida 1984, 168). A schematic of the bid-rigging system of the Saitama Saturday Society, a local industry association, is shown in Figure 1.

Strong disincentives for whistle-blowing also facilitate self-enforcement. The practice of lifetime employment for career employees at large firms makes it irrational for employees to blow the whistle on collusive accords. And the risk of retaliation in a setting defined by a

[Full Size]

Figure 1.
The Saturday Society's Scheme for Bid-Riggng on Public Works
SOURCE : Asahi shinbun, 31 Jan. 1992.

highly regimented system of trade associations discourages whistle-blowing on the activities of competitors. When it does occur, it often takes the form of an anonymous tip from a disgruntled insider. A reporter assigned to the press club at the Ministry of Construction told me that anonymous telephone calls to the newspapers—particularly the Asahi Shinbun , a vocal critic of the dango system—increased perceptibly during the "ice age" of curtailed public construction investment, from 1978 to 1986.

Finally, on occasion, collusive rings utilize coercive incentives to secure compliance. For example, an instance of bid-rigging on a municipal road contract in Shizuoka City came to the attention of the authorities after an infuriated co-conspirator hit a local construction company boss over the head with a metal chair. The incident occurred when the victim voiced displeasure with a particular dango arrangement (Schoenberger 1989). Gangsters (yakuza ) occasionally help enforce selective incentives. In Sakato City in Saitama Prefecture, a contractor desperately desired the contract for an agricultural waterworks project. Several days prior to the date for submitting bids, he telephoned representatives of the other firms designated to bid on the project and attempted to arrange a meeting at a local eatery. Unnerved by the unfolding zenekon scandal, most of the other contractors refused the invitation. As a last resort, the contractor hired two thugs, with ties to a well-known gangster organization, to intimidate the contractors who had declined to collude. When word of yakuza involvement reached the mass media, the bidding was suspended, and the overzealous contractor and the gangsters were arrested (Asahi shinbun , 25 Nov. 1993).

The fourth hurdle for a cartel is devising a mechanism for parceling out the gains (Olson 1982, 20–23). Because of the high transaction costs imposed by the need for consensual bargaining, firms may demand a disproportionate share of the gains from the collusive action in return for their cooperation. Threats to hold out, in turn, will not be credible unless they are sometimes carried out. Reaching agreement on apportioning the costs and benefits may thus take an extraordinary amount of time, and often a broker is needed to orchestrate the accord and create an acceptable formula.

As investigations in the zenekon corruption scandal revealed, for a

number of years, officials of certain large-scale general contractors allegedly parceled out public works projects in the various regions (K. Matsumoto et al. 1993). For example, a senior executive in Obayashi Corporation's Osaka office served as broker in rigging bids for the construction of a levee at Kansai International Airport (Kosei torihiki iinkai 1990, 33–36). And a retired Construction Ministry bureaucrat allegedly performed brokerage in rigging bids for air conditioning and sanitation contracts in offices built for the Ministry of Justice and the Public Prosecutor's Office (Asahi shinbun , 25 Nov. 1987). On occasion, a particular broker will acquire almost mythical powers. Uera Sukemasa, one-time president of Tobishima Corporation, was supposedly so powerful and politically well-connected that he determined the disposition of any public works project he chose (Asahi shinbun , 24 Oct. 1993). Similarly, Kimura Hitoshi came to be known as the "Emperor of the Dango Association" for dam construction. After retiring from a senior post at Taisei Construction, Kimura unabashedly admitted to drawing up a "list of dango for dams" specifying contracts for planned projects over the course of a decade (Asahi shinbun , 2–19 Feb. 1982).

Brokers also devise a formula to apportion the benefits and costs of the collusive action. Commonly, this artifice involves the funneling of an agreed-upon portion of the revenues from a rigged contract, known as "dango money," from the designated winner-to-be to the other firms in the cartel (Minami 1981, 40, 52, and 62). Frequently, an informal agreement concerning the disposition of subsequent contracts effects a compromise. The bid-rigging for a project to dismantle an old factory in Tokyo's Shinagawa District combined several methods. In return for agreeing to submit the low bid of ¥49 million, Goto Dismantling Company allegedly promised to pay ¥3 million to each of the other four conspirator firms and to assist a co-conspirator in obtaining the contract for an upcoming project ("Doboku gyokai" 1979).

The system allegedly employed by public works contractors in Kawasaki City provides further insights into how benefits may be apportioned. To avoid undercutting the predetermined winning bid, representatives of designated firms wishing to win the contract regularly met to work out an acceptable compromise. When such discussions failed to produce a viable modus vivendi, representatives convened

an "adjustment meeting" at the office of the Kawasaki Construction Contractors' Association. If this meeting did not conclude in a suitable accord, a vote or the drawing of lots decided the predetermined low bidder. Over the course of time, this informal process became routinized (Asahi shinbun , 5 Oct. 1993). In Kawasaki, as elsewhere, forcing the bidding to go through several rounds before the prearranged "chosen one" magically triumphs commonly conceals a collusive agreement.

The "shady joint venture" (ura jointo ) is another popular tool for apportioning time fruits of collusive action. The system employed by contractors in Chiba Prefecture illustrates this practice. Following the designation and bidding processes, the contract for a public works project goes to firm 1 as the "nominal" low bidder (meigi ). As the prime contractor, firm 1 is expected to allocate segments of the project to specialized subcontractors. However, in a shady joint venture, the prime contractor tosses the contract as an extralegal "curve ball" (marunage ) to firm 2, nicknamed the "padded mitt" (anko ). Firm 2 then proceeds to throw a curve to firm 3, and firm 3 tosses the ball to firm 4. As nominal prime contractor and subcontractor, firms 1 and 4 can lay just claim for services rendered. But, in a shady joint venture, firms 2 and 3 also receive payment for service charges, even though neither submitted the low bid or performed any work. According to a local contractor, half of all civil engineering works in Chiba Prefecture involved shady joint ventures of this sort (Asahi shinbun , 7 Aug. 1993). A documented instance of a shady joint venture emerged in the bidding for a riparian project in Ibaraki Prefecture in 1978. In that case, a local company submitted the low bid of ¥1.86 billion, while two large-scale Tokyo-based construction firms divided more than two-thirds of the contracted amount. Following a decade of legal maneuvers, the conspirators were "punished" with suspended jail sentences (Asahi shinbun , 8–31 Nov. 1981).

The fifth, and final, problem for a cartel is evading criminal prosecution and the probing eye of government antitrust regulators. Success here requires active or passive complicity on the part of public works bureaucrats, often motivated by pressure from influential legislators and other concerned political actors. Given the increasing levels of international interdependence and aggressive unilateralism, the

survival of collusive accords often depends as well on the apathy or ignorance of trade monitors in other countries (see Keohane and Nye 1977; Bhagwati and Patrick 1993).

In Japan's public construction arena, the supposedly confidential government-set ceiling price for public works provides temptations for officials from the contracting ministries to engage in collusive behavior. The reemployment of large numbers of retired officials in the private sector creates strong incentives for preferential leaks of confidential information. Firms employing retired officials somehow acquire an uncanny ability to divine the government's ceiling price. A survey of major projects undertaken in the mid-1970s concluded that those firms that hired ex-officials submitted bids that were 99.92 percent of the supposedly confidential official estimates (Yamamoto 1975, 138). The amazing calculation skills of Japanese contractors were also displayed in the bidding for the six stages of the ¥1.44-trillion Trans-Tokyo Bay Highway. There, in each instance the winning bid was 99.7 percent of the government ceiling price (McGill 1994, 9).

A case involving the Hokkaido Development Agency (HDA), an organ under the purview of the prime minister's office, offers additional evidence of the connection between retired government officials and strategic leaks of confidential information. Present and former officials of the HDA's Development Bureau admitted that virtually every public works contract handled by the bureau during the preceding quarter of a century went to firms employing ex-officials of the HDA. Prior to the deadline for bidding, the government's secret ceiling price was leaked, and the majority of the leaks seem to have emanated from the bureau's eleven branch offices, which designate which firms will be allowed to submit tenders. The implicated officials sought to justify their actions by claiming that had they not done what they did, local firms would have been victimized by the onslaught of giant Tokyo-based construction companies (Yomiuri shinbun , 2 Mar. 1992).

Another factor facilitating the institutionalized system of collusion is passive complicity on the part of government actors. As a newspaper editorial asserts, "In Japan it seems that government agencies, when commissioning contracts on public works projects, not only look

the other way but also create conditions that make it easy to fix bids" (Asahi shinbun , 6 Oct. 1989). Government complicity includes weak penalties and lax enforcement of the antimonopoly law: "The Japan Fair Trade Commission (JFTC) does not enforce the law in a manner which could be characterized as vigorous or even adequate by U.S. standards (Wolff and Howell 1992, 57).^[10]

The dango system has not only enriched individual politicians but also benefited the LDP as a national force. Along with small business and farmers' organizations, the construction industry serves as one leg in the LDP's grand coalition, and it has consistently ranked at or near the top of the ruling party's list of financial benefactors. Knowledgeable observers estimate that half or more of the LDP's total contributions come frm the construction industry. In the midst of the zenekon scandal, for instance, the president of Taisei Construction publicly admitted that his company alone made ¥8.6 billion in illegal political contributions between 1990 and 1993 (Asahi shinbun , 19 Oct. 1993). Such generosity has bought the construction industry "influence from top to bottom. The steel makers have it only at the top, but construction is strong at the very base of the pyramid" (Hasegawa Tokunosuke, in Schoenberger 1989).

For tax purpose the construction firms lump most of these illegal contributions under "unaccounted-for expenditures." Under Japanese law, companies can avoid providing detailed explanations for such expenditures by paying higher taxes. According to the National Tax Administration Agency, the construction industry typically accounts for 60 to 75 percent of all unaccounted-for expenditures claimed by firms capitalized at ¥100 million or more. During 1992, for example, the tax return's of large-scale construction firms included ¥43.8 billion in such outlays, nearly three-quarters of the nationwide total of ¥59.5 billion. Between 1990 and 1991, three major general contractors—Shimizu, Taisei, and Kajima—reported ¥150 billion in unaccounted-for expenditures (Asahi shinbun , 8 Apr., 31 Aug., 10 Sept., 30 Dec. 1993). And while officially reported political contributions remained relatively unchanged between 1985 and 1991, "shady political donations" nearly doubled.

To generate funds for under-the-table contributions to influential politicians, some contractors founded dummy companies that they

billed for overhead expenses. Other firms formed large pools of "behind the scenes money" from incremental miscellaneous expenses siphoned from networks of branch offices. General contractors also ordered subcontractors to submit inflated bills for their services and compelled them to kick-back a portion of the money to the general contractor. For example, between February 1991 and January 1992, one subcontractor reportedly channeled around ¥50 million to several general contractors; the kickbacks amounted to almost 1 percent of the subcontractor's annual profits (Asahi shinbun, 29 Oct., 8 Nov. 1993).

In short, the dango system is entwined in the mechanisms of the political power in Japan's economy. Construction contractors reap inflated profits, government officials glean administrative power and postretirement security, and legislators harvest political contributions and campaign support. The losers, of course, are the taxpayers: by various estimates, big-rigging and political payoffs inflate the cost of public construction in Japan by 30 to 50 percent.^[11] And inflated bids have been estimated to produce between 16 and 33 percent of the construction industry's total revenues—somewhere between $50 billion to $100 billion annually (McMillan 1991, 201). As noted earlier, the dango system also creates incentives for bribery, graft, and other forms of corruption. Organized crime is believed to control or influence nearly one thousand construction firms (Schoenberger 1989; H. Ito 1987, 58–68; Kaplan and Dubro 1986). Even the cost of private-sector construction is higher in Japan than in any other advanced industrialized country. Although labor costs for construction in the United States are almost double those in Japan, the unit cost of building a suburban office building is nearly three times greater in Tokyo than in Chicago.^[12]

The Logic of Collusive Action

We must not overlook the advantages of the dango system as an alternative to the myopic drive for the immedite and maximal profit in situations of free competition. In this sense, the system allows companies to take a longer view in selecting and pursuing a business strategy, relying all the while on collusive accords for a "fair share" of the mar-

ket. Because of its sensitivity to cyclical economic downturns and to the effects of "lumpiness"—the large size and sporadic frequency of transaction—construction is an industry in which "collusive conduct is as natural as competition (Erickson 1969, 98). Since the costs of idle capacity are high in thus capital-intensive industry, acquiring a large contract at the outset of the building season can often make or break a company. Such cash-flow considerations are all the more significant in the Japanese case, since a substantial portion of the contracted amount is paid in advance.

Given the costs of retaining employees under Japan's lifetime employment system, it is rational for large-scale general contractors to sacrifice potential profits by adhering to a system of collusive action. Workers in subcontracting firms then bear the costs of cyclical economic downturns. And, for entirely different reasons, the system is rational from the perspective of subcontractors. As Takeuchi Yoshio, an official of the Kansai International Airport Corporation notes, "the Japanese subcontractors would be ostracized if they worked with foreigners. They may be able to eat once with a foreign boss, but it's all over for their relationships with Japanese companies" (Schoenberger 1989). Seen in this light, the dango system represents a mutual insurance system for firms in a volatile industry.

Yet the dango system also serves a set of broader interest and functions. Despite a pervasive network of cartels in public construction, the overall quality of Japan's social infrastructure in impressive. In comparison with Italy, where subways, highways, urban transport projects, and other public works were paid for but never built, the state and quality of Japan's public infrastructure is relatively high. Although one might anticipate widespread negligence and venality, instances of shoddily constructed highways, tunnels, and other public structures are remarkably rare. The bidding is rigged, but the contractors uphold high standards of accountability in meeting the exacting government specifications. Indeed, the 1995 Kobe earthquake inflicted the bulk of its devastation on older structures built to less stringent government specifications.

Moreover, inasmuch as the dango system props up the overstaffed and technologically backward firms that constitute the massive underside of the industry, the cartels may be seen as a de facto proxy for

Japan's underdeveloped social welfare and unemployment programs ("Ote kensetsu" 1982, 55; Minami 1981, 71 and 104). Some observers estimate that as many as one-third of Japan's more than half a million licensed construction firms would go bankrupt if the government abolished institutionalized bid-rigging (Schoenberger 1989). That the dango system facilitates political stability, social equality, and mutual reciprocity in government-business relations helps to explain the paradox of Japan's phenomenal and sustained economic growth in a pervasive system of political clientelism.

As long as Japan's public construction market remained insulated from foreign competition, the intricate system of bid-rigging functioned relatively smoothly. It provided the selective incentives required to induce ratonal individuals to engage in otherwise irrational action. And, notwithstanding the imposition of a variety of limited and conditional reforms beginning in 1988, collusive behavior continues to characterize Japan's public construction market. Simply stated, the dango system functions because Japanese political institutions provide ways of circumventing the potential obstacles to successful, sustained collusive action. Specifically. the designated bidder system helps determine the members of bid-rigging cartels, while blocking entry by would-be spoilers. Industry associations enable the self-enforcement of illicit accords and the enforcement and weak penalties for anti-trust violations do little to deter bid-rigging.

This precarious, but relatively stable, equilibrium received a jolt when the U.S. government sought to gain market access for U.S. contractors in the bidding for construction projects at the Kansai Airport. From the Japanese point of view, the opening of the lucrative domestic construction market would not only invite competition from foreign firms but also upset the bid-rigging cartels by introducing competition among domestic firms. In response to this threat to the dango system, government officials, legislators, public works bureaucrats, and the large construction companies all had an enormous stake in resisting any efforts to reform the government procurement policies, no matter how painful the sanctions proposed in Washington, D.C.